| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to Yael Benyamini, Bob Shapell School of Social Work, TelAviv University, Tel Aviv 69978, Israel. E-mail: benyael{at}post.tau.ac.il
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Longevity • Subjective health • Self-assessments of health • Old age • Old-old
The main goal of the current study is to study these gender differences in the association between SRH and mortality. Specifically, there are three objectives: (a) to examine these differences in two different time frames (4 and 9 years of mortality follow-up), (b) to examine these differences among old and old-old people, and (c) to address the question of whether this association is based on the accuracy of poor SRH as a predictor of future decline, and/or of better SRH as a predictor of longevity. The rationale is that learning for whom SRH predicts mortality could provide more information regarding the explanations that have been proposed for this phenomenon (Idler & Benyamini, 1997). If we understand better how SRH is linked to mortality we will also know more about what we can learn from elderly people's SRH. Such findings could advise researchers and practitioners as to when and for whom SRH is in general more (or less) accurate as a predictor of future health, and therefore is likely to be useful.
The study is based on the Cross Sectional and Longitudinal Study (CALAS), with a national sample of elderly Israelis ages 75–94, stratified by age, gender, and place of origin. Thus, there are similar numbers of women and men, and similar numbers of old (75–84) and old-old (85–94) people, enabling rigorous tests of gender and age differences. This is in contrast with studies with nonstratified sampling, in which there are smaller numbers of old-old participants, and in particular of older men.
Gender Differences in the SRH–Mortality Association
Although there are more studies reporting a stronger association for men, compared with women, there are a few studies showing very little gender difference (e.g., Scott, Macera, Cornman, & Sharpe, 1997), and several studies that show a stronger association among women (e.g., Wolinsky & Johnson, 1992). Moreover, although the gender difference is usually apparent for the association between the lowest level of SRH, typically labeled "poor" SRH, and mortality, in some cases the gender difference was found only for the middle range of "fair" and "good" SRH (compared with the reference level of "excellent" SRH). Even these findings are not consistent, as one study reported that the middle range of SRH predicts mortality for women (McCallum, Shadbolt, & Wang, 1994) whereas another study reported that it predicts mortality better for men (Helmer, Barberger-Gateau, Letenneaur, & Dartigues, 1999). To complicate matters more, gender differences sometimes appear only for some causes of death (Kaplan & Camacho, 1983), or in some but not all age groups (e.g., poor SRH significantly predicted mortality only for middle-aged men, Idler & Angel, 1990; only for young-old but not for old-old women or for men, Simons, McCallum, Friedlander, & Simons, 1996). One study reported a stable association of poor SRH with mortality among men over varying follow-up periods, whereas among women it declined over time (from 5 to 32 months of follow-up; Grant, Piotrowski, & Chappell, 1995), another reported declines in the SRH–Mortality association over time for both genders (Scott et al., 1997), and yet another study reported that SRH predicted mortality equally well for both genders in both a short and a longer time frame (Idler & Kasl, 1991).
Proposed Explanations for the SRH–Mortality Association: A Gender Perspective
If we take into account all that is known about physical, psychological, behavioral, and social differences between women and men, it is not surprising that there are differences in the way they form self-ratings of health, or in the association between SRH and other health outcomes. Specifically, there are two types of well-documented health-related gender differences that could be linked to two of the explanations for the SRH–Mortality association suggested by Idler and Benyamini (1997): Differences in the accuracy and prevalence of self-reports of health, and differences in the trajectories of health at old age. Idler and Benyamini (1997) hypothesized that SRH may be a more inclusive and accurate measure of health status (labeled the "sponge" hypothesis by Wolinsky & Tierney, 1998). Women may be more aware of physical symptoms, and thus their SRH may be more accurate. However, their reports of disease, symptoms, and other covariates could also be more accurate, leaving little information to be supplemented by SRH, and thus weakening the SRH–Mortality association for women. Van Doorn (1998) proposed this argument as a possible explanation for her findings that wives' ratings of their husbands' limitations and longevity independently predicted mortality: Wives may be correcting for their husbands' underreporting of conditions, whereas the opposite does not occur. Another study provided evidence that women's SRH is indeed more inclusive (Benyamini, Leventhal, & Leventhal, 2000): It was found to be sensitive to a wider range of health problems and to be related to negative affect that reflects a wider range of life circumstances. The latter difference accounted for the weaker association of SRH to mortality among women, compared with men. Thus, although there are several reasons to believe that women's SRH is more inclusive, it is not clear whether or how this affects the SRH–Mortality association, because greater inclusiveness could result either in greater accuracy of health assessments, or lesser accuracy because of the inclusion of minor issues that are unrelated to mortality.
The second hypothesis regarding the SRH–Mortality association deals with another type of accuracy in judging one's health (Idler & Benyamini, 1997): it states that SRH is a dynamic evaluation, judging the trajectory of health status (labeled the "trajectory" hypothesis by Wolinsky & Tierney, 1998). Data from a large study with a variety of covariates showed that declines in SRH showed a stronger relationship with mortality risk than did a single measure of SRH (Ferraro & Kelley-Moore, 2001). At least at old age, when most deaths occur, the trajectories of women's and men's health are indeed different: Women survive longer but experience more years of disability and ill health (Manton, Woodbury, & Stallard, 1995). Thus, when an elderly man judges his health to be poor, on the average he is more likely to be closer to his death than a woman of the same age who judges her health to be poor. In a similar vein, an elderly woman who judges her health as excellent in on the average more likely to live longer than a man her age who rates his health as excellent. These expectations are congruent with women's and men's trajectories of health at old age (75–84 years) but not at old-old age, when mortality is very high for both genders, and the gender differential in mortality rates diminishes (Manton et al., 1995; Wingard, 1984).
According to this explanation, we could expect men's poor SRH to predict mortality within the next few years better than women's poor SRH (as was found in most studies). Paradoxically, this explanation would also lead us to expect women's excellent SRH to predict longevity within the next few years better than men's excellent SRH. The predictive effect of SRH on mortality is typically calculated by computing the odds of survival for those who rated their health as poor versus those who rated their health as excellent. If the two effects, of greater accuracy of men's poor SRH and of women's excellent SRH, were similar in strength, then they would balance each other, leading to little or no gender difference in the SRH–Mortality association. However, these effects could be less pronounced at certain ages, those in which past trends in health status are less likely to predict future changes. These ages are higher for women than for men, because of the difference in their longevity and trajectories of decline at old age. Therefore we will also repeat our analyses for two age groups, old (75–84 years) and old-old (85–94 years).
Research Questions and Hypotheses
The "sponge" and "trajectory" hypotheses along with the findings reviewed above lead us to formulate two research questions regarding the SRH–Mortality association among women and men:
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
Measures
Sociodemographic Variables
(a) Age (continuous); (b) origin/place of birth (Europe–America, Israel, North Africa, and the Middle East; the latter two were grouped together for sampling purposes, but were separated for the purpose of statistical analyses); (c) education categorized into 2 groups (0–8 years, 9+ years); (d) source of income (only Social Security, or Social Security plus any other pension or income); and (e) living arrangements—in the community alone, with spouse only, with others (with or without spouse), or in an institution.
Measures of Health and Functioning
We assessed SRH by asking "Would you say your health now is excellent, good, fair, or poor?" and coded it numerically so that excellent = 1 and poor = 4; we recorded medication usage, a continuous measure of up to 11 drugs. We asked interviewees if and which drug they took for a specified list of chronic conditions (heart disease, diabetes, hypertension, etc.) and for other reasons. To ensure the validity of the information we asked them to display the containers of all the medications they were currently taking, and we recorded the labels. We measured limitations in activities of daily living (ADL) by asking whether the respondent had difficulty in performing any of seven ADLs: walking across a small room, washing, grooming, dressing, eating, transferring, and toileting (Katz, Downs, Cash, & Grotz, 1970). We recoded this measure as 0, needing no human assistance in any of these measures, or 1, needing human assistance in one or more activity. We measured physical disability by asking whether the respondent had difficulty in bending, stooping, or crouching; walking 1 km; climbing 10 stairs without resting; or carrying 5 kg (a short version based on Harris, Kovar, Suzman, Kleinman, & Feldman, 1989; Nagi, 1976; and Rosow & Breslau, 1966). It was recoded as 0 (no difficulty), 1 (one or two difficulties), or 2 (three or four difficulties). We assessed cognitive status with the Katzman and colleagues (1983) 6-item scale, that included time orientation (measured by three questions that asked about the year, month, and time of day when the interview took place), memory (measured by asking to repeat a certain name and address), and concentration (measured by two questions that asked the participant to count backwards from 20 to 1, and to recite months of the year in reverse order). We scored the scale as normal (0–8), mild impairment (9–19), and cognitively impaired (20+). We measured depressive symptoms by the Center for Epidemiologic Studies Depression Scale (CES-D), a self-report scale that measures depressive symptomatology in the general population (Radloff, 1977). The use of this scale in the old-old Israeli population has been previously reported (Ruskin et al., 1996). Depression was coded into two groups according to the commonly used cut-off point for the CES-D, with a score of 16 or higher indicating high depression.
Statistical Analyses
We predicted mortality from SRH, with an adjustment for sociodemographic and health covariates, by using Cox proportional hazards models. We entered covariates according to the categories listed herein. To avoid exclusion of participants with missing data on only one or two measures, we created dummy variables to represent missing data on education, income, cognitive status, and depression. Data were complete on all other measures.
Procedure
Multilingual trained interviewers interviewed participants in their residence (home or institution). When necessary, they conducted the interview in the native language of the respondent, and in some cases with the aid of a translator (13% of participants were not interviewed in Hebrew). The interview lasted 1.5–2 hr and took place after participants had signed an informed consent form.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
2 = 34.9, p <.0001). Because of the small number of respondents in the "excellent" SRH category, this category was combined with the "good" SRH category for all further analyses.
SRH as a Predictor of Short- and Long-Term Mortality
The association between SRH and mortality was tested for each gender group over the two follow-up periods: First, the short-term, from baseline to the end of 1994; and second, the long-term, 5 years from the beginning of 1995 to the end of 1999 (excluding respondents who died within the first follow-up period). The follow-up periods amount to roughly 4 and 9 years, respectively. Cox proportional hazard models predicting survival were conducted in three steps: SRH alone, with adjustment for sociodemographic measures, and with adjustment also for measures of health and functioning. Table 2 shows the odds ratios (OR) for fair and for poor SRH, the main variable of interest, after each stage of adjustment.
|
The SRH–Mortality Association Among the Old and the Old-Old
Our second question relates to possible age differences in the SRH–Mortality association. Because there was no association between SRH and mortality in the longer follow-up period, these analyses were conducted only for the shorter time frame (see Table 3). For the younger group, SRH predicts survival time better for women, when entered alone and as covariates are added. In the final model, the risk for mortality among women reporting poor SRH is still almost twice that of women reporting excellent or good SRH, although for this smaller sample this increased risk is not significant (p =.11; CI 0.85–4.90). For men the OR for poor SRH is lower than for women. In the final model there is no increase in risk for men with poor compared with excellent or good SRH. In the older age group results for women are similar to those for men: The OR for poor SRH is lower than that for the younger group, and it further diminishes and is nonsignificant after the health and functioning variables are added.
|
|
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
SRH as a Predictor of Short- and Long-Term Mortality
Hazard models take into account the duration of survival, but they do not tell us whether a certain measure predicts mortality similarly for people of varying survival periods. Therefore, we tested the SRH–Mortality association in two time ranges. In the shorter time range, from baseline to approximately 4 years later, the OR for poor SRH is similar for women and men, but it diminishes faster for men with the addition of sociodemographics and health and functioning measures. The final model shows poor SRH, compared with excellent or good SRH, to predict mortality more strongly for the women in our sample, though the difference between the gender groups is not significant.
In the long range, from approximately 4 to 9 years past baseline, poor SRH barely increased the risk for mortality, compared with excellent or good SRH. After the covariates are added, there is virtually no association between SRH and mortality. These findings are in accord with our expectations: If SRH predicts mortality because it is more inclusive and accurate than the covariates used (the "sponge" hypothesis), then it should predict short-term mortality better than long-term mortality, especially at very old age. The findings differ from those of Mossey and Shapiro (1982), who found similar associations for the first 3 years past baseline and for the next 3 years. This may be due to the younger age of their participants (most of them were between 60 and 74, which is younger than the present sample).
Differences in the SRH–Mortality Association Between the Old and the Old-Old
Our second question dealt with the prediction of mortality from the SRH of the old (75–84) compared with the old-old (85–94) participants. The ORs for women's and men's poor SRH compared with good or excellent SRH are higher for the young-old than for the old-old groups. After the health and functioning variables are added to the models, for men in both age groups and for women in the older age group, there is a weak and nonsignificant association between SRH and mortality. Among the younger women, the OR for poor SRH is higher and a twofold risk of mortality (not significant) remains even after health status is controlled for.
Poor Versus Good or Excellent SRH as Predictors of Mortality
Mortality rates within each gender, age, and SRH group suggest that these differences between the young-old and the old-old groups, as well as those between the young-old women and men, are due to the greater accuracy of excellent SRH as a predictor of longevity among the young-old and in particular among the young-old women. The OR for poor SRH in the older group is low, not because people who rated their health as poor survived throughout the follow-up period, but rather because many of the people who rated their health at the higher end did not survive. In the younger group, poor SRH is more accurate in predicting mortality among men compared with women, as expected, but the gender difference in mortality rates among those in the excellent or good SRH group is even higher: Few women in this group died (12%), compared with men (29%). Thus, the higher OR for poor versus excellent or good SRH among women in this age group seems to occur despite the lower accuracy of their poor SRH, and it may be due to the greater accuracy of their excellent or good SRH.
Most studies of the SRH–Mortality association computed OR for poor SRH in reference to excellent SRH as a predictor of mortality (or the other way around; see Idler et al., 2000). SRH, like any predictor that is not perfectly accurate, can show two types of "mistakes": a false positive, equivalent in our case to rating your health as poor but surviving for a long period of time, and a false negative, equivalent to rating your health as excellent or good but not surviving for long after that. The two mistakes are independent of one another. If SRH is not a unidimensional measure, then it is all the more possible that these two mistakes occur at different rates, depending on characteristics of the sample that make one mistake or the other more likely. Indeed, previous studies have provided evidence that SRH is not a biological continuum, but rather a social discontinuity (Smith, Shelley, & Dennerstein, 1994). Patterns of association of excellent health and ill health with various sociodemographic and risk factors were similar (but mirrored) in a Dutch sample, yet the authors conclude that "our understanding of the determinants of ill-health is better than that of the determinants of excellent health" (Mackenbach, Van Den Bos, Joung, Van de Mheen, & Stronks, 1994, p. 1273). The factors underlying elderly people's SRH differ between those with poor or fair SRH on the one hand, and those with good, very good, or excellent health on the other hand (Benyamini, Leventhal & Leventhal, 2003). Another group of researchers who used the CALAS sample argued that positive and negative SRH are not poles of a subjective health continuum (Prager, Walter-Ginzburg, Blumstein, & Modan, 1999). They found stronger correlations of physical functioning with poor SRH among women, compared with men. They also reported that sociodemographic variables were correlated with excellent or good SRH among women, but not among men. They concluded that "the simple health self-evaluation question is not a unitary construct, but rather a complex attitudinal measure" (p. 22).
Recognizing the implications of the distinction between the two poles of the SRH scale may be important in understanding for whom SRH accurately predicts mortality. Gender differences in the accuracy of SRH can be in the accuracy of judging one's health to be poor, in the accuracy of judging one's health to be excellent or very good, or both. If men's poor SRH is indeed more closely linked to serious diseases, or if it corrects for their under-reporting on other health measures, then it should be more accurate than women's SRH in predicting mortality, as found in the present study. Regarding excellent SRH, the "trajectory" hypothesis suggests that it should be more accurate as a predictor of longevity at ages when previous excellent health is likely to continue. These ages are lower for men, because of their faster decline in health at old age and their lower life expectancy. For the very old ages included in our sample, it is more likely that women who rated their health as excellent would indeed survive: For example, a 75-year-old woman who judges herself to be in excellent health has indeed, on the average, a greater probability of surviving the next 10 years, compared with a man her age with the same SRH.
Thus, we propose that at least some of the gender difference in the SRH–Mortality association is due not to differences in the meaning or accuracy of poor SRH but to differences at the high end of the SRH scale. Many studies show that women and men rate their health similarly, but after various indicators of health status are adjusted for, women seem to be more optimistic in their SRH (e.g., Blaxter, 1985; Ferraro, 1980; Fillenbaum, 1979). Yet the differences in objective health status between women and men in the same level of SRH seem to occur mostly at the higher levels of SRH: Women and men with poor SRH often do have similar levels of objective health (Blaxter, 1985; Fillenbaum, 1979; see also Idler, 1993, who interviewed patients at a rehabilitation clinic, all disabled, and who found no significant gender differences in SRH even after health, pain, and disability were controlled for).
Additional Evidence of Gender Differences in the Accuracy of Poor Versus Excellent SRH
There are a few other studies that provide support for gender differences in the accuracy of poor versus excellent health. For example, Kaplan and Camacho (1983) reported a higher OR for poor SRH among women, although among those who rated their health as poor, a higher percentage of men died over the follow-up period (63%) compared with women (55%). The higher OR for women in their sample seems to be due to the higher accuracy of excellent ratings of health—of those participants, only 19% of the women died, compared with 41% of the men. Similarly, data from a two-year follow-up of the Longitudinal Study of Aging sample (Grant et al., 1995) show that poor SRH is linked with more deaths among men (33% compared with 22% among the women), but overall the OR for poor versus excellent or very good SRH as a predictor of mortality is higher for women, probably because only 6% of the women who rated their health at the higher level died, compared with 13% of the men. The authors noted that "the decline in relative risk among women of the LSOA was due to changing mortality not among women with poor and fair SRH, but among women reporting excellent/very good health" (Grant et al., 1995, p. 385).
Idler and Angel (1990) suggested that "one might also observe that the strength of the association in this age group [men 45–64] underscores that fact that, in our society, excellent health is the norm for the non-elderly, and any deviation from the norm indicates a potentially serious risk" (p. 451). For most middle-aged people, excellent health is indeed the norm, and in most cases it is accurate as a predictor of health, functioning, and survival within the next few years. In our youth-driven culture, older people may also aspire for excellent health, and they may strive to maintain and report a sense of good health for as long as possible (at least in comparison with their peers, which seems to be the judgment that people make, whether asked so directly or not). However, at older age, judgments of excellent health are more likely to be inaccurate in the course of a few years. Idler and Angel (1990) also state that older people reported better SRH, and thus controlling for age increased the effect of SRH on mortality; it may be because it diminished the effect of ratings of excellent health that were inaccurate.
In addition to differences between women and men in the prevalence of the two types of mistakes in SRH at different ages, there are other factors that could interact with ratings of poor and/or of excellent SRH and influence the association between SRH and health outcomes. One study reported analyses separately by housing stratum, which revealed an excess of deaths for women living in public housing and reporting excellent health (Idler, Kasl, & Lemke, 1990). Another study repeated the analyses for urban and rural elderly samples. The mortality hazard was significantly higher among urban men who rated their health as poor, compared with all other men, whereas for women, splitting by residence did not change the effect. Among women, survival curves were similar for both urban and rural women at all levels of SRH. The differences that did exist in the SRH–Mortality association between urban and rural women (OR for poor vs. excellent health was 1.63 among urban and 0.77 among rural women) may be due to the fact that the lowest survival rates were in the excellent SRH rural women group (Hays, Schoenfeld, Blazer, & Gold, 1996).
Inconsistencies Regarding Gender Differences in the SRH–Mortality Relation: Conclusions
The inconsistencies in the findings regarding gender differences in the association between SRH and future health and survival may be due to the complexity of this association. If the accuracy of the prediction of mortality from SRH is dependent on the prevalence of the two types of mistakes, those in rating poor health and those in rating excellent health, and these mistakes differ not only between women and men but also by age, and possibly by other factors (such as residence), then we should indeed expect different findings across studies with different demographic distributions. Gender and age distributions vary from study to study: Some studies included a sample of all adult ages; to prevent undersampling of the elderly population, sampling was often stratified by age but ages 65 and over were usually included in the same group. The majority of the studies focused solely on elderly people: When sampling is representative of the population in that age range, only small numbers of the old-old are included in the sample. Our sample included only older adults, aged 75–94, and it was stratified in 5-year groups within this age range. Although the groups used for the current analyses were somewhat smaller as the age increased because of the higher prevalence of proxy responses in the older ages, they were still much larger than those in nonstratified sampling. This strength of the study provided a unique opportunity to study gender and age differences in the SRH–Mortality association, and it suggests a possible explanation for the inconsistencies in the findings from different samples.
Research and Practice Implications
Our findings are in line with two of the explanations suggested for the way SRH is linked to mortality. There are important practical implications to these findings: An older person's SRH summarizes a wealth of information about his or her health that could otherwise be expressed by a long list of measures of the person's physical, psychological, and social status. It is important to ask elderly people for their evaluation of their own health status. Poor SRH typically indicates poor health, and if that is not apparent, further investigation is in place in order to understand why the person rated his or her health as poor. Similarly, a person with very good or excellent SRH is likely to be quite healthy for his or her age. However, at very old age these assessments of better health could become increasingly incorrect, especially for men. If very old people are basing their health assessments on past trajectories and therefore arriving at overly optimistic assessments, we would not necessarily want to contradict those assessments. Psychologically, they may serve a protective purpose. However, the peace of mind acquired by concluding that one is relatively healthy could lead to less motivation for protective health behaviors. We should be aware of this possibility when providing services to very old adults.
It is also important to remember that even though the SRH–Mortality association has been documented in many nations, there may be cultural influences on SRH. In our sample there was a relatively large number of people who rated their health as poor (approximately one third) compared with reports from U.S. studies. Because the sample did not include people of younger ages, it is difficult to tell whether this is due to the very old age of the participants or to a trend that is characteristic of this cultural group. Providers and researchers working within a certain culture should strive to understand the meaning of the two ends of the SRH scale within that culture.
If we learn to interpret and better assess the accuracy of SRH under different conditions, we may be able to use it in certain populations as more than just a crude indicator. Additional research in diverse populations is needed in order to know how to utilize the information captured in SRH and when to rely on it. Our findings emphasize the need to understand the meaning and accuracy of both poor and excellent SRH and to study these in the context of different ages, genders, and life circumstances.
|
Footnotes |
|---|
1 Bob Shapell School of Social Work, Tel Aviv University, Tel Aviv 69978, Israel.
2 Gertner Institute for Epidemiology and Health Policy Research, Chaim Sheba Medical Center, Tel HaShomer, Israel.
Decision Editor: Laurence G. Branch, PhD
Received for publication November 20, 2001. Accepted for publication March 29, 2002.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  T. Blumstein, D. Shmotkin, N. Eyal, A. Shorek, and L. Lerner-Geva A Longitudinal Evaluation of Medication Use Among the Old-Old Population in Israel Research on Aging, January 1, 2008; 30(1): 55 - 73. [Abstract] [PDF]
|
|
|
|
 |
|
 M. Huisman, F. van Lenthe, and J. Mackenbach The predictive ability of self-assessed health for mortality in different educational groups Int. J. Epidemiol., December 1, 2007; 36(6): 1207 - 1213. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. B. Dowd and A. Zajacova Does the predictive power of self-rated health for subsequent mortality risk vary by socioeconomic status in the US? Int. J. Epidemiol., December 1, 2007; 36(6): 1214 - 1221. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Singh-Manoux, A. Dugravot, M. J Shipley, J. E Ferrie, P. Martikainen, M. Goldberg, and M. Zins The association between self-rated health and mortality in different socioeconomic groups in the GAZEL cohort study Int. J. Epidemiol., December 1, 2007; 36(6): 1222 - 1228. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Singh-Manoux, A. Gueguen, P. Martikainen, J. Ferrie, M. Marmot, and M. Shipley Self-Rated Health and Mortality: Short- and Long-Term Associations in the Whitehall II Study Psychosom Med, February 1, 2007; 69(2): 138 - 143. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Singh-Manoux, P. Martikainen, J. Ferrie, M. Zins, M. Marmot, and M. Goldberg What does self rated health measure? Results from the British Whitehall II and French Gazel cohort studies. J Epidemiol Community Health, April 1, 2006; 60(4): 364 - 372. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Han, C. Phillips, L. Ferrucci, K. Bandeen-Roche, M. Jylha, J. Kasper, and J. M. Guralnik Change in Self-Rated Health and Mortality Among Community-Dwelling Disabled Older Women Gerontologist, April 1, 2005; 45(2): 216 - 221. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Meinow, I. Kareholt, M. G. Parker, and M. Thorslund The Effect of the Duration of Follow-Up in Mortality Analysis: The Temporal Pattern of Different Predictors J. Gerontol. B. Psychol. Sci. Soc. Sci., May 1, 2004; 59(3): S181 - S189. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. J. H. Deeg and P. A. Bath Self-Rated Health, Gender, and Mortality in Older Persons: Introduction to a Special Section Gerontologist, June 1, 2003; 43(3): 369 - 371. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
