Secondary Logo

Journal Logo

Mental Health: Original Article

Is Retirement Beneficial for Mental Health?

Antidepressant Use Before and After Retirement

Oksanen, Tuulaa,b; Vahtera, Jussia,c; Westerlund, Hugod; Pentti, Jaanaa; Sjösten, Nooraa; Virtanen, Mariannaa; Kawachi, Ichirob; Kivimäki, Mikaa,e,f

Author Information
doi: 10.1097/EDE.0b013e31821c41bd


Due to population aging, the proportion of retired people is increasing rapidly throughout the industrialized world. It has been projected that if labor force participation remains unchanged, the number of retirees per worker will double by 2050 in most developed countries.1 Many societies have been grappling with the issue of social security solvency, and health in retirement has been widely discussed.2,3

Previous research on the health effects of transition to retirement has provided inconsistent evidence. Some studies suggest that mental health is better in retirement,4–12 whereas some report that it is worse,13–16 and still others found no change in relation to retirement.17–20 Some studies,5,12 but not all,4 suggest that retirement is more beneficial for employees in high-SES compared with low-SES groups. The reasons for these inconsistencies may be due to differences in study design. Indeed, the few prospective studies with information about each worker's health trajectory before the retirement transition, as well as following retirement, consistently suggest that retirement has a beneficial impact on perceived mental well-being and health.4,11,12,21 It is unclear, however, whether this finding is robust with respect to health indicators that do not rely only on self-reports.

To examine potential influences of retirement on objectively assessed mental health, we studied whether transition to retirement is associated with changes in antidepressant medication use as indicated by annual prescription records during the 4 years before and 4 years after the retirement year. We hypothesized that retirement would be associated with a decline in the levels of antidepressant use after taking the secular trend in rising antidepressant prescriptions into account (eFigure 1, As a validity check for our hypothesis, we also examined pre- and postretirement trajectories in prescription use for diabetes medications. Our a priori hypothesis was that there would be no relation between retirement and diabetes medication use. Although diabetes prevalence increases with age, retirement is unlikely to have immediate effects on diabetes risk.4


Context of the Study

The National Health Insurance (NHI) scheme is part of the Finnish social security system covering all citizens regardless of age, wealth, or employment status (ie, employed, unemployed, or retired). Retirement per se has no impact on prescription reimbursements. After a fixed deductible per purchase (10 Euros in 2005), NHI reimburses 50% of the costs of filled prescriptions. This basic reimbursement applies to antidepressants and drugs to treat diabetes. However, diabetic patients are eligible for special reimbursement to get a 100% refund of medicine costs exceeding 5 Euros (in 2005). If the annual medicine costs that are not covered by NHI exceed a certain limit (607 Euros in 2005), the scheme member gets all drugs beyond that limit free during that year. Also, in case of financial difficulties or great burden of diseases, it is possible to apply for extra financial support for medicine costs from local social security authorities.23

The Finnish pension provision consists of an employment-based, earnings-related pension plus a residence-based, national pension that guarantees a minimum income. As an example, the earnings-related pension is about 57% for a person with average wage and a work history of 40 years.24 According to the public sector employees' pension act, the statutory retirement age is generally from 63 to 65 years (63 to 68 years from 2005 onwards), although the age is lower for specific groups (eg, 60 years for primary school teachers, 58 for practical nurses). A disability pension may be granted if, due to an illness, the employee cannot continue working even after attempts at rehabilitation, reeducation, or assistance. Employees may apply for disability pension in the event of more than 300 reimbursed sickness absence days (Sundays excluded) during 2 consecutive years on the basis of the illness causing work disability. Approximately 80% of all disability pension applications are accepted.25

Study Design and Participants

We studied the changes in mental health by assessing trajectories of filled antidepressant prescriptions during a 9-year period around retirement, and comparing the results with trajectories of filled diabetes medication prescriptions in the same cohort. The participants are from the Finnish Public Sector Study cohort: a total of 151,618 employees who have been employed in 10 municipalities or 6 hospital referral districts for at least 6 months in any year from 1991 to 2005.26 Data for cohort members were successfully linked to employers' records and comprehensive national health registers through unique personal identification codes, which are assigned to all citizens in Finland.

Of the cohort members, we identified full-time retirees who retired at statutory retirement age, and those who retired early on health grounds between 1995 and 2004, and who were alive at least 1 year after the retirement—a total of 13,559 persons. These initial inclusion criteria were selected to allow for at least a 3-year follow-up around retirement (ie, from 1 year prior to 1 year after retirement). In case of entitlement to several pensions from different pension schemes at different times, the first one awarded was selected. We excluded the employees of one town, where the medication costs were paid by the employer during employment (n = 2540). Thus, the analytic sample comprised of 11,019 retirees (75% women). The study was approved by the Ethics Committee of the Finnish Institute of Occupational Health.

Assessment of Retirement

Data on retirement were obtained from the Finnish Centre for Pensions, which coordinates all earnings-related pensions for permanent residents in Finland.27 All gainful employment is insured in a pension plan and accrues a pension; thus the pension data with successful linkage were available for all participants. The dates for starting to receive old-age or disability pension were obtained for all participants from 1994 through 2005, irrespective of the participants' employment status or workplace at follow-up. We analyzed those participants whose main medical cause for retirement assigned by the treating physician was mental illness (WHO International Statistical Classification of Diseases and Related Health Problems 10th Revision, ICD-10 codes F00–F99) separately from employees retiring due to physical causes (all other ICD-codes).

Assessment of Antidepressant Use and Use of Diabetes Medication

We determined antidepressant use for each year of the 9-year observation period using the nationwide Drug Prescription Register maintained by the Social Insurance Institution of Finland. The data include the dates of purchase of all drugs in the World Health Organization's Anatomic Therapeutic Chemical (ATC) classification and the corresponding defined daily dosages (DDDs). For each year of observation, we defined antidepressant use as the purchase of antidepressants (ATC code N06A) of at least 30 daily doses.

In a similar way, the annual purchases of diabetes medication (ATC code A10) of at least 30 defined daily doses were retrieved from the Drug Prescription Register. Persons with diabetes can be entitled to special reimbursement for treatment of diabetes, after which the drugs for diabetes are free of charge after a fixed deductible per purchase. Patients who apply for special reimbursement must submit a detailed medical certificate prepared by the treating physician. In all, 74% of employees purchasing diabetes medication during the year of retirement transition were entitled to special reimbursement for the cost of diabetes medication.

Assessment of Preretirement Covariates

From the employers' registers we obtained information on sex, socioeconomic status (SES), geographic area (Southern, Middle, Northern Finland, based on the location of the workplace), and type of employer (town or hospital). SES was categorized according to the occupational-title-based classification of the Statistics Finland to upper-grade nonmanual workers (eg, teachers, physicians), lower-grade nonmanual workers (eg, registered nurses, technicians), and manual workers (eg, cleaners, maintenance workers),28 as determined by the last occupation during the 3 years preceding retirement. Absence because of long-term sickness (yes or no) was assessed by combining records of temporary disability pension with records of sickness absence (lasting at least 10 days) during the year preceding retirement, obtained from the Social Insurance Institution. As in our earlier studies,26 the presence of chronic diseases (yes or no) in retirement was defined by entitlement to special reimbursement for the costs of medication needed to treat diabetes, rheumatoid arthritis, asthma and chronic obstructive pulmonary disease, hypertension, coronary heart disease, or heart failure in retirement, obtained from records at the Social Insurance Institution.

Statistical Analysis

The analyses were based on a 9-year observation window, with the year of retirement as year 0 and the 4 years of observation both before and after retirement as years −4 to −1 and years +1 to +4. A total of 0.9% of the statutory retirees, 2.2% of those who retired due to mental health issues, and 4.4% of those who retired due to physical issues died during years +2 to +4. Mortality was not differential in terms of baseline antidepressant use.

To examine changes in the likelihood of antidepressant treatment, we applied a repeated-measures logistic regression analysis using the generalized estimating equations (GEE) method with autoregressive correlation structure.29 GEE takes into account the intraindividual correlation between measurements and is not sensitive to missing measurements. We chose the autoregressive correlation structure because it assumes correlations between time points to be greater the nearer the measurements are to each other. The autoregressive model also yielded a slightly better fit than the exchangeable working correlation model. We calculated the annual prevalence estimates to plot the trajectory of antidepressant treatment in relation to retirement, adjusting the models for age at retirement and calendar year.22 Similarly, we calculated the annual prevalence estimates of use of diabetes medication, and used the trajectory as a reference.

The life-course approach proposes that the impact of retirement and other life transitions may depend on the proximity to the transition.30,31 Thus, we distinguished the 3 periods of the retirement process as follows: the preretirement period (years −4 to −2), the transition period (years −1 to +1), and the postretirement period (years +2 to +4). This categorization also provided us with sufficient power to test whether there was a change in trend of antidepressant use during each period. We used time as a continuous variable in the analyses, and entered the interaction term (time*period) into the binomial or (where appropriate) Poisson regression model to evaluate whether the trend in antidepressant use was different among the 3 periods.32 We calculated prevalence differences and prevalence ratios and their 95% confidence intervals (CIs) to express change in antidepressant use between the first and last year within each of the 3 periods, adjusting the models for age at retirement and calendar year.

We tested the robustness of the findings first by investigating whether the preretirement covariates modified the shape of the trajectory by entering the second-level interaction term “effect modifier*time*period” into the regression model and calculated the prevalence ratios of the trend in antidepressant use within the preretirement, retirement transition, and postretirement periods. We additionally adjusted the models for preretirement covariates, although stable factors are not plausible confounders for time-dependent trajectories.

Finally, we ran sensitivity analyses to test the following: (1) whether there was a change in trend between the preretirement period and the retirement transition in a subgroup of statutory retirees who had complete data for all years of the preretirement and transition periods (n = 4306), (2) whether the findings were sensitive to the specific dose selected to define antidepressant use (ie, 30 vs. any, at least 90 or 180 defined daily doses per year) or the number of purchases made, and (3) whether there were differences in the number of filled prescriptions by socioeconomic status.

All statistical analyses were conducted with SAS 9.1.3 statistical software (SAS Institute, Inc, Cary, NC).


Between 1995 and 2004, 7138 participants retired at statutory retirement age (mean = 61.2; range = 55 to 67 years) and 3881 retired early on health grounds, 1238 of whom were awarded a disability pension due to mental health issues (mean age = 52.0; range = 24 to 63 years) and 2643 due to physical health issues (mean age = 55.4; range = 23 to 64 years). Women, participants who retired before 60 years of age, and participants who had chronic or disabling diseases were more likely to receive antidepressant treatment before retirement than their male and older counterparts with no chronic or disabling diseases (Table 1). During the follow-up, 1% to 3% of the retirees made only a single purchase of antidepressants, whereas the majority of participants who were on antidepressants filled at least 5 prescriptions (eTable 1, The number of filled prescriptions of antidepressants did not vary between socioeconomic groups. Altogether 61% of those retiring early due to mental disorders were on antidepressant treatment 1 year prior to retirement. The corresponding prevalence was 4% in old age retirees and 14% among those retiring early due to physical causes. The retirement cohorts differed in terms of background characteristics, supporting the rationale for analyzing the 3 cohorts separately (Table 2).

Antidepressant Use by Characteristics of the Study Sample One Year Before Retirement
Type of Retirement and Main Medical Cause of Early Retirement by Background Characteristics 1 Year Before Retirement

Retirement at Statutory Age

Figure 1 displays the estimated annual prevalence of antidepressant and diabetes medication use in relation to statutory retirement, adjusted for retirement age and calendar year. As demonstrated by unadjusted prevalences (eFigures 1 and 2,, the patterns were not driven by the effect of calendar time. Among old-age retirees, there was a 23% decline in the level of antidepressant use after the transition to retirement. The corresponding prevalence ratio for antidepressant use 1 year after versus 1 year before retirement, adjusted for retirement age and calendar year, was 0.77 (95% CI = 0.68 to 0.88). In contrast, there was no substantive change in antidepressant use during the preretirement period (0.93 [0.81 to 1.06]) or during the postretirement period (1.00 [0.88 to 1.14]). We found little evidence to suggest that preretirement factors confounded the observed trajectories, as the results from multivariate adjusted models largely replicated the main findings.

Prevalence of antidepressant and diabetes medication use adjusted for calendar year and retirement age, in relation to year of retirement at statutory age (error bars indicate 95% confidence intervals). Note that the figure is corrected for the increasing secular trend in prescriptions during the study period.

Antidepressant use decreased during the transition period in both sexes. During the postretirement period, the use increased in men but not women (eTable2, Socioeconomic status and the presence of previous sickness absence periods also modified the antidepressant trajectory. After adjusting for retirement age and calendar year, antidepressant use decreased by 32% during the retirement transition period among employees in the highest socioeconomic status group, and less for other socioeconomic groups (15%–16%). The decrease in antidepressant use during the transition to retirement was stronger among statutory retirees who had long-term sickness absence periods during the transition period (adjusted prevalence ratio = 0.65 [CI = 0.55 to 0.77]) than among those with no history of disabling diseases (0.94 [0.80 to 1.11]). Thus, statutory retirement appeared to be particularly beneficial for those with preexisting health problems at work.

Early Retirement on Health Grounds

As shown in Figure 2, a different pattern of antidepressant use trajectories was observed for retirement due to poor health. The shape of the antidepressant trajectory did not depend on calendar time (eFigures 3 and 4, but varied by the medical cause leading to early retirement. Among persons who retired due to mental disorders, antidepressant use increased during the preretirement period (adjusted risk ratio = 1.42 [95% CI = 1.28 to 1.57]), peaking in the year of retirement, and decreasing after retirement (Table 3). Again, the association was dependent on socioeconomic status. Persons in the highest socioeconomic group decreased their antidepressant use after retirement more than other socioeconomic groups (eTable 3,

Prevalence of antidepressant use in relation to year of early retirement due to mental causes and physical causes separately and prevalence of use of drugs for diabetes in both these cohorts combined, adjusted for retirement age and calendar year. Error bars indicate 95% confidence intervals. Note that the figure is corrected for the increasing secular trend in prescriptions during the study period. (Note that the scale for y-axis is different than that in Fig. 1.)
Trends in Antidepressant Use Within the Transition Period and the Pre- and Postretirement Periods by Type and Main Cause of Retirement

We found a simpler antidepressant trajectory among those retiring as a result of physical health issues (Fig. 2). As shown in Table 3, there was an increasing trend in antidepressant use during the preretirement period (prevalence ratio = 1.45 [95% CI = 1.22 to 1.71]), but little change in trend in antidepressant use within the retirement transition period or the postretirement period. Apart from the area of residence, there were no differences in the change in trend between the periods by any background variable (eTable 4,

We found no retirement-related changes in the trajectories of use of drugs for diabetes in any of the 3 groups (Figs. 1 and 2). Instead, the estimated prevalence of annual diabetes medication use increased steadily with increasing age, (eFigure 2, as would be expected for this age-related progressive disorder. Results from sensitivity analyses were consistent with the main analyses (eAppendix,


In this large-scale study, objective repeated data on purchases of prescription medication were used as indicators of the potential effects of retirement on mental health. Data from serial measurements of antidepressant use during the retirement transition were consistent with the hypothesis that retirement may improve mental health among old-age retirees, as well as possibly also those retiring early on mental health grounds. After controlling for the rising secular trend in the prescription of antidepressants,22 the prevalence of antidepressant use decreased by one-fourth from before retirement to afterward. The decrease in antidepressant use was greater among those who stayed employed until they were entitled to old-age pension despite health problems at work.

Retirement at Statutory Age

Our results are in agreement with those from the French GAZEL-cohort and the British Whitehall II study in which statutory retirement was associated with an improvement in self-rated mental health compared with staying employed.4,11,12,21 Several other studies based on self-report data have also provided evidence consistent with these observations.5–10

There are at least 3 possible explanations for these observations. First, release from the demands of work may be beneficial to mental health. This is supported by the observation that those who were encountering problems with disabling diseases benefited more than those without such conditions. Second, having more time to perform activities at home and elsewhere, could allow more autonomy to pursue own interests7 and more time to invest in health such as exercising.20 Increase in personal control8 could reduce need for antidepressant medication. Third, having worked until the age of receiving old-age benefits may lead to feelings of having fulfilled society's expectations.17

Alternative explanations should also be considered. For example, financial difficulties in retirement could decrease the likelihood of a depressed subject seeking medical advice, or lead to a reduction in the number of prescriptions filled by patients. However, the fact that the largest reduction in antidepressant use after retirement was seen in the highest socioeconomic status group suggests that filling fewer prescriptions after retirement is unlikely to be explained by financial concerns alone. A further explanation might be that the demands of work, including the need of medical certificates for longer sickness absences, could provide a stronger incentive for depressed subjects to see a doctor while still in work than in retirement. If the latter is the case, the decrease in purchases of antidepressants may not reflect a decrease of depression at retirement.

Retirement on Health Grounds

Although we observed a decrease in filled prescriptions, the level of antidepressant use remained high among those retiring due to mental health issues. This could reflect the chronic nature of depression and its high relapse rate33 and suggests that work was not the primary cause of depression. As the prevalence of antidepressant use increased leading up to retirement and decreased after transition to retirement, our findings may also be affected by the effort expended to retain the ability to work. A further potential explanation is that anticipation and decision-making in the years before retirement can be stressful, reflected by an increased antidepressant use before retirement. Without further research using other indicators of mental health, it may not be possible to determine whether the observed decline in antidepressant use after retirement was driven by a true improvement in mental health or by the cessation of unsuccessful treatment efforts to continue work.

Those retiring due to physical health issues did not reduce their antidepressant use, which is consistent with previous studies using other health indicators.14,19 The explanation for the observed consistency in the prevalence of antidepressant use in pre- and postretirement could be that retirement did not alter the course of the physical illness leading to early retirement but, instead, pain and low level of physical activity causing poor sleep and poor mental health continued to prevail at the same level.

Strengths and Limitations

The strengths of the study include the large sample size and serial measurements of antidepressant use across a 9-year period centered around retirement and based on comprehensive prescription records obtained from national registers. However, there was some variation in sample size during the first 3 years and the last 3 years of the 9-year observation window. This was because of the limited follow-up for participants who retired at either end of the fixed observation window and for those who died during the follow-up. Nevertheless, selective sample retention is unlikely to explain our findings because the results from sensitivity analyses restricting the sample to those who had complete data at all years of the preretirement and transition periods were consistent with the main results.

The use of antidepressant data circumvented potential response bias from self-reported questionnaires. The validity of our finding was also strengthened by the specificity of the association between retirement and antidepressant use. The purchases of diabetes medication were not altered by retirement (as hypothesized), but, instead, increased over time with increasing age and calendar year, as expected.34,35 Furthermore, our results of a higher prevalence of antidepressant users among women compared with men correspond to sex differences in the prevalence of depression reported for general populations.36

Some caveats should be noted with respect to the data on antidepressants. Filling a prescription does not equate to the actual use of medication, and prescriptions do not capture undiagnosed disease or conditions treated without medication. Thus, using antidepressant prescriptions as a proxy for mental health may have captured only the “tip of the iceberg” of common mental health problems.8 Furthermore, some misclassification is possible because these medications are also used to treat other mental disorders, such as eating or sleeping disorders, and chronic pain. We consider this error to be small because patients with depression and anxiety represent a vast majority of those taking antidepressants.37 We may also have misclassified depressed employees who were treated with atypical antipsychotics, although they are typically used in combination with antidepressants when treating depression with psychotic features. Because we did not have information on the planned duration or recommended dose of the antidepressant treatment, we could not use more fine-grained measures such as discontinuation of antidepressant use.


The observed trajectories of recorded purchases of antidepressant medication are consistent with the hypothesis that retirement may be beneficial for mental health. Future research should investigate the generalizability of the findings to other countries and settings.


1.OECD. Live Longer, Work Longer. Paris: OECD Publishing; 2006.
2.OECD. Ageing and pension system reform: implications for financial markets and economic policies. Financial Market Trends. 2005,2005(suppl 1):i-117.
3.WHO. Gaining Health. The European Strategy for the Prevention and Control of Noncommunicable Diseases. Report of the 56th Session of the WHO Europe Regional Committee for Europe; Copenhagen, Denmark; September 11–14, 2006.
4.Westerlund H, Vahtera J, Ferrie JE, et al. Effect of retirement on major chronic conditions and fatigue: French GAZEL occupational cohort study. BMJ. 2010;341:c6149.
5.Mein G, Martikainen P, Hemingway H, Stansfeld S, Marmot M. Is retirement good or bad for mental and physical health functioning? Whitehall II longitudinal study of civil servants. J Epidemiol Community Health. 2003;57:46–49.
6.Salokangas R, Joukamaa M. Physical and mental health changes in retirement age. Psychother Psychosom. 1991;55:100–107.
7.Mojon-Azzi S, Sousa-Poza A, Widmer R. The effect of retirement on health: a panel analysis using data from the Swiss Household Panel. Swiss Med Wkly. 2007;137:581–585.
8.Drentea P. Retirement and Mental Health. J Aging Health. 2002;14:167–194.
9.Gall TL, Evans DR, Howard J. The retirement adjustment process: changes in the well-being of male retirees across time. Gerontol B Psychol Sci Soc Sci. 1997;52:110–117.
10.Reitzes D, Mutran E, Fernandez M. Does retirement hurt well-being? Factors influencing self-esteem and depression among retirees and workers. Gerontologist. 1996;36:649–656.
11.Jokela M, Ferrie JE, Gimeno D, et al. From midlife to early old age: health trajectories associated with retirement. Epidemiology. 2010;21:284–290.
12.Vahtera J, Westerlund H, Hall M, et al. Effect of retirement on sleep disturbances: the GAZEL prospective cohort study. Sleep. 2009;32:1459–1466.
13.Bossé R, Aldwin CM, Levenson MR, Ekerdt DJ. Mental health differences among retirees and workers: Findings from the normative aging study. Psychol Aging. 1987;2:383–389.
14.Buxton J, Singleton N, Melzer D. The mental health of early retirees: National interview survey in Britain. Soc Psychiatry Psychiatr Epidemiol. 2005;40:99–105.
15.Alavinia S, Burdorf A. Unemployment and retirement and ill-health: a cross-sectional analysis across European countries. Int Arch Occup Environ Health. 2008;82:39–45.
16.Gill S, Butterworth P, Rodgers B, Anstey K, Villamil E, Melzer D. Mental health and the timing of Men's retirement. Soc Psychiatry Psychiatr Epidemiol. 2006;41:515–522.
17.Villamil E, Huppert F, Melzer D. Low prevalence of depression and anxiety is linked to statutory retirement ages rather than personal work exit: a national survey. Psychol Med. 2006;36:999–1009.
18.Butterworth P, Gill SC, Rodgers B, Anstey KJ, Villamil E, Melzer D. Retirement and mental health: analysis of the Australian national survey of mental health and well-being. Soc Sci Med. 2006;62:1179–1191.
19.Melzer D, Buxton J, Villamil E. Decline in common mental disorder prevalence in men during the sixth decade of life. Soci Psychiatry Psychiatr Epidemiol. 2004;39:33–38.
20.Midanik L, Soghikian K, Ransom L, Tekawa I. The effect of retirement on mental health and health behaviors: the Kaiser Permanente Retirement Study. J Gerontol B Psychol Sci Soc Sci. 1995;50:59–61.
21.Westerlund H, Kivimäki M, Singh-Manoux A, et al. Self-rated health before and after retirement in France (GAZEL): a cohort study. Lancet. 2009;374:1889–1896.
22.Moore M, Yuen HM, Dunn N, Mullee MA, Maskell J, Kendrick T. Explaining the rise in antidepressant prescribing: a descriptive study using the general practice research database. BMJ. 2009;339:b3999.
23.The Social Insurance Institution of Finland. Overview of benefit programmes. Accessed January 3, 2011.
24.Hietaniemi M, Ritola S. The Finnish Pension System. Vol 6. Helsinki, Finland: Gummerus Kirjapaino; 2007.
25.OECD. Sickness, Disability and Work: Breaking the Barriers. A Synthesis of Findings Across OECD Countries. Paris: OECD Publishing; 2010.
26.Sjösten N, Vahtera J, Salo P, et al. Increased risk of lost workdays prior to the diagnosis of sleep apnea. Chest. 2009;136:130–136.
27.Suoyrjö H, Oksanen T, Hinkka K, et al. The effectiveness of vocationally oriented multidisciplinary intervention on sickness absence and early retirement among employees at risk: an observational study. Occup Environ Med. 2009;66:235–242.
28.Statistics Finland. Classification of Occupations. Helsinki: Statistics Finland; 1987.
29.Lipsitz S, Kim K, Zhao L. Analysis of repeated categorical data using generalized estimating equations. Stat Med. 1994;13:1149–1163.
30.Moen P. A life course perspective on retirement, gender, and well-being. J Occup Health Psychol. 1996;1:131–144.
31.Kim J, Moen P. Retirement transitions, gender, and psychological well-being: a life-course, ecological model. J Gerontol B Psychol Sci Soc Sci. 2002;57:212–222.
32.Spiegelman D, Hertzmark E. Easy SAS Calculations for risk or prevalence ratios and differences. Am J Epidemiol. 2005;162:199–200.
33.King M, Nazareth I, Levy G, et al. Prevalence of common mental disorders in general practice attendees across Europe. Br J Psychiatry. 2008;192:362–367.
34.King D, Ellis T, Everett C, Mainous AI. Medication use for diabetes, hypertension, and hypercholesterolemia from 1988–1994 to 2001–2006. South Med J. 2009;102:1127–1132.
35.Tabák AG, Jokela M, Akbaraly TN, Brunner EJ, Kivimäki M, Witte DR. Trajectories of glycaemia, insulin sensitivity, and insulin secretion before diagnosis of type 2 diabetes: an analysis from the Whitehall II study. Lancet. 2009;373:2215–2221.
36.Kessler RC, Berglund P, Demler O, et al. The epidemiology of major depressive disorder. JAMA. 2003;289:3095–3105.
37.Mark TL. For what diagnoses are psychotropic medications being prescribed? A nationally representative survey of physicians. CNS Drugs. 2010;24:319–326.

Supplemental Digital Content

© 2011 Lippincott Williams & Wilkins, Inc.