Early Life Circumstances and Healthy Aging in China (Xi Chen, Yale University)

Early Life Circumstances and Healthy Aging in China (Xi Chen, Yale University)



good morning everyone my name is Chen Qi from Yale University thank you very much to Professor winging it for organizing such an important pre-conference workshop on house care reform in China today unfortunately I'm not able to come but I would like to use the last 15 X to 15 minutes to present some of our findings on the past and the present patterns of early life environments and healthy aging in China and some future trends so I would like to also stay on line for Q&A if time allows or otherwise I'd be happy to answer any questions or receive any comments from you halation aging has been very important in China especially in the last couple of decades China is aging very fast in terms of number of elderly and the sheer size of the elderly population and the health of the older people were defined we're being of a population add as it affects not only their other people's themselves but also their next generation and their families through caring needs and the house of the older people also determines the health care costs and Social Security expenditures and also challenge the Chinese economy if you look at this two figures plotted using the China family piano starting which allow us to look at the some dimensions of house over age and the different birth cohort you will find not very promising trend for example the left of figure left the panel shows that for most the ages the later bone cohort had worse house mental house status as measured by the CESD mental house measures and also if you look at the right panel the mental house decline the foremost of people the most ages over time if we come he'll the singing EB Jews who were surveyed in 2010 and come held to the same individual on 2014 you will say the decline so is post declining over age and birth cohort and the same thing happened for self rated house so the China family panel study also allows us to look over a wide range of age cohorts and also for four years apart for the same individuals for each other we also find that declining self-rated health over age and the first cohort the picture is a slightly different for cognition cognitive house for mathematics skills as surveyed by the tested in the china family panel sorry we still find a little bit decline over ages a large decline probably related to the education background and also is the decline over first cohort so the younger per school heart is not doing better than the older cohort which is worrisome but that picture is like over longer age cohort but do not necessarily give us a view of linking the same individuals over the life course so in order to do so some studies looking at for example free to original hypothesis which shows that aging may start in the warm and in utero is a very critical period during which the worse in circumstance made their long-term consequence and there's a large number of studies documenting this and we did a thorough review of the literature and we find that the great Chinese famine has been the most important types of work being done using this as an early shocking prenatal period or postnatal period and people find the consistently that house in later life measured by disability obesity hypertension depression reduced the cognition and cardiovascular disease and all this later life house declining those are very consistent but great relief a means of one exception we don't find a much other studies using credible natural experiments in China to link the two ends of the life course and this he in scenes is very important I mean linking the early life environments and the old age house because not only from the Sharks as we mentioned the house shots but also from the social justice the WHL commission on social determinants of health highlights the row of childhood environments as a primary source of unfair house inequality and in essence is special issue in 2017 also had discussed the house equity and equality as related to the early stage of life and how that widen house radiant over the life course and also there's political salience about this and the Franklin Roosevelt and we're early these also mentioned that individual ability cannot be equalized but we should insists of trying to equalizing the opportunity at the beginning of life so this is related to social justice social equity and Anand Krueger also argue that rising inequality in the United States has been transmitted to unhealthy division of opportunity or late stage of life and which is harmful to economic groups so we have a number of studies trying to looking at the sources of housing quality over life course we try to distinguish two types of housing quality one is due to the childhood circumstance which the children or even the fetus had no control over that which is something critically important for policy makers public policy to address because this is all out the room of the choices of the individuals but the other inequalities due to the efforts for example smoking lifestyle choices so if we miss allocating more resources to the efforts to address that that may need a more distortion e-house behavior and lifestyle over the life course and generate the larger social cost so we really want to understand how much the the health that the childhood circumstance may contribute to long term health especially in all the age and we follow Joan Rivers approach with who is my colleague Yahoo economics department and he is try also trying to to measure the inequality due to childhood circumstance he argues that we really need to like compensates this using the room the limited government resources so in the literature we find that there's all kinds of studies linking the single childhood factors and the later life all caps in different countries there's a little story concurrently combining all these factors to look at a comprehensive picture and how much the overall may contribute and manifest as health inequality in old age so that's what we try to contribute we link the people in two ends of life over of 50 or 60 years apart from each other and we measure comprehensive measure of childhood citizens and and also measure multiple dimensions of health we come he'll commercial measures and machine learning measures to mitigate overfitting issuing the commission of masters and we impose very little structure to the data and let the data tell which is the best model and we look for the best prediction and forecasting model and we use the data helps retirement a survey and the Chinese longitudinal survey of the health and aging and the retirement did the chose to look at those two countries because they were different in many dimensions so we linked for each country survey we linked the panel data of later measured health in old age and their life history survey back to their age five each one age ten and for China and for the US and we measure a number of more than a hundred of childhood circumstances and by age five and we categorize them into seven domains according to the literature and especially the word development a record a report in the 1998 and they reevaluate house in the dimension of self-reported health mental health physical house and the cognitive house and this shows an example of our decomposition and using machine learning the regression tray measured for the physical house we measure physical house by the phreatic measure which is the index of functioning of the body and so we did this for China in the US and the regression trade can tell us the most distinguishable factors that drive the inequality in old age as attributable to childhood and it's like growing a trade the branch at the top shows the most indistinguishable factor so one interesting finding is that for both China USA parental health and health over childhood are commonly very important so housing or early stage of life can transmit into a later life and for China along what's very important is that the regional and rural urban status at the purse are really important that means where were you born in China rural or urban which region may predict what your health would like to be in old age this is very different for the u.s. there's no such findings for the u.s. sample and for the US was more important is the family switching nama status especially raised ethnicity and also early education at the household indicated by the number of books at home so early life cognitive inspiration really helps the individuals to develop and into their old age and shows the divergence in their house in old age and we did you also use another machine learning masters the random forest to look at because Rena Forrester give us a render section of your choice with either 200 trees which form a forest and which can make us for use of the data and we basically find the the same results there and we also did a lot of sample perform the estimates and we use the training sub sample to join the the mass the model and we validate in another nine sub samples and to find the pasture prediction model and overall if we find that the random forest model outperform all other models and which can improve the prediction accuracy by up to 30% so which is very promising not only consider about the outcome we measure but potentially some diseases like dementia we really do not have a effective cure so early prevention and the prediction of risk is really important and using the best model we find that the childhood circumstance may explain 26 to 38 percent of healthy inequality in China in old age and for the u.s. Sampo is even bigger share that means if we can avoid those the this divergence in in the childhood circumstance we can address like one third of the inequality in the old age so which is very surprising and very large salient and we also find that if we use a traditional method by controlling for all these factors and decompose it will largely overestimate so over estimation is not good because it may least lead the policymakers to say that we really have such a large childhood circumstance to address so we really need more accurate estimation so the machine learning masters shows is very promising application and the last today we want to really want to emphasize this is not causal study in the future we need to understand the mechanism whether the this inequality in later life is driven by childhood circumcenter directly always through the intergenerational transmission of unhealthy lifestyles something some different channels we need to further distinguish in order to otherwise public policy and lastly I want to alert the audiences here is we are looking at the life course study so people who are starting now already 65 or older so in the last few decades we all know that there's a rising inequality but it's not affected in factor in our analysis so countries with greater economic in ecology may tend to be countries inequality of opportunity and the disadvantage of being passed on from parents to children so we already from this finger shows that even if we divide our sample for the elderly to different age cohort we already find the younger cohort actually those 45 years to 49 years old they already have a bigger portion of inequality at that stage being attributable to the younger age to the to the childhood even if they have a narrow time window between the childhood and the current stage of life so this is very alarming because with inequality being going up there can be even more inequality in house in old age can be attributable to childhood and the finally China has been doing something to trying to emphasize this life course house in the healthy China to 2013 national blueprint we already say that there's one bigger where big major transition from disease episode to whole life cycle so that means the government wants to manage better the population house from cradle to grave so this is what I want to share thank you very much

Add a Comment

Your email address will not be published. Required fields are marked *