WHERE IS HINDI? WHY HINDI CAN NOT BE IMPOSED AS NATIONAL LANGUAGE? by: Dr. Rajan R Patil, SRM University, Chennai.
Dear friends,
Personally I do subscribe to the view “ jo hindi me hai baat, vo kisi aur me nahi” and for the very same reason have taken on some of my radical madrasi, kannadiga, Marathi friends for their linguistic fanatic stands. I do keep urging them to be pan national to appreciate the beauty of our diversity….some day I would like confront karunanidhi, vatal nagraj and Raj Thakares on the issue .
At the same time Modi is also blissfully being ignorant when he claims Hindi is single largest language spoken by Indians, with UP and Bihar being heartland of hindi Belt. That One more Myth propagated by his ilk.
I have been travelling across Uttar Pradesh(UP) extensively for past few weeks. I have been invited to monitor and evaluate a research study on Dirrhoea control using Zink supplementation in children in UP and Bihar. I was pilot testing their Hindi questionnaire on diarrhea and ORS practices in rural community. To my surprise Simple words like “ kya KAARAN he’, ‘LAKSHAN kya he’ or “PRABHAVI he ki nahi” "goli PARYAPT he ki nahi" etc were not being understood by the rural community in close to Lucknow city. Obviously the questionnaire was prepared in delhi. Worse the there was a real heavy duty word’ NIRJALIKARAN ‘meaning dehydration which I am sure even a pediatrician would find it hard to pronounce and understand let alone a rural community. My first recommendation to the the funders and implementors of this research project was to completely desikaran the questionnaire other wise they will get all funny results.
My point is, Modi’s Hindi starts fizzling within just 25kms radius from capital city Lucknow in so called Hindi heartland UP itself. I encountered Range of dialects starting from, khadi boli, Awadhi, Bhojpuri, Braj bhasha to name few which absolutely have no similarity to the highly sanskritisised chaste hindi that Modi and his group is trying to impose.
I really long for earlier Hindustani language (urdu mixed hindi) that really gave that entoxicating touch to the emotions penned by Kaifi azmi, Mazrooh sultanpuri, gulzar, Hasrat jaipuri etc that made us fall in love with hindi, I still cant go to bed without enjoying one or two of them. Today with increasling political agenda to sankitisize the Hindi and discourage use of urdu ....i feel soul is being lost out of this very beautiful language.
Sarkari Hindi is not the hindi spoken by masses..not even in UP and Bihar then how can it ever by imposed on entire country as national language. India is such a beautiful diverse country any attmept to force, language, religion is neither practical nor ethical. Let all languages,culture, religions flourish and peacefully coexist as they have done for last 5000 years in Bharat. There is absolutely no need to rake these non issue.
Regards,
rajan
==
Dr.Rajan R Patil
Epidemiologist
School of Public Health
SRM University,
Chennai.
Cell :9445811610
Monday, July 28, 2014
Saturday, June 15, 2013
THE NEED FOR NEW BORN SCREENING IN INDIA
By : Dr. Rajan R Patil.
Epidemiologist. School of Public Health,SRM University, Chennai.
Genetics is essentially public
health than than therapeutic.The
strength of genetics lies in
prevention of disease and not its cure. It
therefore needs to be taken up as public health measure for prevention of physical and mental
handicap The uniqueness of genetic disorders is variety and rarity. In medicine
roficiency comes by numbers, whereas most individual genetic disorders are rare
having the prevalence of 1:5000-15000 births, or even more.
INDIAN CONTEXT
Out of every 100 babies born in
this country annually, 6 to 7 have a birth defect. In Indian context, this
would translate to 1.7 million birth defects annually and would account for 9.6
per cent of all newborn deaths.With a large birth cohort of almost 26 million
per year, India would account for the largest share of birth defects in the
world. This would translate to an estimated 1.7 million babies born with birth efects
annually. In the study conducted by National Neonatology Forum, congenital malformations were the second commonest cause
(9.9%) of mortality among stillbirths and the fourth commonest cause (9.6%) of neonatal mortality and that
accounted for 4 per cent of under- ve mortality.
Birth defects account for 9.6%
of all new-born deaths and 4% of under five mortality. According to March of
Dimes, 2006, out of every 100 babies born in this country annually, 6 to 7 have
a birth defect. In Indian context, this would translate to 17 lakhs birth
defects annually. Development delays affect at least 10% children and these
delays if not intercepted timely may lead to permanent disabilities including
cognitive, hearing or vision.
GLOBAL CONTEXT
Globally, about 7.9 million
children are born annually with a serious birth defect of genetic or partially
genetic origin which accounts for 6 percent of the total births . At least 3.3
million children under ve years of age die from birth defects every year and
another 3.2 million of those who survive may be isabled for life. Cutting
across countries and their economic
status, 64.3 infants per thousand live births are born annually with birth
defects
Source :
1.
Aggraval SS. Medical Genetics in India – What
needs to be done? Indian J Med Res 130, October 2009, pp 354-356
2. MoHFW.
GOI. Child Helath Screening and Early Intervention
Services under NRHM.2013n
Services under NRHM.2013
Tuesday, July 26, 2011
RESEARCH IN UNIVERSITIES & INDUSTRY SUPPORT :SYMBIOSIS OR CONFLICT OF INTEREST?
RESEARCH IN UNIVERSITIES & INDUSTRY SUPPORT :SYMBIOSIS OR CONFLICT OF INTEREST?
by -Dr. Rajan R Patil
Higher education comes at cost. Especially the quality higher education. If we don’t give it, then our own ministers complain that our students are brilliant but our faculty are not world class. Conveniently forgetting they have failed to give us the world class infrastructure to give the world class research out put. Lets say, If I were to investigate mysterious viral disease outbreak, I need to be supported by state of the art world virology laboratory back up. Govt simply doesn’t have the money (or the will) to support for this kind of infrastructure for the universities in our country. Then what do we do? Alternatively, either raise the education fees, start charging nauseatingly high capitation fees is one option – but we can not do it, because self appointed civil society will criticize that we are making higher education unaffordable and preach us on virtues of rights of education of every child and equity issues in society. Without good infrastructure back up, there is no motivation for quality faculty to stay back in India, they will migrate to foreign universities, then the same civil society will hit back on us for being we are lured by glitter of “foreign” and brain drain. They misunderstand it to be monetary reasons when it is not. Remember, in higher education salary is never the highest incentive- but it is Research, which is the ultimate product of scholarly activity. Within the universities, a faculty member’s published record is used to guide a most of evaluation-based decisions in science.
In the light of above, how does one expect Academia to not to engage Industry. We can not wait until our Marxist friends bring in total revolution and total overhauling of economic order. At least if they had demonstrated that it was practical and possible in their countries and statess which they ruled for decades , we would still have waited in hope. But it never happened that way in earstwhile soviet union, China, Cubans, Westbengal, Kerala. The universities functioning in leftist and socialist mileu are no better than others in non-leftist system. In fact all we get from China model is how to fake the technology not the innovative research.
what do we do under the given circumstance, hence we go ahead and engage industry in symbiotic relationship. And do take utmost care that the conflict of Interest does not come in play. By and large Teachers do have higher moral and ethical standards in society, so let us have faith in them. Yes there are few exception, like the infamous professor and his stem cell research. But then there are rotten apples in every strata of society.
Morale of the story:
1. Conflict of interest stems from selfish motive/ intent that is converted into action for personal gain and not by simple assoication. As long as one is being fair and upright in his or her coduct-, conflict of interest does not come into play- in spite of potentially having conflict of interest a-priori.
2. Academicians do have ability to handle conflict of Interest. Lets start trusting them. They do know where to draw line, where and when to keep their academic freedom intact. IITs are fine example of exercise of this discretion by researchers while very well engaging industry.
regards,
rajan
==
Dr. Rajan R Patil
Asst Professor.
PhD scholar in Respiratory Health.
Division of Epidemiology
School of Public Health
SRM University
Potheri, Kattankulathur - 603203
Greater Chennai,
INDIA.
Cell: 9445811610
Email : rajanpatil@yahoo.com
by -Dr. Rajan R Patil
Higher education comes at cost. Especially the quality higher education. If we don’t give it, then our own ministers complain that our students are brilliant but our faculty are not world class. Conveniently forgetting they have failed to give us the world class infrastructure to give the world class research out put. Lets say, If I were to investigate mysterious viral disease outbreak, I need to be supported by state of the art world virology laboratory back up. Govt simply doesn’t have the money (or the will) to support for this kind of infrastructure for the universities in our country. Then what do we do? Alternatively, either raise the education fees, start charging nauseatingly high capitation fees is one option – but we can not do it, because self appointed civil society will criticize that we are making higher education unaffordable and preach us on virtues of rights of education of every child and equity issues in society. Without good infrastructure back up, there is no motivation for quality faculty to stay back in India, they will migrate to foreign universities, then the same civil society will hit back on us for being we are lured by glitter of “foreign” and brain drain. They misunderstand it to be monetary reasons when it is not. Remember, in higher education salary is never the highest incentive- but it is Research, which is the ultimate product of scholarly activity. Within the universities, a faculty member’s published record is used to guide a most of evaluation-based decisions in science.
In the light of above, how does one expect Academia to not to engage Industry. We can not wait until our Marxist friends bring in total revolution and total overhauling of economic order. At least if they had demonstrated that it was practical and possible in their countries and statess which they ruled for decades , we would still have waited in hope. But it never happened that way in earstwhile soviet union, China, Cubans, Westbengal, Kerala. The universities functioning in leftist and socialist mileu are no better than others in non-leftist system. In fact all we get from China model is how to fake the technology not the innovative research.
what do we do under the given circumstance, hence we go ahead and engage industry in symbiotic relationship. And do take utmost care that the conflict of Interest does not come in play. By and large Teachers do have higher moral and ethical standards in society, so let us have faith in them. Yes there are few exception, like the infamous professor and his stem cell research. But then there are rotten apples in every strata of society.
Morale of the story:
1. Conflict of interest stems from selfish motive/ intent that is converted into action for personal gain and not by simple assoication. As long as one is being fair and upright in his or her coduct-, conflict of interest does not come into play- in spite of potentially having conflict of interest a-priori.
2. Academicians do have ability to handle conflict of Interest. Lets start trusting them. They do know where to draw line, where and when to keep their academic freedom intact. IITs are fine example of exercise of this discretion by researchers while very well engaging industry.
regards,
rajan
==
Dr. Rajan R Patil
Asst Professor.
PhD scholar in Respiratory Health.
Division of Epidemiology
School of Public Health
SRM University
Potheri, Kattankulathur - 603203
Greater Chennai,
INDIA.
Cell: 9445811610
Email : rajanpatil@yahoo.com
Thursday, July 21, 2011
PHFI- CORPORATE LINK IS NOT CONFLICT OF INTEREST
PHFI- CORPORATE LINK IS NOT CONFLICT OF INTEREST
-Dr. Rajan R Patil
There is criticism from several corners that reputed people who are joining PHFI as board members are lending their reputation for wrong body since PHFI is overloaded with corporate heads.
1) We naeed to engage PHFI, or else the entire space will be taken over by
corporate CEOs- hence we should appreciate that people like Ravi Narayan,
mashelkar, narayan murthy should proactively be part of PHFI.
2) There is nothing wrong in Naryan Murthy joining PHFI because apart from
adding cerebral aura people like him will hep in utilizing the funds in fair and
just manner
I still insist that we need to stop being judgmental about people choosing to
be part of PHFI and certainly not personally attack or question their integrity
or their ethical standards. This was happening repeatedly in these e-groups,
thats how i protested against this uncalled for holier than thou attitude.
Just by not joining PHFI one does not automatically take on the ethical or moral
mantle. Lets facet it, There are unethical people in NGOs and there are also
highly ethical people in private sector. Lets stop brushing everybody in sector
other than ours with same brush.
Different people would like to bring in change in different ways, so these
different approaches need to be respected. Some may be comofortable in trying
to change by criticizing from outside, some may chose to join in and change
the same from inside which calls for greater courage of conviction and
strength of character.
Thats the reason i respect the doers over preachers. Strategy adopted by
people like Ravi Narayan who has been unfairly denigraded in these egroups Â
is in fact more commendable in comparison arm chair moralist preachers, and
thats the reason i continue to hold him in high regard. There is no need to
look down upon like him and other who have already proven their credibility like
Narayan Murthy or Amarthya Sen all of whom are not bunnies. They have achieved
great things in life because they have ALWAYS followed their conviction and not
chartered the their life path on the advice of others. If they are joining
PHFI following their conviction, they are doing with full conscience and there
is absoultely no need to advice them what to and what not to do. They exactly
know what they are doing especially at this stage of life- they are not mercy of
unsolicited advice. Lets have faith in them that the values for which they have
stood all their life,they will continued to be guided by the same value system even when they become part of PHFI.
The same holds true for the people who join UN agencies, world bank or WHO,
corporate bodies- lets not take it for granted that they have dropped their
moral bar (that would povlovian reflex) just by mere joining these institutions
one does not become corrupt-Â one can still keep up to their ethical standards
(unless otherwise proven) all these institutions certainly need people with
values, so let us encourage them as well.
morale of the story:
Conflict of interest stems from selfish motive/ intent that is converted into
action for personal gain and not by simple assoication. As long as one is being fair and upright in his or her coduct-, conflict of interest does not come into play- in spite of potentially having conflict of interest a-priori.
regards,
Dr. Rajan R Patil
Asst Professor
Division of Epidemiology
School of Public Health
SRM University
Chennai, India
9445811610
rajanpatil@yahoo.com
-Dr. Rajan R Patil
There is criticism from several corners that reputed people who are joining PHFI as board members are lending their reputation for wrong body since PHFI is overloaded with corporate heads.
1) We naeed to engage PHFI, or else the entire space will be taken over by
corporate CEOs- hence we should appreciate that people like Ravi Narayan,
mashelkar, narayan murthy should proactively be part of PHFI.
2) There is nothing wrong in Naryan Murthy joining PHFI because apart from
adding cerebral aura people like him will hep in utilizing the funds in fair and
just manner
I still insist that we need to stop being judgmental about people choosing to
be part of PHFI and certainly not personally attack or question their integrity
or their ethical standards. This was happening repeatedly in these e-groups,
thats how i protested against this uncalled for holier than thou attitude.
Just by not joining PHFI one does not automatically take on the ethical or moral
mantle. Lets facet it, There are unethical people in NGOs and there are also
highly ethical people in private sector. Lets stop brushing everybody in sector
other than ours with same brush.
Different people would like to bring in change in different ways, so these
different approaches need to be respected. Some may be comofortable in trying
to change by criticizing from outside, some may chose to join in and change
the same from inside which calls for greater courage of conviction and
strength of character.
Thats the reason i respect the doers over preachers. Strategy adopted by
people like Ravi Narayan who has been unfairly denigraded in these egroups Â
is in fact more commendable in comparison arm chair moralist preachers, and
thats the reason i continue to hold him in high regard. There is no need to
look down upon like him and other who have already proven their credibility like
Narayan Murthy or Amarthya Sen all of whom are not bunnies. They have achieved
great things in life because they have ALWAYS followed their conviction and not
chartered the their life path on the advice of others. If they are joining
PHFI following their conviction, they are doing with full conscience and there
is absoultely no need to advice them what to and what not to do. They exactly
know what they are doing especially at this stage of life- they are not mercy of
unsolicited advice. Lets have faith in them that the values for which they have
stood all their life,they will continued to be guided by the same value system even when they become part of PHFI.
The same holds true for the people who join UN agencies, world bank or WHO,
corporate bodies- lets not take it for granted that they have dropped their
moral bar (that would povlovian reflex) just by mere joining these institutions
one does not become corrupt-Â one can still keep up to their ethical standards
(unless otherwise proven) all these institutions certainly need people with
values, so let us encourage them as well.
morale of the story:
Conflict of interest stems from selfish motive/ intent that is converted into
action for personal gain and not by simple assoication. As long as one is being fair and upright in his or her coduct-, conflict of interest does not come into play- in spite of potentially having conflict of interest a-priori.
regards,
Dr. Rajan R Patil
Asst Professor
Division of Epidemiology
School of Public Health
SRM University
Chennai, India
9445811610
rajanpatil@yahoo.com
Tuesday, July 19, 2011
Academia-Profit making Corporate allignment- is NOT Conflict of Interest
Academia-Profit making Corporate allignment- is NOT Conflict of Interest
-by Dr.Rajan R Patil
There is criticism Dr. R. A. Mashelkar who in spite of being openly aligned with profit-making corporates were also in policy-influencing positions.
which is showcased as conflict of interest.
No we cant blame Mashelkar for Being aligned with profit making corporate. He is very much in align with larger policy that encourages academia and research institutions to court industry and industry courts academia because they are in syboitic relationship. That is how it should be, Industry benefits from R & D of Academia and in turn universities/research Inistitutions get funds- extr mural funding it is called, which even in ICMR and CSIR it is mandatory. In fact it is one of the yardstick to major the credibility of research who can generate funds outside this Institution and most importantly it create jobs for (placements)the students that graduate out of them. University/Research institutions credibility also depends upon the market up take of their trained students.
All progressive Universities/Research institutions have now separate office called corporate Relation office. We too have one. The puprpose is to promote Academia-Industry collaboration. The idea is to keep abreast with latest, our curriculum and syllabus can not be out of tune with market needs. The other day I personally visited fews CROs to get a feel for potential demand for Students with clincial research specialization as we are launching new Post graduate course MS Clinical Trials. It was a eye opener for me as the the existing Curriculam and syllabus is so out of sync. The CROs rued that existing graduate who come to them armed with Clinical Research degree are of no use to them as they simply dont come armed with basic and practical skills all that they have theoretical knowledge and that too outdated curriculum.Hence we took two decisions 1) have six months internship where students will be placed in Industry (CRO)so they get armed with some practical basic skill 2) To invite Industry people to take guest lecture to our stundets so that there is greater synergy between theory and practice.
Similarly for MPH degree we send students for six month intership with GO/NGO organisation all over India, for Hospital Adminsitration degree http://www.princetonreview.com/Careers.aspx?cid=203
we send students Corporate hospitals that is where they will ultimately land up, and for Health care management http://www.mastersinhealthcare.com/ MS Clinical Trial the students are placed in CROs. Remember while GO/NGO may be not for profit, but CRO and Corporate hospitals are ruthlessly for profit. If we teaching students to take charge of these sectors we have to engage this sectors and we do engage these sector proactively in the persuit of our academic interest of our students- in process if i am seen hobnobbing with these for profit making corporates, I can be open to all kind of interpretations and criticism- which is purely stems out of ignorance. We know what we are doing and why we are doing there is no conflict of interest in our intent.
I suppose even Mashelkar too will march forward with same attitude when criticized unfairly for being aligned with profit making corporates.
regards,
rajan
==
Dr. Rajan R Patil
Asst Professor.
PhD scholar in Respiratory Health.
Division of Epidemiology
School of Public Health
SRM University
Potheri, Kattankulathur - 603203
Greater Chennai,
INDIA.
Cell: 9445811610
Email : rajanpatil@yahoo.com
-by Dr.Rajan R Patil
There is criticism Dr. R. A. Mashelkar who in spite of being openly aligned with profit-making corporates were also in policy-influencing positions.
which is showcased as conflict of interest.
No we cant blame Mashelkar for Being aligned with profit making corporate. He is very much in align with larger policy that encourages academia and research institutions to court industry and industry courts academia because they are in syboitic relationship. That is how it should be, Industry benefits from R & D of Academia and in turn universities/research Inistitutions get funds- extr mural funding it is called, which even in ICMR and CSIR it is mandatory. In fact it is one of the yardstick to major the credibility of research who can generate funds outside this Institution and most importantly it create jobs for (placements)the students that graduate out of them. University/Research institutions credibility also depends upon the market up take of their trained students.
All progressive Universities/Research institutions have now separate office called corporate Relation office. We too have one. The puprpose is to promote Academia-Industry collaboration. The idea is to keep abreast with latest, our curriculum and syllabus can not be out of tune with market needs. The other day I personally visited fews CROs to get a feel for potential demand for Students with clincial research specialization as we are launching new Post graduate course MS Clinical Trials. It was a eye opener for me as the the existing Curriculam and syllabus is so out of sync. The CROs rued that existing graduate who come to them armed with Clinical Research degree are of no use to them as they simply dont come armed with basic and practical skills all that they have theoretical knowledge and that too outdated curriculum.Hence we took two decisions 1) have six months internship where students will be placed in Industry (CRO)so they get armed with some practical basic skill 2) To invite Industry people to take guest lecture to our stundets so that there is greater synergy between theory and practice.
Similarly for MPH degree we send students for six month intership with GO/NGO organisation all over India, for Hospital Adminsitration degree http://www.princetonreview.com/Careers.aspx?cid=203
we send students Corporate hospitals that is where they will ultimately land up, and for Health care management http://www.mastersinhealthcare.com/ MS Clinical Trial the students are placed in CROs. Remember while GO/NGO may be not for profit, but CRO and Corporate hospitals are ruthlessly for profit. If we teaching students to take charge of these sectors we have to engage this sectors and we do engage these sector proactively in the persuit of our academic interest of our students- in process if i am seen hobnobbing with these for profit making corporates, I can be open to all kind of interpretations and criticism- which is purely stems out of ignorance. We know what we are doing and why we are doing there is no conflict of interest in our intent.
I suppose even Mashelkar too will march forward with same attitude when criticized unfairly for being aligned with profit making corporates.
regards,
rajan
==
Dr. Rajan R Patil
Asst Professor.
PhD scholar in Respiratory Health.
Division of Epidemiology
School of Public Health
SRM University
Potheri, Kattankulathur - 603203
Greater Chennai,
INDIA.
Cell: 9445811610
Email : rajanpatil@yahoo.com
PHFI- Co-opted by Industry CEOs? ..NOT REALLY!!
PHFI- Co-opted by Industry CEOs? ..NOT REALLY!!
by - Dr. Rajan R Patil
They say, whey an 'event' is inevitable, one has only two choices : (1)either
make futile attempt to fight it out (2) or simply enjoy it and let it happen as
it is anyway inevitable.
However there is third course of action which most of us are beginning to
endorse- ENGAGE THE EVENT PROACTIVELY!! and not simply wish it way. Because the
latter is the wishful thinking, and by engaging proactively there is a
probability that one may change the course event-positively!!
Guys, PHFI is here to stay- lots of money, energy has flown in it and in many
ways it has been "institutionalized" in the sense that it is not linked with
fortunes of UPA govt any more- . So it is inevitable.
One can typically respond in two ways to this inevitable 'event"
1) crib and crib about it-
write- make dizzy long pages after pages ideological monologue on how it is
conspiracy of capitalist take over conspiracy.
2) second join the party anyways- sab loot rahein hai...so be wise khud bhi
behati ganga me haath dholo....
or exercise the THIRD OPTION- CONSTRUCTIVE ENGAGEMENT
....What ever is happening dont just let it happen- by avoiding it or joining
the party status quo will remain same. But by engaging it there is outside
chance of at least influencing it positively- in just and fair way.
Think again the 30 board members listed below- all would have looked same if not
for the reassuring names of Ravi Narayan, Amarthya Sen, Mashelkar and now
Narayan Murthy joining at least some of us are celebrating it- because they are
slowly beginning to occupying the space- So there is now a Hope left of
brining in positive change..... which otherwise if these wonderful souls had not
occupied, the same space
would have been occupied by another set of CEOs whose credentials or the intent
would have been a suspect.
As Dhruv rightly pointed out, Govt is corrupt yet most us choose to engage it -
we choose to be in their committees so that our voice is represented and we
influence its activities positively.
And.... I find it ridiculous when i see chest beating by some of our respected
colleagues regarding PHFI linkage Industry or Industrialists...by the
way....PHFI is not the first academic body that is committing this "sin"Â of
joining hands with industrialsist.....take any academic bodies that we swear of
its reputation- where we dream of sending our children for education-
1.John Hopkins university
2. Harvard university
3 IITs and IIMs in India.
All of these have industry linkage and yet they have reputation in providing
quality education and quality research. Quality education/research comes at
cost- it requires money. Govt does not have enough money or will to provide it.
All universities are encouraged to generate funds to support their research
activies so they have to depend on industry and industrialists to provide.
Industry-Academic feed into each other and benefit out of each other. That
explains why most of academic bodies have corporate CEO as board members. Lets
stop pervert interpretations and desist questioning integrity and intent of
intellects who are choosing consciously to join these bodies fully aware that
their peers will criticise it- yet they have the guts to follow their conviction
without being politically right- so they need to be appreciated for it.
Morale of the story
So wake up friends, by closing eyes reality does not change. We need corporate
heads to ensure uninterrupted flow of funds and we definitely need Ravi Naryans,
Mahselkars, Amarthya sens, Naryan murthys to bring cerebral aura and
credibility to academic bodies and two to ensure that the funds that brought in
are used purposefully and utilized in a fair and just manner.
regards,
rajan
==
Dr. Rajan R Patil
Asst Professor.
PhD scholar in Respiratory Health.
Division of Epidemiology
School of Public Health
SRM University
Potheri, Kattankulathur - 603203
Greater Chennai,
INDIA.
Cell:Â 9445811610
Email : rajanpatil@...
=================================================================================
> Members of the Governing Board of PHFI
>
> * Dr. Montek Singh Ahluwalia
> Deputy Chairman, Planning Commission, Government of India
>
> * Mr. Ashok Alexander
> Director, Avahan - India AIDS Initiative, Bill and Melinda Gates Foundation
>
> * Mr. Mukesh D. Ambani
> Chairman and Managing Director, Reliance Industries Limited
>
> * Mr. K. Chandramouli
> Secretary, Ministry of Health and Family Welfare, Government of India
>
> * Ms. Mirai Chatterjee
> Co-ordinator for Social Security, Self Employed Women's Association (SEWA)
>
> * Dr. Purnendu Chatterjee
> Chairman, The Chatterjee Group
>
> * Dr. Lincoln Chen
> Director of Global Equity Center,
Harvard’s Kennedy School
>
> * Dr. James W. Curran
> Former Chair - Association of Schools of Public Health (ASPH), USA;
> Dean, The Rollins School of Public Health, Emory University, USA
>
> * Dr. Timothy G. Evans
> Dean, James P Grant School of Public Health, BRAC University, Bangladesh
>
> * Mr. Vishwa Mohan Katoch
> Secretary, Department of Health Research, Ministry of Health & Family Welfare
> Director General, Indian Council of Medical Research
>
> * Mr. Uday Nabha Khemka
> Vice Chairman, Sun Group
>
> * Mr. Gautam Kumra
> Director, McKinsey & Company
>
> * Dr. David Lynn
> Head of Strategic Planning & Policy, The Wellcome Trust, UK
>
> * Ms. Kiran Malhotra
> Chairperson, AKM Systems Pvt Ltd
>
> * Dr. Raghunath A. Mashelkar
> CSIR Bhatnagar Fellow, National Chemical
Laboratory
>
> * Mr. Raj Mitta
> Chairman, Essential Value Associates Pvt Ltd.
>
> * Mr. Shiv Nadar
> Founder, HCL
>
> * Mr. T K A Nair
> Principal Secretary to the Honorable Prime Minister of India
>
> * Dr. Ravi Narayan
> Community Health Adviser, Society for Community Health Awareness, Research &
Action (SOCHARA)
>
> * Ms. Rohini Nilekani
> Chairperson, Akshara Foundation, Arghyam Trust, Pratham Books
>
> * Ms. K. Sujatha Rao
> Former Secretary, Ministry of Health & Family Welfare, Government of India
>
> * Mr. J V R Prasada Rao
> Special Advisor to Executive Director, UNAIDS India
>
> * Prof. K Srinath Reddy
> President, Public Health Foundation of India,
> Former HOD, Cardiology, All India Institute of Medical Sciences (AIIMS)
>
> * Dr. Y. Venugopal Reddy
> Former
Governor, Reserve Bank of India
>
> * Dr. Anil Seal
> Director, Cambridge Commonwealth Trust and Cambridge Overseas Trust
>
> * Dr. Amartya Sen
> Professor of Economics and Philosophy, Harvard University
>
> * Dr. Jaime Sepulveda
> Bill & Melinda Gates Foundation
>
> * Dr. A K Shivakumar
> Advisor, UNICEF and Member of National Advisory Council
>
> * Mr Harpal Singh
> Chairman, Ranbaxy Laboratories
>
> * Dr R K Srivastava
> Director General Health Services, Ministry of Health and Family Welfare,
> Government of India
>
> [Non-text portions of this message have been removed]
PHFI- Copted by Capitalist CEOs? NOT REALLY!!
Dr. Rajan R Patil
Asst Professor.
PhD scholar in Respiratory Health.
Division of Epidemiology
School of Public Health
SRM University
Potheri, Kattankulathur - 603203
Greater Chennai,
INDIA.
Cell:Â 9445811610
Email : rajanpatil@...
by - Dr. Rajan R Patil
They say, whey an 'event' is inevitable, one has only two choices : (1)either
make futile attempt to fight it out (2) or simply enjoy it and let it happen as
it is anyway inevitable.
However there is third course of action which most of us are beginning to
endorse- ENGAGE THE EVENT PROACTIVELY!! and not simply wish it way. Because the
latter is the wishful thinking, and by engaging proactively there is a
probability that one may change the course event-positively!!
Guys, PHFI is here to stay- lots of money, energy has flown in it and in many
ways it has been "institutionalized" in the sense that it is not linked with
fortunes of UPA govt any more- . So it is inevitable.
One can typically respond in two ways to this inevitable 'event"
1) crib and crib about it-
write- make dizzy long pages after pages ideological monologue on how it is
conspiracy of capitalist take over conspiracy.
2) second join the party anyways- sab loot rahein hai...so be wise khud bhi
behati ganga me haath dholo....
or exercise the THIRD OPTION- CONSTRUCTIVE ENGAGEMENT
....What ever is happening dont just let it happen- by avoiding it or joining
the party status quo will remain same. But by engaging it there is outside
chance of at least influencing it positively- in just and fair way.
Think again the 30 board members listed below- all would have looked same if not
for the reassuring names of Ravi Narayan, Amarthya Sen, Mashelkar and now
Narayan Murthy joining at least some of us are celebrating it- because they are
slowly beginning to occupying the space- So there is now a Hope left of
brining in positive change..... which otherwise if these wonderful souls had not
occupied, the same space
would have been occupied by another set of CEOs whose credentials or the intent
would have been a suspect.
As Dhruv rightly pointed out, Govt is corrupt yet most us choose to engage it -
we choose to be in their committees so that our voice is represented and we
influence its activities positively.
And.... I find it ridiculous when i see chest beating by some of our respected
colleagues regarding PHFI linkage Industry or Industrialists...by the
way....PHFI is not the first academic body that is committing this "sin"Â of
joining hands with industrialsist.....take any academic bodies that we swear of
its reputation- where we dream of sending our children for education-
1.John Hopkins university
2. Harvard university
3 IITs and IIMs in India.
All of these have industry linkage and yet they have reputation in providing
quality education and quality research. Quality education/research comes at
cost- it requires money. Govt does not have enough money or will to provide it.
All universities are encouraged to generate funds to support their research
activies so they have to depend on industry and industrialists to provide.
Industry-Academic feed into each other and benefit out of each other. That
explains why most of academic bodies have corporate CEO as board members. Lets
stop pervert interpretations and desist questioning integrity and intent of
intellects who are choosing consciously to join these bodies fully aware that
their peers will criticise it- yet they have the guts to follow their conviction
without being politically right- so they need to be appreciated for it.
Morale of the story
So wake up friends, by closing eyes reality does not change. We need corporate
heads to ensure uninterrupted flow of funds and we definitely need Ravi Naryans,
Mahselkars, Amarthya sens, Naryan murthys to bring cerebral aura and
credibility to academic bodies and two to ensure that the funds that brought in
are used purposefully and utilized in a fair and just manner.
regards,
rajan
==
Dr. Rajan R Patil
Asst Professor.
PhD scholar in Respiratory Health.
Division of Epidemiology
School of Public Health
SRM University
Potheri, Kattankulathur - 603203
Greater Chennai,
INDIA.
Cell:Â 9445811610
Email : rajanpatil@...
=================================================================================
> Members of the Governing Board of PHFI
>
> * Dr. Montek Singh Ahluwalia
> Deputy Chairman, Planning Commission, Government of India
>
> * Mr. Ashok Alexander
> Director, Avahan - India AIDS Initiative, Bill and Melinda Gates Foundation
>
> * Mr. Mukesh D. Ambani
> Chairman and Managing Director, Reliance Industries Limited
>
> * Mr. K. Chandramouli
> Secretary, Ministry of Health and Family Welfare, Government of India
>
> * Ms. Mirai Chatterjee
> Co-ordinator for Social Security, Self Employed Women's Association (SEWA)
>
> * Dr. Purnendu Chatterjee
> Chairman, The Chatterjee Group
>
> * Dr. Lincoln Chen
> Director of Global Equity Center,
Harvard’s Kennedy School
>
> * Dr. James W. Curran
> Former Chair - Association of Schools of Public Health (ASPH), USA;
> Dean, The Rollins School of Public Health, Emory University, USA
>
> * Dr. Timothy G. Evans
> Dean, James P Grant School of Public Health, BRAC University, Bangladesh
>
> * Mr. Vishwa Mohan Katoch
> Secretary, Department of Health Research, Ministry of Health & Family Welfare
> Director General, Indian Council of Medical Research
>
> * Mr. Uday Nabha Khemka
> Vice Chairman, Sun Group
>
> * Mr. Gautam Kumra
> Director, McKinsey & Company
>
> * Dr. David Lynn
> Head of Strategic Planning & Policy, The Wellcome Trust, UK
>
> * Ms. Kiran Malhotra
> Chairperson, AKM Systems Pvt Ltd
>
> * Dr. Raghunath A. Mashelkar
> CSIR Bhatnagar Fellow, National Chemical
Laboratory
>
> * Mr. Raj Mitta
> Chairman, Essential Value Associates Pvt Ltd.
>
> * Mr. Shiv Nadar
> Founder, HCL
>
> * Mr. T K A Nair
> Principal Secretary to the Honorable Prime Minister of India
>
> * Dr. Ravi Narayan
> Community Health Adviser, Society for Community Health Awareness, Research &
Action (SOCHARA)
>
> * Ms. Rohini Nilekani
> Chairperson, Akshara Foundation, Arghyam Trust, Pratham Books
>
> * Ms. K. Sujatha Rao
> Former Secretary, Ministry of Health & Family Welfare, Government of India
>
> * Mr. J V R Prasada Rao
> Special Advisor to Executive Director, UNAIDS India
>
> * Prof. K Srinath Reddy
> President, Public Health Foundation of India,
> Former HOD, Cardiology, All India Institute of Medical Sciences (AIIMS)
>
> * Dr. Y. Venugopal Reddy
> Former
Governor, Reserve Bank of India
>
> * Dr. Anil Seal
> Director, Cambridge Commonwealth Trust and Cambridge Overseas Trust
>
> * Dr. Amartya Sen
> Professor of Economics and Philosophy, Harvard University
>
> * Dr. Jaime Sepulveda
> Bill & Melinda Gates Foundation
>
> * Dr. A K Shivakumar
> Advisor, UNICEF and Member of National Advisory Council
>
> * Mr Harpal Singh
> Chairman, Ranbaxy Laboratories
>
> * Dr R K Srivastava
> Director General Health Services, Ministry of Health and Family Welfare,
> Government of India
>
> [Non-text portions of this message have been removed]
PHFI- Copted by Capitalist CEOs? NOT REALLY!!
Dr. Rajan R Patil
Asst Professor.
PhD scholar in Respiratory Health.
Division of Epidemiology
School of Public Health
SRM University
Potheri, Kattankulathur - 603203
Greater Chennai,
INDIA.
Cell:Â 9445811610
Email : rajanpatil@...
Saturday, June 4, 2011
CRUDE BIRTH RATE (CBR) & TOTAL FERTILITY RATE (TFR): CORRELATED? - -by DR. RAJAN PATIL
CRUDE BIRTH RATE (CBR) & TOTAL FERTILITY RATE (TFR): CORRELATED?
-by DR. RAJAN PATIL
This is complilation of n intense but interesting debate was initiated in a public health e-group on a question on relationship between CBR and TFR- on the plausibility of TFR being on increasing trend while CBR was declining. The debate evoked plethora of responses arguing for and against the topic. Since the issue under discussion was a demographic topic, three expert demographers were consulted and their views enlisted.
Question Posted: The TFR of Nagaland according to NFHS 3 is 3.7 in comparison to national average of 2.7 At the same time birth rate of the state is 17.5 in comparison to national average of 22.8 as per SRS October 2009.
I am unable to understand the plausibility of these two figures. I will be grateful if some body can throw light on this. Other features of the state are that median age at first birth is 21.8 in comparison to 19.8 nationally and sex ratio also favours women.
Response :Surely TFR takes into account the rates of the last 30 years while current rates are just point figures.Of course different sources or their inherent biases might also influence things(NFHS is trying to prove something, whereas SRS is a more neutral and very efficient way of collecting information)
Response from a professional Demographer
TFR is an 'exante' (expected) measure. TFR is the total number of births likely to occur for a woman, who enters reproduction at a particular point of time, which depends on the median age at marriage/first
birth, current level of practice of family planning and prevalent natural wastage (abortions) rate. Birth rate or crude birth rate is an 'exposed'
measure calculated as the total number of births in a year to 1000 mid-year population in that year.  The latter measure depends to a large extent on the number of women who have given birth to in that year. The size of the population of women who are currently in reproductive age depends on earlier
fertility level and sex ratio of the population at reproductive age group. For this reason, the two measures could exhibit differing patterns in relation to national average.
With best: Jayaraj.
Respondent: even then (and that being understood), average age of marriage is higher than national avg and unmet need for contraception is lower. sex ratios are higher for women than national avg. unless there are a huge number of abortions, this discrepency should not occur. it is worth questioning.
Respondent:I think that the best explanation would be the possibility of the figures being fictitious as SRS people - by their own honest admission in private at Guwahati - hardly leave their offices to make new data and NFHS partners in NE have been less than sincere in collecting data.
Respondent :Another point is that if you have children more times (ie have the last child much later on) the annual rate will be low but the TFR will remain high. This would be even more possible if Nagaland has low MMR/ low number of women's deaths and India has a high death rate - especially affecting later pregnancies
Dr. Rajan R Patil : The question was essentially exploratory in nature- of the inherent logic - on possibility of occurrence of apparent incongruity between  the  two said demographic indicators and if so, was it conceptually plausible the response from our demopher most credible and convincing for the reason of its conceptual clarity and inherent logical base on which it is grounded. Hence like any constructs it is best verified be application of inductive or deductive logic and not by contextualizing to any geographic point. conceptually when any constructs passes validity tests (convergent and discriminant) then generally it is accepted for its coherence.
The fact that Demographic indicators are quantitative in nature and measurable to decimal level, that makes life simple for its cross validity examination-once verified it can be very well be universalized.
Morale of the story
1. Since both question and the answer are on conceptual plane- hence  the geographic setting- be it Nagaland or India and hence the idiosyncrasies associated Indian data are irrelevant here.
2. The strength of any given construct stems from its ability to hold ground in its abstractness independent of its setting- be it Nagaland in India or wonderland in Timbaktu- It will still be true.
Dr. Rajan R Patil : if you insist that your interest here is to seek the answer especially contextualised and geographicalized to Nagaland in particular there is one hurdle- only validity of figures concerning to Nagaland could be answered by cross checking source, authenticity issues of data collection behaviour as already suggested by our other colleagues. BUT YOU CAN NEVER COMPARE it with national averages.
As a cardinal rule of epidemiology( and i hope it is true for any quantitative sciences like Bio-statistics & Demogrpahy included) It is wrong to compare and make inference by comparing two point prevalence figures of two heterogeneous communities (Nagaland and India), because SRS or NFHS give only crude rates.
TWO CRUDE RATES SHOULD NEVER BE COMPARED . crude rates are meant for only description of single group characteristics and not for two population comparison- it will be miserably misleading.
However If you still insist that your basic interest is in camparison of nagaland figures with National averages then we ill have to invoke 2nd cardinal rule of group comparison- to always convert the given crude rates into standardized rates by direct standardization method (or any alternative suggested by demographers). ONLY The resultant ADJUSTED STANDARDIZED RATES should be compared between two hetrogeneos populations and inference drawn.
Comparison of crude rates can be done not for drawing comparative inference but for different purposes- to varify the issues brought out by Sunil Koul- to check authenticiy, honesty of data collection behaviour in North east. To do this it is best to first homogenize the groups i.e, compare Nagaland figures with NE region averages rather than national averages. So that data collection and validity issues as stand alone issues mentioned by sunil as specific to NE can be discounted for in final conclusions- because then it become intragroup croup comparison within NE region and hence Nagaland and NE region can be considered homogeneos, at least relatively speaking to National comparison.
In conclusion you me need get one demographer to convert the given crude figures into standardized figure and only then make attempt to compare with Nagaland vs national comparisons and draw conclusions
Respondent :the TFR is an age standardised artificial cohort rate and lends itself to comparision across time and space. I am a trained demographer (as suggested by the previous writer -Dr.Patil, and this is one of the fundamental lessons of technical demography) and there is very little scope for me being wrong about this. In so far as comparisions of crude rates are concerned, yes indeed one needs to standardise for age as that could cause distortions and therefore any attempt at comparison should be with standardised rates.
Respondent: I understand the limitations of data which you are trying to highlight. It is evident already, but we still use point prevalence figures for comparison and rank communities accordingly. If we do not do this, then, we can never set goals and never aspire that everybody should have highest attainable status of health as characteristics for every society or for that matter every individual is different. Wrt homogenization, it is always arbitrary, as you know that there is heterogeneity/variability in the same village/neighborhood based on inclusions and exclusions. So rates are based on averages built on averages and I am unable to understand why averages should not be compared. This is a different matter that based on comparison you go deeper to find the reasons for variations.
Now the main question which is still unanswered is the coherence between CBR and TFR as both relate to same activity.
Dr. Rajan R Patil :I though your query was already answered long back by our good demographer friend that CBR and TFR can not have any association they are independent of each other.
-Let me try to Demystify the rationale in simple words….
(and dhruv... with this i am responding to your request for simplifying complex concepts in simple words for advocacy groups)
Rates are the game of numerators and Denominator
Your confusion (as i read) is that when numerator for CBR and TFR- is same ( live births) but only denominators are differing yet why NO coherence?
Let me dissect it layer by layer:
lets first take Numerator characteristics
Please understand, Numerator only appears to be same for both BUT it is actually not. Please note the numerator CBR is TOTAL NO. OF ALL LIVE BIRTHS in a region (every single live birth is accounted for) but for TFR only live births selected in selected sample of households (and not every single live birth region)
Now let us look at the denominator side.
While denominator for CBR is all population in given region (its a simple frequency count of every individual in the region) the story is finished here for CBR.
Where as for TFR denominator COMBINATION of inter play of - Total count number (1) AND OTHER charecteristics of population in denominator (2nd, 3rd & 4th point).
1. Total count of individual in denominator ( like denominator for CBR)
2. proportion of women in different age groups from 15yrs to 44years
3. Mean Age at marriage of these woman in denominator.
4. birth order in the community
There fun side to denominator for CBR is that all population means( ALL MEN+WOMEN) in the community Now, please note CBR denominator is in fact inclusive of denominator of TFR -sub group WOMEN in 14-44 yrs age group is also included.
Because propoortion (%) of this 15-44years women in Total population will be different in every state, so also Mean age at marriage of these women and so is birth order. Every state will show different permutation and combination
EVIDENCE FOR INDEPENDENCE BETWEEN CBR and TFR
To help your self to put the whole story in perspective please do this home work
Compare TFR of TN and Kerala. CBR of Kerala is low but has High TFR on the other hand. In TN has High CBR but Low TFR. Thats the evidence for you on why TFR and CBR show no coherence
HOWEVER if you follow golden rule in epidemiology ALWAYS compare LIKES WITH LIKES you will spared of many misgivings. Example – you are most welcome to compare CBR of nagaland with CBR of India (but as I said standardize the rates first then compare), so also you can jolly well compare TFR of nagaland with TFR of India (as mala suggested here standardization is not required because TFR is any way ready made age specific rate and is not crude like CBR).
BUT never make mistake of cross comparison of CBR of one NAGALAND with TFR of India. It is absurd.
Hope that helps.
(And also hope that expert demographers like Mala Ramanathan will find this technically right as well. if not, i am willing to stand corrected with all humility- since demography is not my subject and my efforts here are simplyfying in lay terms- i may have missed out out on finer nuances of technicality)
Respondent: Thanks for elucidating the issues regarding some of demographic indicators. It will be worthwhile to construct an article on Challenges of Technical Demography for Public Health  Advocacy or some such for the bulletin.
Can we make it even simpler some of the fundamental issues, you have raised?- Often figures are flown and floated to gain some edge but are 'bad' from scientific reasons. I think we need to rectify such anomaly. First, those who are using such figures must know that they are 'bad' figures but useful for
advocacy (after politics allows you to some extent if it is favor of the pro-people policies!) I had initiated a column for the bulletin called Armchair (Re)view. Can you guys something on fallacies that often we face?
Respondent: Thanks for very valuable input of so many friends. But at the end of the day, I am still unclear to in interpreting them huge diversity in two indicators which relate to births in the same population
of more or less with the same reference period. I understand that one is 'expected' and the other one is 'exposed' measure and may show differing trends in different cohorts. But then are these indicators incoherent to each other and can they be used for comparison in different population cohorts as is
normally done by us all and by the Govt. of India while ranking different states and deciding financial allocations. My question is simple "Is it possible to have low birth rate with high TFR and vice versa". Is there any normativity?
Another issue foxing me about Nagaland is that decadal growth rate between 1991 and 2001 census. It is whopping 64.53% while all India is 21.54% (Surprisingly, Punjab's growth rate which has huge inmigration is 21.54% only and its birth rate is also 17.3/1000 population). Again the question is with
CBR like 17.5, is it possible to have decadal growth rate of 64.53% in Nagaland where immigration is almost nil (my assumption).
>Demographers and others, please enlighten. I am not ready to completely believe What
Sunil has said. According to many demographers, data generated by the Census, NSSO, NFHS(IIPS)
India is considered quite credible.
Respondent :To add my bit to the knowledge/confusion:I heard this when I was in nagaland..... they find it difficult to reach 100% immunisation - since there just not as many children around as expected from population projections. At the time of the census there was some competitive spirit in many tribes/villages trying to boast of high populations. So census figures are very high- and whenever you use that as the denominator every other numerator would come to grief. This year some one is being very honest in reporting immunisation honestly and figures are dopping to about half- and well soon I expect the figures will go up- perhaps. But all this is rumor - not fact. And like for all rumors it should be disbelieved till otherwise proven. Certainly not to be quoted- hence not posted on the egroups. If one of you wants to take credit for hearing such a rumor and post it I have no objections, as long as I am not the source of the information.
Respondent: But then this is a common story everywhere from the service providers. I recall, several years back I went in a primary school of a very remote village in Bandikui block of Dausa Dt. in Rajasthan. There were 7 girls and 41 boys enrolled in the school and school authorities claimed 100% enrollment. I could not believe them that how can there be only 7 girls against 41 boys of the same age group. Truth came out on house visits. ANMs and other health workers use these pretexts very commonly to show 100% coverage. But can we believe it? Other side is that now we have to look towards RGI, NSSO, IIPS with caution? Leave aside manual checks, didn't they do statistical checks before all this data is published? It is becoming like HMIS collected by the health department.
Respondent :You dont get the point- I am saying that the population figure and growth rate is inflated. Take your example and let me reconstruct it. There are only 7 boys and 7 girls in the school and they have claimed there are 41 boys so that they get a higher mid day meal allocation!! That was the allegation. And hey dont laugh at the HMIS ofthe health dept. You have to know to read it. Remember Poirot in one of Agatha Christies books. A lady asks him- what is your method. Poirot replies- well I just talk to people and they just tell me. The lady is amazed. How silly she says-- "they could easily lie to you." And Poirot replies- that is what is the most instructive part!!! The errors in HMIS dear sir as instructive as the numbers in it. Do join us on the HMIS workshop - March 23 to 27th or so.
Respondent :This is a general point abt most govt data collection. How do you estimate the amount of exaggeration? School attendance is exaggerated because if the attendance is low, the school will get cancelled in the next year's budget.
Dr.Rajan R Patil : Am responding to your specific query over Data quality issues in our national programmes -detecting, estimation  and ways to correct them. would like to emphasize that enough work has been done on this issue and solutions also has been already proposed.
In a nutshell  a data validity check can be carried out on a sample of existing data by both qualitative and quantitative methodology. Correction factor worked out and applied. Please note there can not be One Generalized correctionfactor applicable to all – since data quality issues varying degrees of data quality issues- so it has to be worked out specifically to each state and be applied to there only.
Am saying this with some convicition since we were concerned and have done some ground
work in investigating of data quality issues especially when I was with  UNDP-WHO working on setting up disease surveillances systems in Bihar and Orissa and elswhere in country and have
published our findings too. Please ref to our work in following journals
1) Deepa TM, Venkata Rao, Rajan R Patil,Reuben Samuel. Community Reporting Community reporting model in Disease Surveillance - National Medical Jounal of India. Vol 21, No4, 2008.Â
2) Venkatarao E, Rajan R Patil, Anita Anasuya, Deepa TM, Reuben Samuel. Monitoring data quality in syndromic surveillance learnings from resource limited setting. (under review)
Also I had investigated data collection and quality issues in cotton farmers suicide in Warangal dist in Andhra pradesh. Findings available at :
3. Rajan Patil. Suicide deaths among farmers. British Medical Journal (S.A ed ) June 2002;18(4):310
Morale of the story
No need to re-invent the wheel etc, most of the issues/techniques are well studied, documented in
India and globally by many researchers. What is needed is simply apply and incorporate their findings in our practice.
Respondent : I would add some simplifications:
1. TFR is a hypothetical and CBR is real. (Probability of inaccuracy > of both may be same since the process is the same for measuring number of births - so may cancel each other). So variable discrepancies is not the cause of its non comparativeness. It is what is the nature of indicatpr - what is it really indicating? Efforts are made to corrleate statistically the FRs and CBR
2. The TFR is a measure of the fertility of an /*imaginary*/ woman who passes through her reproductive life subject to /all/ the age-specific fertility rates for ages 15–49 that were recorded for a given population in a given year.
3. TFR is the average number of children a woman /would/ have were she to fast-forward through all her childbearing years in a single year.According to the 2008 Revision Population Database, crude birth
rate (CBR) is the number of births over a given period divided by the person-years lived by the population over that period. It is expressed as number of births per 1,000 population.
http://www.pregnantpause.org/numbers/fertility.htm has some explanations given quite simply.
Respondent :Your link is very interesting. However from what I could make out fromthe extant literature, CBR is not a bad proxy for TFR -see links below. Narendra's original question still is not answered (pardon us dumkopfs Dr Patil Sir) - atleast I do not understand it - and to me Sunil Kaul's explanation and that of Sundar seems to be the simplest. Hypotheses non fingo.
See inter alia
1.CBR versus TFR in cross-national fertility research.
Entwisle B. Demography. 1981 Nov;18(4):635-43.PMID: 7308541 [PubMed - indexed for
MEDLINE
2.In (cautious) defense of the crude birth rate.
Kent MM, Haub C.Popul Today. 1984 Feb;12(2):6-7.PMID: 12265925 [PubMed - indexed for MEDLINE
Respondent : I understand the limitations of data which you are trying to highlight. It is evident already, but we still use point prevalence figures for comparison and rank communities accordingly. If we do not do this, then, we can never setgoals and never aspire that everybody should have highest attainable status of health as characteristics for every society or for that matter every individual is different. Wrt homogenization, it is always arbitrary, as you know that there is heterogeneity/variability in the same village/neighborhood based on inclusions and exclusions. So rates are based on averages built on averages and I am unable to understand why averages should not be compared. This is a different matter that based on comparison you go deeper to find the reasons for variations.
Now the main question which is still unanswered is the coherence between CBR and TFR as both relate to same activity.
Dr. Rajan R Patil : I though your query was already answered long back by demographer that CBR and TFR can not have any association they are independent of each other.
-Let me try to Demystify the rationale in simple words….
(and dhruv... with this i am responding to your request for simplifying complex concepts in simple words for advocacy groups)
Rates are the game of numerators and Denominator
Your confusion (as i read) is that when numerator for CBR and TFR- is same ( live births) but only denominators are differing yet why NO coherence?
Let me dissect it layer by layer:
lets first take Numerator characteristics
Please understand, Numerator only appears to be same for both BUT it is actually not. Please note the numerator CBR is TOTAL NO. OF ALL LIVE BIRTHS in a region (every single live birth is accounted for) but for TFR only live births selected in selected sample of households (and not every single live birth region)
Now let us look at the denominator side.
While denominator for CBR is all population in given region (its a simple frequency count of every individual in the region) the story is finished here for CBR.
Where as for TFR denominator COMBINATION of inter play of - Total count number (1) AND OTHER charecteristics of population in denominator (2nd, 3rd & 4th point).
1. Total count of individual in denominator ( like denominator for CBR)
2. proportion of women in different age groups from 15yrs to 44years
3. Mean Age at marriage of these woman in denominator.
4. birth order in the community
There fun side to denominator for CBR is that all population means( ALL MEN+WOMEN) in the community Now, please note CBR denominator is in fact inclusive of denominator of TFR -sub group WOMEN in 14-44 yrs age group is also included.
Because propoortion (%) of this 15-44years women in Total population will be different in every state, so also Mean age at marriage of these women and so is birth order. Every state will show different permutation and combination
EVIDENCE FOR INDEPENDENCE BETWEEN CBR and TFR
To help your self to put the whole story in perspective please do this home work
Compare TFR of TN and Kerala. CBR of Kerala is low but has High TFR on the other hand. In TN has High CBR but Low TFR. Thats the evidence for you on why TFR and CBR show no coherence
HOWEVER if you follow golden rule in epidemiology ALWAYS compare LIKES WITH LIKES you will spared of many misgivings. Example – you are most welcome to compare CBR of nagaland with CBR of India (but as I said standardize the rates first then compare), so also you can jolly well compare TFR of nagaland with TFR of India (as mala suggested here standardization is not required because TFR is any way ready made age specific rate and is not crude like CBR).
BUT never make mistake of cross comparison of CBR of one NAGALAND with TFR of India. It is absurd.
Hope that helps.
(And also hope that expert demographers like Mala Ramanathan will find this technically right as well. if not, i am willing to stand corrected with all humility- since demography is not my subject and my efforts here are simplyfying in lay terms- i may have missed out out on finer nuances of technicality)
Respondent : My take (from a non-demographer's perspective) -
If you compare CBR and TFR from NFHS 3, there is hardly any diversity of the kind that worries Narendra - the relative variability for the two indicators across states is similar. Please see attached graphed data (attachments not available on the mfc e-group) - and please note the different y-axis scales for the two indicators. What then is the diversity that we are discussing? The "low" and "high" TFRs and CBRs that we are discussing are relative to what?
There may still be two "facts" that intrigue:
1. the difference between the 2005 (NFHS3) CBR of 28.5, and the SRS figure of 17.5 for 2009. I notice also that Nagaland birth rates from SRS do not seem to be available for earlier than 2006. In 2006, CBR there was reported as 13.9 (less than half of the 2005 NFHS figure). For 2007, and 2008 it is reported as 17.3 and 17.4. According to SRS, the CBR is rising, if at all changing. The difference between NFHS and SRS could be because of a number of sampling and non-sampling issues, some already discussed in this debate. I would begin with a close look at the relative methods used for computation of CBR, and if they are the same, look at what the sampling strategies of the two were (remembering that SRS is a "panel"). If the samples of both are reasonably representative of the state, non-sampling errors (data quality issues) are the reason for variability.
2. the issue of large decadal growth rate that Narendra has quoted (I have not confirmed this) - it could simply be that between 1991 and 2001, the population did increase substantially, due to a relatively high birth rates and plummeting death rates then (or simply, more complete enumeration in 2001!)
Re definitional issues - pl check out http://www.measuredhs.com/help/Datasets/Fertility.htm for how these indicators are computed by DHS (NFHS).
- DHS measures CBR also as an age-adjusted rate, but I am not sure how SRS does it. In any case, it is best to compare all fertility rates from the same data set (NFHS), instead of across data sets.
- CBR and Age-Sp FR (from which CBR is computed) are based on a count of births in the three years preceding the sample survey (both are based on sample data, irrespective of source - except when computed from the decadal census). Thus, as computed, they represent the same time period of fertility performance of a population ("last three years"). (TFR is "hypothetical" only in the sense that it purports to represent the fertility history of an average woman, but is not hypothetical if you see it for what it actually is - an age-weighted sum of the previous three years' birth rates)
- CBR would theoretically be agnostic to age - it should not be restricted to ages 15-45/49, as age-sp TFR is. Births to women younger and older would be included. (However, in the case of NFHS/DHS, it probably excludes these groups, because the sampling is from a frame of 15-49 year women - I am not sure how SRS handles this)
- CBR has the entire population for the denominator, not just women
- Any variability between the two indices could therefore arise from any or all of the following variabilities:
-- An unusually distorted sex ratio of the 15-49 age group in the population (affects denominator)
-- An unusually distorted proportion of women 15-49 to the rest of the population (affects denominator)
-- Different non-sampling errors (data quality issues) - if the indices being compared are from two different surveys (if from same survey, this would not be a major variable) (affects numerator/denominator)
But that is only theoretical, if in reality there is not much diversity in the first place.
If TFR is computed using a hypothetically correct 15-49 y population despite overall population inflation - would that not be more valid than CBR? It evidently would be, if your numerators (births) are correct.
I believe, though, that birth reporting is grossly inadequate in the routine reporting system, probably in all states. For instance, we calculated CBR from routine reports from a clutch of districts in AP some years ago, and found it was around 8 - not more than half of expected. Post-JSY, I would expect more complete reporting - but still largely only for JSY eligible births. In other words, the assumption that numerators (births) are correctly reported may be misplaced. In the case of Nagaland, it is possible that the proportion of unreported deliveries is larger than in other states because of the population inflation that you describe. I have no idea about the level of sincerity in recording and reporting pregnancies in Nagaland - if they are "naturally" good at it, maybe the numerators are reasonably accurate.
You also need to somehow confirm population inflation: the 70% under-reporting may be attributable to simply a number of uncovered areas - blind spots with no reporting - in a hilly / forested state with inaccessible areas. For instance, what is the proportion of PHC's / SC's that are reporting the data in the HMIS fact sheet of Nagaland? If that is not the problem, a PHC-level break-up of population compared to census figures should help understand the distribution of the claimed population inflation.
Finally you need to somehow confirm the differential population inflation that you suspect: I do not fully understand how numbers can be increased by over-reporting ages of certain age-groups. A comparison of population pyramids between the census and the NFHS might help in this.
There is another problem with computing TFR from routine data: your would need to know the age of each delivering mother, and the HMIS needs to organize itself to collate the number of births by each age group (15-19, 20-24 etc), so as to first derive Age-sp FR. Without this, TFR would be impossible to compute.
The age-restricted CBR that NFHS seems to use may be a better understood indicator than TFR for monitoring purposes - just a suggestion.
As for the TFR-CBR discrepancy, as I have indicated in previous mails, I believe we saw a discrepancy because we were comparing TFR from NFHS to CBR from SRS. If we look at NFHS alone, there is no discrepancy. The question of why SRS is reporting lower than NFHS remains - need to compare computing methods and sampling methods to understand this. One possibility is that either the NFHS sample of the SRS panel is non-representative.
The HMIS fact sheet is very interesting and informative, by the way. I need to understand more of that - separately.
One final question - how much of the incentives / bonus linked to CBR/TFR is a true story? There are many more outcomes worthy of being incentivized!
Respondent : Data collectors in Jharkhand have been found to do such things during DLHS work. So it is good to find out how strong the quality checks are. You may also remember that we found major discrepancies on a Census sitedespite the original Census data from Pakur being reasonably accurate.
In that sense Sundar's MIS workshop - that deals with how you cross check data (the dog that didn't bark was it?) and Naren's original question are useful
Prabir
Dr. Rajan R Patil :a picture is worth 1000 words- so goes the cliché.
The line graph shared by Sridhar’s essentially communicates the same thing pictorially what demographer said as general concept- CBR and TFR are independent of each other- There need not necessarily have any correlation positive or negative). This thing is well depicted comparative line graphs evidence in states likie UP, Nagaland and Meghalaya.
CAUTION
Wait a minute. Should we accept above arguments just because the statistics and in form this line graph is showing so. For a moment let me be devil's advocate and declare that statistics can be deceptive-ONE CAN’T ALWAYS LIE AROUND WITH STATISTICS -why should we accept this unsourced graph?- We don’t even who were behind preparing this graph?, what statistics were take it into consideration?-what were their exclusion and inclusion criteria in
preparing this particular graph. All these crucial information is not accompanied with the graph. So it can very well deserves to be junked.
If statistics and graphs are our basis of accepting concepts then one can always prepare beautiful grpahs by smartly tweaking inclusion and exclusion criteria of data we are taking in for anaylysis  and can show exactly the opposite that TFR and CBR are correlated - apart from showing this beautifully on graphy- one can even work out quantitatively the Coefficient of correlations between TFR and CBR so smartly and convincingly that IT WILL show statistical significance of P < 0.00000000000000000000000000000000000000000000 to Infinity
Now tell me- shoulde we accept this abonoxious statisical evidence just because is worked out by a maverick genius and who has worked it out all the statisitics on a USA imported advance super computer and latest statistical packages??? The answr is big NO - when we clearly know that there can not be association between CBR and TFR conceptually.
As it is emphasized - statistics should be used for our enlightening and not as lamp post for support as used by drunkards. Statistics should be subservient to us and not we be subservient to subservient to statistics. Else we all will become top floor khaali robots who can churn out numbers at press button without making head or tail of it.
MORALE OF THE STORY
BASIS OF OUR CONCLUSION AND DECISION Â MAKING IN PUBLIC HEALTH SHOULD ALWAYS BE
BASED ON SOUND CONCEPTUAL CALRITY Â WITH STRONG LOGICAL REASONING
I dont want people to believe us just because i and or 10 others are repeatedly saying so. Or just believe it because dumb statistics and numbers are saying so.. Let him appreciate logically why these indicators were invented what they were conceptually meant and what they ommunicate individually. Then look into relationship between indicators.
I would like people accept only if he appreciate concept and logical reasoning behind given by demographer
RESPONDENT: I just happen to be one of those - I suspect there are many - who have not quite followed your explanation about the conceptual link between TFR and CBR, and why they must be independent. I also do not understand your fury - either at author of the graph or at statistics in general.
I happen to be the obnoxious maverick genius who generated that evil graph – but using a simple 2007-make 12-inch laptop, this morning sitting at home, not a supercomputer in the US. It took me about 7 minutes to copy paste the TFR and CBR columns of data from Table 4.3 of NFHS3/India report (p. 83-84) into a ppt slide (in a 2003 version of that most outdated of software - MS Office), and for it to be graphed with two different y-axis scales - I did not even adjust the scale. Most of the 7 minutes was spent finding the table and copy-pasting the columns from the PDF file. There was no statistical manipulation involved. Using two different y-axis scales is the equvalent of multiplying all the TFR values
by 10 to make it visually comparable to CBR values - it does not distort reality.
I did not bother to provide the source reference because I thought it was self-evident in this informal dialogue - the title of the slide says so, and the source data in the slide is available for all to see.
To my eye, as to Prabir's, the two indices match up closely - they are not independent of each other. I do not need a statistical test to tell me that. Since you are convinced otherwise, please do run your tests and let us know how indpendent these two data points are.
The reason why TFR and CBR covary is also simple - which I described in the previous mail. I understand (from the reference link to the Measure-DHS site provided in that mail) that DHS uses Age-Sp FR to compute CBR as well as TFR - the numerators are identical. The little variation you see is because the
denominators vary somewhat for CBR and TFR - which also I described. Thisvariation is probably a reflection of the different sex ratios in the 15-49 age group in different states.
I re-read Dr Jayaraj's and other explanations. They are not inconsistent with all this - they are only saying that if the two indices appear to vary differently, there is nothing inherently unexpected. No one has said they should not covary closely. If you see them in the spirit of what they are supposed to represent, they can certainly vary independently of each other, but the way DHS computes the indices, they are nearly identical.
The main point I was trying to make was not this at all. The point was answering Narendra's original question: "The TFR of Nagaland according to NFHS 3 is 3.7 in comparison to national average of 2.7 At the same time birth rate of the state is 17.5 in comparison to national average of 22.8 as per SRS October 2009. I am unable to understand the plausibility of these two figures. "
He is bothered by the inconsistency between Nagaland and all India figures. What I pointed out is that there is no inconsistency if you looked at both TFR andCBR from one source - NFHS. (The first data point on that unspeakable graph isfor all-India, and you can see that the all-India numbers are not what is
reported by SRS, and the relationship between the two indices for all India is pretty much the same as for most states - or simply look up the NFHS3 report table if this graph cannot be trusted).
Frankly, I had never bothered about this matter until Narendra raised the issue, and then I started looking around to learn how TFR and CBR are actually computed by DHS. I then realized that they should closely covary, so I went back and looked at the NFHS data, and found that they do covary closely. And that the original mistake was to look at TFR from one source and CBR from another. We all make such mistakes, and learn from them.
I would like to be corrected if I am wrong about how DHS computes CBR - and that might open up the debate again. Also, someone knowledgeable might want to comment on how SRS computes CBR.
Dr. Rajan Patil :I had begun my mail with the appreciation of the graph you shared (even unaware of source) because it was very much in sync with what i had to say. I have nothing against numbers and statistics- in fact you must have noted i have used statistical language of numerators and denominators etc (in the appended mail below) So i personally very much love statistics (in fact I would be out of job without it ;). That way be assured we are all on the same wave length.
Subsequently, i was being devil's advocate(i had said so in my earlier mail too (for those who hate statisitics) and was only echoing their likely response to unusual deluge of statistics, numbers and graphs being exchanged in this debate.
But yes, i would still stand by one thing. Statistics is only a tool. If our statistics are not in coherence with the conceptual facts it only means there is anomalies in the data (as we all have agreed and accepted with our demographic data collection)
That being so, In examining the conceptual facts- If our statistics is not in syn with it, we may need to change the data sets (e.g, instead of Indian data - we may choose more valid date from other country) and re-examine the issue, but not vice versa - Never change conceptual facts just because our bad statistics (out of bad data) is not supporting it.
We need to use statistics to validate our conceptual facts But Never use statistics (eg. our NFHS/SRS if we have so many doubt on its quality) and (never) use the same to accept or reject conceptual facts.
Morale of the story - (and my final take on this issue..)
1) We need to agree collectively that quality of our demographic data generated in India is not good enough to test demographic conceptual facts.
2) For clarification on any demographic indicators, or in understanding conceptual relationship between different demographic indicators- Let us revisit our textbooks rather than questioning it with bad data in our hands.
Respondent: I cannot respond to all the questions immediately as I have a deadline to meet.But one question - can there be different ways of computing CBR? According to NFHS 3 the CBR of Nagaland is 28.5 while of the same period i.e. 2006, it is 16.4 according to SRS (SRS bulletin). Any thoughts.
Respondent : Here- to add to the confusion is the HMIS data of 2009-2010 for Nagaland. I have it in this format for all districts and states in the country.
Respondent : That is what I am hoping someone will enlighten us about. Until now, I had assumed CBR is computed as usually described - the number births in a year per 1000 mid year population. The limitations of this are known- comparisons are affected by age/sex distribution of the populations, and byyear-to-year variation. To make estimates more stable, births over several years are taken together, and there are several ways of doing this, all of which give similar results. DHS uses 3-year counts, I have no idea what SRS uses. DHS also seems to use age-adjustment, which is generally not used for CBRs as far as Iknow.
There must be more knowledgeable members on these groups, who could provide more definitive answers.
Just fyi - the April 2006 SRS bulletin reports the CBR as 13.9 for Nagaland (and Oct as 16.4)! Transcription error? Variation in data?
Respondent : Let me join into this discussion with a further set of questions.
I got a call from the Family Planning(FP) officer in Nagaland. He had heard of this debate - he was confused and he was desperate. This is what he asked:
1. " I have been telling everyone that we have achieved population stabilisation But is that true? Or do I have to reduce birth rates much further?" Please Rajan - Please Sridhar - what do I reply?
2. The FP officer continued " If your answer is that - dont worry- you have achieved success- then can I tell my minister- 'that dont worry sir- we have achieved success - but it will take some years for it to show up on the data reports on TFR' . Would that be correct ? "" Please Rajan- Please Sridhar what do I reply.
3. The FP officer continued : "I told my minister that we have achieved success, since you had not replied. But then I went to the Planning Commission for our annual plan presentation with the minister. And they told us that the extra performance based bonus plan allocation we promised if you reach fertility targets will not be given to you. I told them about your reply? And they said- 'listen the bonus was linked to the TFR and not to the CBR and that when TFR comes down we will see? CBR is not good enough? Till then you must work harder and lower fertility rates at a faster rate.' Then what do I reply?" Please Rajan- Please Sridhar help me- How do I reply to this officer?
4. The officer went on: "In desperation I said to the planning commission 'I confess that in the last census many of our tribes had falsely inflated their figures to show a higher population. Part of inter-tribal competition. And therefore such a growth rate?'. But the demographer there was not impressed. He said... 'in that case your TFR should be even higher?' Is that correct ? Or was he pulling a fast one and pulling my leg?' Please Rajan- Please Sridhar- I am getting desperate now. I thought I knew the answers but am all confused now. Tell me - what should I reply?
Respondent :Just in case you are not up to answering my funny questions. Here is the line of thought:
If CBR and TFR represented merely two diffierent ways of measuring the same thing- it would not matter for the programme manager that they are divergent- for the action point is the same viz CBR being the real figure, one need not take action as the birth rate is under control. The programme has clearly succeeded.
2. However if the denominator has been inflated by a process of inflating population figures in the census which I think is what is happenning than the numerator is true and the denominator is falsely increased and therefore the birth rate is misleadingly low. Therefore TFR would be a better indicator in such case. The rub is that I still dont understand what implication the increase in population would have for TFR. If the denominator is drawn from a census estimate of the 15 to 45 women in the population and that was not inflated when the figures of the population were inflated-- since inflation was done( ? ) by increasing the age of older persons or children - then the TFR for the same birth rate would be much higher- as we are seeing.
3.I am not at all sure that this is the way the denominator for TFR is done. If however this hypothesis is correct, it would help me solve the problem of reading the Nagaland HMIS data where unaccounted deliveries are surprisingly high.
That was the Poirot Allusion.
I do suspect that in Nagaland they have figured this out- for they shared the problem of the inflated population with me, only when I was pressing to find an explanation to the HMIS data interpretation problems. I had of course at that time, not noticed the TFR and CBR discrepancy. The point is to look at it from the point of view of what use we put information too. And from that perspective new possibilities open up. That is what I thought in the beginning and kept it to myself. However you have all given such convincing explanations - that I dont quite know whether what i say is consistent with all the rest that has been said and agreed upon.
Now would I get a response. Is there a flaw in this logic?
Dr Rajan R Patil : I think you have certainly hit the holy grail of 'eye of the storm' of this debate by first raising set of questionings through your tet`-e tet` with Nagaland officals and responding to your own questions. That really helped.
Personally i find it most logically coherent explanation to the aparent descrepency pointed out by - which also concurrently and very aptly addressed specific data related questions raised by chinu,dhruv,sridhar in the context of Nagaland and i finally i get what i was looking for- the logical & rationale that could possible explain this deviance from conceptual fact. At least i am convinced.
At this stage, rather than going round and round, i feel that we should get few expert demographers to scrutinize and authenticate all that has been said thus far by us.
I take the initaitive with all your kind permistion to share this entire string with Demographer, demography text book author& mos importantly as a person involved in planning and coordinating demographic data collection activity in India In his capacity as director of IIPS india (1978-92). And also has been involved with NFHS and DHS planning. Incidentally he is our visiting professor of demography at our university.
Dr. Rajan R Patil : Sundar, I really thought your funny questions were largely rhetorical - that you were trying to make the point that real situations are complex and not academic, so I did not respond.
I agree partly with your line of thought:
If TFR is computed using a hypothetically correct 15-49 y population despite overall population inflation - would that not be more valid than CBR? It evidently would be, if your numerators (births) are correct.
I believe, though, that birth reporting is grossly inadequate in the routine reporting system, probably in all states. For instance, we calculated CBR from routine reports from a clutch of districts in AP some years ago, and found it was around 8 - not more than half of expected. Post-JSY, I would expect more complete reporting - but still largely only for JSY eligible births. In other words, the assumption that numerators (births) are correctly reported may be misplaced. In the case of Nagaland, it is possible that the proportion of unreported deliveries is larger than in other states because of the population inflation that you describe. I have no idea about the level of sincerity in recording and reporting pregnancies in Nagaland - if they are "naturally" good at it, maybe the numerators are reasonably accurate.
You also need to somehow confirm population inflation: the 70% under-reporting may be attributable to simply a number of uncovered areas - blind spots with no reporting - in a hilly / forested state with inaccessible areas. For instance, what is the proportion of PHC's / SC's that are reporting the data in the HMIS fact sheet of Nagaland? If that is not the problem, a PHC-level break-up of population compared to census figures should help understand the distribution of the claimed population inflation.
Finally you need to somehow confirm the differential population inflation that you suspect: I do not fully understand how numbers can be increased by over-reporting ages of certain age-groups. A comparison of population pyramids between the census and the NFHS might help in this.
There is another problem with computing TFR from routine data: your would need to know the age of each delivering mother, and the HMIS needs to organize itself to collate the number of births by each age group (15-19, 20-24 etc), so as to first derive Age-sp FR. Without this, TFR would be impossible to compute.
The age-restricted CBR that NFHS seems to use may be a better understood indicator than TFR for monitoring purposes - just a suggestion.
As for the TFR-CBR discrepancy, as I have indicated in previous mails, I believe we saw a discrepancy because we were comparing TFR from NFHS to CBR from SRS. If we look at NFHS alone, there is no discrepancy. The question of why SRS is reporting lower than NFHS remains - need to compare computing methods and sampling methods to understand this. One possibility is that either the NFHS sample of the SRS panel is non-representative.
The HMIS fact sheet is very interesting and informative, by the way. I need to understand more of that - separately.
One final question - how much of the incentives / bonus linked to CBR/TFR is a true story? There are many more outcomes worthy of being incentivized.
Respondent :Thanks for very valuable input of so many friends. But at the end of the day, I am still unclear to in interpreting the huge diversity in two indicators which relate to births in the same population of more or less with the same reference period. I understand that one is 'expected' and the other one is 'exposed' measure and may show differing trends in different
cohorts. But then are these indicators incoherent to each other and can they be used for comparison in different population cohorts as is normally done by us all and by the Govt. of India while ranking different states and deciding financial allocations. My question is simple "Is it possible to have low birth
rate with high TFR and vice versa". Is there any normativity?
Another issue foxing me about Nagaland is that decadal growth rate between 1991 and
2001 census. It is whopping 64.53% while all India is 21.54% (Surprisingly, Punjab's growth rate which has huge inmigration is 21.54% only and its birth rate is also 17.3/1000 population). Again the question is with CBR like 17.5, is it possible to have decadal growth rate of 64.53% in Nagaland where immigration is almost nil (my assumption).
Demographers and others, please enlighten. I am not ready to completely believe What Sunil has said. According to many demographers, data generated by the Census, NSSO, NFHS(IIPS)
India is considered quite credible.
Respondent: I have attached an excel file calculating TFR, CBR of Nagaland actually and hypothetically.
During this calculation Pop and no of births are actual figures (HMIS)I have kept no of births constant. AND no of women of reproductive age(18-21 for simplifying the calculation) and percentage of women of the specific age who delivered also constant.
Later I have changed population only. Doing so, the CBR varies drastically (since the denominator is population but TFR remains constant. Although, the variance may not be so great thats why there is some trend of consistency as Sridhar is saying but scientifically, they cannot be as the figures may swing drastically if there is great variance. Statistical correlation may point toward its probability - after all statistics is THE science of probability. While epidemiology wants accuracy to predict and plan. CBR cannot represent fertility beyond a certain range. Thats what the paper referred to by Chinu is saying.
mfc may not be able to see the attached file. I can send it.
Pls comment if the calculations are correct.
Pop 2931412
Total Births 15282
CBR 5.2
Age group No of women of this age group Out of which who Delivered babies Delivered babies
% of women of this age group
18 5098 1020 20
19 13200 5280 40
20 16105 8053 50
21 3100 930 30
15282
37503 15282 140
0
TFR 1.4
Pop 5931412
Total Births 15282
CBR 2.6
Age group No of women Delivered babies Delivered babies
18 5098 1020 20
19 13200 5280 40
20 16105 8053 50
21 3100 930 30
15282
15282 140
0
TFR 1.4
Pop 1931412
Total Births 15282
CBR 7.9
Age group No of women Delivered babies Delivered babies
% of women of this age group
18 5098 1020 20
19 13200 5280 40
20 16105 8053 50
21 3100 930 30
15282
15282 140
0
TFR 1.4
Dr. Rajan R Patil: I am appending below response from Prof. K Srinivasan one of theleading demographer ofour country who headed IIPS bombay wasinvolved in designing and planningNFHS, DLHS in India.Besideshas written textbook on demography
Dear Dr. Patil,
Any qualified demographer should be in a postion to clarify your doubts. Here is a brief explanation.
Crude Birth Rate ( CBR) is an index influenced not only by fertility levels but also by the age distribution of the population. That is why it is called Crude Birth Rate . On the other hand, Total fertility Rate (TFR) is independent of the age distribution ofthe population and is considered as a pure measure of fertility and is used for comparison of fertilitylevels across populations and over the same population over time . Of course there are more refined measures of fertility but TFR has been used widely as a good measure of fertility.
Coming to Nagaland case that you have suggested, the state hasa smaller proportion of women in the reproductive ages in its populationcompared to the country as a whole ; hence a lower CBR
compared to its TFR levels. In many other populations the differences may be even dramatic.
Consider a population of 1000 persons where there are only 60 women in the age group 15-44 , equally distributed across individual ages , ; 2 women in age 15, 2 in age 16,.... and 2 in age 44. Suppose in one year one woman ineach age give birth to a child , there will be 30 births with age specific fertility rate for each age being 0.5. TFR being the sum of the age specific fertility rates will be 15 ( 30 *0.5) ,On the other hand the Crude Birth Rate will be only 30. ( 30 births per 1000 population). if we reduce the age specific ferility rate for each age to 0.25 the TFR will be 7.5 andthe CBR will be 17.5. Such extreme examples prove the point that the age distribution of a population can considerably distort the CBR measure and hence this measure cannot be
used to study fertility trneds and differentials.
I hope I ahve clarified your issue.
K. Srinivasan,
Emeritus Professor,
International Institute for Population Sciences,
Deonar, Mumbai, India, 400 088
Respondent : Response of Prof. K. Srinivasan establishes the relationship between CBR and TFR. His example shows that high TFR will result into high birth rate. He has neither refuted the Nagaland data nor has he explained that how high TFR of Nagaland has resulted into low birth
rate. Then there is different between CBR calculated by the NFHS andthe SRS.
I am more unclear after reading the response of Prof K. Srinivasan. First he has not cast any doubt over the CBR and TFR figures of Nagaland. He agrees that these are dramatic and can possibly be there. But then what he describes is just opposite of the
Dr.Rajan R Patil :I am really concerned when you make convoluted interpretation from the example and ascribe it to Prof. Srinivasan.
Please note He never said " High TFR will result into High CBR" that is entirely your isreading from his example.. He had given that example only to show that CBR and TFR can be totally independent of each other and there need not any correlation between them. That is exactly what many of us were saying and were convinced about about it- any apparent correlation is only coincidental.
The fact that CBR and TFR are not related and are are infact independent of each other is a basic and fundamental concept which we as public health students have learnt and should be able to explain with reasons and logic with our exisiting elemenatry demography knowlege BUT i still I referred this to Prof Srinivasan only to get independent expert opinion for not only a qualified demogrpaher BUT more importantly because he was in thick of action in planning, designing and coordinating of Indian demographic and health data as longest serving director of premier insititution IIPS bombay, that is the nodal agency for any national data collection over four decades (lest even he would dismmissed as theoretical academician not knowing what happens in the field visa vie abut data quality, data fudging etc)
Respondent : Rajan,u keep emphasising the fact that CBR and TBR are "completely unrelated"!! i find this most ridiculous. they are not directly related mathematically, but they are throwing light on the same public health pehnomena in the same population. pls do pot go overboard and throw out babies with the bath water. public health is not supposed to be some mystical science that defies all reason and logic and common sense. ofcourse there is correlation between CBR and TFR which is not a statistical correlation and depends on a host of other dependent variables such as sex ratio, percentage of women on repr age group, sampling mehtods etc etc as every one (inlcuding narendra) has been saying. but to say they are "completely unrelated" is very strange.
from my readings, none of the highly qualified people on this discussion have said that.
Dr.Rajan R Patil: Vandana, you said...".....ofcourse there is correlation between CBR and TFR which is not a statistical correlation and depends on a host of other dependent variables...."
Thats where the catch is...
Any genuine correlation between two variables should necessarily show statistical significance for it to be basis of our public health decision making.
Any observed correlation which is statistically non significant can be explained by
1. Spurious Association ( selection bias, sampling method, data quality/fudging)
2. Chance (random variability)
3. Co-founding effect (Age, Sex composition)
Whatever little observed correlation between CBR & TFR observed is because of confounding effects i.e, co-variation of third factor as you rightly pointed eg., sex ratio, percentage of women on repr age group, sampling mehtods etc.
In that case, this debate is needless because :
1. Normally if data is well collected (valid and reliable) then CBR-TFR WILL BE STATIFICALLY NON-SIGNIFICANT.
2. Any partial/observed/apareent statistically non significant correlation observed in CBR-TFR only indicate weak associations between diverse variables which is of only theoretical interest- (for making multiple conjectures-as we are currently doing)
3. Public health decision making can not be based on conjectures as we can not plan our intervention on something that can not demonstrate credible cause-effect relationship with strong correlation(statistically significant) - (after ruling out Chance and confounding effect). That would be against tenets of Evidence based Public health .
Respondent :the point is also that we would be very poor public health experts if we did not ask questions based on discrepencies that may or may not be statistically significant after we have finished our exploration. narendra did well to find some data that seemed incongruous (i havent heard any one question what is going on nagaland before). he was also perfectly aware of the types of answers given by people (we had a telephone covnersation before he posted this question). these indicators are not some sacrosanct things - they are to help us to understnd a situation in different ways. the phenomena 'fertility' and 'birth' surely are interrelated; sorry i cannot accept that they are not, since fertility is not currently being understood as how many eggs and sperms are produced - it is still being measured by babies. therefore their measures must be interrelated. we use different methods to understand the different aspects of the same picture, and though we do not capture the 'truth' by any of these, we do expect that all our methods should broady lead us to some logic. what the experts have told us merely this - in their experience, the discrepency we have come across is mathematically possible because of the various reasons we have talked about. may i submit that there is a statistical relationship? i would ask the mathematical question, suppose the trend in CBR was steadily upwards and the trend in TFR be steadily downwards, would not the difference between the two graphs, at some point become statistically significant? till what point would they continue to consider the incongruity a mere statistical anomaly? someone (lay person like me) may answer this question by saying - it is not possible for this to happen - and the only way it becomes possible is if the two are related - they are positively correlated, or SHOULD be. and if they are not, there is something wrong and they are not helpful indicators.
anyway. i think what is more to the point is that i do not agree with your denigration of the question itself. i maintain it is worth asking and worth thinking whether one should be concerned that one or the other is off. better that than thinking they are not inter related, and therefore it is silly to even ask the question. but if you choose to continue with what i consider a fairly nit-picking appraoch to catching hold of a word here and there and attacking its originator in fairly personal, insulting and agressive terms - well that is your prerogative.
Respondent: I am not an demographic expert - but then very few amongst us are! I was a meer computer programming myself due to suggestions from a few experts.
Our demographer friend has pointed out with examples similar to one I had estimated. The contention is that TFR is a projected indicator - what may happen. CBR is what actually happens.
In this context, it would have been more fruitful if we get an expert opinion about Nagaland data rather than theoretical one. SO to that extent narendra's questions still need to answered.
I am sure all the ifs and buts issues raised have authentic census data eg age wise sex distribution etc.! Rajan, can we work on both the formulae with Nagaland data (not sample - SRS.NFHS3)
Respondent : gain if the birth rates all moved up together there would be an apparent link. Most important, if the real population pyramid resembled the ideal one CBR and TFR would be directly related. So I think we need to understand numbers and statistics more broadly. The tool of Hill's postulates apply to the study of hidden cause and effect relationships. They are useful tools. You have analysed this question using them and they nearly fit. We miss larger math relationships if we use a butter knife for dissection.
Respodent : let me try once more.
Look at the figures for Nagaland and India:
NFHS 3:
Nagaland: CBR: 28.5, TFR: 3.77
India: CBR: 23.1, TFR: 2.68
SRS, 2006:
Nagaland: CBR: 13.9 (April), 16.4 (October)
India: CBR: 24.1 (April), 23.8 (October)
I understand the following from the above:
1. Within NFHS, the relative proportions of TFR and CBR are fairly close - there is only a small inconsistency, probably explained by Prof Srinivasan's explanation that the crude rate is lower relative to the TFR in Nagaland as compared to all of India because there is a smaller proportion of women in the reproductive age in Nagaland as compared to India.
2. There is probably an error in the way Nagaland SRS data was collected / computed in 2006 - there are no such fluctuations in the CBR of other states between April and October of 2006. This could mean that there is something wrong in the data from Nagaland.
3. The only major discrepancy is when you compare SRS CBR to NFHS TFR for Nagaland and India and see divergent trends. This is what made you say " ... unless we assume that there are extremely very few women in Nagaland and producing on an average 3.7 kids in their reproductive life time". There is no way to solve this "puzzle" since there is no TFR data available from SRS - and going by NFHS data alone, the discrepancy seems to be small enough to be explained as above. I don't think there is a way to compute crude CBR from NFHS either. So, I am afraid the existence of a puzzle is difficult to establish. We are just looking at data that are not good enough for comparing. Or, either the NFHS or the SRS data is plain wrong.
Disclaimer: this has nothing to do with Rajan's passionate claim that TFR and CBR are unrelated!
Respodent :Please find two more papers with regard to relationship between
CBR and TFR besides what Chinu had sent.
The link of first one is:
http://books.google.co.in/books?id=-
p2OIDPr5jkC&pg=PA228&lpg=PA228&dq=Relationship+between+CBR+and+TFR&sour
ce=bl&ots=Xh36QBWSvB&sig=21u42ONeLqdGa6ie6luURymKcI4&hl=en&ei=4ZtPTcjRK
MfIrQf1vJDaBg&sa=X&oi=book_result&ct=result&resnum=4&ved=0CCcQ6AEwAw#v=
onepage&q=Relationship%20between%20CBR%20and%20TFR&f=false
Second is the Family Planning DemographicIndices. It is available on
the NIHFW web site. I am attaching a copy of it. Please read
highlighted part on page 20.
Third, I wish if somebody could unravel the Nagaland puzzle because
demographer example does not help unless we assume that there are
extremely very few women in Nagaland and producing on an average 3.7
kids in their reproductive life time.
Dr. Rajan R Patil: I wanted to reassure you am treating this discourse purely at technical plane – absolutely nothing personal!.
Since many of our learned friends are convinced and subscribe to your view and hypothesis, there must be some truth in it. I find it little difficult to relate to it as it would require me to unlearn some of the very fundamental concepts that form the building blocks on which entire discipline of demography and to large extent public health stands. So before we rock the boat from the bottom, let us get our arguments peer reviewed. I have done it earlier got my line of thougth verified.
Now I call upDon you and other colleagues who share your line argument to get the rationale and logic behind your/their respective argument peer reviewed. Here I don’t mean publication in journal that’s two year process, but at the least, corrobation by atleast one expert demographer who is willing to endorse your views and vouch for it publicly. I will be happy to relearn every thing. We are talking science, where nothing is sancro-sanct, it has to keep evolving, sometime even turning upside down.
Let us remember, Earth would have continued to be flat and been centre of universe until somebody questioned by setting the logic right.
At personal relationship front I generally maintain..
Kabira (me) Khara Bazaar Mein, Mange Sabki Khair
Na Kahu Se Dosti, Na Kahu Se Bair
So no offence! I will be happy to be part of this churn around, will always be proud of my association with you. I have enjoyed and continue to cherish my memories of your warm hospitality during my visit to chittorgarh/prayas, same I would like to reciprocate when you hit the southern shores..
Respondent :No love lost, Rajan. Your erudite comments frenzied debate - a commoner like me is enriched. Seems from many postings that this has been discussed even earlier and some algorithms (spurious?) worked out – refer the papers Chinu and I cited earlier. Dr. Srinivasan in his response wrote “the age distribution of a population can considerably distort the CBR measure and hence this measure cannot be used to study fertility trends and differentials.� But prior to making this comment he showed us how ASFR and TFR influence CBR… there could be dramatic results and then his conclusion. If NFHS & SRS data about Nagaland is correct then we should assume that there are few women in reproductive age group with high fertility – Dr. Srinivasan corroborates.. We do not have
access to life tables of Nagaland but other evidences do not support this conjecture. Do we accept the anomalies of record keeping … Sunil described. Perhaps, better would be to get hold of life tables.
Now, what inference one can draw from Nagalandâs reported CBR & TFR is population of the state booming (far away from replacement fertility) or perhaps not growing in proportional manner owing to lesser CBR.
I fondly recall your visit and we all are “Kabiras� in this wonderful group
of ours.
Correspondence:
Dr. Rajan R Patil
Division of Epidemiology
School of Public Health
SRM University
Potheri, Kattankulathur - 603203
Greater Chennai,
INDIA.
Cell: 9445811610 & 9025378036
Email : rajanpatil@yahoo.com, rajanpat@gmail.com
-by DR. RAJAN PATIL
This is complilation of n intense but interesting debate was initiated in a public health e-group on a question on relationship between CBR and TFR- on the plausibility of TFR being on increasing trend while CBR was declining. The debate evoked plethora of responses arguing for and against the topic. Since the issue under discussion was a demographic topic, three expert demographers were consulted and their views enlisted.
Question Posted: The TFR of Nagaland according to NFHS 3 is 3.7 in comparison to national average of 2.7 At the same time birth rate of the state is 17.5 in comparison to national average of 22.8 as per SRS October 2009.
I am unable to understand the plausibility of these two figures. I will be grateful if some body can throw light on this. Other features of the state are that median age at first birth is 21.8 in comparison to 19.8 nationally and sex ratio also favours women.
Response :Surely TFR takes into account the rates of the last 30 years while current rates are just point figures.Of course different sources or their inherent biases might also influence things(NFHS is trying to prove something, whereas SRS is a more neutral and very efficient way of collecting information)
Response from a professional Demographer
TFR is an 'exante' (expected) measure. TFR is the total number of births likely to occur for a woman, who enters reproduction at a particular point of time, which depends on the median age at marriage/first
birth, current level of practice of family planning and prevalent natural wastage (abortions) rate. Birth rate or crude birth rate is an 'exposed'
measure calculated as the total number of births in a year to 1000 mid-year population in that year.  The latter measure depends to a large extent on the number of women who have given birth to in that year. The size of the population of women who are currently in reproductive age depends on earlier
fertility level and sex ratio of the population at reproductive age group. For this reason, the two measures could exhibit differing patterns in relation to national average.
With best: Jayaraj.
Respondent: even then (and that being understood), average age of marriage is higher than national avg and unmet need for contraception is lower. sex ratios are higher for women than national avg. unless there are a huge number of abortions, this discrepency should not occur. it is worth questioning.
Respondent:I think that the best explanation would be the possibility of the figures being fictitious as SRS people - by their own honest admission in private at Guwahati - hardly leave their offices to make new data and NFHS partners in NE have been less than sincere in collecting data.
Respondent :Another point is that if you have children more times (ie have the last child much later on) the annual rate will be low but the TFR will remain high. This would be even more possible if Nagaland has low MMR/ low number of women's deaths and India has a high death rate - especially affecting later pregnancies
Dr. Rajan R Patil : The question was essentially exploratory in nature- of the inherent logic - on possibility of occurrence of apparent incongruity between  the  two said demographic indicators and if so, was it conceptually plausible the response from our demopher most credible and convincing for the reason of its conceptual clarity and inherent logical base on which it is grounded. Hence like any constructs it is best verified be application of inductive or deductive logic and not by contextualizing to any geographic point. conceptually when any constructs passes validity tests (convergent and discriminant) then generally it is accepted for its coherence.
The fact that Demographic indicators are quantitative in nature and measurable to decimal level, that makes life simple for its cross validity examination-once verified it can be very well be universalized.
Morale of the story
1. Since both question and the answer are on conceptual plane- hence  the geographic setting- be it Nagaland or India and hence the idiosyncrasies associated Indian data are irrelevant here.
2. The strength of any given construct stems from its ability to hold ground in its abstractness independent of its setting- be it Nagaland in India or wonderland in Timbaktu- It will still be true.
Dr. Rajan R Patil : if you insist that your interest here is to seek the answer especially contextualised and geographicalized to Nagaland in particular there is one hurdle- only validity of figures concerning to Nagaland could be answered by cross checking source, authenticity issues of data collection behaviour as already suggested by our other colleagues. BUT YOU CAN NEVER COMPARE it with national averages.
As a cardinal rule of epidemiology( and i hope it is true for any quantitative sciences like Bio-statistics & Demogrpahy included) It is wrong to compare and make inference by comparing two point prevalence figures of two heterogeneous communities (Nagaland and India), because SRS or NFHS give only crude rates.
TWO CRUDE RATES SHOULD NEVER BE COMPARED . crude rates are meant for only description of single group characteristics and not for two population comparison- it will be miserably misleading.
However If you still insist that your basic interest is in camparison of nagaland figures with National averages then we ill have to invoke 2nd cardinal rule of group comparison- to always convert the given crude rates into standardized rates by direct standardization method (or any alternative suggested by demographers). ONLY The resultant ADJUSTED STANDARDIZED RATES should be compared between two hetrogeneos populations and inference drawn.
Comparison of crude rates can be done not for drawing comparative inference but for different purposes- to varify the issues brought out by Sunil Koul- to check authenticiy, honesty of data collection behaviour in North east. To do this it is best to first homogenize the groups i.e, compare Nagaland figures with NE region averages rather than national averages. So that data collection and validity issues as stand alone issues mentioned by sunil as specific to NE can be discounted for in final conclusions- because then it become intragroup croup comparison within NE region and hence Nagaland and NE region can be considered homogeneos, at least relatively speaking to National comparison.
In conclusion you me need get one demographer to convert the given crude figures into standardized figure and only then make attempt to compare with Nagaland vs national comparisons and draw conclusions
Respondent :the TFR is an age standardised artificial cohort rate and lends itself to comparision across time and space. I am a trained demographer (as suggested by the previous writer -Dr.Patil, and this is one of the fundamental lessons of technical demography) and there is very little scope for me being wrong about this. In so far as comparisions of crude rates are concerned, yes indeed one needs to standardise for age as that could cause distortions and therefore any attempt at comparison should be with standardised rates.
Respondent: I understand the limitations of data which you are trying to highlight. It is evident already, but we still use point prevalence figures for comparison and rank communities accordingly. If we do not do this, then, we can never set goals and never aspire that everybody should have highest attainable status of health as characteristics for every society or for that matter every individual is different. Wrt homogenization, it is always arbitrary, as you know that there is heterogeneity/variability in the same village/neighborhood based on inclusions and exclusions. So rates are based on averages built on averages and I am unable to understand why averages should not be compared. This is a different matter that based on comparison you go deeper to find the reasons for variations.
Now the main question which is still unanswered is the coherence between CBR and TFR as both relate to same activity.
Dr. Rajan R Patil :I though your query was already answered long back by our good demographer friend that CBR and TFR can not have any association they are independent of each other.
-Let me try to Demystify the rationale in simple words….
(and dhruv... with this i am responding to your request for simplifying complex concepts in simple words for advocacy groups)
Rates are the game of numerators and Denominator
Your confusion (as i read) is that when numerator for CBR and TFR- is same ( live births) but only denominators are differing yet why NO coherence?
Let me dissect it layer by layer:
lets first take Numerator characteristics
Please understand, Numerator only appears to be same for both BUT it is actually not. Please note the numerator CBR is TOTAL NO. OF ALL LIVE BIRTHS in a region (every single live birth is accounted for) but for TFR only live births selected in selected sample of households (and not every single live birth region)
Now let us look at the denominator side.
While denominator for CBR is all population in given region (its a simple frequency count of every individual in the region) the story is finished here for CBR.
Where as for TFR denominator COMBINATION of inter play of - Total count number (1) AND OTHER charecteristics of population in denominator (2nd, 3rd & 4th point).
1. Total count of individual in denominator ( like denominator for CBR)
2. proportion of women in different age groups from 15yrs to 44years
3. Mean Age at marriage of these woman in denominator.
4. birth order in the community
There fun side to denominator for CBR is that all population means( ALL MEN+WOMEN) in the community Now, please note CBR denominator is in fact inclusive of denominator of TFR -sub group WOMEN in 14-44 yrs age group is also included.
Because propoortion (%) of this 15-44years women in Total population will be different in every state, so also Mean age at marriage of these women and so is birth order. Every state will show different permutation and combination
EVIDENCE FOR INDEPENDENCE BETWEEN CBR and TFR
To help your self to put the whole story in perspective please do this home work
Compare TFR of TN and Kerala. CBR of Kerala is low but has High TFR on the other hand. In TN has High CBR but Low TFR. Thats the evidence for you on why TFR and CBR show no coherence
HOWEVER if you follow golden rule in epidemiology ALWAYS compare LIKES WITH LIKES you will spared of many misgivings. Example – you are most welcome to compare CBR of nagaland with CBR of India (but as I said standardize the rates first then compare), so also you can jolly well compare TFR of nagaland with TFR of India (as mala suggested here standardization is not required because TFR is any way ready made age specific rate and is not crude like CBR).
BUT never make mistake of cross comparison of CBR of one NAGALAND with TFR of India. It is absurd.
Hope that helps.
(And also hope that expert demographers like Mala Ramanathan will find this technically right as well. if not, i am willing to stand corrected with all humility- since demography is not my subject and my efforts here are simplyfying in lay terms- i may have missed out out on finer nuances of technicality)
Respondent: Thanks for elucidating the issues regarding some of demographic indicators. It will be worthwhile to construct an article on Challenges of Technical Demography for Public Health  Advocacy or some such for the bulletin.
Can we make it even simpler some of the fundamental issues, you have raised?- Often figures are flown and floated to gain some edge but are 'bad' from scientific reasons. I think we need to rectify such anomaly. First, those who are using such figures must know that they are 'bad' figures but useful for
advocacy (after politics allows you to some extent if it is favor of the pro-people policies!) I had initiated a column for the bulletin called Armchair (Re)view. Can you guys something on fallacies that often we face?
Respondent: Thanks for very valuable input of so many friends. But at the end of the day, I am still unclear to in interpreting them huge diversity in two indicators which relate to births in the same population
of more or less with the same reference period. I understand that one is 'expected' and the other one is 'exposed' measure and may show differing trends in different cohorts. But then are these indicators incoherent to each other and can they be used for comparison in different population cohorts as is
normally done by us all and by the Govt. of India while ranking different states and deciding financial allocations. My question is simple "Is it possible to have low birth rate with high TFR and vice versa". Is there any normativity?
Another issue foxing me about Nagaland is that decadal growth rate between 1991 and 2001 census. It is whopping 64.53% while all India is 21.54% (Surprisingly, Punjab's growth rate which has huge inmigration is 21.54% only and its birth rate is also 17.3/1000 population). Again the question is with
CBR like 17.5, is it possible to have decadal growth rate of 64.53% in Nagaland where immigration is almost nil (my assumption).
>Demographers and others, please enlighten. I am not ready to completely believe What
Sunil has said. According to many demographers, data generated by the Census, NSSO, NFHS(IIPS)
India is considered quite credible.
Respondent :To add my bit to the knowledge/confusion:I heard this when I was in nagaland..... they find it difficult to reach 100% immunisation - since there just not as many children around as expected from population projections. At the time of the census there was some competitive spirit in many tribes/villages trying to boast of high populations. So census figures are very high- and whenever you use that as the denominator every other numerator would come to grief. This year some one is being very honest in reporting immunisation honestly and figures are dopping to about half- and well soon I expect the figures will go up- perhaps. But all this is rumor - not fact. And like for all rumors it should be disbelieved till otherwise proven. Certainly not to be quoted- hence not posted on the egroups. If one of you wants to take credit for hearing such a rumor and post it I have no objections, as long as I am not the source of the information.
Respondent: But then this is a common story everywhere from the service providers. I recall, several years back I went in a primary school of a very remote village in Bandikui block of Dausa Dt. in Rajasthan. There were 7 girls and 41 boys enrolled in the school and school authorities claimed 100% enrollment. I could not believe them that how can there be only 7 girls against 41 boys of the same age group. Truth came out on house visits. ANMs and other health workers use these pretexts very commonly to show 100% coverage. But can we believe it? Other side is that now we have to look towards RGI, NSSO, IIPS with caution? Leave aside manual checks, didn't they do statistical checks before all this data is published? It is becoming like HMIS collected by the health department.
Respondent :You dont get the point- I am saying that the population figure and growth rate is inflated. Take your example and let me reconstruct it. There are only 7 boys and 7 girls in the school and they have claimed there are 41 boys so that they get a higher mid day meal allocation!! That was the allegation. And hey dont laugh at the HMIS ofthe health dept. You have to know to read it. Remember Poirot in one of Agatha Christies books. A lady asks him- what is your method. Poirot replies- well I just talk to people and they just tell me. The lady is amazed. How silly she says-- "they could easily lie to you." And Poirot replies- that is what is the most instructive part!!! The errors in HMIS dear sir as instructive as the numbers in it. Do join us on the HMIS workshop - March 23 to 27th or so.
Respondent :This is a general point abt most govt data collection. How do you estimate the amount of exaggeration? School attendance is exaggerated because if the attendance is low, the school will get cancelled in the next year's budget.
Dr.Rajan R Patil : Am responding to your specific query over Data quality issues in our national programmes -detecting, estimation  and ways to correct them. would like to emphasize that enough work has been done on this issue and solutions also has been already proposed.
In a nutshell  a data validity check can be carried out on a sample of existing data by both qualitative and quantitative methodology. Correction factor worked out and applied. Please note there can not be One Generalized correctionfactor applicable to all – since data quality issues varying degrees of data quality issues- so it has to be worked out specifically to each state and be applied to there only.
Am saying this with some convicition since we were concerned and have done some ground
work in investigating of data quality issues especially when I was with  UNDP-WHO working on setting up disease surveillances systems in Bihar and Orissa and elswhere in country and have
published our findings too. Please ref to our work in following journals
1) Deepa TM, Venkata Rao, Rajan R Patil,Reuben Samuel. Community Reporting Community reporting model in Disease Surveillance - National Medical Jounal of India. Vol 21, No4, 2008.Â
2) Venkatarao E, Rajan R Patil, Anita Anasuya, Deepa TM, Reuben Samuel. Monitoring data quality in syndromic surveillance learnings from resource limited setting. (under review)
Also I had investigated data collection and quality issues in cotton farmers suicide in Warangal dist in Andhra pradesh. Findings available at :
3. Rajan Patil. Suicide deaths among farmers. British Medical Journal (S.A ed ) June 2002;18(4):310
Morale of the story
No need to re-invent the wheel etc, most of the issues/techniques are well studied, documented in
India and globally by many researchers. What is needed is simply apply and incorporate their findings in our practice.
Respondent : I would add some simplifications:
1. TFR is a hypothetical and CBR is real. (Probability of inaccuracy > of both may be same since the process is the same for measuring number of births - so may cancel each other). So variable discrepancies is not the cause of its non comparativeness. It is what is the nature of indicatpr - what is it really indicating? Efforts are made to corrleate statistically the FRs and CBR
2. The TFR is a measure of the fertility of an /*imaginary*/ woman who passes through her reproductive life subject to /all/ the age-specific fertility rates for ages 15–49 that were recorded for a given population in a given year.
3. TFR is the average number of children a woman /would/ have were she to fast-forward through all her childbearing years in a single year.According to the 2008 Revision Population Database, crude birth
rate (CBR) is the number of births over a given period divided by the person-years lived by the population over that period. It is expressed as number of births per 1,000 population.
http://www.pregnantpause.org/numbers/fertility.htm has some explanations given quite simply.
Respondent :Your link is very interesting. However from what I could make out fromthe extant literature, CBR is not a bad proxy for TFR -see links below. Narendra's original question still is not answered (pardon us dumkopfs Dr Patil Sir) - atleast I do not understand it - and to me Sunil Kaul's explanation and that of Sundar seems to be the simplest. Hypotheses non fingo.
See inter alia
1.CBR versus TFR in cross-national fertility research.
Entwisle B. Demography. 1981 Nov;18(4):635-43.PMID: 7308541 [PubMed - indexed for
MEDLINE
2.In (cautious) defense of the crude birth rate.
Kent MM, Haub C.Popul Today. 1984 Feb;12(2):6-7.PMID: 12265925 [PubMed - indexed for MEDLINE
Respondent : I understand the limitations of data which you are trying to highlight. It is evident already, but we still use point prevalence figures for comparison and rank communities accordingly. If we do not do this, then, we can never setgoals and never aspire that everybody should have highest attainable status of health as characteristics for every society or for that matter every individual is different. Wrt homogenization, it is always arbitrary, as you know that there is heterogeneity/variability in the same village/neighborhood based on inclusions and exclusions. So rates are based on averages built on averages and I am unable to understand why averages should not be compared. This is a different matter that based on comparison you go deeper to find the reasons for variations.
Now the main question which is still unanswered is the coherence between CBR and TFR as both relate to same activity.
Dr. Rajan R Patil : I though your query was already answered long back by demographer that CBR and TFR can not have any association they are independent of each other.
-Let me try to Demystify the rationale in simple words….
(and dhruv... with this i am responding to your request for simplifying complex concepts in simple words for advocacy groups)
Rates are the game of numerators and Denominator
Your confusion (as i read) is that when numerator for CBR and TFR- is same ( live births) but only denominators are differing yet why NO coherence?
Let me dissect it layer by layer:
lets first take Numerator characteristics
Please understand, Numerator only appears to be same for both BUT it is actually not. Please note the numerator CBR is TOTAL NO. OF ALL LIVE BIRTHS in a region (every single live birth is accounted for) but for TFR only live births selected in selected sample of households (and not every single live birth region)
Now let us look at the denominator side.
While denominator for CBR is all population in given region (its a simple frequency count of every individual in the region) the story is finished here for CBR.
Where as for TFR denominator COMBINATION of inter play of - Total count number (1) AND OTHER charecteristics of population in denominator (2nd, 3rd & 4th point).
1. Total count of individual in denominator ( like denominator for CBR)
2. proportion of women in different age groups from 15yrs to 44years
3. Mean Age at marriage of these woman in denominator.
4. birth order in the community
There fun side to denominator for CBR is that all population means( ALL MEN+WOMEN) in the community Now, please note CBR denominator is in fact inclusive of denominator of TFR -sub group WOMEN in 14-44 yrs age group is also included.
Because propoortion (%) of this 15-44years women in Total population will be different in every state, so also Mean age at marriage of these women and so is birth order. Every state will show different permutation and combination
EVIDENCE FOR INDEPENDENCE BETWEEN CBR and TFR
To help your self to put the whole story in perspective please do this home work
Compare TFR of TN and Kerala. CBR of Kerala is low but has High TFR on the other hand. In TN has High CBR but Low TFR. Thats the evidence for you on why TFR and CBR show no coherence
HOWEVER if you follow golden rule in epidemiology ALWAYS compare LIKES WITH LIKES you will spared of many misgivings. Example – you are most welcome to compare CBR of nagaland with CBR of India (but as I said standardize the rates first then compare), so also you can jolly well compare TFR of nagaland with TFR of India (as mala suggested here standardization is not required because TFR is any way ready made age specific rate and is not crude like CBR).
BUT never make mistake of cross comparison of CBR of one NAGALAND with TFR of India. It is absurd.
Hope that helps.
(And also hope that expert demographers like Mala Ramanathan will find this technically right as well. if not, i am willing to stand corrected with all humility- since demography is not my subject and my efforts here are simplyfying in lay terms- i may have missed out out on finer nuances of technicality)
Respondent : My take (from a non-demographer's perspective) -
If you compare CBR and TFR from NFHS 3, there is hardly any diversity of the kind that worries Narendra - the relative variability for the two indicators across states is similar. Please see attached graphed data (attachments not available on the mfc e-group) - and please note the different y-axis scales for the two indicators. What then is the diversity that we are discussing? The "low" and "high" TFRs and CBRs that we are discussing are relative to what?
There may still be two "facts" that intrigue:
1. the difference between the 2005 (NFHS3) CBR of 28.5, and the SRS figure of 17.5 for 2009. I notice also that Nagaland birth rates from SRS do not seem to be available for earlier than 2006. In 2006, CBR there was reported as 13.9 (less than half of the 2005 NFHS figure). For 2007, and 2008 it is reported as 17.3 and 17.4. According to SRS, the CBR is rising, if at all changing. The difference between NFHS and SRS could be because of a number of sampling and non-sampling issues, some already discussed in this debate. I would begin with a close look at the relative methods used for computation of CBR, and if they are the same, look at what the sampling strategies of the two were (remembering that SRS is a "panel"). If the samples of both are reasonably representative of the state, non-sampling errors (data quality issues) are the reason for variability.
2. the issue of large decadal growth rate that Narendra has quoted (I have not confirmed this) - it could simply be that between 1991 and 2001, the population did increase substantially, due to a relatively high birth rates and plummeting death rates then (or simply, more complete enumeration in 2001!)
Re definitional issues - pl check out http://www.measuredhs.com/help/Datasets/Fertility.htm for how these indicators are computed by DHS (NFHS).
- DHS measures CBR also as an age-adjusted rate, but I am not sure how SRS does it. In any case, it is best to compare all fertility rates from the same data set (NFHS), instead of across data sets.
- CBR and Age-Sp FR (from which CBR is computed) are based on a count of births in the three years preceding the sample survey (both are based on sample data, irrespective of source - except when computed from the decadal census). Thus, as computed, they represent the same time period of fertility performance of a population ("last three years"). (TFR is "hypothetical" only in the sense that it purports to represent the fertility history of an average woman, but is not hypothetical if you see it for what it actually is - an age-weighted sum of the previous three years' birth rates)
- CBR would theoretically be agnostic to age - it should not be restricted to ages 15-45/49, as age-sp TFR is. Births to women younger and older would be included. (However, in the case of NFHS/DHS, it probably excludes these groups, because the sampling is from a frame of 15-49 year women - I am not sure how SRS handles this)
- CBR has the entire population for the denominator, not just women
- Any variability between the two indices could therefore arise from any or all of the following variabilities:
-- An unusually distorted sex ratio of the 15-49 age group in the population (affects denominator)
-- An unusually distorted proportion of women 15-49 to the rest of the population (affects denominator)
-- Different non-sampling errors (data quality issues) - if the indices being compared are from two different surveys (if from same survey, this would not be a major variable) (affects numerator/denominator)
But that is only theoretical, if in reality there is not much diversity in the first place.
If TFR is computed using a hypothetically correct 15-49 y population despite overall population inflation - would that not be more valid than CBR? It evidently would be, if your numerators (births) are correct.
I believe, though, that birth reporting is grossly inadequate in the routine reporting system, probably in all states. For instance, we calculated CBR from routine reports from a clutch of districts in AP some years ago, and found it was around 8 - not more than half of expected. Post-JSY, I would expect more complete reporting - but still largely only for JSY eligible births. In other words, the assumption that numerators (births) are correctly reported may be misplaced. In the case of Nagaland, it is possible that the proportion of unreported deliveries is larger than in other states because of the population inflation that you describe. I have no idea about the level of sincerity in recording and reporting pregnancies in Nagaland - if they are "naturally" good at it, maybe the numerators are reasonably accurate.
You also need to somehow confirm population inflation: the 70% under-reporting may be attributable to simply a number of uncovered areas - blind spots with no reporting - in a hilly / forested state with inaccessible areas. For instance, what is the proportion of PHC's / SC's that are reporting the data in the HMIS fact sheet of Nagaland? If that is not the problem, a PHC-level break-up of population compared to census figures should help understand the distribution of the claimed population inflation.
Finally you need to somehow confirm the differential population inflation that you suspect: I do not fully understand how numbers can be increased by over-reporting ages of certain age-groups. A comparison of population pyramids between the census and the NFHS might help in this.
There is another problem with computing TFR from routine data: your would need to know the age of each delivering mother, and the HMIS needs to organize itself to collate the number of births by each age group (15-19, 20-24 etc), so as to first derive Age-sp FR. Without this, TFR would be impossible to compute.
The age-restricted CBR that NFHS seems to use may be a better understood indicator than TFR for monitoring purposes - just a suggestion.
As for the TFR-CBR discrepancy, as I have indicated in previous mails, I believe we saw a discrepancy because we were comparing TFR from NFHS to CBR from SRS. If we look at NFHS alone, there is no discrepancy. The question of why SRS is reporting lower than NFHS remains - need to compare computing methods and sampling methods to understand this. One possibility is that either the NFHS sample of the SRS panel is non-representative.
The HMIS fact sheet is very interesting and informative, by the way. I need to understand more of that - separately.
One final question - how much of the incentives / bonus linked to CBR/TFR is a true story? There are many more outcomes worthy of being incentivized!
Respondent : Data collectors in Jharkhand have been found to do such things during DLHS work. So it is good to find out how strong the quality checks are. You may also remember that we found major discrepancies on a Census sitedespite the original Census data from Pakur being reasonably accurate.
In that sense Sundar's MIS workshop - that deals with how you cross check data (the dog that didn't bark was it?) and Naren's original question are useful
Prabir
Dr. Rajan R Patil :a picture is worth 1000 words- so goes the cliché.
The line graph shared by Sridhar’s essentially communicates the same thing pictorially what demographer said as general concept- CBR and TFR are independent of each other- There need not necessarily have any correlation positive or negative). This thing is well depicted comparative line graphs evidence in states likie UP, Nagaland and Meghalaya.
CAUTION
Wait a minute. Should we accept above arguments just because the statistics and in form this line graph is showing so. For a moment let me be devil's advocate and declare that statistics can be deceptive-ONE CAN’T ALWAYS LIE AROUND WITH STATISTICS -why should we accept this unsourced graph?- We don’t even who were behind preparing this graph?, what statistics were take it into consideration?-what were their exclusion and inclusion criteria in
preparing this particular graph. All these crucial information is not accompanied with the graph. So it can very well deserves to be junked.
If statistics and graphs are our basis of accepting concepts then one can always prepare beautiful grpahs by smartly tweaking inclusion and exclusion criteria of data we are taking in for anaylysis  and can show exactly the opposite that TFR and CBR are correlated - apart from showing this beautifully on graphy- one can even work out quantitatively the Coefficient of correlations between TFR and CBR so smartly and convincingly that IT WILL show statistical significance of P < 0.00000000000000000000000000000000000000000000 to Infinity
Now tell me- shoulde we accept this abonoxious statisical evidence just because is worked out by a maverick genius and who has worked it out all the statisitics on a USA imported advance super computer and latest statistical packages??? The answr is big NO - when we clearly know that there can not be association between CBR and TFR conceptually.
As it is emphasized - statistics should be used for our enlightening and not as lamp post for support as used by drunkards. Statistics should be subservient to us and not we be subservient to subservient to statistics. Else we all will become top floor khaali robots who can churn out numbers at press button without making head or tail of it.
MORALE OF THE STORY
BASIS OF OUR CONCLUSION AND DECISION Â MAKING IN PUBLIC HEALTH SHOULD ALWAYS BE
BASED ON SOUND CONCEPTUAL CALRITY Â WITH STRONG LOGICAL REASONING
I dont want people to believe us just because i and or 10 others are repeatedly saying so. Or just believe it because dumb statistics and numbers are saying so.. Let him appreciate logically why these indicators were invented what they were conceptually meant and what they ommunicate individually. Then look into relationship between indicators.
I would like people accept only if he appreciate concept and logical reasoning behind given by demographer
RESPONDENT: I just happen to be one of those - I suspect there are many - who have not quite followed your explanation about the conceptual link between TFR and CBR, and why they must be independent. I also do not understand your fury - either at author of the graph or at statistics in general.
I happen to be the obnoxious maverick genius who generated that evil graph – but using a simple 2007-make 12-inch laptop, this morning sitting at home, not a supercomputer in the US. It took me about 7 minutes to copy paste the TFR and CBR columns of data from Table 4.3 of NFHS3/India report (p. 83-84) into a ppt slide (in a 2003 version of that most outdated of software - MS Office), and for it to be graphed with two different y-axis scales - I did not even adjust the scale. Most of the 7 minutes was spent finding the table and copy-pasting the columns from the PDF file. There was no statistical manipulation involved. Using two different y-axis scales is the equvalent of multiplying all the TFR values
by 10 to make it visually comparable to CBR values - it does not distort reality.
I did not bother to provide the source reference because I thought it was self-evident in this informal dialogue - the title of the slide says so, and the source data in the slide is available for all to see.
To my eye, as to Prabir's, the two indices match up closely - they are not independent of each other. I do not need a statistical test to tell me that. Since you are convinced otherwise, please do run your tests and let us know how indpendent these two data points are.
The reason why TFR and CBR covary is also simple - which I described in the previous mail. I understand (from the reference link to the Measure-DHS site provided in that mail) that DHS uses Age-Sp FR to compute CBR as well as TFR - the numerators are identical. The little variation you see is because the
denominators vary somewhat for CBR and TFR - which also I described. Thisvariation is probably a reflection of the different sex ratios in the 15-49 age group in different states.
I re-read Dr Jayaraj's and other explanations. They are not inconsistent with all this - they are only saying that if the two indices appear to vary differently, there is nothing inherently unexpected. No one has said they should not covary closely. If you see them in the spirit of what they are supposed to represent, they can certainly vary independently of each other, but the way DHS computes the indices, they are nearly identical.
The main point I was trying to make was not this at all. The point was answering Narendra's original question: "The TFR of Nagaland according to NFHS 3 is 3.7 in comparison to national average of 2.7 At the same time birth rate of the state is 17.5 in comparison to national average of 22.8 as per SRS October 2009. I am unable to understand the plausibility of these two figures. "
He is bothered by the inconsistency between Nagaland and all India figures. What I pointed out is that there is no inconsistency if you looked at both TFR andCBR from one source - NFHS. (The first data point on that unspeakable graph isfor all-India, and you can see that the all-India numbers are not what is
reported by SRS, and the relationship between the two indices for all India is pretty much the same as for most states - or simply look up the NFHS3 report table if this graph cannot be trusted).
Frankly, I had never bothered about this matter until Narendra raised the issue, and then I started looking around to learn how TFR and CBR are actually computed by DHS. I then realized that they should closely covary, so I went back and looked at the NFHS data, and found that they do covary closely. And that the original mistake was to look at TFR from one source and CBR from another. We all make such mistakes, and learn from them.
I would like to be corrected if I am wrong about how DHS computes CBR - and that might open up the debate again. Also, someone knowledgeable might want to comment on how SRS computes CBR.
Dr. Rajan Patil :I had begun my mail with the appreciation of the graph you shared (even unaware of source) because it was very much in sync with what i had to say. I have nothing against numbers and statistics- in fact you must have noted i have used statistical language of numerators and denominators etc (in the appended mail below) So i personally very much love statistics (in fact I would be out of job without it ;). That way be assured we are all on the same wave length.
Subsequently, i was being devil's advocate(i had said so in my earlier mail too (for those who hate statisitics) and was only echoing their likely response to unusual deluge of statistics, numbers and graphs being exchanged in this debate.
But yes, i would still stand by one thing. Statistics is only a tool. If our statistics are not in coherence with the conceptual facts it only means there is anomalies in the data (as we all have agreed and accepted with our demographic data collection)
That being so, In examining the conceptual facts- If our statistics is not in syn with it, we may need to change the data sets (e.g, instead of Indian data - we may choose more valid date from other country) and re-examine the issue, but not vice versa - Never change conceptual facts just because our bad statistics (out of bad data) is not supporting it.
We need to use statistics to validate our conceptual facts But Never use statistics (eg. our NFHS/SRS if we have so many doubt on its quality) and (never) use the same to accept or reject conceptual facts.
Morale of the story - (and my final take on this issue..)
1) We need to agree collectively that quality of our demographic data generated in India is not good enough to test demographic conceptual facts.
2) For clarification on any demographic indicators, or in understanding conceptual relationship between different demographic indicators- Let us revisit our textbooks rather than questioning it with bad data in our hands.
Respondent: I cannot respond to all the questions immediately as I have a deadline to meet.But one question - can there be different ways of computing CBR? According to NFHS 3 the CBR of Nagaland is 28.5 while of the same period i.e. 2006, it is 16.4 according to SRS (SRS bulletin). Any thoughts.
Respondent : Here- to add to the confusion is the HMIS data of 2009-2010 for Nagaland. I have it in this format for all districts and states in the country.
Respondent : That is what I am hoping someone will enlighten us about. Until now, I had assumed CBR is computed as usually described - the number births in a year per 1000 mid year population. The limitations of this are known- comparisons are affected by age/sex distribution of the populations, and byyear-to-year variation. To make estimates more stable, births over several years are taken together, and there are several ways of doing this, all of which give similar results. DHS uses 3-year counts, I have no idea what SRS uses. DHS also seems to use age-adjustment, which is generally not used for CBRs as far as Iknow.
There must be more knowledgeable members on these groups, who could provide more definitive answers.
Just fyi - the April 2006 SRS bulletin reports the CBR as 13.9 for Nagaland (and Oct as 16.4)! Transcription error? Variation in data?
Respondent : Let me join into this discussion with a further set of questions.
I got a call from the Family Planning(FP) officer in Nagaland. He had heard of this debate - he was confused and he was desperate. This is what he asked:
1. " I have been telling everyone that we have achieved population stabilisation But is that true? Or do I have to reduce birth rates much further?" Please Rajan - Please Sridhar - what do I reply?
2. The FP officer continued " If your answer is that - dont worry- you have achieved success- then can I tell my minister- 'that dont worry sir- we have achieved success - but it will take some years for it to show up on the data reports on TFR' . Would that be correct ? "" Please Rajan- Please Sridhar what do I reply.
3. The FP officer continued : "I told my minister that we have achieved success, since you had not replied. But then I went to the Planning Commission for our annual plan presentation with the minister. And they told us that the extra performance based bonus plan allocation we promised if you reach fertility targets will not be given to you. I told them about your reply? And they said- 'listen the bonus was linked to the TFR and not to the CBR and that when TFR comes down we will see? CBR is not good enough? Till then you must work harder and lower fertility rates at a faster rate.' Then what do I reply?" Please Rajan- Please Sridhar help me- How do I reply to this officer?
4. The officer went on: "In desperation I said to the planning commission 'I confess that in the last census many of our tribes had falsely inflated their figures to show a higher population. Part of inter-tribal competition. And therefore such a growth rate?'. But the demographer there was not impressed. He said... 'in that case your TFR should be even higher?' Is that correct ? Or was he pulling a fast one and pulling my leg?' Please Rajan- Please Sridhar- I am getting desperate now. I thought I knew the answers but am all confused now. Tell me - what should I reply?
Respondent :Just in case you are not up to answering my funny questions. Here is the line of thought:
If CBR and TFR represented merely two diffierent ways of measuring the same thing- it would not matter for the programme manager that they are divergent- for the action point is the same viz CBR being the real figure, one need not take action as the birth rate is under control. The programme has clearly succeeded.
2. However if the denominator has been inflated by a process of inflating population figures in the census which I think is what is happenning than the numerator is true and the denominator is falsely increased and therefore the birth rate is misleadingly low. Therefore TFR would be a better indicator in such case. The rub is that I still dont understand what implication the increase in population would have for TFR. If the denominator is drawn from a census estimate of the 15 to 45 women in the population and that was not inflated when the figures of the population were inflated-- since inflation was done( ? ) by increasing the age of older persons or children - then the TFR for the same birth rate would be much higher- as we are seeing.
3.I am not at all sure that this is the way the denominator for TFR is done. If however this hypothesis is correct, it would help me solve the problem of reading the Nagaland HMIS data where unaccounted deliveries are surprisingly high.
That was the Poirot Allusion.
I do suspect that in Nagaland they have figured this out- for they shared the problem of the inflated population with me, only when I was pressing to find an explanation to the HMIS data interpretation problems. I had of course at that time, not noticed the TFR and CBR discrepancy. The point is to look at it from the point of view of what use we put information too. And from that perspective new possibilities open up. That is what I thought in the beginning and kept it to myself. However you have all given such convincing explanations - that I dont quite know whether what i say is consistent with all the rest that has been said and agreed upon.
Now would I get a response. Is there a flaw in this logic?
Dr Rajan R Patil : I think you have certainly hit the holy grail of 'eye of the storm' of this debate by first raising set of questionings through your tet`-e tet` with Nagaland officals and responding to your own questions. That really helped.
Personally i find it most logically coherent explanation to the aparent descrepency pointed out by - which also concurrently and very aptly addressed specific data related questions raised by chinu,dhruv,sridhar in the context of Nagaland and i finally i get what i was looking for- the logical & rationale that could possible explain this deviance from conceptual fact. At least i am convinced.
At this stage, rather than going round and round, i feel that we should get few expert demographers to scrutinize and authenticate all that has been said thus far by us.
I take the initaitive with all your kind permistion to share this entire string with Demographer, demography text book author& mos importantly as a person involved in planning and coordinating demographic data collection activity in India In his capacity as director of IIPS india (1978-92). And also has been involved with NFHS and DHS planning. Incidentally he is our visiting professor of demography at our university.
Dr. Rajan R Patil : Sundar, I really thought your funny questions were largely rhetorical - that you were trying to make the point that real situations are complex and not academic, so I did not respond.
I agree partly with your line of thought:
If TFR is computed using a hypothetically correct 15-49 y population despite overall population inflation - would that not be more valid than CBR? It evidently would be, if your numerators (births) are correct.
I believe, though, that birth reporting is grossly inadequate in the routine reporting system, probably in all states. For instance, we calculated CBR from routine reports from a clutch of districts in AP some years ago, and found it was around 8 - not more than half of expected. Post-JSY, I would expect more complete reporting - but still largely only for JSY eligible births. In other words, the assumption that numerators (births) are correctly reported may be misplaced. In the case of Nagaland, it is possible that the proportion of unreported deliveries is larger than in other states because of the population inflation that you describe. I have no idea about the level of sincerity in recording and reporting pregnancies in Nagaland - if they are "naturally" good at it, maybe the numerators are reasonably accurate.
You also need to somehow confirm population inflation: the 70% under-reporting may be attributable to simply a number of uncovered areas - blind spots with no reporting - in a hilly / forested state with inaccessible areas. For instance, what is the proportion of PHC's / SC's that are reporting the data in the HMIS fact sheet of Nagaland? If that is not the problem, a PHC-level break-up of population compared to census figures should help understand the distribution of the claimed population inflation.
Finally you need to somehow confirm the differential population inflation that you suspect: I do not fully understand how numbers can be increased by over-reporting ages of certain age-groups. A comparison of population pyramids between the census and the NFHS might help in this.
There is another problem with computing TFR from routine data: your would need to know the age of each delivering mother, and the HMIS needs to organize itself to collate the number of births by each age group (15-19, 20-24 etc), so as to first derive Age-sp FR. Without this, TFR would be impossible to compute.
The age-restricted CBR that NFHS seems to use may be a better understood indicator than TFR for monitoring purposes - just a suggestion.
As for the TFR-CBR discrepancy, as I have indicated in previous mails, I believe we saw a discrepancy because we were comparing TFR from NFHS to CBR from SRS. If we look at NFHS alone, there is no discrepancy. The question of why SRS is reporting lower than NFHS remains - need to compare computing methods and sampling methods to understand this. One possibility is that either the NFHS sample of the SRS panel is non-representative.
The HMIS fact sheet is very interesting and informative, by the way. I need to understand more of that - separately.
One final question - how much of the incentives / bonus linked to CBR/TFR is a true story? There are many more outcomes worthy of being incentivized.
Respondent :Thanks for very valuable input of so many friends. But at the end of the day, I am still unclear to in interpreting the huge diversity in two indicators which relate to births in the same population of more or less with the same reference period. I understand that one is 'expected' and the other one is 'exposed' measure and may show differing trends in different
cohorts. But then are these indicators incoherent to each other and can they be used for comparison in different population cohorts as is normally done by us all and by the Govt. of India while ranking different states and deciding financial allocations. My question is simple "Is it possible to have low birth
rate with high TFR and vice versa". Is there any normativity?
Another issue foxing me about Nagaland is that decadal growth rate between 1991 and
2001 census. It is whopping 64.53% while all India is 21.54% (Surprisingly, Punjab's growth rate which has huge inmigration is 21.54% only and its birth rate is also 17.3/1000 population). Again the question is with CBR like 17.5, is it possible to have decadal growth rate of 64.53% in Nagaland where immigration is almost nil (my assumption).
Demographers and others, please enlighten. I am not ready to completely believe What Sunil has said. According to many demographers, data generated by the Census, NSSO, NFHS(IIPS)
India is considered quite credible.
Respondent: I have attached an excel file calculating TFR, CBR of Nagaland actually and hypothetically.
During this calculation Pop and no of births are actual figures (HMIS)I have kept no of births constant. AND no of women of reproductive age(18-21 for simplifying the calculation) and percentage of women of the specific age who delivered also constant.
Later I have changed population only. Doing so, the CBR varies drastically (since the denominator is population but TFR remains constant. Although, the variance may not be so great thats why there is some trend of consistency as Sridhar is saying but scientifically, they cannot be as the figures may swing drastically if there is great variance. Statistical correlation may point toward its probability - after all statistics is THE science of probability. While epidemiology wants accuracy to predict and plan. CBR cannot represent fertility beyond a certain range. Thats what the paper referred to by Chinu is saying.
mfc may not be able to see the attached file. I can send it.
Pls comment if the calculations are correct.
Pop 2931412
Total Births 15282
CBR 5.2
Age group No of women of this age group Out of which who Delivered babies Delivered babies
% of women of this age group
18 5098 1020 20
19 13200 5280 40
20 16105 8053 50
21 3100 930 30
15282
37503 15282 140
0
TFR 1.4
Pop 5931412
Total Births 15282
CBR 2.6
Age group No of women Delivered babies Delivered babies
18 5098 1020 20
19 13200 5280 40
20 16105 8053 50
21 3100 930 30
15282
15282 140
0
TFR 1.4
Pop 1931412
Total Births 15282
CBR 7.9
Age group No of women Delivered babies Delivered babies
% of women of this age group
18 5098 1020 20
19 13200 5280 40
20 16105 8053 50
21 3100 930 30
15282
15282 140
0
TFR 1.4
Dr. Rajan R Patil: I am appending below response from Prof. K Srinivasan one of theleading demographer ofour country who headed IIPS bombay wasinvolved in designing and planningNFHS, DLHS in India.Besideshas written textbook on demography
Dear Dr. Patil,
Any qualified demographer should be in a postion to clarify your doubts. Here is a brief explanation.
Crude Birth Rate ( CBR) is an index influenced not only by fertility levels but also by the age distribution of the population. That is why it is called Crude Birth Rate . On the other hand, Total fertility Rate (TFR) is independent of the age distribution ofthe population and is considered as a pure measure of fertility and is used for comparison of fertilitylevels across populations and over the same population over time . Of course there are more refined measures of fertility but TFR has been used widely as a good measure of fertility.
Coming to Nagaland case that you have suggested, the state hasa smaller proportion of women in the reproductive ages in its populationcompared to the country as a whole ; hence a lower CBR
compared to its TFR levels. In many other populations the differences may be even dramatic.
Consider a population of 1000 persons where there are only 60 women in the age group 15-44 , equally distributed across individual ages , ; 2 women in age 15, 2 in age 16,.... and 2 in age 44. Suppose in one year one woman ineach age give birth to a child , there will be 30 births with age specific fertility rate for each age being 0.5. TFR being the sum of the age specific fertility rates will be 15 ( 30 *0.5) ,On the other hand the Crude Birth Rate will be only 30. ( 30 births per 1000 population). if we reduce the age specific ferility rate for each age to 0.25 the TFR will be 7.5 andthe CBR will be 17.5. Such extreme examples prove the point that the age distribution of a population can considerably distort the CBR measure and hence this measure cannot be
used to study fertility trneds and differentials.
I hope I ahve clarified your issue.
K. Srinivasan,
Emeritus Professor,
International Institute for Population Sciences,
Deonar, Mumbai, India, 400 088
Respondent : Response of Prof. K. Srinivasan establishes the relationship between CBR and TFR. His example shows that high TFR will result into high birth rate. He has neither refuted the Nagaland data nor has he explained that how high TFR of Nagaland has resulted into low birth
rate. Then there is different between CBR calculated by the NFHS andthe SRS.
I am more unclear after reading the response of Prof K. Srinivasan. First he has not cast any doubt over the CBR and TFR figures of Nagaland. He agrees that these are dramatic and can possibly be there. But then what he describes is just opposite of the
Dr.Rajan R Patil :I am really concerned when you make convoluted interpretation from the example and ascribe it to Prof. Srinivasan.
Please note He never said " High TFR will result into High CBR" that is entirely your isreading from his example.. He had given that example only to show that CBR and TFR can be totally independent of each other and there need not any correlation between them. That is exactly what many of us were saying and were convinced about about it- any apparent correlation is only coincidental.
The fact that CBR and TFR are not related and are are infact independent of each other is a basic and fundamental concept which we as public health students have learnt and should be able to explain with reasons and logic with our exisiting elemenatry demography knowlege BUT i still I referred this to Prof Srinivasan only to get independent expert opinion for not only a qualified demogrpaher BUT more importantly because he was in thick of action in planning, designing and coordinating of Indian demographic and health data as longest serving director of premier insititution IIPS bombay, that is the nodal agency for any national data collection over four decades (lest even he would dismmissed as theoretical academician not knowing what happens in the field visa vie abut data quality, data fudging etc)
Respondent : Rajan,u keep emphasising the fact that CBR and TBR are "completely unrelated"!! i find this most ridiculous. they are not directly related mathematically, but they are throwing light on the same public health pehnomena in the same population. pls do pot go overboard and throw out babies with the bath water. public health is not supposed to be some mystical science that defies all reason and logic and common sense. ofcourse there is correlation between CBR and TFR which is not a statistical correlation and depends on a host of other dependent variables such as sex ratio, percentage of women on repr age group, sampling mehtods etc etc as every one (inlcuding narendra) has been saying. but to say they are "completely unrelated" is very strange.
from my readings, none of the highly qualified people on this discussion have said that.
Dr.Rajan R Patil: Vandana, you said...".....ofcourse there is correlation between CBR and TFR which is not a statistical correlation and depends on a host of other dependent variables...."
Thats where the catch is...
Any genuine correlation between two variables should necessarily show statistical significance for it to be basis of our public health decision making.
Any observed correlation which is statistically non significant can be explained by
1. Spurious Association ( selection bias, sampling method, data quality/fudging)
2. Chance (random variability)
3. Co-founding effect (Age, Sex composition)
Whatever little observed correlation between CBR & TFR observed is because of confounding effects i.e, co-variation of third factor as you rightly pointed eg., sex ratio, percentage of women on repr age group, sampling mehtods etc.
In that case, this debate is needless because :
1. Normally if data is well collected (valid and reliable) then CBR-TFR WILL BE STATIFICALLY NON-SIGNIFICANT.
2. Any partial/observed/apareent statistically non significant correlation observed in CBR-TFR only indicate weak associations between diverse variables which is of only theoretical interest- (for making multiple conjectures-as we are currently doing)
3. Public health decision making can not be based on conjectures as we can not plan our intervention on something that can not demonstrate credible cause-effect relationship with strong correlation(statistically significant) - (after ruling out Chance and confounding effect). That would be against tenets of Evidence based Public health .
Respondent :the point is also that we would be very poor public health experts if we did not ask questions based on discrepencies that may or may not be statistically significant after we have finished our exploration. narendra did well to find some data that seemed incongruous (i havent heard any one question what is going on nagaland before). he was also perfectly aware of the types of answers given by people (we had a telephone covnersation before he posted this question). these indicators are not some sacrosanct things - they are to help us to understnd a situation in different ways. the phenomena 'fertility' and 'birth' surely are interrelated; sorry i cannot accept that they are not, since fertility is not currently being understood as how many eggs and sperms are produced - it is still being measured by babies. therefore their measures must be interrelated. we use different methods to understand the different aspects of the same picture, and though we do not capture the 'truth' by any of these, we do expect that all our methods should broady lead us to some logic. what the experts have told us merely this - in their experience, the discrepency we have come across is mathematically possible because of the various reasons we have talked about. may i submit that there is a statistical relationship? i would ask the mathematical question, suppose the trend in CBR was steadily upwards and the trend in TFR be steadily downwards, would not the difference between the two graphs, at some point become statistically significant? till what point would they continue to consider the incongruity a mere statistical anomaly? someone (lay person like me) may answer this question by saying - it is not possible for this to happen - and the only way it becomes possible is if the two are related - they are positively correlated, or SHOULD be. and if they are not, there is something wrong and they are not helpful indicators.
anyway. i think what is more to the point is that i do not agree with your denigration of the question itself. i maintain it is worth asking and worth thinking whether one should be concerned that one or the other is off. better that than thinking they are not inter related, and therefore it is silly to even ask the question. but if you choose to continue with what i consider a fairly nit-picking appraoch to catching hold of a word here and there and attacking its originator in fairly personal, insulting and agressive terms - well that is your prerogative.
Respondent: I am not an demographic expert - but then very few amongst us are! I was a meer computer programming myself due to suggestions from a few experts.
Our demographer friend has pointed out with examples similar to one I had estimated. The contention is that TFR is a projected indicator - what may happen. CBR is what actually happens.
In this context, it would have been more fruitful if we get an expert opinion about Nagaland data rather than theoretical one. SO to that extent narendra's questions still need to answered.
I am sure all the ifs and buts issues raised have authentic census data eg age wise sex distribution etc.! Rajan, can we work on both the formulae with Nagaland data (not sample - SRS.NFHS3)
Respondent : gain if the birth rates all moved up together there would be an apparent link. Most important, if the real population pyramid resembled the ideal one CBR and TFR would be directly related. So I think we need to understand numbers and statistics more broadly. The tool of Hill's postulates apply to the study of hidden cause and effect relationships. They are useful tools. You have analysed this question using them and they nearly fit. We miss larger math relationships if we use a butter knife for dissection.
Respodent : let me try once more.
Look at the figures for Nagaland and India:
NFHS 3:
Nagaland: CBR: 28.5, TFR: 3.77
India: CBR: 23.1, TFR: 2.68
SRS, 2006:
Nagaland: CBR: 13.9 (April), 16.4 (October)
India: CBR: 24.1 (April), 23.8 (October)
I understand the following from the above:
1. Within NFHS, the relative proportions of TFR and CBR are fairly close - there is only a small inconsistency, probably explained by Prof Srinivasan's explanation that the crude rate is lower relative to the TFR in Nagaland as compared to all of India because there is a smaller proportion of women in the reproductive age in Nagaland as compared to India.
2. There is probably an error in the way Nagaland SRS data was collected / computed in 2006 - there are no such fluctuations in the CBR of other states between April and October of 2006. This could mean that there is something wrong in the data from Nagaland.
3. The only major discrepancy is when you compare SRS CBR to NFHS TFR for Nagaland and India and see divergent trends. This is what made you say " ... unless we assume that there are extremely very few women in Nagaland and producing on an average 3.7 kids in their reproductive life time". There is no way to solve this "puzzle" since there is no TFR data available from SRS - and going by NFHS data alone, the discrepancy seems to be small enough to be explained as above. I don't think there is a way to compute crude CBR from NFHS either. So, I am afraid the existence of a puzzle is difficult to establish. We are just looking at data that are not good enough for comparing. Or, either the NFHS or the SRS data is plain wrong.
Disclaimer: this has nothing to do with Rajan's passionate claim that TFR and CBR are unrelated!
Respodent :Please find two more papers with regard to relationship between
CBR and TFR besides what Chinu had sent.
The link of first one is:
http://books.google.co.in/books?id=-
p2OIDPr5jkC&pg=PA228&lpg=PA228&dq=Relationship+between+CBR+and+TFR&sour
ce=bl&ots=Xh36QBWSvB&sig=21u42ONeLqdGa6ie6luURymKcI4&hl=en&ei=4ZtPTcjRK
MfIrQf1vJDaBg&sa=X&oi=book_result&ct=result&resnum=4&ved=0CCcQ6AEwAw#v=
onepage&q=Relationship%20between%20CBR%20and%20TFR&f=false
Second is the Family Planning DemographicIndices. It is available on
the NIHFW web site. I am attaching a copy of it. Please read
highlighted part on page 20.
Third, I wish if somebody could unravel the Nagaland puzzle because
demographer example does not help unless we assume that there are
extremely very few women in Nagaland and producing on an average 3.7
kids in their reproductive life time.
Dr. Rajan R Patil: I wanted to reassure you am treating this discourse purely at technical plane – absolutely nothing personal!.
Since many of our learned friends are convinced and subscribe to your view and hypothesis, there must be some truth in it. I find it little difficult to relate to it as it would require me to unlearn some of the very fundamental concepts that form the building blocks on which entire discipline of demography and to large extent public health stands. So before we rock the boat from the bottom, let us get our arguments peer reviewed. I have done it earlier got my line of thougth verified.
Now I call upDon you and other colleagues who share your line argument to get the rationale and logic behind your/their respective argument peer reviewed. Here I don’t mean publication in journal that’s two year process, but at the least, corrobation by atleast one expert demographer who is willing to endorse your views and vouch for it publicly. I will be happy to relearn every thing. We are talking science, where nothing is sancro-sanct, it has to keep evolving, sometime even turning upside down.
Let us remember, Earth would have continued to be flat and been centre of universe until somebody questioned by setting the logic right.
At personal relationship front I generally maintain..
Kabira (me) Khara Bazaar Mein, Mange Sabki Khair
Na Kahu Se Dosti, Na Kahu Se Bair
So no offence! I will be happy to be part of this churn around, will always be proud of my association with you. I have enjoyed and continue to cherish my memories of your warm hospitality during my visit to chittorgarh/prayas, same I would like to reciprocate when you hit the southern shores..
Respondent :No love lost, Rajan. Your erudite comments frenzied debate - a commoner like me is enriched. Seems from many postings that this has been discussed even earlier and some algorithms (spurious?) worked out – refer the papers Chinu and I cited earlier. Dr. Srinivasan in his response wrote “the age distribution of a population can considerably distort the CBR measure and hence this measure cannot be used to study fertility trends and differentials.� But prior to making this comment he showed us how ASFR and TFR influence CBR… there could be dramatic results and then his conclusion. If NFHS & SRS data about Nagaland is correct then we should assume that there are few women in reproductive age group with high fertility – Dr. Srinivasan corroborates.. We do not have
access to life tables of Nagaland but other evidences do not support this conjecture. Do we accept the anomalies of record keeping … Sunil described. Perhaps, better would be to get hold of life tables.
Now, what inference one can draw from Nagalandâs reported CBR & TFR is population of the state booming (far away from replacement fertility) or perhaps not growing in proportional manner owing to lesser CBR.
I fondly recall your visit and we all are “Kabiras� in this wonderful group
of ours.
Correspondence:
Dr. Rajan R Patil
Division of Epidemiology
School of Public Health
SRM University
Potheri, Kattankulathur - 603203
Greater Chennai,
INDIA.
Cell: 9445811610 & 9025378036
Email : rajanpatil@yahoo.com, rajanpat@gmail.com
Subscribe to:
Posts (Atom)