Received 4 February 2016; accepted 6 March 2016; published 9 March 2016
Civil servants do not plan their retirement while they are actively employed. When time comes for retirement, most civil servants do not welcome it. The period preceding retirement is made of ambivalent feelings; it is the time of divestments and anxiety longing expectations  . The transition to retirement is a life event perceived as a major crisis  which deserves to be apprehended from a bio-psychosocial perspective. For 20% of the population under study, this transition is experienced as a growth crisis against the original anxiety of castration and death. “Workers’ health condition at the end of active life and beyond depends on work conditions and more globally it depends on the onerous nature of their previous work”  . Grant Schellenberg et al.  by demonstrating in 2005 that 22% of pensioners grab hold of their employment and 14% believed that they were still needed, aroused the interest in this period of the civil servants life.
2. Population and Methods
It was a cross-sectional, descriptive and analytical study with prospective data collection from June 1 to August 30, 2013 conducted by a pre-trained team of interviewers. 233 pensioners were included in the study through a comprehensive census and recruited consecutively at the Public Treasury’s Revenue Office in Parakou (Republic of Benin) where they usually receive their pension. A self-administered questionnaire on vulnerability criteria, an interview and Hamilton Depression Rating Scale for assessing pensioners’ level of anxiety and depression were used in data collection process in accordance with the ethical principles contained in the World Medical Association’s Declaration of Helsinki  .
The dependent variable was civil servants’ onerous retirement. In legal terms, “onerousness would be characterized by a kind of the employee’s attrition due to the specific work conditions”  . The independent variables were represented by: socio-demographic characteristics, the psycho-emotional aspects of retirement, physical disorders and family burdens.
Data processing and analysis were conducted using Epi info software version 7. The qualitative variables were expressed in frequency with 95% confidence gap. The comparison of frequencies was made using Chi-square test of Karl Pearson or that of Fischer depending on the case with a statistical significance threshold for p less than 5%.
The prevalence of male pensioners was 85.4%. Sex ratio was 5.85 and the average age was 61.1 ± 6.7 years.
3.1. Anxiety and Somatic Disorders of Pensioners
Table 1 below presents the distributions of pensioners per: the level of anxiety and somatic disorders.
Table 1. Distribution of pensioners per level of anxiety and somatic disorders.
3.2. Factors Associated with the Prevalence of Anxiety among Retired Civil Servants in Parakou
3.3. Factors Associated with the Prevalence of Depressed Mood of Retired Civil Servants in Parakou
The prevalence of pensioners with depressed mood was 87/233 (37.33%).
Table 5 below presents the factors associated with the prevalence of depressed mood of retired civil servants in Parakou
3.4. Psychological Gradient of Health
Figure 1 below shows graphs in connection with depressed mood, anxiety and presence of chronic disease per pensioners’ different occupational categories.
The average age in the study was 61.1 ± 6.7 years. The most represented age group was 55 and 60 years. This average age is different from the one reported by Byers et al.  in 2010 in the United States, which was 68 ± 9.2 years. This difference could be explained by the socio-economic context which is completely different in the two countries. In Benin, retirement age is 55 years for middle Managers and lower, whilst it is 60 years for Executives. The little representation of elderly people over 70 years in the study was be due to the fact that, senior citizens are mostly affected by age-related diseases, therefore they are represented by a third party in view of receiving their pensions. This may also be due to the fact that the global population is in the turmoil of a major phenomenon; aging  and which does not spare Benin where life span is around 59.2 years. Another reason is the fact that many of these pensioners do not live beyond 60 years.
Male prevalence was 85.4%. Sex-ratio was 5.85. This strong representation of men in the environment where the study was conducted could be justified by the delay in girls schooling. By way of contrast, the prevalence of anxiety is more important among females (88.2%) than males (73.0%) with no significantly statistical difference. Women are generally more emotional than men. Once confronted with the same stress factor such as retirement, women will have more difficulties in managing associated psychological disorders thereto. Moreover, retirement
Figure 1. In connection with depressed mood, anxiety and presence of chronic disease per pensioners’ different occupational categories.
period often matching with menopause exposes women to a double stress factor. The prevalence of married pensioners involved in the study was 90.1%. Claim of anxiety symptoms was registered with 75.2% of these married pensioners (see Table 2). This seems paradoxical since marriage is a social protection factor against loneliness. According to Harris  , the problems seem to be due to lack of communication in the marriage.
The respondent pensioners were represented in the occupational categories as follow: A 18.9%; B 26.6%; C 27.9%; 23.6%; E 3% (see Table 2, Table 3 and Table 4). The low representation of category E is probably due to the fact that it is a category comprising workers of all kinds, manual workers and cargo handlers that the public administration employs without really formalizing their recruitment. Regarding the low proportion of category A, it is important to consider the fact that during colonial era, working in the public administration was not conditioned by a very high intellectual level. At that time and during the post-colonial era higher education was perceived as luxury. The administration opted for training middle and junior Managers instead of senior Executives.
The anxiety state of respondent pensioners was significantly influenced by the number of year spent in re-
Table 2. Distribution of anxiety prevalence per gender, marital status and the socio-occupational category.
Table 3. Distribution of anxiety prevalence per the feelings upon the announcement of retirement and the prevalence of feelings upon the announcement of retirement based on the occupational category.
Table 4. Distribution of the prevalence of depressed mood per age and occupational category.
tirement (p = 0.00002). Anxiety prevalence increased from 72.7% for 0 to 3 years in retirement to 81.9% for 4 years to 10 years after retirement and then changed to 71.6% for 10 years and above in retirement (see Table 5). The duration of 0 to 3 years of retirement is a period of adjustment when the pensioner is subject to significant disruption of his daily life especially in Benin context, where receiving the first pension may take several months or even several years, plunging the pensioner in a real precarious life. The increase in anxiety prevalence for 4 to 10 years in retirement could be linked to the worsening of defects of age-related disorders. The decrease with stabilization of anxiety prevalence after 10 years in retirement is the result of a discreet adjustment to retirement.
4.1. Onerous Psycho-Emotional Nature of Retirement
The feeling upon the announcement of retirement significantly influenced the psychological condition of respondent pensioners (88.9%; p = 0.03). Upon the announcement of retirement, 81.3% were joyful (see Table 3). This joy would seem out of denial, because at that moment, pensioners were 2.9 times more likely to be anxious, apprehending health risks (stretching age-related disorders), isolation and loneliness from retirement. 28.8% of respondent pensioners claimed insomnia (see Table 1). This result is higher than that of Vahtera et al.  in 2009 who reported sleep disorders prevalence of 18.8% to 20.6% from the first to the seventh year after retirement. These sleep disorders among pensioners could also be related to the sudden change of lifestyle induced by retirement. A retired employee has more time to rest and sleep during most of the day, thus reducing the duration of his night sleep.
Pensioners state of anxiety was significantly associated with the presence of somatic disorders (p = 0.002); 86.2% of anxious people were suffering from a chronic disease (see Table 5). According to Rice et al.  in 2004, chronic diseases reduce self-dependence and the functioning of senior citizens and also interfere with their daily activities. Moser et al.  in 2010 reported higher levels of anxiety among pensioners with organic decompensation of age-related disorders. Westerlund et al.  in 2010 showed that the prevalence, when cumulated with respiratory and cardio-vascular disease as well as diabetes increases with age.
4.2. Factors Associated with Mood Disorders among Pensioners
Depressed Mood was significantly associated with age and 48.6% of pensioners aged 50 to 55 years had depressed mood (see Table 4). Factors associated with the depressed mood of respondent pensioners could be apprehended through factors associated with the prevalence of anxiety (see Table 2, Table 3 and Table 5). The majority (66.2%) of pensioners in the study presented a slight anxiety. It is the most recurrent psychiatric disorder among elderly pensioners. The prevalence of anxiety among pensioners who planned their retirement was
Table 5. Distribution of anxiety prevalence per: retirement planning, presence of chronic disease, health insurance plan, number of dependents, feeling of abandon and seniority in retirement.
84.2%. Those who were suffering from chronic disease were 86.2%. Pensioners who did not have health insurance were 24.6%. Those who had more than 5 dependants were anxious (77.6%). Pensioners with a depressed mood were 37.33 %. This prevalence is greater than that of Byers et al.  , in 2010 who showed that mood disorders prevalence decreases with age and varies depending on gender. These same authors point out that the prevalence of anxiety disorders is as high as mood disorders and that, race or ethnicity did not influence this anxiety prevalence. Bekhet et al.  reported that among people who had negative emotions within a population of pensioners, 23.9% were anxious.
In descending order, the prevalence of anxiety was higher among civil servants in categories D and E than those in category A which itself is higher than those in categories B and C (see Table 2). This leads Harris  to say that retirement means different challenges for the people. This difference between the categories although not statistically significant (p = 0.27), could be due to a loss of category A Executives’ social status and related benefits thereto. Regarding the other lower categories, the deterioration of their economic situation that was already not very rosy could be an explanatory hypothesis.
Anxiety prevalence is higher among those who planned their retirement (84.2%) than among those who did not plan (28%) (see Table 5). Preparing retirement would mean greater awareness of the retirement-event, and this will likely be reason of more anxiety.
On the eve of his retirement, one of the worst worries of the civil servant is the guarantee of a secure employment for his children. In a dampened economic environment where the rate of unemployment of young people is on the rise, the already financially weak and aging pensioner, without health insurance is obliged to ensure the support of his unemployed graduate. The state of pensioners anxiety was significantly associated with the feeling of abandon (p = 0.02) regarding 87.8% of the studied population who had a feeling of being abandoned by their relatives. One of the determinants of pensioners’ anxiety was the abandon by their family, their former friends from their socio-professional environment. This feeling of loneliness is also observed by Bekhet et al.  in 2012, who showed that it significantly influences the state of anxiety. The more an individual feels lonely, the more he is anxious. They revealed that 28.7% of pensioners who had negative emotions were reported to be lonely and had a feeling of abandon, and that the prevalence of loneliness among elderly pensioners was estimated at 40%. This figure has been relatively constant over the past 25 years according to Cohen- Mansfield et al.  . So, committing suicidal is a serious threat.
4.3. Psycho-Social Gradient of the Pensioners’ Health
The occupational category did not significantly influence the presence or absence of chronic disease. Employees of category A (27.3%) had at least one chronic disease against 51.6% for category B, 35.4% for category C, 43.6% for category D and 42.9% for category E. These results reveal that Executives from category A have chronic diseases prevalence lower than that of all other categories. Similarly, it was noted that occupational category did not significantly influence the presence or absence of depressed mood within the studied population. According to employees categories, the proportions of pensioners with depressed mood were respectively 36.4% for category A, 43.5% for category B, 29.2% for category C; 38.2% for category D and 57.1% for category E. Furthermore, pensioners in category E presented higher prevalence of depressed mood (57.1%). These results can be explained by the fact that senior Executives were better off, and therefore could care for their health needs and by so doing enjoy better quality of life. According to Lagacé et al.  , “Health-related social inequalities are strongly dependent on work, morbidity and mortality, both decreasing from bottom to top of the social ladder (health gradient)”. Thus, we observe a gradient according to socio-occupational categories  . This gradient is recorded for all aspects of health studied  . Health condition declarations are marked by a North-South gradient in northern Europe  . Moreover, there are large inequalities of chances to live in good health during retirement  . This major fact is barely considered in retirement discussions.
In Parakou, retired female pensioners live this period of their life as an existential crisis more than men do. Anxiety is more common among pensioners in categories A and E. The feeling of fear upon the announcement of retirement, its planning or not, the apprehension of abandon by relatives, seniority in retirement and age-related disorders decompensating are significantly associated with anxiety among pensioners. The absence of health insurance in case of illness reinforces depressed mood disorders whose prevalence is higher among pensioners in category E, underpinning an ascending health-related social gradient when progressing to category A pensioners. It is important that policy makers and labor unions collaborate in terms of passing better laws on the prevention of onerous retirement.
 World Medical Association (2012) WMA Declaration of Helsinki: Ethical Principles for Medical Research Involving Human Subjects. 2013. Dostupno. http://www.wma.net
 Byers, A.L., Yaffe, K., Covinsky, K.E., Friedman, M.B. and Bruce, M.L. (2010) High Occurrence of Mood and Anxiety Disorders among Older Adults. Archives of General Psychiatry, 67, 489-496.
 Rice, D.P. and Fineman, N. (2004) Economic Implications of Increased Longevity in the United States. Annual Review of Public Health, 25, 1-17.
 Moser, D.K., Dracup, K., Evangelista, L.S., et al. (2010) Comparison of Prevalence of Symptoms of Depression, Anxiety, and Hostility in Elderly Patients with Heart Failure, Myocardial Infarction, and a Coronary Artery Bypass Graft. Heart & Lung: The Journal of Acute and Critical Care, 39, 378-385. http://dx.doi.org/10.1016/j.hrtlng.2009.10.017
 Westerlund, H., Vahtera, J., Ferrie, J.E., Singh-Manoux, A., Pentti, J., Melchior, M., et al. (2010) Effect of Retirement on Major Chronic Conditions and Fatigue: French GAZEL Occupational Cohort Study. BMJ, 341, 1136-1149. http://dx.doi.org/10.1136/bmj.c6149
 Bekhet, A.K. and Zauszniewski, J.A. (2012) Mental Health of Elders in Retirement Communities: Is Loneliness a Key Factor? Archives of Psychiatric Nursing, 26, 214-224.
 Cohen-Mansfield, J. and Parpura-Gill, A. (2007) Loneliness in Older Persons: A Theoretical Model and Empirical Findings. International Psychogeriatrics, 19, 279-294.
 Debrand, T. and Lengagne, P. (2007) Work Onerousness and Senior Citizens Health Conditions in Europe. Economics & Statistics, 403, 19-38. http://dx.doi.org/10.3406/estat.2007.7086