Data quality in jamaica

2,825 views

Published on

ABOUT THE AUTHOR

Paul Andrew Bourne is the Director of Socio-Medical Research Institute, Jamaica. He has co-written monographs on Corruption, Political Culture in Jamaica, Other subjects, and authored books on Growing Old in Jamaica, Analyzing Quantitative Data, Understanding Health and Health Measurement, and Sexual Expressions in Jamaica. Dr. Bourne has authored and co-authored plethora of journal articles on health status, health measurement, sexual and reproductive health, and ageing matters. His works have been published in top journals, and recently his thrust has been on data quality in national surveys, particularly in Jamaica.

Published in: Health & Medicine
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
2,825
On SlideShare
0
From Embeds
0
Number of Embeds
2
Actions
Shares
0
Downloads
15
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Data quality in jamaica

  1. 1. Data Quality in Jamaica Paul Andrew Bourne
  2. 2. Data Quality in Jamaica Paul Andrew Bourne i
  3. 3. First published in Jamaica, 2011 by Paul A. Bourne© Paul A BourneISBNAll rights of this book are reserved. No part of this publication may be reproduced (electronicallyor otherwise), stored in retrieval system, or transmitted in any other form (photocopying,recording or otherwise) with the prior permission of the publisher. ii
  4. 4. TABLE OF CONTENTS pageList of Tables vList of Figures ixPreface xAcknowledgement xiiDedication xivPART I: HEALTH STATUS: USAGE OF HEALTH DATA 1 Introduction 1 A theoretical framework of good health status of Jamaicans: using econometric analysis to model good health status 5 2 An Epidemiological Transition of Health Conditions, and Health Status of the Old- Old-To Oldest-Old in Jamaica: A comparative analysis using two cross-sectional surveys 26 3 Self-evaluated health and health conditions of rural residents in a middle-income nation 56 4 Disparities in self-rated health, health care utilization, illness, chronic illness and other socio-economic characteristics of the Insured and Uninsured 83 5 Variations in social determinants of health using an adolescence population: By different measurements, dichotomization and non-dichotomization of health 113 6 Self-rated health status of young adolescent females in a middle-income developing country 140 7 Health of females in Jamaica: using two cross-sectional surveys 159 8 Health of children less than 5 years old in an Upper Middle Income Country: Parents’ views 179 9 Health of males in Jamaica 204 iii
  5. 5. PART II: ERRORS IN DATA Introduction 230 10 Dichotomising poor self-reported health status: Using secondary cross-sectional survey data for Jamaica 232 11 Paradoxes in self-evaluated health data in a developing country 253 12 The validity of using self-reported illness to measure objective health 278 13 The image of health status and quality of life in a Caribbean society 298 Paul A. Bourne, Donovan A. McGrowder, Christopher A.D. Charles, Cynthia G. Francis 14 The quality of sample surveys in a developing nation 317 Paul A. Bourne, Christopher A.D. Charles, Neva South-Bourne, Chloe Morris, Denise Eldemire-Shearer, Maureen D. Kerr-CampbellPart III: DATA QUALITY 15 Practices, Perspectives and Traditions 349 iv
  6. 6. List of Tables pageTable 1.1.1: Good Health Status of Jamaicans by Some Explanatory Variables 22Table 1.1.2: Good Health Status of Elderly Jamaicans by Some Explanatory Variables 23Table 1.1.3: Good Health Status of Middle Age Jamaicans by Some Explanatory Variables 24Table 1.1.4: Good Health Status of Young Adults Jamaicans by Some Explanatory Variables 25Table 2.2.1. Socio-demographic characteristics of sample 43Table 2.2.2. Self-reported illness by sex of respondents, 2002 and 2007 44Table 2.2.3. Self-reported illness by marital status, 2002 45Table 2.2.4. Self-reported illness by marital status, 2007 46Table 2.2.5. Self-reported illness by Age cohort, 2002 and 2007 47Table 2.2.6. Mean age of oldest-old with particular health conditions 48Table 2.2.7. Diagnosed Health Conditions by Aged cohort 49Table 2.2.8. Self-reported illness (in %) by health status. 50Table 2.2.9. Health care-seeking behaviour and health status, 2007 51Table 2.2.10. Health care-seeking behaviour by health status controlled for aged cohort 52Table 2.2.11. Logistic regression on Good Health status by variables 53Table 3.3.1. Demographic characteristics, 2002 and 2007 65Table 3.3.2: Self-reported health conditions by particular social variables 67Table 3.3.3. Health care-seeking behaviour by sex, self-reported illness, health coverage,social hierarchy, education, age and length of illness, 2002 and 2007 69Table 3.3.4. Stepwise Logistic regression: Social and psychological determinants ofself-evaluated health, 2002 and 2007 71 v
  7. 7. Table 3.3.5. Stepwise Logistic regression: R-squared for Social and psychologicaldeterminants of self-evaluated health, 2002 and 2007 72Table 4.4.1. Crowding, expenditure, income, age, and other characteristicsby health insurance status 102Table 4.4.2. Health, health care seeking behaviour, illness and particulardemographic characteristics by health insurance status 103Table 4.4.3. Age cohort by diagnosed illness 104Table 4.4.4. Illness by age of respondents controlled for health insurance status 105Table 4.4.5. Age cohort by diagnosed health condition, and health insurance status 106Table 4.4.6. Logistic regression: Explanatory variables of self-ratedmoderate-to-very good health 107Table 4.4.7. Logistic regression: Explanatory variables of self-reported illness 108Table 4.4.8. Logistic regression: Explanatory variables of health care seeking behaviour 109Table 4.4.9. Logistic regression: Explanatory variables of self-reported diagnosedchronic illness 110Table 4.4.10. Multiple regression: Explanatory variables of income 111Table 4.4.11. Logistic regression: Explanatory variables of health insurance status 112Table 5.5.1: Demographic characteristic of studied population 134Table 5.5.2: Particular demographic variables by area of residence 136Table 5.5.3: Logistic regression: Variables of antithesis of illness amongadolescence population 137Table 5.5.4: Logistic and Ordinal Logistic regression: Factors explainingself-reported health status of adolescents 138Table 5.5.5: Self-rated health status and antithesis of illness 139Table 6.6.1: Descriptive analysis of variables of target cohort 157Table 6.6.2: Socio-demographic and psychological variables of self-relatedhealth status of the sample 158 vi
  8. 8. Table 7.7.1. Sociodemographic characteristics of sample by area of residence, 2002 and 2007 174Table 7.7.2. Self-rated health status by self-reported illness, 2007 175Table 7.7.3. Self-rated health status by income quintile, 2007 177Table 7.7.4. Self-reported diagnosed health condition by per capita income 178Table 8.8.1. Socio-demographic characteristic of sample, 2002 and 2007 196Table 8.8.2. Health status by self-reported illness 197Table 8.8.3. Health status by self-reported diagnosed illness 198Table 9.9.1. Sociodemographic characteristics of sample, 2002 and 2007 222Table 9.9.2. Health status and self-rated illness 223Table 9.9.3. Predictors of poor self-reported illness by some explanatory variables, 2002 224Table 9.9.4. Predictors of not self-reporting an illness by some explanatory variables, 2007 225Table 9.9.5. Model summary for 2002 logistic regression analysis 226Table 9.9.6. Model summary for 2007 logistic regression analysis 227Table 10.10.1. Socio-demographic characteristic of sample 249Table 10.10.2. Very poor or poor and moderated-to-very poorself-reported health status of sexes (in %) 250Table 10.10.3. Odds ratios for very poor or poor and moderate-to-very poorself-reported health of sexes by particular variables 251Table 10.10.4. Odds ratios of poor health status by age cohorts 252Table 11.11.1. Socio-demographic characteristic of sample by sex of respondents 273Table 11.11.2. Socio-demographic characteristic of sample by educational level 274Table 11.11.3. Socio-demographic characteristic of sample by self-reported illness 275Table 11.11.4. Stepwise Logistic Regression: Good self-rated health statusby socio-demographic, economic and biological variables 276 vii
  9. 9. Table 11.11.5. Stepwise Logistic Regression: Self-reported illness bysocio-demographic and biological variables 277Table 12.12.1. Life expectancy at birth for the sexes, self-reported illness, andmortality, 1989-2007 292Table 12.12.2. Life expectancy at birth of population and sex of children byself-reported illness 293Table 13.13.1 Demographic characteristics of sample for CLG and JSLC, 2007 312Table 13.13.2 Quality of life and health status by gender of respondents, CLG and JSLC 313Table 13.13.3 Quality of Life and health status by area of residence, CLG and JSLC 314Table 13.13.4 Quality of life, health status and standardized health status 315Table 13.13.5 QoL by economic situation of individual and family, CLG 316Table 14.14.1. Health and curative care visits: 2000-2007 344Table 14.14.2: Proportion of Survey (Sample) vs. Proportion of Population 345Table 14.14.3. Descriptive characteristic of samples: Sub-national and National surveys 346Table 14.14.4. Characteristic of samples: Sub-national and National surveys 347 viii
  10. 10. List of Figures pageFigure 2.2.1. Diagnosed health conditions, 2002 and 2007 54Figure 2.2.2. Self-reported illness (in %) by Income Quintile, 2002 and 2007 55Figure 7.7.1. Mean scores for self-reported diagnosed health conditions, 2002 and 2007 176Figure 8.8.1. Mean age of health conditions of children less than 5 years old 199Figure 8.8.2. Health status by Parent-reported illness (in %) examined by gender 200Figure 8.8.3. Health status by parent-reported illness (in %) examined by area of residence 201Figure 8.8.4. Health status by parent-reported illness (in %) examined by social classes 202Figure 8.8.5. Health status by health care-seeking behaviour 203Figure 9.9.1. Mean age for males with particular self-reported diagnosed illness 228Figure 12.12.1. Life expectancy at birth for the population by self-reported illness (in %) 294Figure 12.12.2. Life expectancy at birth for female by self-reported illness of female (in %) 295Figure 12.12.3. Life expectancy at birth for male by self-reported illness of male (in %) 296Figure 12.12.4. Mortality (in No of people) and self-reported illness/injury (in %) 297 ix
  11. 11. PREFACEFor centuries, academics, researchers, government agents and policy specialists have relied oncross-sectional data, results and statistics from International Agencies (World Bank; WorldHealth Organization, WHO; United Nations, UN; International Labour Organization, ILO; etcetera), Statistical Institute of Jamaica (STATIN) and Planning Institute of Jamaica (PIOJ) aswell as reputable Universities (Oxford, Cambridge, Harvard, Yale, Stanford, University of theWest Indies, et cetera). The fundamental assumption is that the quality of the data is high,reliable and accurate for usage. Since 1989, STATIN and PIOJ have been collecting self-reported data from Jamaicans to guide and formulate policies. The data are published in theJamaica Survey of Living Conditions (JSLC). Although the JSLC is a collection of results from amodified questionnaire of World Bank’s Living Standard Household Survey, academics,researchers and governmental agencies have been using the data, there is a fundamentalassumption that the data quality is reliable, valid and accurate for usage. Relying on anassumption of data quality is unscientific, non-verification, cannot detect and correct errors. One of the basic tasks of demography is the production of reliable demographicestimates. Despite the available demographic tools available to demographers, epidemiologists,and statisticians, they have been using Survey Data published by the STATIN and PIOJ, withoutdata quality verification (ie. Content Error Testing). Data quality in Jamaica may be good (ie Census and JCLC), but this is based on lowcoverage errors. There are two main types of errors in data, coverage and content, but muchattention has been placed on coverage errors examinations. Coverage errors refer to thecompleteness of inclusion of people or events in a sample. This error can be rectified throughbetter sampling selection, sampling frame, which has been done for the selection of samples forthe JSLC. The gradual development and consistent updating of sampling frames, from which thepeople are drawn for the JSLC, reduced the coverage errors from identification, modification andrectification. Thus the statistical methods relating to coverage errors have been utilized asrecently as on the 2007-2009 JSLC, making the errors lesser and increasing the completeness ofthe sample. x
  12. 12. Demographers and Epidemiologists are concerned about pursuing reliable data in order toincrease the quality of their estimates. As such, they evaluate the ‘Content’ of the collected data,to identify and correct any ‘Content Errors’. This is performed using matching census recordswith records from surveys, as apart of the data quality verification and reliability process. In aneffort to correct errors in age data, demographers (such as Preston, Elo, Rosenwaike and Hill;Caldwell; Ewanks) have used matching studies to assess content errors, testing the consistency ofthe data. The assumption here is that data are not of a high quality because they have beencollected from the source(s). The same holds true in Jamaica. It is within the aforementionedcontext that we must examine the quality of surveys, censuses, and other data collection methodsin Jamaica and not hide behind tradition, credentials, status and past reputation. By acceptingthat data are of a high quality denotes that we are failing to continuously utilize science in thepursuit of truth as truth is not constant over time (or indefinitely continuous). This volume is designed primarily to clarify the quality of sample survey data in Jamaica,particularly the Jamaica Survey of Living Conditions (JSLC). Science is about inquiry, whichmeans that it can be used to question the cosmology and foundations of current epistemology.The JSLC publishes collected data on different issues reported by Jamaicans, suggesting that theestimates from this could be incorrect, unreliable or of low quality without content verification.Quality is data is critical to the quality estimates, indicating that low quality data can result inerroneous findings (or estimates). The gradual development of health science cannot rest on thepillows of unsubstantiated data. It is this unscientific and crucial assumption that can createfundamental flaws in policy formulation and intervention programmes. This book recognizes thelikeliness of such a situation and seeks to evaluate the content of health data, because theimportance of the health is critical to national development and so cannot be felt to unverifieddata. Readers who seek supplementary coverage of areas which are in this volume can reviewodds ratio, confidence interval, multivariate analysis (logistic and multiple regressions),theoretical and conceptual framework, as these will provide more information on technical issuesused in this book. The majority of the chapters were taken from publications in different journals. All thechapters were carefully selecting in keeping with the general theme and focus of the volume, xi
  13. 13. “Data Quality in Jamaica’. Initially the materials appearing in these pages were rehearsed in agraduate class in Public Health at the University of the West Indies, Mona and with otherscholars in health sciences. Chapters 12 and 13 were co-authored with other Caribbean andInternational scholars, aiding in the coverage of the material and the scope of the volume. All theother chapters were solely written by the author. Paul Andrew Bourne xii
  14. 14. ACKNOWLEDGEMENTThe pursuit of science cannot rest on unsubstantiated (or unverified) data. Science is about thepursuit of truth, which denotes that nothing is with verification. Facts cannot be established onunverifiable information (such as myth, tradition, customs, religious cosmology), but it about isreaching out to establish truths that are based on logic, gradual development, reliability,generality, and validity. Thus, the use of health data cannot rely on tradition, authority, andcredibility as the health affects development, which makes it reliability. Effective policies cannotbe fashioned around inaccurate and lowly reliable data as this will void the cost of datacollection. While science is a gradually developed with trial and errors, verification of data paramountthe final results. Thus, quality data is crux upon which science holds its value. As the quality ofthe data collected holds more of the depth of the scientific estimates than the logic and otherscientific approaches. For decades (from 1989), in Jamaica, we have been using survey data,relying accuracy of the data collector and institutions. This denotes that while we advanceestimates and fact from the data, there exists a scientific unanswered question “How is the dataquality of survey, particularly the JSLC?” Within the value of science, unanswered questions are good as they for the basis upon whichfuture studies are conducted, as this will advance knowledge on health matters in Jamaica. Thequestion of ‘How is data quality in Jamaica?” in respect to the content errors are stillunanswered. This book, therefore, owes itself to the pursuit of truth more that the establishmentof tradition and/or the sanctioning of authority. Thus, the author acknowledges the search fortruth as this the birth of knowledge that can guide effective policy and intervention programmes. xiii
  15. 15. DEDICATION This book is dedicated to the‘Pursuit of Truth’ xiv
  16. 16. Part IHEALTH STATUS: USAGE OF HEALTH DATA 1
  17. 17. INTRODUCTIONMany researchers, scholars and academics utilize secondary cross-sectional survey data, becauseof the high cost and time allocation in conducting primary research. Secondary cross-sectionalsurveys are in response to affordability and time, which create a barrier to primary data collection.The question that is frequently asked, therefore, by user of those data is “How reliable is thecontent and coverage of the already collected data?” Some researchers rely on the credibility ofthe data collectors (such as WHO, UN, ILO, World Bank, Statistical Agencies, NASA,established Universities) in answering the aforementioned question. While those Organizationsare of a high standard, science is not about the non-verification of objects, events and dataestimates, particularly data collected from other sources. The reliance on the reliability and validity of data source go to the crux of trustworthinessand not science. This assumption violates the premise of science, verification of issues. Althoughscience rest on gradual development of issues before conclusions are finalized, many of theaforementioned Organizations have been in existence for some time and have access to moreresources than single scholar (or researcher), particularly in developing nations, but this does notdenote an arbitrary and unquestionable reliance on them, their data, estimates and findings. Themeaning of unquestionable facts destroys all the pillows upon which science are based, retardlogic and further scientific discoveries. Science is about the pursuit of truths, indicating thatquestioning is a normal component in validation, consistency and reliability. Outside of theverification of truths, there can be no science as everything is mere proposition. It is the logic,gradual development, continuous inquiry, verification, validity, consistency and reliability thatdistinguish sciences from mythology, customs, traditions and opinions. In Jamaica, researchers, scholars, academics and ordinary citizens rely on the estimatesand results of STATIN, PIOJ, the University of the West Indies and other establishedInternational Organizations. There is an undeniable reality that those Agencies have longcontributed to scientific estimates, results and cosmologies, but this is not sufficient to endscientific inquiry on their conceptualizations and results. Many discoveries emerged out of thequestioning of the establishments, epistemology, cosmology, customs, traditions, authority, and 2
  18. 18. not accepting things because they were stated. Knowledge is not consistent over all time intervals,making its changeable on new information at a specified interval. Science is about the continuation of truth searching, making it a persistent quest of allthings including the establishments, customs, tradition, knowledge, authority, and ‘naturalphilosophy’. Facts and knowledge are changeable with logic, gradual development of new facts,justification of knowledge, refutation of the old knowledge, testing of old and the establishmentof new principles, laws, and methods. Science cannot co-exist indefinitely with unanswered, non-justifiable and opinionated issues as “What is in an interval (i.e. in time)?” can change withsystematic, logical and conceptual inquiry. Simply put, cosmologies (or world views) are basedon a set of propositions that are flexible. With more knowledge about something, the truthchanges and different paradigms are established to explain events, object, situations andknowledge. Hence, knowledge is only hidden in time, changeable with time and empiricism. Ifknowledge is not stationary throughout time, then the reliability of result can be questioned,irrespective of the credibility of the data source. Since 1989, Jamaicans and other scholars have been using the data of STATIN and/orPIOJ, with some never questioning the content of the results. However, statisticians havequestioned the coverage of the data source that has led to modifications of the sampling framesand the decreasing of coverage errors. This has increased the generalizability of sample frame,size and data estimates. Clearly we should question issue to advance science, knowledge of whatis. With the lowering of coverage errors in JSLC, this does not frame any purity about the content.Because the instrument of the JSLC is a modification of the World Bank’s Living StandardHousehold Survey, this does not mean unquestionable estimates, results and content. While theinstrument provide some reliability about questionnaire, reliability does not end withquestionnaire and sample design. Caribbean demographers (such as Paul Bourne, SharonPriestley, Julian Devonish), who are cognizant of content errors in surveys as well as censuses,have neglected to provide a framework for understanding data quality in Jamaica. They as well asother non-demographers have relied on traditions, authority, agencies and the industrializednations to stipulate data quality, without questioning estimates, results and data sources. The author, who is a trained demographer, has published plethora of articles from healthusing the JSLC. Because science is about the pursuit of truths, the author is therefore concerned 3
  19. 19. about data quality, particularly in the JSLC, as the correctness of the estimates relies on datasource. There are two main types of data errors (such as coverage and content errors). On manyoccasions statisticians have evaluated coverage errors that have increased the quality of thesample estimates. Their efforts and works have increased the generality of the sample survey, butdo not ruled out other errors. This means that the quality of the JSLC data is currently higher inJamaica, increasing reliability and provision for better generalizability of the population. Likestatisticians, the author is questioning the quality of the JSLC data. This in no way speaks of thequestioning of the credibility of workers – including data gathers, statisticians and workers.Instead of the author’s questioning of the content of the JSLC data on health, is just an inquirythat validate and/or improve the estimates and results. Prior to beginning a comprehensive inquiry of the data quality, the author presents worksthat use the data on health. This volume is separated into two Parts. Part I is the presentation ofdifferent topics on health using the JSLC dataset. It is worth adding here chapters on health forreaders to understanding the estimates and results and how this volume will enhance thoseestimates and findings. 4
  20. 20. CHAPTER 1A Theoretical Framework of Good Health Status of Jamaicans: UsingEconometric Analysis to Model Good Health StatusThe socio-psychological and economic factors produced inequalities in health and need to beconsidered in health development. In spite of this, extensive review of health Caribbean revealedthat no study has examined health status over the life course of Jamaicans. With the value ofresearch to public health, this study is timely and will add value to understanding the elderly,middle age and young adults in Jamaica. The aim of this study is to develop models that can beused to examine (or evaluate) social determinants of health of Jamaicans across the life course,elderly, middle age and young adults. Eleven variables emerged as statistically significantpredictors of current good health Status of Jamaicans (p<0.05). The factors are retirementincome (95%CI=0.49-0.96), logged medical expenditure (95% CI =0.91-0.99), marital status(Separated or widowed or divorced: 95%CI=0.31-0.46; married: 95%CI=0.50-0.67; Nevermarried), health insurance (95%CI=0.029-0.046), area of residence (other towns:, 95%CI=1.05-1.46; rural area:), education (secondary: 95%CI=1.17-1.58; tertiary: 95%CI=1.47-2.82;primary or below: OR=1.00), social support (95%CI=0.75-0.96), gender (95%CI=1.281-1.706),psychological affective conditions (negative affective: 95%CI=0.939-0.98; positive affective:95%CI:1.05-1.11), number of males in household (95%CI:1.07-1.24), number of children inhousehold (95%CI=1.12-1.27) and previous health status. There are disparities in the socialdeterminants of health across the life course, which emerged from the current findings. Thefindings are far reaching and can be used to aid policy formulation and how social determinantsof health are viewed in the future.I NTRODUCTIONHealth is a multidimensional construct which goes beyond dysfunctions (illnesses, ailment orinjuries) [1-14]. Although World Health Organization (WHO) began this broaden conceptualframework in the late 1940s [1], Engel [3] was the first to develop the biopsychosocial model thatcan be used to examine and treat health of mentally ill patient. Engel’s biopsychosocial modelwas both in keeping with WHO’s perspective of health and again a conceptual model of health.Both WHO and Engel’s works were considered by some scholar as too broad and as such difficult 5
  21. 21. to measure [15]; although this perspective has some merit, scholars have ventured into usingdifferent proxy to evaluate the ideal conceptual definition forwarded by WHO for some time now. Psychologists have argued that the use of diseases to proxy health is unidirectional (ornegative) [2], and that the inclusion of social, economic and psychological conditions in health isbroader and more in keeping with the WHO’s definition of health than diseases. Diener was thefirst psychologist to forward the use of happiness to proxy health (or wellbeing) of an individual[16, 17]. Instead of debating along the traditional cosmology health, Diener took the discussioninto subjective wellbeing. He opined that happiness is a good proxy for subjective wellbeing of aperson, and embedded therein is wider scope for health than diseases. Unlike classical economistswho developed Gross Domestic Product per capita (GDP) to examine standard of living (orobjective wellbeing) of people as well this being an indicator of health status along with otherindicators such as life expectancy, Diener and others believe that people are the best judges oftheir state. This is no longer a debate, as some economists have used happiness as a proxy ofhealth and wellbeing [18-20]; and they argued that it is a good measurement tool of the concept.Theoretical Framework Whether the proxy of health (or wellbeing) is happiness, self-reported health status, self-rated health conditions, life satisfaction or ill-being, it was not until in the 1970s that econometricanalyses were employed to the study of health. Grossman [9] used econometric to capture factorsthat simultaneously determine health stock of a population. Grossman’s work transformed theconceptual framework outlined by WHO and Engel to a theoretical framework for the study ofhealth. Using data for the world, Grossman established an econometric model that capturesdeterminants of health. The model read (Model 1): 6
  22. 22. H t = ƒ (H t-1 , G o , B t , MC t , ED) ……………………………………………….. Model (1) where H t – current health in time period t , stock of health (H t-1 ) in previous period , B t –smoking and excessive drinking, and good personal health behaviours (including exercise – G o ),MC t ,- use of medical care, education of each family member (ED), and all sources of householdincome (including current income). Grossman’s model was good at the time; however, one of the drawbacks to this model wasthe fact that some crucible factors were omitted by the aforementioned model. Based on thatlimitation, using literature, Smith and Kington [10] refined, expanded and modified Grossman’swork as it omitted important variables such as price of other inputs and family background orgenetic endowment which are crucible to health status. They refined Grossman’s work to includesocioeconomic variables as well as some other factors [Model (2)]. H t = H* (H t-1 , P mc , P o , ED, Et , R t , A t , G o ) ………………………..…………… Model (2) Model (2) expresses current health status H t as a function of stock of health (H t-1 ), priceof medical care P mc , the price of other inputs P o , education of each family member (ED), allsources of household income (Et ), family background or genetic endowments (G o ), retirementrelated income (R t ), asset income (A t ). It is Grossman’s work that accounts for economists like Veenhoven’s [20] and Easterlin’s[19] works that used econometric analysis to model factors that determine subjective wellbeing.Like Veenhoven [20], Easterlin [19] and Smith and Kington [10], Hambleton et al. [6] used thesame theoretical framework developed by Grossman to examine determinants of health of elderly(ages 65+ years) in Barbados. Hambleton et al.’s work refined the work of Grossman and addedsome different factors such as geriatric depression index; past and current nutrition; crowding; 7
  23. 23. number of children living outside of household; and living alone. Unlike Grossman’s study, hefound that current disease conditions accounted for 67.2% of the explained variation in healthstatus of elderly Barbadians, with life style risks factors accounting for 14.2%, and social factors18.6%. One of the additions to Grossman’s work based on Hambleton et al.’s study was actualproportion of each factor on health status and life style risk factors. A study published in 2004, using life satisfaction and psychological wellbeing to proxywellbeing of 2,580 Jamaicans, Hutchinson et al. [21] employed the principles in econometricanalysis to examine social and health factors of Jamaicans. Other studies conducted by Bourne ondifferent groups and sub-groups of the Jamaican population have equally used the principles ofeconometric analysis to determine factors that explain health, quality of life or wellbeing [5, 8, 22,23]. Despite the contribution of Hutchinson et al’s and Bourne’s works to the understanding ofwellbeing, there is a gap in the literature on a theoretical framework explains good health status ofthe life course of Jamaicans. The current study will model predictors of good health status ofJamaicans as well as good health status of young adults, middle age adults and elderly in order toprovide a better understanding of the factors that influence each cohort.METHODSParticipants and questionnaireThe current research used a nationally cross-sectional survey of 25,018 respondents from the 14parishes in Jamaica. The survey used stratified random probability sampling technique to drawthe 25,018 respondents. The non-response rate for the survey was 29.7% with 20.5% who did notrespond to particular questions, 9.0% did not participated in the survey and another 0.2% wasrejected due to data cleaning. The study used secondary cross-sectional data from the Jamaica 8
  24. 24. Survey of Living Conditions (JSLC). The JSLC was commissioned by the Planning Institute ofJamaica (PIOJ) and the Statistical Institute of Jamaica (STATIN). These two organizations areresponsible for planning, data collection and policy guideline for Jamaica.The JSLC is a self-administered questionnaire where respondents are asked to recall detailedinformation on particular activities. The questionnaire covers demographic variables, health,immunization of children 0 to 59 months, education, daily expenses, non-food consumptionexpenditure, housing conditions, inventory of durable goods, and social assistance. Interviewersare trained to collect the data from household members. The survey is conducted between Apriland July annually.ModelThe multivariate model used in this study is a modification of those of Grossman and Smith &Kington which captures the multi-dimensional concept of health, and health status. The presentstudy further refine the two aforementioned works and in the process adds some new factors suchas psychological conditions, crowding, house tenure, number of people per household and adeconstruction of the numbers by particular characteristics i.e. males, females and children (ages≤ 14 years). Another fundamental difference of the current research and those of Grossman, andSmith and Kington is that it is area specific as it is focused on Jamaican residents. The proposed model that this research seeks to evaluate is displayed below [Model (3)]:H t = f(H t-1 ,P mc , ED i , R t , A t , Q t , HH t , C i , En i , MS i , HI i , HT i , SS i , LL i ,X i , CRi , D i , O i , Σ(NP i ,PP i ), M i ,N i , FS i , A i , W i , ε i )…..Model (3) The current health status of a Jamaica, H t , is a function of 23 explanation variables, whereH t is current health status of person i, if good or above (i.e. no reported health conditions four 9
  25. 25. week leading up to the survey period), 0 if poor (i.e. reported at least one health condition); H t-1 isstock of health for previous period; lnPmc is logged cost of medical care of person i; ED i iseducational level of person i, 1 if secondary, 1 if tertiary and the reference group is primary andbelow; Rt is retirement income of person i, 1 if receiving private and/or government pension, 0 ifotherwise; HI i is health insurance coverage of person i, 1 if have a health insurance policy, 0 ifotherwise; HT i is house tenure of person i, 1 if rent, 0 if squatted; Xi is gender of person i, 1 iffemale, 0 if male; CRi is crowding in the household of person i; Σ(NPi,PPi) NPi is the summationof all negative affective psychological conditions and PPi is the summation of all positiveaffective psychological conditions; M i is number of male in household of person i and Fi isnumber of female in household of person i; Ai is the age of the person i and N i is number ofchildren in household of person i; LLi is living arrangement where 1= living with familymembers or relative, and 0=otherwise and social standing (or social class), W i .Statistical analysisStatistical analyses were performed using Statistical Packages for the Social Sciences (SPSS) forWindows, Version 16.0 (SPSS Inc; Chicago, IL, USA). A single hypothesis was tested, whichwas ‘health status of rural resident is a function of demographic, social, psychological andeconomic variables.’ The enter method in logistic regression was used to test the hypothesis inorder to determine those factors that influence health status of rural residents if the dependentvariable is a binary one; and linear multiple regression in the event the dependent variable was anormally distributed metric variable . The final model was established based on those variablesthat are statistically significant (ie. p < 0.05) – ie 95% confidence interval (CI), and all othervariables were removed from the final model (p>0.05). Continuing, categorical variables werecoded using the ‘dummy coding’ scheme. 10
  26. 26. The predictive power of the model was tested using Omnibus Test of Model and Hosmerand Lemeshow [24] was used to examine goodness of fit of the model. The correlation matrix wasexamined in order to ascertain whether autocorrelation (or multi-collinearity) existed betweenvariables. Cohen and Holliday [25] stated that correlation can be low/weak (0 to 0.39); moderate(0.4-0.69), or strong (0.7-1.0). This was used in this study to exclude (or allow) a variable in themodel. Where collinearity existed (r > 0.7), variables were entered independently into the modelto determine those that should be retained during the final construction of the model. To deriveaccurate tests of statistical significance, we used SUDDAN statistical software (Research TriangleInstitute, Research Triangle Park, NC), and this was adjusted for the survey’s complex samplingdesign. Finally, Wald statistics was used to determine the magnitude (or contribution) of eachstatistically significant variables in comparison with the others, and the odds ratio (OR) for theinterpreting each significant variables.Results: Modelling Current Good Health Status of Jamaicans, Elderly, Middle Age andYoung adultsPredictors of current Good Health Status of Jamaicans. Using logistic regression analyses, elevenvariables emerged as statistically significant predictors of current good health status of Jamaicans(p<0.05, see Model 4). The factors are retirement income, logged medical expenditure, maritalstatus, health insurance, area of residence, education, social support, gender, psychologicalaffective conditions, number of males in household, number of children in household andprevious health status (Table 1.1.1). Ht = f(H t-1, Rt , P mc , ED i , MSi, HI i , SS i,AR i, X i , Σ(NP i,PP i), M i,N i, ε i)...……………………………..... Model(4) 11
  27. 27. The model [ie Model (4)] had statistically significant predictive power (χ2 (27) =1860.639,p < 0.001; Hosmer and Lemeshow goodness of fit χ2=4.703, p = 0.789) and overall correctlyclassified 85.7% of the sample (correct classified 98.3% of cases of good health status andcorrectly classified 33.9% of cases of dysfunctions). There was a moderately strong statistical correlation between age, marital status,education, retirement income, per capita income quintiles, property ownership, and so these wereomitted from the initial model (ie model 3). Based on that fact, three age groups were classified(young adults – ages 15 to 29 years; middle age adults – ages 30 to 59 years; and elderly – ages60+ years) and the initial model was once again tested. There were some modifications of theinitial model in keeping with the age group. For young adults the initial model was amended byexcluding retirement income, property ownership, divorced, separated or widowed, number ofchildren in household, and house tenure. The exclusion was based on the fact that more than 15%of cases missing in some categories and a high correlation between variables.Predictors of current Good Health Status of elderly Jamaicans. From the logistic regressionanalyses that were used on the data, eight variables were found to be statistically significant inpredicting good health Status of elderly Jamaicans (P < 0.5) (see Model 5). These factors wereeducation, marital status, health insurance, area of residence, gender, psychological conditions,number of males in household, number of children in household and previous health status (seeTable 1.1.2). Ht = f(H t-1, ED i, MSi , HI i, ,ARi , X i, Σ(PP i), M i,N i, ε i)...…………………………………………………..... Model (5) The model had statistically significant predictive power (model χ2 (27) =595.026, P <0.001; Hosmer and Lemeshow goodness of fit χ2=5.736, p = 0.677) and overall correctly 12
  28. 28. classified 75.5% of the sample (correctly classified 94.6% of cases of good or beyond healthstatus and correct classified 44.7% of cases of dysfunctions). Predictors of current Good Health Status of middle age Jamaicans. Using logisticregression, six variables emerged as statistical significant predictors of current good health statusof middle age Jamaican (p < 0.05) (Model 6). These factors are logged medical expenditure,physical environment, health insurance, gender of respondents, psychological condition, andnumber of children in household and previous health status (see Table 1.1.3) Ht = f(Ht-1, P mc , En i , HI i, X i , Σ(NP i),N i, εi)...........................................……………………………..... Model(6) Based on table 3, the model had statistically significant predictive power (model χ2 (27)=547.543, p < 0.001; Hosmer and Lemeshow goodness of fit χ2=4.318, p = 0.827) and overallcorrectly classified 87.2% of the sample (correctly classified 98.3% of cases of good or beyondhealth status and correct classified 28.2% of cases of dysfunctions).Predictors of current Good Health Status of young adult in Jamaica. Using logistic regression, twovariables emerged as statistically significant predictors of current good health status of youngadults in Jamaica (p<0.05) (Model 7). These are health insurance coverage, psychologicalcondition, social class and previous health status (Table 1.1.4). Ht = f(H t-1, W i, HI i, Σ(NP i), εi )...............................................…………………………….....Model (7) From table 3, the model had statistically significant predictive power (model χ2 (19) =453.733, p < 0.001; Hosmer and Lemeshow goodness of fit χ2=5.185, p = 0.738) and overallcorrectly classified 92.6% of the sample (correctly classified 99.0% of cases of good or beyondhealth status and correct classified 28.2% of cases of dysfunctions). 13
  29. 29. Limitations to the Models Good Health Status of Jamaicans [ie Model (4)], elderly [ie Model (5)], middle age adults[ie Model (6)], and young adults [ie Model (7) are derivatives of Model (3). Good Health Status[ie Model (4) – Model (7)] cannot be distinguished and tested over different time periods, persondifferential, and these are important components of good health. H t = f(H t-1 , R t , P mc , ED i , MS i , HI i , SS i ,AR i , X i , Σ(NP i ,PP i ), M i ,N i , ε i )...………………………..... Model (4) H t = f(H t-1 , ED i , MS i , HI i , ,AR i , X i , Σ(PP i ), M i ,N i , ε i )...………………………………………..... Model (5) H t = f(H t-1 , P mc , En i , HI i , X i , Σ(NP i ),N i , ε i )....................................……………………………..... Model (6) H t = f(H t-1 , Wi , HI i , Σ(NP i ), ε i ).......................................................……………………….…….......Model (7) H t = f(H t-1 ,P mc , ED i , R t , A t , Q t , HH t , C i , En i , MS i , HI i , HT i , SS i , LL i ,X i , CR i , D i , O i , Σ(NP i ,PP i ), M i ,N i , FS i , A i , Wi ,ε i )………………………………………………………………………..Model (3) The current work is a major departure from Grossman’s theoretical model as he assumedthat factors affecting good health Status over the life course are the same, this study disagreedwith this fundamental assumption. This study revealed that predictors of good health status arenot necessarily the same across the life course, and differently from that of the general populace.Despite those critical findings, healthy time gained can increase good health status directly andindirectly but this cannot be examined by using a single cross-sectional study. Health does notremain constant over any specified period, and to assume that this is captured in age is to assumethat good or bad health change over year (s). Health stock changes over short time intervals, andso must be incorporated within any health model. 14
  30. 30. People are different even across the same ethnicity, nationality, next of kin andsocialization. This was not accounted for in the Grossman’s or the current work, as this is one ofthe assumptions. Neither Grossman’s study nor the current research recognized the importance ofdifferences in individuals owing to culture, socialization and genetic composition. Eachindividual’s is different even if that person’s valuation for good health Status is the same assomeone else who share similar characteristics. Hence, a variable P representing the individualshould be introduced to this model in a parameter α (p). Secondly, the individual’s good (or bad)health is different throughout the course of the year and so time is an important factor. Thus, theresearcher is proposing the inclusion of a time dependent parameter in the model. Therefore, thegeneral proposition for further studies is that the function should incorporate α (p, t) a parameterdepending on the individual and time. An unresolved assumption of this work which continues from Grossman’s model is thatpeople choose health stock so that desired health is equal to actual health. The current data cannottest this difference in the aforementioned health status and so the researcher recommends thatfuture study to account for this disparity so we can identify factors of actual health and differencebetween the two models.Discussions This study has modelled current good status of Jamaicans. Defining health into twocategories (ie good – not reported an acute or illness; or poor – reported illness or ailment), thisstudy has found that using logistic regression health status can be modeled for Jamaicans. Thefindings revealed that the probability of predicting good health status of Jamaicans was 0.789,using eleven factors; and that approximately 86% of the data was correctly classified in this study.Continuing, in Model (4) approximately 98% of those who had reported good health status were 15
  31. 31. correctly classified, suggesting that using logistic regression to examine good health status of theJamaican population with the eleven factors that emerged is both a good predictive model and agood evaluate or current good health status of the Jamaican population. This is not the first studyto examine current good health status or quality of life in the Caribbean or even Jamaica [6, 21-23, 26], but that none of those works have established a general and sub-models of good healthover the life course. In Hambleton et al’s work, the scholars identified the factors (ie historical, current, lifestyle, diseases) and how much of health they explain (R2=38.2%). However, they did not examinethe goodness of fit of the model or the correctness of fit of the data. Bourne’s works [12,13] weresimilar to that of Hambleton et al’s study, as his study identified more factors (psychologicalconditions; physical environment, number of children or males or females in household and socialsupport) and had a greater explanatory power (adjusted r square = 0.459) but again the goodnessof fit and correctness of fit of the data were omitted. Again this was the case in Hutchinson et al.’sresearch. Like previous studies in the Caribbean that have examined health status [6, 21-23, 26],those conducted by the WHO and other scholars [27-32] did not explore whether socialdeterminants of health vary across the life course. Because this was not done, we have assumedthat the social determinants are the same across the life. However, a study by Bourne andEldemire-Shearer [33] introduced into the health literature that social determinants differ acrosssocial strata for men. Such a work brought into focus that there are disparities in the socialdeterminants of health across particular social characteristic and so researchers should notarbitrarily assume that they are the same across the life course. While Bourne and Eldemire-Shearer’s work [33] was only among men across different social strata in Jamaica (poor and 16
  32. 32. wealthy), the current study shows that there are also differences in social and psychologicaldeterminants of health across the life course. The current study has concluded that the factors identified to determine good health statusfor elderly, had the lowest goodness of fit (approximately 68%) while having the greatestexplanatory power (R2= 35%). The findings also revealed low explanatory powers for youngadults (R2=22.6%) and middle age adults (R2=23%), with latter having a greater goodness of fitfor the data as this is owing to having more variables to determine good health. Such a findinghighlights that we know more about the social determinants for the elderly than across other agecohorts (middle-aged and young adults). And that using survey data for a population to ascertainthe social determinants of health is more about those for the elderly than across the life course ofa population. Another important finding is of the eleven factors that emerge to explain good healthstatus of Jamaicans, when age cohorts were examine it was found that young adults had the leastnumber of predictors (ie health insurance, social class and negative affective psychologicalconditions). This suggests that young adult’s social background and health insurance areimportant factors that determine their good health status and less of other determinants that affectthe elderly and middle age adults. It should be noted that young adult is the only age cohort withwhich social standing is a determinant of good health. Even though the good health status modelthat emerged from this study is good, the low explanatory power indicates that young adults areunique and further study is needed on this group in order to better understand those factors thataccount for their good health. Furthermore, this work revealed that as people age, the socialdeterminants of health of the population are more in keeping with those of the elderly than atyounger ages. Hence, the social determinants identified by Grossman [9], Smith and Kington [10] 17
  33. 33. and purported by Abel-Smith [11] as well as the WHO [27] and affiliated researchers [28-32] aremore for the elderly population than the population across the life course.ConclusionsThere are disparities in the social determinants of health across the life course, which emergedfrom the current findings. The findings are far reaching and can be used to aid policy formulationand how we examine social determinants of health. Another issue which must be researched iswhether there are disparities in social determinants of health based on the conceptualization andmeasurement of health status (using self-reported health, and health conditions).DisclosuresThe author reports no conflict of interest with this work.DisclaimerThe researcher would like to note that while this study used secondary data from the JamaicaSurvey of Living Conditions (JSLC), none of the errors in this paper should be ascribed to thePlanning Institute of Jamaica (PIOJ) and/or the Statistical Institute of Jamaica (STATIN), but tothe researcher.AcknowledgementThe author thanks the Data Bank in Sir Arthur Lewis Institute of Social and Economic Studies,the University of the West Indies, Mona, Jamaica for making the dataset (2002 JSLC) availablefor use in this study, and the National Family Planning Board for commissioning the survey. 18
  34. 34. REFERENCES1. World Health Organization, WHO. Preamble to the Constitution of the World HealthOrganization as adopted by the International Health Conference, New York, and June 19-22,1946; signed on July 22, 1946 by the representatives of 61 States (Official Records of the WorldHealth Organization, no. 2, p. 100) and entered into force on April 7, 1948. “Constitution of theWorld Health Organization, 1948.” In Basic Documents, 15th ed. Geneva, Switzerland: WHO,1948.2.Brannon L, Feist J. Health psychology. An introduction to behavior and health 6th ed. LosAngeles: Thomson Wadsworth; 2007.3. Engel G. A unified concept of health and disease. Perspectives in Biology and Medicine.1960; 3:459-485.4. Engel GL. The clinical application of the biopsychosocial model. Am J of Psychiatry.1980; 137:535-544.5. Bourne P. Using the biopsychosocial model to evaluate the wellbeing of the Jamaican elderly.West Indian Med J. 2007b; 56: (suppl 3), 39-40.6. Hambleton IR, Clarke K, Broome HL, Fraser HS, Brathwaite F, Hennis AJ. 2005. Historicaland current predictors of self-reported health status among elderly persons in Barbados. Rev PanSalud Public. 2005; 17:342-352.7. Longest BB, Jr. Health Policymaking in the United States, 3rd. Chicago, IL: HealthAdministration Press; 2002.8. Bourne PA. Quality of Life of Jamaican women. West Indian Med J. 2008; 57: (suppl 4), 49.9. Grossman M. The demand for health- a theoretical and empirical investigation. New York:National Bureau of Economic Research; 1972.10. Smith JP, Kington R. Demographic and economic correlates of health in old age.Demography. 1997; 34:159-170.11. Abel-Smith B. An introduction to health: Policy, Planning and Financing. Harlow: PearsonEducation; 1994.12. Bourne PA. Health Determinants: Using Secondary Data to Model Predictors of Well-beingof Jamaicans. West Indian Med J. 2008; 57:476-481.13. Bourne PA. Medical Sociology: Modelling Well-being for elderly People in Jamaica. WestIndian Med J. 2008; 57:596-604.14. Bourne PA. Health measurement. Health 2010;2(5):465-476. 19
  35. 35. 15. Bok S. Rethinking the WHO definition of health. Working Paper Series, 14; 2004. Retrievedfrom http://www.golbalhealth.harvard.edu/hcpds/wpweb/Bokwp14073.pdf (accessed May 26,2007).16. Diener E. Subjective wellbeing. Psychological Bulletin. 1984; 95:542-575.17. Diener E. Subjective wellbeing. The science of happiness and proposal for a national index.Am Psychol Ass. 2000; 55:34-43.18. Crisp R. Wellbeing. The Stanford Encyclopedia of Philosophy (winter edition) E N Zalta ed;2005. Retrieved from http://plato.stanford.edu/archives/win2005/entries/wellbeing (accessedAugust 22, 2006).19. Easterlin RA. Building a better theory of well-being. Prepared for presentation at theconference Paradoxes of Happiness in Economics, University of Milano-Biococca, March 21-23,2003. Retrieved from http://www-rcf.use.edu/ easterl/papers/BetterTheory.pdf (accessed August26, 2006).20. Veenhoven R. Happiness in nations, subjective appreciation of in 56 nations 1946-1992.Rotterdam, Netherlands: Erasmus University; 1993.21. Hutchinson G, Simeon DT, Bain BC, Wyatt GE, Tucker MB, LeFranc E. Social and healthdeterminants of wellbeing and life satisfaction in Jamaica. International J of Social Psychiatry.2004; 50:43-53.22. Bourne PA. Modelling social determinants of self-evaluated health of poor old people in amiddle-income developing nation. J Biomedical Sci and Engineering, 2010;3: 700-710.23. Bourne PA. Social and environmental correlates of self-evaluated health of poor agedJamaicans. HealthMed journal 2010;4(2):284-296.24. Homer D, Lemeshow S. Applied Logistic Regression, 2nd edn. John Wiley & Sons Inc., NewYork, 2000.25. Cohen L, Holliday M. Statistics for Social Sciences. London, England: Harper and Row,1982.26. Asnani MR, Reid ME, Ali SB, Lipps G, Williams-Green P. Quality of life in patients withsickle cell disease in Jamaica: rural-urban differences. Rural and Remote Health. 2008; 8: 1-9.27. WHO. The Social Determinants of Health; 2008. Available athttp://www.who.int/social_determinants/en/ (accessed April 28, 2009).28. Kelly M, Morgan A, Bonnefog J, Beth J, Bergmer V. The Social Determinants of Health:developing Evidence Base for Political Action, WHO Final Report to the Commission; 2007. 20
  36. 36. 29. Wilkinson RG, Marmot M. Social Determinants of Health. The Solid Facts, 2nd ed.Copenhagen: World Health Organization; 2003.30. Solar O, Irwin A. A Conceptual Framework for Analysis and Action on the SocialDeterminants of Health. Discussion paper for the Commission on Social Determinants of HealthDRAFT April 2007. Available fromhttp://www.who.int/social_determinants/resources/csdh_framework_action_05_07.pdf (AccessedApril 29, 2009).31. Graham H. Social Determinants and their Unequal Distribution Clarifying PolicyUnderstanding The MilBank Quarterly 2004;82 (1), 101-124.32. Pettigrew M, Whitehead M, McIntyre SJ, Graham H, Egan M. Evidence for Public HealthPolicy on Inequalities: 1: The Reality According To Policymakers. Journal of Epidemiology andCommunity Health 2004;5, 811 – 816.33. Bourne PA, Eldemire-Shearer D. Differences in social determinants of health between men inthe poor and the wealthy social strata in a Caribbean nation. North Am J of Med Sci 2010;2(6):267-275. 21
  37. 37. Table 1.1.1: Good Health Status of Jamaicans by Some Explanatory Variables CI (95%) Wald statistic Odds Variable P Ratio Coefficient Std Error. Lower Upper Middle Quintile -0.03 0.10 0.09 0.764 0.97 0.81 1.17 Two Wealthiest Quintiles -0.11 0.10 1.26 0.261 0.90 0.74 1.09 Poorest-to-poor Quintiles* Retirement Income -0.38 0.17 4.88 0.027 0.68 0.49 0.96 Household Head 0.17 0.29 0.37 0.543 1.19 0.68 2.08 Logged Medical Expenditure -0.05 0.02 5.10 0.024 0.95 0.91 0.99 Average Income 0.00 0.00 1.56 0.212 1.00 1.00 1.00 Average Consumption 0.00 0.00 0.16 0.689 1.00 1.00 1.00 Environment 0.01 0.07 0.02 0.891 1.01 0.88 1.16 Separated or Divorced or Widowed -0.97 0.10 87.36 0.000 0.38 0.31 0.46 Married -0.55 0.08 53.05 0.000 0.58 0.50 0.67 Never married* Health Insurance -3.31 0.12 776.64 0.000 0.04 0.03 0.05 Other Towns 0.21 0.08 6.64 0.010 1.24 1.05 1.46 Urban Area -0.01 0.13 0.00 0.952 0.99 0.78 1.27 Rural Area* House Tenure - Rent -1.08 0.88 1.48 0.224 0.34 0.06 1.93 House Tenure - Owned -0.42 0.55 0.58 0.447 0.66 0.23 1.93 House Tenure- Squatted* Secondary Education 0.31 0.08 15.81 0.000 1.36 1.17 1.58 Tertiary Education 0.71 0.17 18.09 0.000 2.03 1.45 2.82 Primary and below* Social Support -0.17 0.07 6.33 0.012 0.85 0.75 0.96 Living Arrangement -0.06 0.13 0.20 0.659 0.95 0.73 1.22 Crowding -0.01 0.04 0.08 0.772 0.99 0.91 1.07 Land ownership -0.07 0.07 0.90 0.342 0.93 0.81 1.08 Gender 0.39 0.07 28.67 0.000 1.48 1.28 1.71 Negative Affective -0.04 0.01 14.96 0.000 0.96 0.94 0.98 Positive Affective 0.07 0.01 26.26 0.000 1.08 1.05 1.11 Number of males in household 0.14 0.04 13.36 0.000 1.15 1.07 1.24 Number of females in household 0.06 0.04 2.36 0.124 1.06 0.98 1.14 Number of children in household 0.17 0.03 29.16 0.000 1.19 1.12 1.27 Constant 1.89 0.65 8.31 0.004 6.59χ2 (27) =1860.639, p < 0.001; n = 8,274-2 Log likelihood = 6331.085Hosmer and Lemeshow goodness of fit χ2=4.703, p = 0.789.Nagelkerke R2 =0.320Overall correct classification = 85.7% (N=7,089)Correct classification of cases of good or beyond health status =98.3% (N=6,539)Correct classification of cases of dysfunctions =33.9% (N=550); *Reference group 22
  38. 38. Table 1.1.2: Good Health Status of Elderly Jamaicans by Some Explanatory Variables Std Wald Odds Coefficient Error statistic P Ratio CI (95%) Lower Upper Middle Quintile -0.10 0.15 0.47 0.495 0.90 0.67 1.22 Two Wealthiest Quintiles 0.12 0.17 0.47 0.491 1.12 0.81 1.56 Poorest-to-poor quintiles Retirement Income -0.22 0.22 1.00 0.317 0.81 0.53 1.23 Household Head 0.89 0.65 1.86 0.172 2.44 0.68 8.76 Logged Medical Expenditure -0.06 0.04 2.16 0.142 0.95 0.88 1.02 Average Income 0.00 0.00 0.93 0.335 1.00 1.00 1.00 Environment -0.16 0.12 1.80 0.180 0.86 0.68 1.08 Separated or Divorced or -0.49 0.15 11.00 0.001 0.61 0.46 0.82 Widowed Married -0.33 0.15 4.82 0.028 0.72 0.54 0.97 Never married* -3.35 0.22 241.88 0.000 0.04 0.02 0.05 Health Insurance Other Towns 0.33 0.14 5.32 0.021 1.39 1.05 1.83 Urban 0.40 0.21 3.48 0.062 1.49 0.98 2.27 Rural areas* House tenure - rented -20.37 40192.9 0.00 1.000 0.00 0.00 House tenure - owned 1.22 1.24 0.96 0.327 3.38 0.30 38.60 House tenure – squatted* Secondary Education -0.46 0.11 16.06 0.000 0.63 0.51 0.79 Tertiary Education 0.81 0.35 5.45 0.020 2.26 1.14 4.47 Primary or below* Social support -0.08 0.11 0.47 0.495 0.93 0.75 1.15 Living arrangement 0.26 0.18 2.11 0.146 1.30 0.91 1.84 Crowding -0.05 0.09 0.29 0.593 0.95 0.80 1.14 Landownership 0.17 0.13 1.72 0.190 1.19 0.92 1.54 Gender 0.47 0.12 14.67 0.000 1.60 1.26 2.04 Negative Affective -0.03 0.02 1.97 0.160 0.97 0.94 1.01 Positive Affective 0.07 0.02 9.26 0.002 1.07 1.03 1.12 Number of male 0.18 0.07 6.75 0.009 1.19 1.04 1.36 Number of females 0.05 0.07 0.49 0.485 1.05 0.91 1.21 Number of children 0.22 0.06 12.09 0.001 1.24 1.10 1.40 Constant -1.32 1.44 0.83 0.362 0.27χ2 (27) =595.026, p < 0.001; n = 2,002-2 Log likelihood = 2,104.66Hosmer and Lemeshow goodness of fit χ2=5.736, p = 0.677.Nagelkerke R2 =0.347Overall correct classification = 75.5% (N=1.492)Correct classification of cases of good or beyond health status =94.6% (N=1,131)Correct classification of cases of dysfunctions =44.7% (N=361);*Reference group 23
  39. 39. Table 1.1.3: Good Health Status of Middle Age Jamaicans by Some Explanatory Variables Std Wald Odds Coefficient Error statistic P Ratio CI (95%) Lower Upper Middle Quintile 0.03 0.15 0.04 0.834 1.03 0.76 1.40 Two Wealthiest Quintiles -0.29 0.15 3.67 0.055 0.75 0.56 1.01 Poorest-to-poor Quintiles* Retirement Income -0.57 0.36 2.44 0.119 0.57 0.28 1.16 Household Head 0.50 0.45 1.24 0.265 1.66 0.68 4.01 Logged Medical Expenditure -0.09 0.04 6.44 0.011 0.91 0.85 0.98 Average Income 0.00 0.00 0.53 0.465 1.00 1.00 1.00 Environment 0.31 0.12 7.41 0.006 1.37 1.09 1.71 Separated or Divorced or Widowed -0.20 0.23 0.77 0.380 0.82 0.53 1.28 Married -0.18 0.11 2.68 0.102 0.84 0.68 1.04 Never married* Health Insurance -3.04 0.17 320.76 0.000 0.05 0.03 0.07 Other Towns 0.11 0.12 0.75 0.387 1.11 0.87 1.42 Urban -0.01 0.19 0.00 0.963 0.99 0.68 1.44 Rural areas* House tenure - rented 17.94 20029.78 0.00 0.999 0.00 House tenure - owned -1.33 1.12 1.43 0.232 0.26 0.03 2.35 House tenure – squatted* Secondary education 0.19 0.13 2.11 0.146 1.20 0.94 1.55 Tertiary education 0.34 0.23 2.23 0.135 1.41 0.90 2.21 Primary or below* Social support -0.08 0.10 0.57 0.450 0.93 0.76 1.13 Living Arrangement -0.19 0.21 0.87 0.351 0.83 0.55 1.24 Crowding -0.05 0.06 0.65 0.419 0.95 0.85 1.07 Landownership -0.13 0.11 1.47 0.226 0.88 0.71 1.08 Gender 0.51 0.11 21.41 0.000 1.66 1.34 2.06 Negative Affective -0.08 0.02 24.66 0.000 0.92 0.90 0.95 Positive Affective 0.05 0.02 4.51 0.034 1.05 1.00 1.10 Number of males in house 0.03 0.06 0.23 0.630 1.03 0.92 1.14 Number of female in house 0.08 0.06 2.09 0.149 1.08 0.97 1.21 Number of children in house 0.10 0.04 5.47 0.019 1.11 1.02 1.21 Constant 3.29 1.25 6.89 0.009 26.77χ (27) =547.543, p < 0.001; n = 3,799 2-2 Log likelihood = 2,776.972Hosmer and Lemeshow goodness of fit χ2=4.318, p = 0.827.Nagelkerke R2 =0.230Overall correct classification = 87.2% (N=3,313)Correct classification of cases of good or beyond health status =98.3% (N=3,143)Correct classification of cases of dysfunctions =28.2% (N=170);*Reference group 24
  40. 40. Table 1.1.4: Good Health Status of Young Adults Jamaicans by Some Explanatory Variables CI (95%) Wald Odds Coefficient Std Error statistic P Ratio Lower Upper Middle Quintile -0.06 0.19 0.10 0.747 0.94 0.65 1.37 Two Wealthiest Quintiles -0.59 0.18 11.10 0.001 0.55 0.39 0.78 Poorest-to-poor quintiles* Household Head -0.25 0.39 0.41 0.520 0.78 0.36 1.68 Logged Medical Expenditure 0.01 0.04 0.09 0.760 1.01 0.93 1.10 Average Income 0.00 0.00 3.29 0.070 1.00 1.00 1.00 Environment -0.03 0.13 0.04 0.840 0.97 0.75 1.26 Health Insurance -3.73 0.21 321.51 0.000 0.02 0.02 0.04 Other Towns 0.23 0.15 2.42 0.120 1.26 0.94 1.69 Urban -0.05 0.18 0.07 0.788 0.95 0.68 1.34 Rural area* Secondary education -0.06 0.41 0.02 0.886 0.94 0.43 2.09 Tertiary education -0.39 0.47 0.70 0.405 0.68 0.27 1.69 Primary and below* Social support -0.14 0.13 1.22 0.269 0.87 0.68 1.12 Crowding 0.04 0.06 0.65 0.420 1.05 0.94 1.16 Gender 0.19 0.15 1.60 0.206 1.20 0.90 1.60 Negative Affective -0.04 0.02 4.22 0.040 0.96 0.93 1.00 Positive Affective 0.07 0.03 6.81 0.009 1.07 1.02 1.13 Number of males in house 0.13 0.07 3.67 0.055 1.13 1.00 1.29 Number of females in house 0.06 0.06 0.87 0.351 1.06 0.94 1.20 Married 0.08 0.22 0.13 0.717 1.09 0.70 1.68 Never married* Constant 2.75 0.67 16.62 0.000 15.57χ2 (19) =453.733, p < 0.001; n = 4,174-2 Log likelihood = 2,091.88Hosmer and Lemeshow goodness of fit χ2=5.185, p = 0.738.Nagelkerke R2 =0.226Overall correct classification = 92.6% (N=3,864)Correct classification of cases of good or beyond health status =99.0% (N=3,757)Correct classification of cases of dysfunctions =28.2% (N=107);*Reference group 25
  41. 41. CHAPTER 2An Epidemiological Transition of Health Conditions, and Health Status of theOld-Old-To-Oldest-Old in Jamaica: A comparative analysis using two cross-sectional surveysThere is a paucity of information on the old-old-to-oldest-old in Jamaica. In spite of studies onthis cohort, there has never been an examination of the epidemiological transition in healthcondition affect this age cohort. The aims of the current study are 1) provide an epidemiologicalprofile of health conditions affecting Jamaicans 75+ years, 2) examine whether there is anepidemiological transition in health conditions affecting old-old-to-oldest-old Jamaicans, 3)evaluate particular demographic characteristics and health conditions of this cohort, 4) assesswhether current self-reported illness is strongly correlated with current health status, 5) meanage of those with particular health conditions, 6) model health status and 7) provide valuableinformation upon which health practitioners and public health specialists can make moreinformed decisions. In 2007, 44% of old-to-oldest-old Jamaicans were diagnosed withhypertension, which represents a 5% decline over 2002. The number of cases of diabetes mellitusincreased over 570% in the studied period. The poor indicated having more health conditionsthan the poorest 20% of the sample. The implications of the shift in health conditions will createa health disparity between 75+ year adults and the rest of the population.IntroductionThe elderly population (ages 60+ years) constituted 10.9% of Jamaica’s population, which meansthat this age cohort is vital in public health planning [1]. Eldemire [2] opined that “The majorityof Jamaican older persons are physically and mentally well and living in family units”. This viewwas substantiated in an early study; when Eldemire [3] found that approximately 81 percent ofthe seniors reported that they were physically competent to care for themselves, without anyform of external intervention. Eldemire’s work revealed that 88.5 percent being physiologicallyindependent. 26
  42. 42. Many elderly persons are more than physically independent as Eldemire [3] found 65.5percent of them supported themselves, with males reporting a higher self-support (82.6%)compared to females, 47.7%. A study conducted by Franzini and colleague [4] found that socialsupport was directly related to self-reported health, which is collaborated by Okabayashi et al’sstudy [5]. The aforementioned situation can explain why many elderly are offered socio-economic support. Eldemire [3] found that approximately 71 percent of children were willing toaccept responsibility for their parents, with seniors who were older than 75 years being likely toneed support. Seniors ages 75-84 years are referred to as old-old and those 85+ are referred asoldest-old. The 2001 Population Census of Jamaica found approximately 66 percent of the elderlylive in private households [6], which imply that the aged are physically and mentally competent.This is in keeping with Eldemire’s studies [2, 3]. The functional independence of the elderly isnot atypical to Jamaica as DaVanzo and Chan [7], using data from the Second Malaysian FamilyLife Survey which includes 1,357 respondents of age 50 years and older living in privatehouseholds, noted that some benefits of co-residence range from emotional support,companionship, physical and financial assistance [8]. Embedded in DaVanzo and colleague’swork is the issue of ‘Is it functional independence or stubbornness?’ that accounts for the elderlypersons’ report that they are physically and mentally well in order that family and onlookers willnot request that they live in home care facilities. This brings into focus the issues of health statusand health conditions of elderly Jamaicans. Physical disability and health problems increase with age [9]. Bogue [9] opined thatdemand for medical care increases with ageing and that this is owing to health deteriorations. He[9] also noted that as an individual age, the demands on their children increases and likewise 27
  43. 43. their demand on the public services also increases. Statistics revealed that 15.5% of Jamaicansreported suffering from an illness/injury in 2007; this was 2.8 times more for individuals ages65+ and 2.4 times for those people ages 60+ years [10]. This further goes to concurs withBogue’s perspective that ageing is associated with increased illness. Concurrently, in 2007,51.9% of Jamaicans who reported an illness, in the 4-week period of the survey, indicated thatthis was recurring compared to 75.1% of the elderly. The elderly also sought more medical care(72%) compared to the general population (66%), purchased more medication (78.3% comparedto the general population, 73.3%) and had more health insurance coverage (27.8%) compared tothe general population (21.1%) [10]. The aforementioned findings only concur with the work ofBogue, and still does not provide us with changing in health conditions of the elderly inparticular the old-old-to-oldest old. Using a sub-sample of 3,009 elderly Jamaicans, Bourne [11] found that the generalwellbeing was low; but, within the context of Bogue’s work, raised the question of the old-old orthe oldest-old’s health status. Bourne [12], using a sub-sample of 1,069 respondents ages 75+years, found that 51.3% of those 75-84 years had poor health status compared to 52.6% of theoldest-old. There was no significant statistical difference between the poor health status of old-old and oldest-old Jamaicans. While poor health status comprised of health conditions, Bourne’sworks do not provide us with an understanding of the health conditions over time and whetherthese are changing or not. A study on elderly Barbadians by Hambleton and colleagues [13]found that current health conditions (diseases) were the most influential predictor of currenthealth status and adds value to discourse that health conditions provide some understanding ofhealth status. However, this finding does not clarify the epidemiological transition of healthconditions affecting the old-old-to-oldest-old Caribbean nationals, in particular Jamaicans. 28
  44. 44. An extensive review of health and ageing literature in the Caribbean revealed no studythat has examined an epidemiological transition of health conditions of people 75+ years. InJamaica, 4% of the population in 2007 were older than 75+ years, indicating that over 100,000Jamaicans have reached 75 years or older. This is a critical group that must be studied for publichealth planning as more elderly have chronic dysfunctions than any other age cohort in thepopulation. The aims of the current study are 1) provide an epidemiological profile of healthconditions affecting Jamaicans 75+ years, 2) examine whether there is an epidemiologicaltransition in health conditions affecting old-old-to-oldest-old Jamaicans, 3) evaluate particulardemographic characteristic and health conditions of this cohort, 4) assess whether current self-reported illness is strongly correlated with current health status, 5) mean age of those withparticular health conditions, 6) model health status and 7) provide valuable information uponwhich health practitioners and public health specialists can make more informed decisions.Materials and MethodsThe current study utilized a sub-sample of approximately 4% from each nationally cross-sectional survey that was conducted in 2002 and 2007. The sub-sample was 282 people ages 75+years from the 2007 cross-sectional survey (6,783 respondents) and 1,069 people ages 75+ yearsfrom the 2002 survey (25,018 respondents). The survey is known as the Jamaica Survey ofLiving Conditions which began in 1989. The survey was drawn using stratified random sampling. This design was a two-stagestratified random sampling design where there was a Primary Sampling Unit (PSU) and aselection of dwellings from the primary units. The PSU is an Enumeration District (ED), whichconstitutes a minimum of 100 residences in rural areas and 150 in urban areas. An ED is anindependent geographic unit that shares a common boundary. This means that the country was 29
  45. 45. grouped into strata of equal size based on dwellings (EDs). Based on the PSUs, a listing of all thedwellings was made, and this became the sampling frame from which a Master Sample ofdwelling was compiled, which in turn provided the sampling frame for the labour force. Onethird of the Labour Force Survey (i.e. LFS) was selected for the JSLC [14, 15]. The sample wasweighted to reflect the population of the nation. The JSLC 2007 [14] was conducted May and August of that year; while the JSLC 2002was administered between July and October of that year. The researchers chose this survey basedon the fact that it is the latest survey on the national population and that that it has data on self-reported health status of Jamaicans. A self-administered questionnaire was used to collect thedata, which were stored and analyzed using SPSS for Windows 16.0 (SPSS Inc; Chicago, IL,USA). The questionnaire was modelled from the World Bank’s Living Standards MeasurementStudy (LSMS) household survey. There are some modifications to the LSMS, as JSLC is morefocused on policy impacts. The questionnaire covered areas such as socio-demographic variables– such as education; daily expenses (for past 7-day; food and other consumption expenditure;inventory of durable goods; health variables; crime and victimization; social safety net andanthropometry. The non-response rate for the survey for 2007 was 26.2% and 27.7%. The non-response includes refusals and rejected cases in data cleaning.MeasuresAge: The length of time that one has existed; a time in life that is based on the number of yearslived; duration of life. Or it is the total number of years which have elapsed since birth [16].Elderly (or aged, or seniors): The United Nations defined this as people ages 60 years and older[17]. 30
  46. 46. Old-Old. An individual who is 75 to 84 years old [9]Oldest-old. A person who is 85+ years old [9].Health conditions (i.e. self-reported illness or self-reported dysfunction): The question wasasked: “Is this a diagnosed recurring illness?” The answering options are: Yes, Cold; Yes,Diarrhoea; Yes, Asthma; Yes, Diabetes; Yes, Hypertension; Yes, Arthritis; Yes, Other; and No.Self-rated health status: “How is your health in general?” And the options were very good; good;fair; poor and very poor.Good health status is a dummy variable, where 1=good to very good health status, 0 = otherwiseIncome Quintile can be used to operationalize social class. Social class: The upper classes werethose in the wealthy quintiles (quintiles 4 and 5); middle class was quintile 3 and poor those inlower quintiles (quintiles 1 and 2).Health care-seeking behaviour. This is a dichotomous variable which came from the question“Has a doctor, nurse, pharmacist, midwife, healer or any other health practitioner been visited?”with the option (yes or no).Statistical AnalysisDescriptive statistics, such as mean, standard deviation (± SD), frequency and percentage wereused to analyze the socio-demographic characteristics of the sample. Chi-square was used toexamine the association between non-metric variables, and Analysis of Variance (ANOVA) wasused to test the relationships between metric and non-dichotomous categorical variables whereasindependent sample t-test was used to examine a statistical correlation between a metric variable 31
  47. 47. and a dichotomous categorical variable. The level of significance used in this research was 5%(i.e. 95% confidence interval).ResultSociodemographic characteristics of sampleOf the sample for 2002, 57.6% was female compared to 57.4% females in 2007. The mean age in2002 was 81.3 years (SD = 5.6 years), and this was 81.4 years (SD = 5.4 years) in 2007. Morethan two-thirds of the 2002 sample dwelled in rural areas, 20.8%. In 2007, the percent of samplewho resided in urban areas increased by 169.7%, and a reduction by 25.9% of those who dwelledin rural zones compared to a marginal reduction of 4.3% in semi-urban areas (Table 2.2.1).Concurrently, in 2007, 51.6% of sample reported suffering from an illness which was a 22%increase over 2002. Five percent more people sought medical care in 2007 over 2002 (ie 69.2%).Illness (or health conditions)A number of shifts in diagnosed health conditions were observed in this study. The number ofcases of hypertension and arthritis were observed between the two periods. The most significantincrease in health conditions was in diabetes mellitus cases (i.e. 576%) (Figure 2.2.1). A cross tabulation between self-reported illness and sex revealed that there was nosignificant statistical correlation between the two variables (Table 2.2.3). Although no statisticalassociated existed between the self-reported illness and sex, the percent of men who reported anillness in 2007 over 2002 increased by 30.5% compared to 16.4% for females. No significant statistical relationship existed between self-reported illness and maritalstatus (Tables 2.2.4, 2.2.5). In spite of the aforementioned situation, the divorced sample 32
  48. 48. reported the greatest percentage of increased in self-reported illness (16.7%) followed to marriedpeople (15.7%); separated individuals (11.6%), widowed (5.8%) and those who were nevermarried reported the least increase in self-reported illness (5.2%). No significant statistical correlation existed between self-reported illness and age cohortof respondents – P >0.05 – (Table 2.2.5). Although the aforementioned is true, the percent ofold-old who reported illness in 2007 over 2002 increased by 23.6% compared to a 16.6%increased in the oldest-old cohort over the same period. A cross tabulation between diagnosed self-reported health conditions and age ofrespondents revealed a significant association between the two variables (Table 2.2.6). Onexamination, in 2002, the lowest mean age was recorded by people who indicated that they hadarthritis. However, for 2007, the mean age was the lowest for old-old-to-oldest-old who hadreported the common cold. A shift which is evident from the finding is the mean age of thosewith diabetes mellitus in 2002 (79.5 yrs. ± 2.5 yrs), which was the second lowest age of personwith illness in 2002 to the greatest mean age for people with the same dysfunction in 2007 (90.20yrs ± 3.54 yrs) (Table 2.2.6). Based on Table 2.2.7, no significant statistical association was found between diagnosedhealth conditions and age cohort of the sample – P >0.05. In spite of this reality, some interestingfindings are embedded in the data across the two years. The findings revealed an exponentialincrease in diabetes mellitus and the common cold. However, the most significant increaseoccurred in diabetic cases in the oldest-old. Reductions were recorded in hypertension, arthritisand unspecified categorization. A cross-tabulation between self-reported illness (in %) and Income Quintile revealed asignificant statistical correlation between both variables for 2002 (χ2 (df = 4) = 11.472, P =0.022) 33
  49. 49. and 2007 (χ2 (df = 4) = 10.28, P < 0.05). Based on Figure 2.2.2, the poor had highest self-reported cases of illness compared to the other social groups. Although this was the case for2002 and 2007, the wealthy reported more illnesses than the wealthiest 20% of sample.Concurrently, the poorest 20% reported the greatest increase in self-reported illness for 2007over 2002 (19.4%) with the wealthy segment of the sample reported the least increase (2.7%). The first time that the Jamaica Survey of Living Conditions (JSLC) collected informationon self-reported illness and general health status (health status) of Jamaicans was in 2007. Basedon that fact, this study will not be able to compare the health status of the sample for the twostudied years; however, this will be the basis upon which future studies can compare. The cross-tabulation between the two aforementioned variables was a significantly correlated one (χ2 (df =2) = 39.888, P < 0.001) (Table 2.2.8).Health care-seeking behaviourA cross tabulation of health care seeking behaviour and aged cohort revealed no statisticalrelationship between the two variables for 2002 (χ2(df=1) = 0.004, P = 0.947) and for 2007(χ2(df=1) = 1.308, P = 0.253). Table 2.2.9 revealed that there is a significant statistical relationship between health care-seeking behaviour and health status of the sample (χ2 (df = 2) = 10.539, P = 0.005, cc=0.265).Further examination showed that 57.1% of old-old-to-oldest-old sought medical care, and ashealth status decreases the percent of sample seeking medical care increases. Of those whoreported poor health, 86.7% of them have sought medical care in the 4-week period of thesurvey. When the aforementioned association was further investigated by aged cohort, thedifference was explained by old-old (χ2 (df = 2) = 11.296, P = 0.004, cc=0.305) and not oldest-old (χ2 (df = 2) = 0.390, P = 0.823) (Table 2.2.10). 34

×