A comparative analysis of the health status of men aged 60–72 years and men aged 73+ years in Jamaica: Are there differences across municipalities?

ABSTRACT Background Since 1990, the number of older men (60+ years) in Jamaica has increased to in excess of 100 000, while there are 30 000 men aged 73+ years. This is despite the fact that men have higher mortality and morbidity rates than women and seek medical treatment less frequently than women. There exists, however, a dearth in literature regarding this phenomenon and, therefore, this study has endeavoured to reduce this gap. Objectives The aims of this study were to, (1) model the health status of men aged 60–72 years in Jamaica, (2) model the health status of men aged 73+ years in Jamaica and (3) examine the disparity in health status of the two groups in order to ascertain the factors that influence the good health status of elderly men. Method A sample of 1432 men aged 60+ years were extracted from a survey of 25 018 Jamaicans. Secondly, a sub-sample of 633 men aged 73+ years was extracted from the 1432 men aged over 60. The data from which those samples were extracted is called the Jamaica survey of living conditions (JSLC). The JSLC began in 1988 from a model of the World Bank's Living conditions survey and is a nationally cross-sectional probability sample. The current study used descriptive statistics to provide background information on the sociodemographic characteristics of the sample and logistic regressions were utilised to examine the factors that predict good health of men aged 60–72 years and men aged 73+ years in Jamaica. Results The average age of the sample of men aged 60+ years was 71.14 years (SD = 5.64 years) and 78.5 years for the sample of men aged 73+ years (SD = 7.97 years). Approximately 63% of men aged 60–72 years indicated that their health was good compared to 53.3% for men aged 73+ years. Rural men recorded the least health status across the age cohorts. With regard to predictors of good health, the same factors were found to determine approximately 27% of the variability in good health. Ownership of health insurance was found to be the most influential predictor of good health and positive affective psychological condition the least significant predictor of good health for men aged 73+ years and second to last of five factors for older men. Conclusion The highest self-reported good health was indicated by men aged 60–72 years and men aged 73+ years who dwelled in towns other than Kingston, the capital. The least good health was experienced by rural men. For older men, health insurance coverage does not indicate preventative health, but preparation for curative ill-health.


INTRODUCTION
From 1880 to 1882, life expectancy at birth for women in Jamaica was 39.8 years compared to 37.02 years for men (Table 1). In 2004, more than a century later, women were outliving men by 6 years (Table 1). In Jamaica, population ageing is a feminised phenomenon. This is typically the same around the world. From 1950 to 1955, world statistics showed that life expectancy at birth for women was 47.9 years, compared to 45.2 years for men, indicating that former sex was outliving the latter by 2.7 years. 1,2 The disparity in life expectancy at birth between the sex cohorts increased to 4.2 years between 2000 and 2005. 1 According to Jamaica's Demographic statistics 3 , 10.9% of females were aged 60+ years, compared to 10.3% of males. For the world, in 2000, 11.1% of the female population was older than 60 years, compared to 8.9% of men. Concomitantly, world statistics indicated that a woman who is 60 years old is likely to live for an additional 20.4 years, compared to 17 years for men. 1 Life expectancy is one of the indicators of the health status of an individual or population, which implies that women are enjoying a better health status than men.
Courtenay 4 noted, from research conducted by the Department of Health and Human Services and Centers for Disease Control, that from the 15 leading causes of death (except Alzheimer's disease), the death rates were higher for men and boys in all age cohorts, compared to women and girls. Embedded within this theorising are the differences in fatal diseases explained by gender constitution, 5 which Courtenay 5 contributed to behavioural practices of the sexes that cause men to die approximately 6 years earlier than women. 6 Studies have shown, however, that women have a higher propensity to contract particular conditions such as depression, osteoporosis and osteoarthritis. 7, 8 Herzog 8 noted that '… it appears that older women are more likely to be impaired by their health problems, while older men (aged 60+ years) are more likely to die from them.' A study that was conducted by Schoen et al. 9 on a group of adolescents, revealed a different finding from what was reported by the other studies. 7,8 They found that men were more likely than women to feel stressed, 'overwhelmed', or 'depressed'; they attributed this to the limited nature of men's social networks.
Schoen et al. 9 found that men in general tend to be more stressed and less healthy than women and further argued that men are more likely to use denial, distraction, alcoholism, and other social strategies to conceal their illness or disabilities. 10,11,12,13 On the other hand, Herzog 8 , referring to studies from a number of experts, wrote that women had higher rates of depression than their male counterparts. Could suicide among the aged be the result of depression? This is likely to be underreported, because other illnesses are often present and given as cause of death. Hertzog 8 noted that data on suicide and depression yielded different results and, therefore, suicide is not necessarily an indicator of depression.
Along with a longer life expectancy, particularly for women, the number of years spent being unhealthy are on the rise. 14 In an attempt to calculate the 'quality of lived years', the World Health Organization (WHO) 14 developed the disability adjusted life expectancy (DALE) scale in order to account for unhealthy years in relation to life expectancy. The DALE does not only use length of years to indicate health and well-being status of an individual or a nation, but also incorporates the number of years lived without disabilities. The institution found that these accounted for a 14% reduction in life expectancy for poorer countries and 9% for more developed nations.
Jamaica is a developing country, which means that, according to the DALE, both sexes are experiencing 14 years of unhealthy life expectancy. In spite of this, yearly on average (since 1990), there are 565 men who cross the threshold of the life expectancy in Jamaica (72.3 years at birth). This figure is included in the 1842 men who cross the 60-year-old threshold annually; 30 of whom (8%) are older than their life expectancy at birth. Men and women are living longer, but the former seek health care less frequently (Table 2). Table 2 shows that men reported less illness/injury than women, sought less medical care and spent more time in health care facilities, all of which accounts for the disparity in life expectancy between the sexes.
Irrespective of the self-reported health conditions given by men, they experience higher rates of morbidity and mortality than women in Jamaica. 15 The Jamaican Ministry of Health's publication 15 showed that, of the five leading causes of death -malignant neoplasms, cerebrovascular disease, heart disease, diabetes mellitus and homicide -men outnumbered women in three of them. The risk of developing malignant neoplasms is 39% higher for men than for women and, similarly, 71.2 per 100 000 men develop heart disease, compared to 66.1 per 100 000 women. On the other hand, the risk of cerebrovascular disease is 14% higher, and diabetes mellitus is 64% higher, for women than for men. 15 In 2007, approximately 11% of men were older than 60 years (N = 132 931, Table 3). Using the results from Table 3 and Bourne's 16 earlier study, it was calculated that 40.2% of elderly Jamaicans reported suffering from at least one dysfunction (N = 118 603), 13.1% of men reported ill-health (N = 173 135), 75.1% of elderly people who reported ill-health had recurring ill-health (N = 89 071) and 72% of the elderly who had self-rated ill-health sought medical care (N = 85 394). It can be extrapolated from the data that approximately 5% of the 13.1% of self-assessed health conditions are accounted for by elderly men. Furthermore, it  can be extrapolated from these statistics that 3.8% of elderly men expressed having recurring ill-health. Has the rationale for not studying older men's ill health conditions been due to the fact that only 5% were subject to such conditions?
Many studies have been done on elderly Caribbean nationals and Jamaicans in particular. 16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33 An extensive review of the literature, however, showed that none have specifically examined men's health, or those factors that influence the good health of older men (aged 60+ years) in Jamaica. In spite of pressure by the WHO and some scholars in a drive to examine the social determinants of health 33,34,35,36,37,38 in the Caribbean and, in particular in Jamaica, no work has been done on this subject area. This study is innovative, as it seeks to investigate the social, psychological and environmental determinants of the health status of older men. Studies on older Caribbean nationals are not the same as an investigation of the health status of older men (aged 60+ years) in Jamaica. The aims of this study were to, (1) ascertain factors that influence good health status of elderly men (aged 60+ years) in Jamaica, (2) ascertain the determinants of good health status of older men, (3) determine the potency of each variable and (4) distinguish between determinants of the men.

Theoretical framework
Many studies have employed multivariate analyses in the examination of health status. 16,17,26,39,40,41,42,43,44 The use of econometric analysis in the study of health was developed by Michael Grossman 44 . This approach simultaneously captures biomedical and non-biomedical variables, unlike the bivariate analysis that is only able to investigate two variables. Based on the WHO's definition, 45 health is inclusive of biomedical, socio-economic and psychological factors. Health, therefore, is determined by many factors and the use of an econometric model makes it possible to identify these. A multivariate model has a fundamental advantage over bivariate relations, as health is a multidimensional phenomenon; this model is able to capture more variables and without excluding some variables that cannot be accommodated in a bivariate association.
The theoretical framework that underlines the present work was developed by Bourne 17 and is a modification of the work of Grossman 44 and Smith and Kingston 43 . Grossman was the first to establish an econometric model to evaluate the health status of people. The model encapsulates some variables that determine health status of people in the world and can be represented in the following equation: In [Eqn 1], H t represents the current health in relation to time period t. The stock of health in the previous period is shown by (H t-1 ), health behaviours, such as smoking, excessive drinking and good personal health is represented by B t , with exercise included as a separate variable -G o . The use of medical care MC t , education of each family member (ED) and all sources of household income (including current income) are also included.
Grossman's model was further expanded upon by Smith and Kingston to include socio-economic variables, as seen in the following equation: This second equation expresses current health status H t as a function of stock of health (H t-1 ), price of medical care P mc , the price of other inputs P o , education of each family member (ED), all sources of household income (E t ), family background or genetic endowments (G o ), retirement related income (R t ) and asset income (A t ).
Given that particular conditions influence the elderly differently from other age cohorts, Bourne used an econometric analysis to build a model that captures variables that influence the subjective well-being of elderly Jamaicans. This is represented by the following equation:

Design
This research study used secondary data collected jointly by the Planning Institute of Jamaica (PIOJ) and the Statistical Institute 3 of Jamaica (STATIN). For this paper, the sub-sample was 1423 elderly men (aged 60+ years). The mean age of the sub-sample was 71.14 years (SD = 7.97 years). Another sub-sample was extracted from the survey, which comprised 633 men aged 73+ years (i.e. men living beyond the life expectancy, which is 72 years in Jamaica). The two sub-samples were extracted from a larger, nationally prevalent study, conducted between June and October 2002, of some 25 018 respondents. Stratified random sampling techniques were employed to design the survey (i.e. the Jamaica survey of living conditions [JSLC]) and detailed selfadministered questionnaires were used to collect the data from the respondents. The questionnaire was modelled from the World Bank's Living standards measurement study (LSMS) household survey. There were some modifications made to the LSMS as the JSLC is more focused on policy impacts. The questionnaire covered questions on sociodemographic, economic and wealth variables, crime and victimisation, social welfare, health status, health services, nutrition, housing and physical environment.
Interviewers who collected the data were trained to address the questions and concerns of interviewees. Data were stored and retrieved in the SPSS program (SPSS Inc; Chicago, USA) and, for the present research, descriptive statistics were used to provide certain sociodemographic characteristics of the sub-sampled population.
Based on the principles of parsimony (i.e. all variables that should be included were included), the final model only constituted those variables that were statistically significant (i.e. p < 0.05). This was attained by the using the health literature and the variables that were included within the framework of the current data set.
Demographic characteristics were provided for the sample and the sub-sample of men aged 60+ years and 73+ years. Logistic regression was used to establish, (1) a model for good health status of elderly men in Jamaica, (2) Wald statistics to examine the contribution of each significant variable in the model and (3) the odds ratios interpreted to address the difference within each variable.
Multivariate analysis (logistic regression) was used because the researcher wanted to test a number of variables simultaneously, and the fact that the dependent variable was binary; the most fitting statistical technique was logistic regression. The model that was tested in this study is adapted from Bourne's model [Eqn 3], to represent the following: The results were presented using unstandardised coefficients, Wald statistics, odds ratio (OR) and confidence interval (95% CI). The predictive power of the model was tested using the Hosmer and Lemeshow test, 46 to examine goodness of fit of the model. The correlation matrix was examined in order to ascertain if autocorrelation (or multi-collinearity) existed between variables. Based on Cohen and Holliday 47 , the correlation can be weak (0-0.39), moderate (0.4-0.69), or strong (0.7-1.0). This matrix was used to exclude (or allow) a variable in the model. Wald statistics were used to determine the magnitude (or contribution) of each statistically significant variable in comparison with the others, and the OR for the interpretation of each significant variable.

Analysis Key terms
Health: This is defined as the self-rated health status of an individual.
Good health: This variable is derived from a number of questions that enquired about particular health conditions. It is a binary variable where, 1 = no reported ill-health and 0 = reported at least one health condition. 39 Age: This is the total number of years lived since birth, measured from one birthday to the next. Married: This is a binary variable, where 1 = those who indicated they were married and 0 = otherwise.
Poverty level: This is also a binary variable, where 1 = those people who are in the two poor quintiles (i.e. poorest and poor) and 0 = otherwise (i.e. those wealthier individuals in quintiles 3-5).
Crime index: In this study, the crime index was calculated using the following equation: This equation represents the frequency with which an individual witnessed or experience a crime, where i denotes 0, 1 and 2 (0 indicates not witnessing or experiencing a crime, 1 indicates witnessing one or two crimes and 2 indicates seeing three or more crimes). The variable T j denotes the degree of the different typologies of crime witnessed or experienced by an individual, where j is define by the following scale: 1 = valuables stolen, 2 = attacked with or without a weapon, 3 = threatened with a gun and 4 = sexually assaulted or raped. The summation of the frequency of crime by the degree of the incident ranges from 0 to a maximum of 51.

Area of residence:
This is the general geographic locale in which an individual resides, where 1 = Kingston metropolitan areas (which are all the areas that are 100% urban) and 0 = otherwise, and 1 = peri-urban areas (which are places that are not 100% urban) and 0 = otherwise. The reference group is from a rural area.
population was married (50%), owned their own homes (85%), resided in rural areas (68%) and reported good health (63%). The majority of men had, at most, primary level education (62%); however, 3% had attained tertiary level education and 98.6% reported that they were the head of their household. Per capita income was evenly distributed, with 23.8% being in the wealthiest quintile. Furthermore, crime seemed to have minimal effects on the respondents.
In terms of area of residence, 70.9% (n = 449) resided in rural areas, 19.3% (n = 122) in peri-urban areas; and 9.8% (n = 62) in urban areas. A little more than half (53.5%, n = 333) reported good health. The majority of these elderly men had, at most, primary level education (67.4%, n = 402), while only 2.7% had attained tertiary level education. Almost all of the men (98.4%) indicated that they were the head of their households.
Per capita income was evenly distributed, with marginally more individuals being in the wealthiest quintile (23.4%, n = 148) and the poorest quintile (21.6%, n = 137). The average consumption per person in the men aged 73+ years was $JM77 877.07 (SD = $JM72 014) -at the time of the study the exchange rate was $US1.00 = $JM50.97. In addition, crime seems to have a minimal affects on men aged 73+ years.

Analysis of logistic regression on good health of men aged 60-72 years
Of the 16 predisposed variables that were used in the model ( With respect to factors that have a positive effect on health, positive affective psychological conditions (Wald statistic = 11.67, 95% CI = 1.04-1.16) accounted for more variability than area of residence. On examining positive affective psychological conditions, it was found that when an elderly man experienced a positively affective condition, he was 1.1 times more likely to report good health. Findings revealed that men who reside in rural areas suffer from diminished good health. This means that those in peri-urban areas are 1.5 times (odds ratio, 95% CI = 1.06-2.13) more likely to report good health compared to an elderly man who dwelled in rural Jamaica. The elderly men who resided in the Kingston metropolitan area were 1.6 times (odds ratio, 95% CI = 1.02-2.52) more likely to report good health compared to those in rural Jamaica.

Analysis of logistic regression on good health of men aged 73+ years
What factors account for good health in the men aged 73+ years? Table 5 shows that, of the 15 predisposed factors that tested for the initial model (good health of men aged 73+ years), five explain the variability in good health. These determine 27.7% of the variability in good health (chi-square (18) = 132.21, p = 0.001, Nagelkerke R-square = 0.277, -2 log Likelihood = 653.92). These five variables are: age, secondary level education, ownership of health insurance, area of residence and positive affective psychological conditions. Three of the explanatory variables negatively contribute to good health (age, secondary level education and health), and two positively affect good health (area of residence and positive affective psychological conditions).
The model had statistically significant predictor power (model χ 2 = 132.21, p < 0.001, Hosmer and Lemeshow goodness of fit χ 2 = 14.474, p = 0.070) and correctly classified 71% of the sample (correctly classified 84.9% of those who had good health and 55.1% of those did not report poor health).
Ownership of health insurance carried the most weight in determining good health (Wald statistic = 53.6, 95% CI = 0.029-0.129), followed by secondary level education with reference to primary level education (Wald statistic = 8.38, 95% CI = 0.357-0.820), living in peri-urban areas (Wald Statistic = 7.609, 95% CI = 1.230-3.396) and the least, dwelling in Kingston metropolitan area (Wald statistic = 4.396, 95% CI = 1.053-4.577). Embedded in these findings are the realisations that, (1) good health of men aged 73+ years are eroded with years of life, (2) those with primary level education enjoy a better self-reported health than those with secondary and tertiary level education, (3) owning health insurance does not positively contribute to good health, it is only an indicator of those who are likely to have poorer health, (4) men aged 73+ years who dwell in peri-urban areas are more likely to enjoy greater self-reported good health, followed by those who reside in the Kingston metropolitan area and, lastly, by those in rural areas and (5) men aged 73+ years that are experiencing more positive affective psychological conditions are 1.1 times more likely to report good health.

DISCUSSION
All of the reports by the United Nations 1,2 and the WHO 14 , coupled with those of the Jamaican Ministry of Health 15 and the JSLC that have been published on population, ageing, health or gender issues, have shown that women outlive men. The disparity in the life expectancy rate between the sexes is 6 years in Jamaica and 8 years using data on the world. Living longer means having to defy the odds of mortality for more years. This occurrence is accounted for by healthy lifestyle practices, implying that unhealthy lifestyle practices lead to higher mortality and morbidity in men than women. In spite of these realities, there are men living beyond the life expectancy in their respective geopolitical areas of residence. In Jamaica, the life expectancy for men is 72.3 years. The term the researcher has coined that refers to men who are alive beyond the life expectancy of their nation is men aged 73+ years.
Over an 18-year period (ending 2007), there were 40 948 men aged 73+ years in Jamaica and the average yearly increase of men aged 73+ years the period was 565. This means that there are men living beyond the expected life expectancy, mortality and disease-causing mortality rates. In spite of the high mortality and morbidity of men, this study provides information on what constitutes good health for older men.
Literature has shown that the good health status of people lessens as they become older. 48,49,50,51,52 This study concurs with this finding, with 63% of men aged 60-72 years reporting good health, compared to 53.5% of men aged 73+ years, indicating that health decreases with ageing. This point was further reinforced by the finding in which age was a factor of good health in each of the models. Age as a factor was ranked as the second most influential in determining 'good' health (or lack thereof) for men aged 60-72 years compared to being one of the four influential factors for men aged 73+ years. However, age is not the only factor that affects good health of elderly men.
Other studies have shown that education influences good health and that tertiary level education is positively associated with better health. 4,16,17,24,26,33,34,35,36,37,38,39,43,44 This study agrees that education influences good health and that there is no difference between the health status of elderly men with primary and tertiary level education. Another interesting finding to emerge from this study is the fact that older men with primary level education in Jamaica have better self-reported health than those with secondary education. This contravenes other research that have shown that better quality education determines higher quality of health, but this is not the case for men who are living beyond 59 years.
Poverty, overcrowding, consumption and marital status in other studieshave shown to influence good health 17,19,24,26,43 ; however, this is not the case for elderly men in this study. The fact that living beyond a particular year of birth (60 years) means that the individual has surpassed the need for certain material possessions and appetite for some foods; therefore, having . Studies done about the elderly have indicated that overcrowding, consumption and marital status determine well-being, but in this study it was found not to be the case. The current work has refined the factors which effect elderly men and established that they are somewhat different from those that influence the health and well-being of elderly Jamaicans.
Although poverty does not directly relate to good health of elderly Jamaicans, good health has been found to differ based on area of residence. In 1997, the prevalence of poverty in the country was 9.9% and, 10 years later (2007), it had increased to 19.9%. 53 Despite this exponential increase in the prevalence of poverty, 71.3% of the stock of poverty is accounted for by rural areas. Based on the data for 2007, 46.6% of elderly Jamaicans dwelled in rural areas, 20.9% in peri-urban areas and 32.5% in the Kingston metropolitan area. 53 Within the context of the current study, it was found that the state of health of elderly men in rural areas were worse than other areas of residence and that poverty indirectly influences health. Furthermore, the best health is likely to be experienced by other town dwellers, but not those who dwell in the Kingston metropolitan area (100% cities). The stock of poverty for elderly residents of urban areas was 2.2 times (19.9%) greater than that of the distribution of poverty in peri-urban areas (8.9%), indicating that poverty indirectly determines good health of aged residents.
It has well established that positive affective psychological conditions are correlated to health 17,54,55,56,57,58,59,60 and this was concurred by the current study. Lyubomirsky 54 shows that happier people view life in a positive manner. This attitudinal state explains how decisions are influenced and moods experienced. With a positive attitude, a better quality of life is experienced, as the individual thinks, acts, builds and carries out their life's task with more self-assurance. 55 The contrary is also true -a pessimistic individual is more likely to have a lower self-esteem, less self-fulfilment and less self-actualisation than someone who is optimistic. DeNeve and Cooper 56 have found that happier people are more optimistic and positive in nature. Diener and Seligman 57 point out that moods are not stationary, thus happy people can have negative moods, but they do not dwell on the negatives indefinitely. Harris and Lightsey 58 have established that negative affective conditions such as guilt, fear, anger and disgust inversely affect subjective well-being -just as positive factors directly influence well-being. 59 However, this was not the case for elderly men in Jamaica. The literature has shown that the elderly seek more health care than any other age cohort, so their psychological state is directly influenced by their physical condition. If an elderly individual does not perceive that they have control over an illness or disability, it may result in self-destructive behaviour, 60 which will negatively influence well-being. McCarthy 60 offers a further justification for the correlation between psychological state and subjective well-being, when he writes that diabetic patients are six to seven times more likely to suffer from psychiatric illnesses, anxiety and depression than non-diabetic patients. In the current work, it was found that aged men who are experiencing positive affective psychological conditions are 1.1 times more likely to report good health and that this variable minimally contributes to good health for elderly Jamaican men.
Ownership of health insurance coverage does not only indicate health-seeking behaviour, it also means that men who have surpassed 60 years purchased more health insurance if they believed that they were more likely to become ill. Hence, health insurance is not a preventative measure; instead it is a product that is more demanded by this cohort of men who are more likely to report ill-health. This finding denotes that men aged 60+ years who own this product are using it as a cost reduction mechanism because they are aware that, as a result of their illhealth, the frequency with which they will need to visit health facilities will increase.

CONCLUSION
In summary, the good health of men aged 60+ years deteriorates as they become older. This study has shown that there is no difference between the factors that determine good health of men aged 60-72 years and men aged 73+ years. Good health is strongly influenced by ownership of health insurance coverage, but not by positive affective psychological conditions. Men aged 60-72 years and men aged 73+ years who resided in rural Jamaica reported the least good health and the greatest selfreported good health was experienced by those in peri-urban areas. This study is the first of its kind as no literature exists with which to conduct a comparative study. This limitation, however, does not hamper it from providing insight into the health status of men aged 60-72 years and the factors which predict good health for this group, as well as men aged 73+ years.