- Email: [email protected]
Why are suicidal thoughts less prevalent in older age groups? Age differences in the correlates of suicidal thoughts in the English Adult Psychiatric Morbidity Survey 2007
Journal of Affective Disorders 177 (2015) 42–48
Contents lists available at ScienceDirect
Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad
Research report
Why are suicidal thoughts less prevalent in older age groups? Age differences in the correlates of suicidal thoughts in the English Adult Psychiatric Morbidity Survey 2007 Claudia Cooper a,n, Khadija Rantell a, Martin Blanchard a, Sally McManus b, Michael Dennis c, Traolach Brugha d, Rachel Jenkins e, Howard Meltzer d,†, Paul Bebbington a a
Division of Psychiatry, 2nd Floor, Charles Bell House, 67-73 Riding House St., London W1W 7EJ, UK NatCen for Social Research, London EC1V 0AX, UK College of Medicine, Swansea University, Singleton Park, Swansea, Wales, UK d Department of Health Sciences, University of Leicester, 22-28 Princess Road West, Leicester LE1 6TP UK e Kings College London, Institute of Psychiatry, De Crespigny Park, London, UK b c
art ic l e i nf o
a b s t r a c t
Article history: Received 1 November 2014 Received in revised form 9 February 2015 Accepted 11 February 2015 Available online 19 February 2015
Background: Suicidal ideation is more strongly associated with suicidal intent in later life, so risk factors may also differ by age. We investigated whether the relationship between suicidal ideation and established correlates varied by age in a representative population. Methods: We used data from the 2007 Adult Psychiatric Morbidity Survey of England to assess the relationship between age and suicidal thoughts across 20-year age bands, using logistic regression, adjusted for survey weights. We used mediation analyses to assess the extent to which other factors mediate the relationship between suicidal thoughts and age. Results: Reports of previous-year suicidal thoughts decreased with age. This was partly explained by (1) lower rates of reported child abuse (in those aged 75þ ), of depression, and of anxiety symptoms (in those aged 55þ ), factors all strongly associated with suicidal thoughts, and (2) higher rates of protective factors in people aged 35þ, specifically homeownership and cohabitation. Rates of phobias, irritability and compulsions also decreased with age, and the association of these symptoms with suicidal thoughts was particularly strong in the youngest (16–34) age group. People who reported experiencing childhood abuse in all age groups reported more suicidal thoughts, suggesting abuse has lifelong negative effects on suicidal ideation. Limitations: The response rate was 57%. Older people may be less likely to recall childhood abuse. Conclusions: Sexual and physical abuse in childhood are associated with suicidal ideas throughout the lifespan, so screening for suicidal ideas in younger and older people should be routine and vigorous, and cover experiences in early life: management may require appropriate psychological interventions. & 2015 Elsevier B.V. All rights reserved.
Keywords: Suicide Epidemiology Aging
1. Background Suicide is the 10th leading cause of death worldwide (Hawton and van Heeringen, 2009). Effective risk prevention requires identification of those most likely to experience suicidal ideation and make non-lethal suicide attempts, as these thoughts and behaviours greatly increase the risk of suicide. The risk factors for suicidal ideation and actions differ from those of completed suicide. In particular, middle aged men are at greatest risk of completed suicide (Office of National Statistics, 2013b), while suicidal thoughts, plans and intent are more frequently reported
n
†
Corresponding author. Deceased.
http://dx.doi.org/10.1016/j.jad.2015.02.010 0165-0327/& 2015 Elsevier B.V. All rights reserved.
by younger women. The ratio of self-harm to completed suicide falls with increasing age from more than 200 to one in teenagers to less than 10 to one in people over 60 years of age (Hawton and Harriss, 2008). Physical frailty and a greater determination to die (Conwell et al., 1998) may conspire to increase the risk of death from each episode of self-harm in older people. Research in this age group has consistently shown that self-harm frequently involves strong suicidal intent (Salib et al., 2001). Thus, suicidal ideation may differ qualitatively in older people and their younger counterparts, and the established predictors might vary correspondingly. Some people exhibiting suicidal ideation in younger and midlife die by suicide before reaching older age, but most do not, as completed suicide is rare while suicidal thoughts are common. Thus surviving to old age cannot significantly explain the reduced suicidality among older people.
C. Cooper et al. / Journal of Affective Disorders 177 (2015) 42–48
Mental disorders are important predictors of suicidal ideation and completed suicide (Gunnell and Lewis, 2005). Epidemiological studies indicate that the prevalence of psychiatric symptoms and disorders (other than dementia) decreases with age (Alonso et al., 2004;McBride et al., 2013;McManus et al., 2009). The decrease in mental disorders in older age is concordant with the decrease in reported suicidal ideation. However, physical illness and disability are more prevalent in older age groups (Russell et al., 2009), and as they are associated with suicidal ideation this would tend to counteract the effect of reduced susceptibility to common mental disorders. Other consistent cross-national risk factors for suicidal ideation, plans and attempts include female gender, less education, not being married, and stressful life events (Nock et al., 2008). Recent papers on childhood abuse, psychiatric disorder, and suicide attempts and ideation detail the impact of early trauma on adult psychiatric morbidity and behaviour, particularly suicidal acts and self-harm (Bebbington et al., 2009; Jonas et al., 2011). Although the quality of childhood care may be extremely important in determining the way in which older adults cope with the threats and losses to independence from ageing and its associated life events (Martindale, 2007), there is very little research linking childhood trauma, mental state and behaviour in older people. Some might question the ability of older people to recall their childhood relationships as accurately as younger adults, for whom the experiences are more recent. However, memories with powerful, personal emotional significance are usually maintained over the lifespan (Holland and Kensinger, 2010). While some studies have suggested that up to 40% of childhood abuse may be forgotten or repressed for a time, this forgetting does not appear to be associated with age (Berntsen and Rubin, 2002). While around 5% of people aged over 60 have dementia (Ferri et al., 2005), dementia often provokes the re-experiencing of trauma from childhood sexual assault, because early, implicit, traumatic memories become relatively more important as explicit, more recent memories are lost (Australian Institute of Family Studies, 2010). Thus, why should a seventy-year-old adult recall childhood trauma in any way differently from a forty-year-old? There are nonetheless more general problems relating to the investigation of childhood abuse, due to the pervasive stigma that attaches to it (Coffey et al., 1996). Collecting accurate information about it in epidemiological surveys remains methodologically challenging. The 2007 English Adult Psychiatric Morbidity Survey offers advantages in this respect, seeking to overcome the effects of stigma by using computer assisted self interview (CASI) in the section relating to abuse. The survey thereby provided information on suicidal thoughts, childhood adversity and mental health in a large sample of the English adult household population. We investigated the correlates of thoughts of suicide in this representative population in order to identify disparities in their origins between younger and older age groups (16–34 years, 35– 54 years, 55–74 years and 75þ ). Based on the existing literature, we hypothesised that there would be differences in the relationship between suicidal thoughts and age group moderated by the effects of mental illness, substance misuse, sociodemographic characteristics (Nock et al., 2008), physical ill health, social networks (Handley et al., 2014) and childhood abuse (Bebbington et al., 2009; Jonas et al., 2011).
43
households. The survey adopted a multi-stage stratified probability sampling design. The sampling frame was the small user Postcode Address File. One adult aged 16 years or over was selected for interview in each eligible household using the Kish grid method (Kish, 1965). Full details of the sampling method, procedure and quality control are published elsewhere (McManus et al., 2009). Ethical approval for the survey was obtained from one of the Research Ethics Committees of the National Research Ethics Service appropriate for non-clinical populations. The interviews involved computerassisted personal interviewing (CAPI), with answers entered by the interviewers directly into a laptop. Particularly sensitive information was collected by computer assisted self-completion interview (CASI). The laptop was given to the participant for this. The respondents knew beforehand that the interviewer was unable to see the results of the self-completed parts of the interview, which included questions about childhood abuse. Fifty seven per cent of the eligible sample took part, and full interviews were successfully carried out on 7403 people; 7353 completed the CASI section on childhood abuse, while 50 people refused or were unable to complete this. Reasons given included: sight impairment, not being able to read, and refusal. Where this was the case respondents were offered the option of having the CASI questions read aloud. 2.2. Variables We selected sociodemographic and illness characteristics previously identified as predictors of suicidal ideation (e.g. Nock et al., 2008). 2.2.1. Sociodemographic information Standardised questions provided information about age, gender, cohabitation status and home ownership. Alcohol use in the past 6 months was recorded using the Alcohol Use Disorders Identification Test (AUDIT) (Saunders et al., 1993). Respondents with an AUDIT score of 8 or more were classed as having hazardous use of alcohol. Participants were asked about the number of adults they felt close to as a measure of their primary support network. This was dichotomised for analysis at a network size of 43 people, as recommended in previous literature (Brugha et al., 2003). 2.2.2. Mental illness Non-psychotic psychiatric disorders were assessed in relation to the past week, using the Clinical Interview Schedule-Revised (CIS-R) (Lewis et al., 1992). This can be administered by nonclinically trained interviewers. It provides scores for symptoms relating to common mental disorders. We analysed the subscales individually (see Table 1). Scores of 2 or more were used to denote the presence of individual symptoms (McBride et al., 2013).
2.1. Sample
2.2.3. Physical health/disability We included two variables measuring this. The first was a subjective item “My health limits moderate activities” yes/no from the SF-12 quality of life measure (Ware et al., 1996). Participants were also asked whether they needed help with any of seven Activities of Daily Living (ADLs). The ADLs were: personal care; mobility; medical care such as taking pills, having injections or changes of dressing; preparing meals, shopping, laundry and housework; practical activities such as gardening, decorating, or household repairs; dealing with paperwork; and managing money. This was developed from a previous measure (Bebbington et al., 2000).
We used data from the 2007 Adult Psychiatric Morbidity Survey. Unlike previous surveys in this programme (Jenkins et al., 2009), it covered only England and there was no upper age limit. The sample was designed to be representative of people living in private
2.2.4. Childhood sexual and physical abuse Respondents were asked about different types and levels of abuse. It was possible to distinguish abuse occurring in childhood and adolescence (i.e. before 16 years) from that occurring in
2. Method
44
C. Cooper et al. / Journal of Affective Disorders 177 (2015) 42–48
Table 1 Association of the presence of suicidal thoughts in the past year and age group with socio-demographic and clinical characteristics. n (%)
Suicidal thoughts, last year, n (%)
Age group n (%) 16–34 (n¼ 1603)
Socio-demographic Gender Lives with partner Home owner Childhood sexual and physical abuse Child sexual abuse reported Child physical abuse reported Physical health Reported disability Needs help with any ADLs Mental health Somatic symptoms Fatigue Concentration/forgetfulness Sleep Irritability Worries about physical health Depression Depressive ideas Worry Anxiety (n ¼7389) Phobias Panic Compulsions Obsessions Hazardous alcohol use (n¼ 7378) More than 3 close family or friends (n¼ 7269)
35–54 (n ¼2543)
55–74 (n¼ 2307)
75þ (n¼ 950)
Female Male Yes No Yes No
4197 3192 4129 3260 5167 2156
(51.4) (48.6) (63.1) (36.9) (70.4) (29.6)
224 115 119 220 170 162
(5.2) (3.4) (2.9) (6.7) (3.2) (6.7)
918 685 726 877 848 735
(49.9) (50.1) (43.0) (57.0) (54.1) (45.9)
1435 (50.5) 1108 (49.5) 1655 (75.6) 888 (24.4) 1874 (76.2) 654(23.8)
1272 1035 1436 871 1791 505
(51.5) (48.5) (74.8) (25.2) (80.2) (19.8)
581 369 316 634 661 269
(60.2) (39.8) (46.6) (63.4) (73.7) (26.3)
Yes No Yes No
652 6628 351 6962
(8.5) (91.5) (4.7) (95.3)
79 249 54 277
(12.2) (3.5) (16.0) (3.7)
137 1446 87 1500
(7.5) (92.5) (5.0) (95.0)
290 2226 142 2383
(10.3) (89.7) (5.3) (94.7)
201 2080 100 2185
(8.9) (91.1) (4.2) (95.8)
27 885 22 894
(2.9) (97.1) (2.6) (97.4)
Yes No Yes No
1460 5942 2795 4608
(15.9) (84.1) (32.9) (67.1)
111 228 228 111
(7.7) (3.7) (8.5) (2.3)
83 1520 351 1252
(4.6) (95.4) (20.5) (79.5)
343 2200 721 1822
(12.3) (87.7) (27.1) (72.9)
581 1725 1042 1265
(23.7) (76.3) (42.7) (57.3)
453 497 681 269
(48.0) (52.0) (71.4) (28.6)
o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ Yes No Yes No
6933 470 5217 2186 6641 762 5028 2375 6187 1216 6843 560 6500 903 6680 723 6000 1403 6769 634 7002 401 7194 209 7111 292 6997 406 1603 5789 6807 474
(94.0) (6.0) (72.2) (27.8) (90.5) (9.5) (69.9) (30.1) (82.7) (17.3) (93.3) (6.7) (88.5) (11.5) (90.7) (9.3) (81.3) (18.7) (92.0) (8.0) (94.5) (5.5) (97.4) (2.6) (96.0) (4.0) (94.8) (5.2) (24.2) (75.8) (94.5) (5.5)
257 (3.5) 82 (17.6) 82 (1.5) 257 (11.5) 174 (2.5) 165 (21.3) 113 (2.2) 226 (9.3) 153 (2.2) 186 (14.3) 250(3.5) 89 (15.1) 138 (2.0) 201 (22.2) 129(1.8) 210(29.0) 111(1.9) 228(14.9) 178 (2.6) 161 (23.8) 248(3.3) 91(22.1) 257(3.4) 82(38.7) 275(3.6) 64(21.5) 243(3.3) 96(23.2) 107 (6.3) 231 (3.7) 259 (3.7) 71 (13.9)
1488 115 1131 472 1434 169 1109 494 1216 387 1511 92 1399 204 1393 210 1226 377 1450 153 1475 128 1547 56 1501 102 1491 112 512 1089 1480 97
(93.5) (6.5) (73.5) (26.5) (90.6) (9.4) (90.6) (29.1) (77.8) (22.2) (95.0) (5.0) (88.1) (11.9) (88.4) (11.6) (88.2) (21.8) (91.4) (8.6) (92.3) (7.7) (97.2) (2.8) (94.3) (5.7) (93.7) (6.3) (34.0) (66.0) (94.7) (5.3)
2348 195 1739 804 2223 320 1736 807 1987 556 2337 206 2186 357 2217 326 1954 589 2257 286 2360 183 2445 98 2428 115 2382 161 606 1936 2367 141
(93.1) (6.9) (70.3) (29.7) (89.0) (11.0) (70.0) (30.0) (79.0) (21.0) (92.8) (7.2) (87.5) (12.5) (89.0) (11.0) (88.6) (21.4) (90.0) (10.0) (93.9) (6.1) (96.7) (3.3) (95.9) (4.1) (94.5) (5.5) (24.0) (76.0) (95.2) (4.8)
2189 118 1671 636 2124 183 1541 766 2080 227 2125 182 2059 248 2159 148 1976 331 2151 156 2229 78 2258 49 2249 58 2203 104 390 1916 2144 137
(95.3) (4.7) (73.9) (26.1) (92.7) (7.3) (68.5) (31.5) (90.0) (10.0) (92.6) (7.4) (90.1) (9.9) (94.3) (5.7) (86.3) (13.7) (94.0) (6.0) (96.6) (3.4) (97.9) (2.1) (97.6) (2.4) (95.9) (4.1) (17.1) (82.9) (94.8) (5.2)
908 42 676 274 860 90 642 308 904 46 870 80 856 94 911 39 844 106 911 39 938 12 944 6 933 17 921 29 95 848 816 99
(95.1) (4.9) (70.7) (29.3) (90.3) (9.7) (67.7) (32.3) (94.1) (5.9) (91.3) (8.7) (89.9) (10.1) (95.6) (4.4) (88.6) (11.4) (95.7) (4.3) (98.9) (1.1) (92.1) (7.9) (98.4) (1.6) (96.9) (3.1) (10.5) (89.5) (90.1) (9.9)
n gives unweighted numbers, while percentages are weighted.
adulthood. Reported physical abuse in childhood was defined as “a severe beating by parent or carer before the age of 16 years”. Reported sexual abuse in childhood combined responses to two items and was defined in the analysis as “sexual abuse ranging from inappropriate touching to non-consensual sexual intercourse before the age of 16 years”. Intercourse before 16 reported as consensual (which is legally defined as child sexual abuse) was not included in the definition of abuse for this survey.
reported thinking about taking their own life, based on the question “Have you ever thought of taking your life, even though you would not actually do it?” A positive response to suicidal thoughts was followed up by asking whether this last occurred in the past week, past year or longer ago. Thoughts in the previous year were used in the analysis. Participants were also asked whether they had attempted suicide in the past year. 2.3. Analysis
2.2.5. Dependent variable: ‘suicidal thoughts in the previous year’ The term ‘suicidal thoughts’ used in the analysis was given a narrow definition – it included only those respondents who
We weighted data by the strata, sampling units and sampling weights. The survey data were weighted to take account of survey design and non-response, so that the results were representative
C. Cooper et al. / Journal of Affective Disorders 177 (2015) 42–48
of the household population aged 16 years and over. Weighting was necessarily complex, and more details are available in the main report (McManus et al., 2009). We used Stata (Statacorp, 2013) and the SVY (survey data) commands with these weights for both descriptive and inferential statistics. We used methods accounting for the survey design to approximate counts and standard errors. We performed survey weighted multivariable logistic regressions to investigate the association between suicidal thoughts and age as a categorical variables (16–34 years, 35–54 years, 55–74 years and 75 þ), adjusting for other factors. The factors considered in the multivariable analyses were: demographic factors (gender, qualifications, being a home owner, and living with a partner), disability (reporting disability and needing help with ADLs), child abuse (reporting child sexual or physical abuse) and mental health (CIS-R subscales). We used the ‘medeff’ module in Stata (Hicks and Tingley, 2011, 2012) to assess the extent to which each of these factors mediates the association between age and suicidal thoughts. The ‘medeff’ command estimates mediation effects; it provides the percentage of the indirect effect accounted for by the factors studied. Finally, for variables where there was significant variation in the total effect mediated across age groups, we used the userwritten Stata command ‘ldecomp’ (Buis, 2014) to estimate factual (e.g. the actual proportion of 16–34 year olds reporting suicidal ideation) and counterfactual proportions (e.g. the proportion of 16–34 year olds who would report suicidal ideation if they were exposed to the same risk factors, operating in the same way as people aged 75 þ). This method enabled us to estimate the direct effect of the variables studied on suicidal ideation and the indirect effect (amount of these effects explained by age), and total effects such that the direct and indirect effects add up to the total effect.
45
3. Results 3.1. Suicidal thoughts and attempts in the year prior to interview. The sample ranged in age from 16 to 95 years. Thoughts about suicide in the previous year (recent suicidal thoughts) were reported by 339 (4.3%) of the sample. Recent suicidal thoughts were significantly more common in younger age groups. They were reported by 88 (5.1%) participants aged 16–34, 165 (5.4%) of participants aged 35–54, 66 (2.4%) of participants aged 55–74 and 20 (2.2%) of those aged 75 and above (Table 1). Only 52 people reported suicide attempts in the last year, 13 of those aged 16–34, 33 of those aged 35–54, six of those aged 55–74 and none in the oldest age group. All but one of the people reporting a suicide attempt also reported suicidal thoughts. 3.2. Effect of socio-demographic factors Suicidal thoughts in the past year were more likely to be reported by female respondents, and those who drank more alcohol, but less likely by those who owned their own homes, lived with a partner or had more than three close personal relationships. Respondents aged 16–34 were less likely to be living with a partner or living in a home they owned than older respondents. The protective effect of these two demographic characteristics was significantly modulated by age, being relatively greater in people aged 16–34 (as demonstrated by the higher % total mediated effect in these age categories for these characteristics, shown in Table 2). Table 3 sets out the factual and counterfactual proportions for these relationships. It shows that the true proportion of people aged 16–34 who report suicidal thoughts was
Table 2 Association between patients characteristics and suicidal thoughts and percentage of total effect mediated by age group. Characteristic (reference group)
Unadjusted oddsa
% of Total effect mediatedb 16–34 (n¼1601)
35–54 (n¼2538)
55–74(n¼2304)
75þ (n¼ 946)
Demographic characteristics Female gender Living with partner Home owner Hazardous alcohol use More than 3 close family or friends
1.6 (1.2–2.1) 0.4 (0.3–0.5) 0.5 (0.4–0.6) 1.8 (1.3–2.3) 0.2 (0.2–0.3)
1 90 70 27 0.4
0 53 21 2 4
1 19 16 9 0.1
7 25 4 13 18
Child abuse Reported child sexual abuse Reported child physical abuse
3.8 (2.8–5.1) 5.0 (3.5–7.0)
9 8
18 9
4 2
17 9
Physical health Reported moderate disability Needs help with any ADLs
2.2 (1.7–2.9) 4.0 (3.1–5.1)
75 120n
16 40
16 26
54 86
Mental health Somatic symptoms Fatigue Concentration/forgetfulness Sleep Irritability Worries about physical health Depression Depressive ideas Worry Anxiety Phobias Panic Compulsions Obsessions
6.0 (4.4–8.2) 8.3 (6.2–11.2) 10.4 (7.9–13.6) 4.7 (3.6–6.1) 7.2 (5.6–9.4) 4.9 (3.6–6.6) 13.8 (10.5–18.1) 22.3 (17.0–29.2) 9.1 (7.0–11.9) 11.7 (8.9–15.3) 8.3 (6.2–11.2) 17.9 (12.6–25.5) 7.3 (5.2–10.4) 8.9 (6.7–11.8)
11 9 1 4 69 22 9 71 48 18 49 8 34 27
13 20 27 1 40 8 20 44 34 39 13 20 2 6
8 5 17 7 39 6 13 45 29 19 17 8 12 9
5 9 3 8 55 1 10 55 37 30 31 25 17 15
a b
Estimates obtained from a logistic regression without additional covariates. Estimates obtained using ‘Medeff’ Stata command; total effect mediated is a theoretical construct and is occasionally more than 100%.
46
C. Cooper et al. / Journal of Affective Disorders 177 (2015) 42–48
Table 3 Predicted and counterfactual proportionsa shown as percentages for significant predictors of suicidal thoughts. Characteristic
How 16–34 age group suicidal thoughts might change (%)
Sociodemographic …if same % Living 16–34 with partner as: 35–54 55–74 75þ Home ownerb 16–34 35–54 55–74 75þ Physical health Moderate disability
16–34 35–54 55–74 75þ Needing help with 16–34 ADLs 35–54 55–74 75þ
Mental health Irritability
Phobias
Compulsions
16–34 35–54 55–74 75þ 16–34 35–54 55–74 75þ 16–34 35–54 55–74 75þ
How 35–54 age group suicidal thoughts might change (%)
How 55–74 age group suicidal thoughts might change (%)
How 75þ age group suicidal thoughts might change (%)
5.2 3.8 3.8 5.0 5.0 4.2 4.1 4.3
7.4 5.5 5.5 7.1 6.3 5.4 5.2 5.5
3.3 2.4 2.4 3.2 3.0 2.5 2.5 2.6
2.2 1.6 1.7 2.2 2.5 2.1 2.0 2.1
5.2 5.9 7.0 9.3 5.2 5.9 7.6 10.8
4.8 5.5 6.5 8.6 4.8 5.5 7.1 10.0
1.8 2.0 2.4 3.3 1.6 1.9 2.4 3.5
1.2 1.3 1.6 2.2 1.0 1.2 1.5 2.2
5.2 5.0 3.7 3.2 5.2 4.9 4.3 3.9 5.2 4.9 4.6 4.4
5.6 5.5 4.0 3.5 5.8 5.5 4.9 4.4 5.7 5.5 5.1 5.0
3.5 3.3 2.4 2.1 3.0 2.8 2.4 2.2 2.8 2.6 2.4 2.4
3.6 3.5 2.5 2.2 3.0 2.8 2.4 2.2 2.6 2.4 2.3 2.2
a Percentage by age group reporting suicidal thoughts, and (in grey cells) the % predicted to report suicidal thoughts if the distribution of each of the characteristics studied changed to that found in a different age group. b True proportions vary due to slightly different population for which this information available.
5.2%, but if their likelihood of living with a partner equalled that of people aged 35–54 or 55–74, only 3.8% would report these thoughts, because living with a partner is protective. Similarly, the proportion of respondents aged 16–34 who reported suicidal ideation would be around one percentage point lower (4% vs 5%) if their rate of home ownership was as high as that of people in the older age groups. 3.3. Effects of child abuse Childhood sexual and physical abuse were both strong predictors of recent suicidal thoughts. Childhood sexual abuse was associated with a fourfold, and physical abuse a five fold increase in risk of experiencing suicidal ideation in unadjusted analyses (Table 2). Abuse was less frequently reported by people aged 75 þ, but the relationship between reporting these events and having recent suicidal thoughts varied very little with age. 3.4. Effects of physical health characteristics As expected, the proportion of people reporting moderate disability or needing help with ADLs rose sharply with age. These characteristics were associated with reporting more suicidal thoughts, and this relationship was stronger in the youngest (16–34) and oldest (aged 75 þ) age groups. In Table 3, the factual and counterfactual proportions demonstrate that the proportion of people aged 75 þ reporting suicidal ideation would theoretically reduce from 2.2% to 1.0% if they had the same level of need for help
with ADLs as people aged 16–34. Similarly, the proportion of the youngest people reporting suicidal ideation would theoretically be 10.8% rather than 5.2% if they had the same level of need for ADL help as people aged 75 þ. 3.5. Effects of mental health symptoms Depressive ideas and symptoms of worry were strongly associated with reporting suicidal thoughts across all age groups, as expected. These symptoms were more frequently reported by younger respondents (aged under 55 years). The strength of the relationship between these symptoms and suicidal ideation was comparable across the age groups. Symptoms of irritability, phobias and compulsions also declined with age. As well as occurring more frequently, these symptoms were all more strongly associated with reporting suicidal ideas in the youngest (16–34) than in older groups. In Table 3, the counterfactual proportions demonstrate that if people aged 16–34 had levels of irritability, phobias and compulsions comparable to those aged 75þ, then instead of 5.2% their rates of reported suicidal ideation would be 3.2%, 3.9% and 4.4% respectively.
4. Discussion Psychiatric morbidity, childhood sexual and physical abuse, physical illness and disability, hazardous alcohol use, small social networks and not living with a partner or owning one's home were
C. Cooper et al. / Journal of Affective Disorders 177 (2015) 42–48
all associated with recent suicidal thoughts across all age groups. The primary aim of our study was to clarify why suicidal thoughts decline in frequency across the life span. Our findings suggest that the lower frequency of reported depression and anxiety symptoms in people aged 55 and over, and of child abuse reported by people aged 75 and over might explain the lower rates of suicidal ideas reported by older people, given that these characteristics are strongly related to suicidal ideation across all ages in our sample. People in the youngest age group also had the highest rates of irritability, compulsions and phobias, and these symptoms were particularly strongly related to suicidal ideation in this group. Younger people (aged 16–34) also had the lowest rates of home ownership and co-habitation, both of which protected against suicidal ideation; among those younger people who did own their home or co-habit, these characteristics were relatively more protective against suicidal thoughts than in older age groups. We speculate that people in this age group who were single or renting their accommodation may have been more likely to view these situations as undesirable (and be trying to change them), whereas in older age groups these circumstances may have been life choices for some, and, where they were not, at least more accepted. This may also be a cohort effect, as home ownership levels rose steadily from 23% in 1918 to 69% in 2001, thus becoming an expected achievement for many (Office of National Statistics, 2013a). The recent decline in home ownership in Britain following the financial crisis of 2008 post-dated this survey, and might conduce to a subsequent increase in suicidal ideation. We were particularly interested in why people in older age groups report low levels of suicidal ideation. The high rates of disability in people aged 75þ are striking. Our findings that psychiatric morbidity and ADL impairments were associated with suicidal thoughts across the age groups accord with previous studies (Dennis et al., 2007; Dennis et al., 2009). Disability was an important predictor of suicidal thoughts in the youngest and the oldest age groups, so the lower rate of suicidal thoughts in people aged 75þ cannot be explained as a greater tolerance of disability and impairment. This oldest age group were also less likely than those in middle age to be living with a partner, an important protector against suicidal ideation. It has been suggested that rates of mental illness symptoms are lower in older people because of under-reporting, and our study cannot rule this out. However, if this were so, older people should be more likely to have suicidal ideation at a given level of mental symptoms than their younger counterparts, but we did not find this. The association between childhood sexual abuse and suicidal ideation has been highlighted in previous publications from this survey (Jonas et al., 2011). We found in the current paper that the association between child physical abuse and suicidal ideation was even greater than that between child sexual abuse and suicidal ideation, although the definition used for physical abuse, a severe beating, may have only detected more severe physical abuse. This is the first study to include childhood abuse in a comparison of the correlates of suicidal thinking across age groups: it appears to indicate that the negative effects of abuse on suicidality are lifelong. Many people who are abused as children end up in difficult relationships as adults (Jonas et al., 2011) and they may also have distorted ideas about their own self-worth and the way others feel about them, although we did not measure adult relationship quality or self-worth in the current study. The significance of strong and enduring relationships and their association with good mental health and well-being is well recognised in both genders (Gibb et al., 2011). Such an association must be the result of complex interactions, with choice of partner, the ability to work within a relationship, and the buffering and support provided by the relationship during times of stress all being important. The breakdown of a relationship has enormous effects on self-esteem and mental health, and is a well-recognised reason for people to state that they no longer have reasons to live.
47
4.1. Limitations The British National Psychiatric Morbidity Programme involves repeated representative household surveys which have had decreasing response rates over time; the current response rate of 57% is fairly low, although comparable with other recent national surveys (Mindell et al., 2012) (and 70% of people successfully contacted did then agree to take part). We weighted the data to correct for nonresponse on a range of sociodemographic and area characteristics. Non-response weighting had little effect on the results, showing that non-responders seem to be similar to responders (Jonas et al., 2011). Moreover, the current study explored associations, so findings are relatively less sensitive to response bias than studies of prevalence. Just 20 people aged 75þ in the sample reported suicidal thoughts, and this is a significant limitation. Great efforts were made in the survey to maintain the quality of information by using confidential data entry. The reported prevalence of childhood abuse decreased in older age. Some of this may be explained by shorter life expectancy in those experiencing adverse childhood events, by age-related increases in memory problems or by “positivity” of memories (Schlagman et al., 2006). The greater acceptance in recent years that childhood abuse, especially sexual abuse is common (Smart, 2000) may have had an effect, as abusive experiences, especially those at the less severe end of the spectrum may be more likely to be reported as abuse if they have been named as such at the time. However, even though reasons for lower levels of reported childhood abuse with increasing age remain obscure, the relationship between childhood abuse and lack of well-being remains strong in the older adult group. Around 15% (n¼52) of those reporting suicidal ideation also reported a suicide attempt in the last year. We did not distinguish those who reported acting on their thoughts from those reporting suicidal thought without associated actions. We could not analyse actual suicidal acts as the dependent variable because of the small number of participants reporting these actions, although we have previously reported that suicidal thinking represents a strong indicator of vulnerability to suicidal acts from the National Psychiatric Morbidity Surveys (Bebbington et al., 2010). Nonetheless as most people who have suicidal thoughts do not attempt suicide, knowledge about risk factors that can predict whether people with suicidal ideation progress to actual self harm are of critical importance for suicide prevention strategies, and other recent studies have completed such analyses (Glenn and Nock, 2014). 4.2. Implications Our results provide further support for vigorous opportunistic screening for suicidal ideas in both young and old, as poor mental and physical health, disability, and social isolation are associated with suicidal ideation across all age groups. The detrimental effect of childhood abuse appears to be maintained in older adults, probably operating through effects on selfesteem, and seems to increase the risk of thoughts of suicide. This indicates that clinicians should continue to ask about childhood experiences, as older people may benefit from appropriate psychological work for this, and for other relationship difficulties. Where there is insufficient family or peer support, it should be recognised that, in contradiction to current trends in mental and primary health care, there may be a need for longer term support from professional services. The alert listening ear of a regularly seen and trusted health worker may be crucial for older people. Role of the funding source No external funding was received for this work.
48
C. Cooper et al. / Journal of Affective Disorders 177 (2015) 42–48
Conflict of interest None of the authors has any financial involvement or affiliation with any organisation whose financial interests may be affected by material in this manuscript, or which potentially bias it.
Acknowledgements None.
References Alonso, J., Angermeyer, M.C., Bernert, S., Bruffaerts, R., Brugha, I.S., Bryson, H., de Girolamo, G., De Graaf, R., Demyttenaere, K., Gasquet, I., Haro, J.M., Katz, S.J., Kessler, R.C., Kovess, V., Lepine, J.R., Ormel, J., Polidori, G., Russo, L.J., Vilagut, G., Almansa, J., Arbabzadeh-Bouchez, S., Autonell, J., Bernal, M., Buist-Bouwman, M.A., Codony, M., Domingo-Salvany, A., Ferrer, M., Joo, S.S., Martinez-Alonso, M., Matschinger, H., Mazzi, F., Morgan, Z., Morosini, R., Palacin, C., Romera, B., Taub, N., Vollebergh, W.A.M., 2004. Prevalence of mental disorders in Europe: results from the european study of the epidemiology of mental disorders (ESEMeD) project. Acta Psychiatr. Scand. 109, 21–27. Australian Institute of Family Studies, 2010. Ageing, dementia and adult survivors of childhood sexual assault, Report. Bebbington, P., Brugha, T., Meltzer, H., Farrell, M., Ceresa, C., Jenkins, R., Lewis, G., 2000. Psychiatric disorder and dysfunction in the UK national survey of psychiatric morbidity. Soc. Psychiatry Psychiatr. Epidemiol. 35 (5), 191–197. Bebbington, P.E., Cooper, C., Minot, S., Brugha, T.S., Jenkins, R., Meltzer, H., Dennis, M., 2009. Suicide attempts, gender, and sexual abuse: data from the 2000 British psychiatric morbidity survey. Am. J. Psychiatry 166 (10), 1135–1140. Bebbington, P.E., Minot, S., Cooper, C., Dennis, M., Meltzer, H., Jenkins, R., Brugha, T., 2010. Suicidal ideation, self-harm and attempted suicide: results from the British psychiatric morbidity survey 2000. Eur. Psychiatry 25 (7), 427–431. Berntsen, D., Rubin, D.C., 2002. Emotionally charged autobiographical memories across the life span: the recall of happy, sad, traumatic, and involuntary memories. Psychol. Aging 17 (4), 636–652. Brugha, T.S., Morgan, Z., Bebbington, P., Jenkins, R., Lewis, G., Farrell, M., Meltzer, H., 2003. Social support networks and type of neurotic symptom among adults in British households. Psychol. Med. 33 (2), 307–318. Buis, M., 2014. LDECOMP: Stata module decomposing the total effects in a logistic regression into direct and indirect effects. Stat. Softw. Compon.. Coffey, P., Leitenberg, H., Henning, K., Turner, T., Bennett, R.T., 1996. Mediators of the long-term impact of child sexual abuse: perceived stigma, betrayal, powerlessness, and self-blame. Child Abuse Negl. 20 (5), 447–455. Conwell, Y., Duberstein, P.R., Cox, C., Herrmann, J., Forbes, N., Caine, E.D., 1998. Age differences in behaviors leading to completed suicide. Am. J. Geriatr. Psychiatry 6 (2), 122–126. Dennis, M., Baillon, S., Brugha, T., Lindesay, J., Stewart, R., Meltzer, H., 2009. The influence of limitation in activity of daily living and physical health on suicidal ideation: results from a population survey of Great Britain. Soc. Psychiatry Psychiatr. Epidemiol. 44 (8), 608–613. Dennis, M., Baillon, S., Brugha, T., Lindesay, J., Stewarvand, R., Meltzer, H., 2007. The spectrum of suicidal ideation in Great Britain: comparisons across a 16–74 years age range. Psychol. Med. 37 (6), 795–805. Ferri, C.P., Prince, M., Brayne, C., Brodaty, H., Fratiglioni, L., Ganguli, M., Hall, K., Hasegawa, K., Hendrie, H., Huang, Y.Q., Jorm, A., Mathers, C., Menezes, P.R., Rimmer, E., Scazufca, M., 2005. Global prevalence of dementia: a Delphi consensus study. Lancet 366 (9503), 2112–2117. Gibb, S.J., Fergusson, D.M., Norwood, L.J., 2011. Relationship duration and mental health outcomes: findings from a 30-year longitudinal study. Br. J. Psychiatry 198 (1), 24–30. Glenn, C.R., Nock, M.K., 2014. Improving the short-term prediction of suicidal behavior. Am. J. Prev. Med. 47 (3S2), 176–180.
Gunnell, D., Lewis, G., 2005. Studying suicide from the life course perspective: implications for prevention. Br. J. Psychiatry 187, 206–208. Handley, T.E., Hiles, S.A., Inder, K.J., Kay-Lambkin, F.J., Kelly, B.J., Lewin, T.J., McEvoy, M., Peel, R., Attia, J.R., 2014. Predictors of suicidal ideation in older people: a decision tree analysis. Am. J. Geriatr. Psychiatry 22 (11), 1325–1335. Hawton, K., Harriss, L., 2008. How often does deliberate self-harm occur relative to each suicide? A study of variations by gender and age. Suicide Life-Threat. Behav. 38 (6), 650–660. Hawton, K., van Heeringen, K., 2009. Suicide. Lancet 373 (9672), 1372–1381. Hicks, R., Tingley, D., 2011. Causal mediation analysis. Stata J. 11, 1–15. Hicks, R., Tingley, D., 2012. Mediation: Stata Module for Causal Mediation Analysis and Sensitivity Analysis, Computer Program. Holland, A.C., Kensinger, E.A., 2010. Emotion and autobiographical memory. Phys. Life Rev. 7 (1), 88–131. Jenkins, R., Meltzer, H., Bebbington, P., Brugha, T., Farrell, M., McManus, S., Singleton, N., 2009. The British mental health survey programme: achievements and latest findings. Soc. Psychiatry Psychiatr. Epidemiol. 44 (11), 899–904. Jonas, S., Bebbington, P., McManus, S., Meltzer, H., Jenkins, R., Kuipers, E., Cooper, C., King, M., Brugha, T., 2011. Sexual abuse and psychiatric disorder in England: results from the 2007 adult psychiatric morbidity survey. Psychol. Med. 41 (4), 709–719. Kish, L., 1965. Survey Sampling. Wiley and Sons, London. Lewis, G., Pelosi, A.J., Araya, R., Dunn, G., 1992. Measuring psychiatric disorder in the community: a standardized assessment for use by lay interviewers. Psychol. Med. 22 (2), 465–486. Martindale, B., 2007. Resilience and vulnerability in later life. Br. J. Psychother. 23, 205–216. McBride, O., Bebbington, P., Cooper, C., 2013. Could the lower prevalence of affective disorder in older people be due to measurement error? Reliability of the revised clinical interview schedule in younger and older adults. J. Affect. Disord. 148 (2–3), 310–315. McManus, S., Meltzer, H., Brugha, T.S., Bebbington, P.E., Jenkins, R., 2009. Adult Psychiatric Morbidity in England, 2007: Results of a Household Survey, HSCIC, Report. Mindell, J., Biddulph, J.P., Hirani, V., Stamatakis, E., Craig, R., Nunn, S., Shelton, N., 2012. Cohort profile: the health survey for England. Int. J. Epidemiol. 41 (6), 1585–1593. Nock, M.K., Borges, G., Bromet, E.J., Alonso, J., Angermeyer, M., Beautrais, A., Bruffaerts, R., Chiu, W.T., de Girolamo, G., Gluzman, S., De Graaf, R., Gureje, O., Haro, J.M., Huang, Y., Karam, E., Kessler, R.C., Lepine, J.P., Levinson, D., Medina-Mora, M.E., Ono, Y., Posada-Villa, J., Williams, D., 2008. Cross-national prevalence and risk factors for suicidal ideation, plans and attempts. Br. J. Psychiatry 192 (2), 98–105. Office of National Statistics, 2013a. A Century Home Ownersh, Online Source. Office of National Statistics, 2013b. Suicides United Kingdom, 2011, Report. Russell, D., Turner, R.J., Joiner, T.E., 2009. Physical disability and suicidal ideation: a community-based study of risk/protective factors for suicidal thoughts. Suicide Life-Threat. Behav. 39 (4), 440–451. Salib, E., Tadros, G., Cawley, S., 2001. Elderly suicide and attempted suicide: one syndrome. Med. Sci. Law 41 (3), 250–255. Saunders, J.B., Aasland, O.G., Babor, T.F., Delafuente, J.R., Grant, M., 1993. Development of the alcohol-use disorders identification test (audit) – who collaborative project on early detection of persons with harmful alcohol-consumption 2. Addiction 88 (6), 791–804. Schlagman, S., Schulz, J., Kvavilashvili, L., 2006. A content analysis of involuntary autobiographical memories: examining the positivity effect in old age. Memory 14 (2), 161–175. Smart, C., 2000. Reconsidering the recent history of child sexual abuse, 1910–1960. J. Soc. Policy 29, 55–71. Statacorp, 2013. Stata Statistical Software: Release 13. College Station, TX, Statacorp, Computer Program. Ware, J.E., Kosinski, M., Keller, S.D., 1996. A 12-item short-form health survey – construction of scales and preliminary tests of reliability and validity. Med. Care 34 (3), 220–233.
Contents lists available at ScienceDirect
Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad
Research report
Why are suicidal thoughts less prevalent in older age groups? Age differences in the correlates of suicidal thoughts in the English Adult Psychiatric Morbidity Survey 2007 Claudia Cooper a,n, Khadija Rantell a, Martin Blanchard a, Sally McManus b, Michael Dennis c, Traolach Brugha d, Rachel Jenkins e, Howard Meltzer d,†, Paul Bebbington a a
Division of Psychiatry, 2nd Floor, Charles Bell House, 67-73 Riding House St., London W1W 7EJ, UK NatCen for Social Research, London EC1V 0AX, UK College of Medicine, Swansea University, Singleton Park, Swansea, Wales, UK d Department of Health Sciences, University of Leicester, 22-28 Princess Road West, Leicester LE1 6TP UK e Kings College London, Institute of Psychiatry, De Crespigny Park, London, UK b c
art ic l e i nf o
a b s t r a c t
Article history: Received 1 November 2014 Received in revised form 9 February 2015 Accepted 11 February 2015 Available online 19 February 2015
Background: Suicidal ideation is more strongly associated with suicidal intent in later life, so risk factors may also differ by age. We investigated whether the relationship between suicidal ideation and established correlates varied by age in a representative population. Methods: We used data from the 2007 Adult Psychiatric Morbidity Survey of England to assess the relationship between age and suicidal thoughts across 20-year age bands, using logistic regression, adjusted for survey weights. We used mediation analyses to assess the extent to which other factors mediate the relationship between suicidal thoughts and age. Results: Reports of previous-year suicidal thoughts decreased with age. This was partly explained by (1) lower rates of reported child abuse (in those aged 75þ ), of depression, and of anxiety symptoms (in those aged 55þ ), factors all strongly associated with suicidal thoughts, and (2) higher rates of protective factors in people aged 35þ, specifically homeownership and cohabitation. Rates of phobias, irritability and compulsions also decreased with age, and the association of these symptoms with suicidal thoughts was particularly strong in the youngest (16–34) age group. People who reported experiencing childhood abuse in all age groups reported more suicidal thoughts, suggesting abuse has lifelong negative effects on suicidal ideation. Limitations: The response rate was 57%. Older people may be less likely to recall childhood abuse. Conclusions: Sexual and physical abuse in childhood are associated with suicidal ideas throughout the lifespan, so screening for suicidal ideas in younger and older people should be routine and vigorous, and cover experiences in early life: management may require appropriate psychological interventions. & 2015 Elsevier B.V. All rights reserved.
Keywords: Suicide Epidemiology Aging
1. Background Suicide is the 10th leading cause of death worldwide (Hawton and van Heeringen, 2009). Effective risk prevention requires identification of those most likely to experience suicidal ideation and make non-lethal suicide attempts, as these thoughts and behaviours greatly increase the risk of suicide. The risk factors for suicidal ideation and actions differ from those of completed suicide. In particular, middle aged men are at greatest risk of completed suicide (Office of National Statistics, 2013b), while suicidal thoughts, plans and intent are more frequently reported
n
†
Corresponding author. Deceased.
http://dx.doi.org/10.1016/j.jad.2015.02.010 0165-0327/& 2015 Elsevier B.V. All rights reserved.
by younger women. The ratio of self-harm to completed suicide falls with increasing age from more than 200 to one in teenagers to less than 10 to one in people over 60 years of age (Hawton and Harriss, 2008). Physical frailty and a greater determination to die (Conwell et al., 1998) may conspire to increase the risk of death from each episode of self-harm in older people. Research in this age group has consistently shown that self-harm frequently involves strong suicidal intent (Salib et al., 2001). Thus, suicidal ideation may differ qualitatively in older people and their younger counterparts, and the established predictors might vary correspondingly. Some people exhibiting suicidal ideation in younger and midlife die by suicide before reaching older age, but most do not, as completed suicide is rare while suicidal thoughts are common. Thus surviving to old age cannot significantly explain the reduced suicidality among older people.
C. Cooper et al. / Journal of Affective Disorders 177 (2015) 42–48
Mental disorders are important predictors of suicidal ideation and completed suicide (Gunnell and Lewis, 2005). Epidemiological studies indicate that the prevalence of psychiatric symptoms and disorders (other than dementia) decreases with age (Alonso et al., 2004;McBride et al., 2013;McManus et al., 2009). The decrease in mental disorders in older age is concordant with the decrease in reported suicidal ideation. However, physical illness and disability are more prevalent in older age groups (Russell et al., 2009), and as they are associated with suicidal ideation this would tend to counteract the effect of reduced susceptibility to common mental disorders. Other consistent cross-national risk factors for suicidal ideation, plans and attempts include female gender, less education, not being married, and stressful life events (Nock et al., 2008). Recent papers on childhood abuse, psychiatric disorder, and suicide attempts and ideation detail the impact of early trauma on adult psychiatric morbidity and behaviour, particularly suicidal acts and self-harm (Bebbington et al., 2009; Jonas et al., 2011). Although the quality of childhood care may be extremely important in determining the way in which older adults cope with the threats and losses to independence from ageing and its associated life events (Martindale, 2007), there is very little research linking childhood trauma, mental state and behaviour in older people. Some might question the ability of older people to recall their childhood relationships as accurately as younger adults, for whom the experiences are more recent. However, memories with powerful, personal emotional significance are usually maintained over the lifespan (Holland and Kensinger, 2010). While some studies have suggested that up to 40% of childhood abuse may be forgotten or repressed for a time, this forgetting does not appear to be associated with age (Berntsen and Rubin, 2002). While around 5% of people aged over 60 have dementia (Ferri et al., 2005), dementia often provokes the re-experiencing of trauma from childhood sexual assault, because early, implicit, traumatic memories become relatively more important as explicit, more recent memories are lost (Australian Institute of Family Studies, 2010). Thus, why should a seventy-year-old adult recall childhood trauma in any way differently from a forty-year-old? There are nonetheless more general problems relating to the investigation of childhood abuse, due to the pervasive stigma that attaches to it (Coffey et al., 1996). Collecting accurate information about it in epidemiological surveys remains methodologically challenging. The 2007 English Adult Psychiatric Morbidity Survey offers advantages in this respect, seeking to overcome the effects of stigma by using computer assisted self interview (CASI) in the section relating to abuse. The survey thereby provided information on suicidal thoughts, childhood adversity and mental health in a large sample of the English adult household population. We investigated the correlates of thoughts of suicide in this representative population in order to identify disparities in their origins between younger and older age groups (16–34 years, 35– 54 years, 55–74 years and 75þ ). Based on the existing literature, we hypothesised that there would be differences in the relationship between suicidal thoughts and age group moderated by the effects of mental illness, substance misuse, sociodemographic characteristics (Nock et al., 2008), physical ill health, social networks (Handley et al., 2014) and childhood abuse (Bebbington et al., 2009; Jonas et al., 2011).
43
households. The survey adopted a multi-stage stratified probability sampling design. The sampling frame was the small user Postcode Address File. One adult aged 16 years or over was selected for interview in each eligible household using the Kish grid method (Kish, 1965). Full details of the sampling method, procedure and quality control are published elsewhere (McManus et al., 2009). Ethical approval for the survey was obtained from one of the Research Ethics Committees of the National Research Ethics Service appropriate for non-clinical populations. The interviews involved computerassisted personal interviewing (CAPI), with answers entered by the interviewers directly into a laptop. Particularly sensitive information was collected by computer assisted self-completion interview (CASI). The laptop was given to the participant for this. The respondents knew beforehand that the interviewer was unable to see the results of the self-completed parts of the interview, which included questions about childhood abuse. Fifty seven per cent of the eligible sample took part, and full interviews were successfully carried out on 7403 people; 7353 completed the CASI section on childhood abuse, while 50 people refused or were unable to complete this. Reasons given included: sight impairment, not being able to read, and refusal. Where this was the case respondents were offered the option of having the CASI questions read aloud. 2.2. Variables We selected sociodemographic and illness characteristics previously identified as predictors of suicidal ideation (e.g. Nock et al., 2008). 2.2.1. Sociodemographic information Standardised questions provided information about age, gender, cohabitation status and home ownership. Alcohol use in the past 6 months was recorded using the Alcohol Use Disorders Identification Test (AUDIT) (Saunders et al., 1993). Respondents with an AUDIT score of 8 or more were classed as having hazardous use of alcohol. Participants were asked about the number of adults they felt close to as a measure of their primary support network. This was dichotomised for analysis at a network size of 43 people, as recommended in previous literature (Brugha et al., 2003). 2.2.2. Mental illness Non-psychotic psychiatric disorders were assessed in relation to the past week, using the Clinical Interview Schedule-Revised (CIS-R) (Lewis et al., 1992). This can be administered by nonclinically trained interviewers. It provides scores for symptoms relating to common mental disorders. We analysed the subscales individually (see Table 1). Scores of 2 or more were used to denote the presence of individual symptoms (McBride et al., 2013).
2.1. Sample
2.2.3. Physical health/disability We included two variables measuring this. The first was a subjective item “My health limits moderate activities” yes/no from the SF-12 quality of life measure (Ware et al., 1996). Participants were also asked whether they needed help with any of seven Activities of Daily Living (ADLs). The ADLs were: personal care; mobility; medical care such as taking pills, having injections or changes of dressing; preparing meals, shopping, laundry and housework; practical activities such as gardening, decorating, or household repairs; dealing with paperwork; and managing money. This was developed from a previous measure (Bebbington et al., 2000).
We used data from the 2007 Adult Psychiatric Morbidity Survey. Unlike previous surveys in this programme (Jenkins et al., 2009), it covered only England and there was no upper age limit. The sample was designed to be representative of people living in private
2.2.4. Childhood sexual and physical abuse Respondents were asked about different types and levels of abuse. It was possible to distinguish abuse occurring in childhood and adolescence (i.e. before 16 years) from that occurring in
2. Method
44
C. Cooper et al. / Journal of Affective Disorders 177 (2015) 42–48
Table 1 Association of the presence of suicidal thoughts in the past year and age group with socio-demographic and clinical characteristics. n (%)
Suicidal thoughts, last year, n (%)
Age group n (%) 16–34 (n¼ 1603)
Socio-demographic Gender Lives with partner Home owner Childhood sexual and physical abuse Child sexual abuse reported Child physical abuse reported Physical health Reported disability Needs help with any ADLs Mental health Somatic symptoms Fatigue Concentration/forgetfulness Sleep Irritability Worries about physical health Depression Depressive ideas Worry Anxiety (n ¼7389) Phobias Panic Compulsions Obsessions Hazardous alcohol use (n¼ 7378) More than 3 close family or friends (n¼ 7269)
35–54 (n ¼2543)
55–74 (n¼ 2307)
75þ (n¼ 950)
Female Male Yes No Yes No
4197 3192 4129 3260 5167 2156
(51.4) (48.6) (63.1) (36.9) (70.4) (29.6)
224 115 119 220 170 162
(5.2) (3.4) (2.9) (6.7) (3.2) (6.7)
918 685 726 877 848 735
(49.9) (50.1) (43.0) (57.0) (54.1) (45.9)
1435 (50.5) 1108 (49.5) 1655 (75.6) 888 (24.4) 1874 (76.2) 654(23.8)
1272 1035 1436 871 1791 505
(51.5) (48.5) (74.8) (25.2) (80.2) (19.8)
581 369 316 634 661 269
(60.2) (39.8) (46.6) (63.4) (73.7) (26.3)
Yes No Yes No
652 6628 351 6962
(8.5) (91.5) (4.7) (95.3)
79 249 54 277
(12.2) (3.5) (16.0) (3.7)
137 1446 87 1500
(7.5) (92.5) (5.0) (95.0)
290 2226 142 2383
(10.3) (89.7) (5.3) (94.7)
201 2080 100 2185
(8.9) (91.1) (4.2) (95.8)
27 885 22 894
(2.9) (97.1) (2.6) (97.4)
Yes No Yes No
1460 5942 2795 4608
(15.9) (84.1) (32.9) (67.1)
111 228 228 111
(7.7) (3.7) (8.5) (2.3)
83 1520 351 1252
(4.6) (95.4) (20.5) (79.5)
343 2200 721 1822
(12.3) (87.7) (27.1) (72.9)
581 1725 1042 1265
(23.7) (76.3) (42.7) (57.3)
453 497 681 269
(48.0) (52.0) (71.4) (28.6)
o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ o2 2þ Yes No Yes No
6933 470 5217 2186 6641 762 5028 2375 6187 1216 6843 560 6500 903 6680 723 6000 1403 6769 634 7002 401 7194 209 7111 292 6997 406 1603 5789 6807 474
(94.0) (6.0) (72.2) (27.8) (90.5) (9.5) (69.9) (30.1) (82.7) (17.3) (93.3) (6.7) (88.5) (11.5) (90.7) (9.3) (81.3) (18.7) (92.0) (8.0) (94.5) (5.5) (97.4) (2.6) (96.0) (4.0) (94.8) (5.2) (24.2) (75.8) (94.5) (5.5)
257 (3.5) 82 (17.6) 82 (1.5) 257 (11.5) 174 (2.5) 165 (21.3) 113 (2.2) 226 (9.3) 153 (2.2) 186 (14.3) 250(3.5) 89 (15.1) 138 (2.0) 201 (22.2) 129(1.8) 210(29.0) 111(1.9) 228(14.9) 178 (2.6) 161 (23.8) 248(3.3) 91(22.1) 257(3.4) 82(38.7) 275(3.6) 64(21.5) 243(3.3) 96(23.2) 107 (6.3) 231 (3.7) 259 (3.7) 71 (13.9)
1488 115 1131 472 1434 169 1109 494 1216 387 1511 92 1399 204 1393 210 1226 377 1450 153 1475 128 1547 56 1501 102 1491 112 512 1089 1480 97
(93.5) (6.5) (73.5) (26.5) (90.6) (9.4) (90.6) (29.1) (77.8) (22.2) (95.0) (5.0) (88.1) (11.9) (88.4) (11.6) (88.2) (21.8) (91.4) (8.6) (92.3) (7.7) (97.2) (2.8) (94.3) (5.7) (93.7) (6.3) (34.0) (66.0) (94.7) (5.3)
2348 195 1739 804 2223 320 1736 807 1987 556 2337 206 2186 357 2217 326 1954 589 2257 286 2360 183 2445 98 2428 115 2382 161 606 1936 2367 141
(93.1) (6.9) (70.3) (29.7) (89.0) (11.0) (70.0) (30.0) (79.0) (21.0) (92.8) (7.2) (87.5) (12.5) (89.0) (11.0) (88.6) (21.4) (90.0) (10.0) (93.9) (6.1) (96.7) (3.3) (95.9) (4.1) (94.5) (5.5) (24.0) (76.0) (95.2) (4.8)
2189 118 1671 636 2124 183 1541 766 2080 227 2125 182 2059 248 2159 148 1976 331 2151 156 2229 78 2258 49 2249 58 2203 104 390 1916 2144 137
(95.3) (4.7) (73.9) (26.1) (92.7) (7.3) (68.5) (31.5) (90.0) (10.0) (92.6) (7.4) (90.1) (9.9) (94.3) (5.7) (86.3) (13.7) (94.0) (6.0) (96.6) (3.4) (97.9) (2.1) (97.6) (2.4) (95.9) (4.1) (17.1) (82.9) (94.8) (5.2)
908 42 676 274 860 90 642 308 904 46 870 80 856 94 911 39 844 106 911 39 938 12 944 6 933 17 921 29 95 848 816 99
(95.1) (4.9) (70.7) (29.3) (90.3) (9.7) (67.7) (32.3) (94.1) (5.9) (91.3) (8.7) (89.9) (10.1) (95.6) (4.4) (88.6) (11.4) (95.7) (4.3) (98.9) (1.1) (92.1) (7.9) (98.4) (1.6) (96.9) (3.1) (10.5) (89.5) (90.1) (9.9)
n gives unweighted numbers, while percentages are weighted.
adulthood. Reported physical abuse in childhood was defined as “a severe beating by parent or carer before the age of 16 years”. Reported sexual abuse in childhood combined responses to two items and was defined in the analysis as “sexual abuse ranging from inappropriate touching to non-consensual sexual intercourse before the age of 16 years”. Intercourse before 16 reported as consensual (which is legally defined as child sexual abuse) was not included in the definition of abuse for this survey.
reported thinking about taking their own life, based on the question “Have you ever thought of taking your life, even though you would not actually do it?” A positive response to suicidal thoughts was followed up by asking whether this last occurred in the past week, past year or longer ago. Thoughts in the previous year were used in the analysis. Participants were also asked whether they had attempted suicide in the past year. 2.3. Analysis
2.2.5. Dependent variable: ‘suicidal thoughts in the previous year’ The term ‘suicidal thoughts’ used in the analysis was given a narrow definition – it included only those respondents who
We weighted data by the strata, sampling units and sampling weights. The survey data were weighted to take account of survey design and non-response, so that the results were representative
C. Cooper et al. / Journal of Affective Disorders 177 (2015) 42–48
of the household population aged 16 years and over. Weighting was necessarily complex, and more details are available in the main report (McManus et al., 2009). We used Stata (Statacorp, 2013) and the SVY (survey data) commands with these weights for both descriptive and inferential statistics. We used methods accounting for the survey design to approximate counts and standard errors. We performed survey weighted multivariable logistic regressions to investigate the association between suicidal thoughts and age as a categorical variables (16–34 years, 35–54 years, 55–74 years and 75 þ), adjusting for other factors. The factors considered in the multivariable analyses were: demographic factors (gender, qualifications, being a home owner, and living with a partner), disability (reporting disability and needing help with ADLs), child abuse (reporting child sexual or physical abuse) and mental health (CIS-R subscales). We used the ‘medeff’ module in Stata (Hicks and Tingley, 2011, 2012) to assess the extent to which each of these factors mediates the association between age and suicidal thoughts. The ‘medeff’ command estimates mediation effects; it provides the percentage of the indirect effect accounted for by the factors studied. Finally, for variables where there was significant variation in the total effect mediated across age groups, we used the userwritten Stata command ‘ldecomp’ (Buis, 2014) to estimate factual (e.g. the actual proportion of 16–34 year olds reporting suicidal ideation) and counterfactual proportions (e.g. the proportion of 16–34 year olds who would report suicidal ideation if they were exposed to the same risk factors, operating in the same way as people aged 75 þ). This method enabled us to estimate the direct effect of the variables studied on suicidal ideation and the indirect effect (amount of these effects explained by age), and total effects such that the direct and indirect effects add up to the total effect.
45
3. Results 3.1. Suicidal thoughts and attempts in the year prior to interview. The sample ranged in age from 16 to 95 years. Thoughts about suicide in the previous year (recent suicidal thoughts) were reported by 339 (4.3%) of the sample. Recent suicidal thoughts were significantly more common in younger age groups. They were reported by 88 (5.1%) participants aged 16–34, 165 (5.4%) of participants aged 35–54, 66 (2.4%) of participants aged 55–74 and 20 (2.2%) of those aged 75 and above (Table 1). Only 52 people reported suicide attempts in the last year, 13 of those aged 16–34, 33 of those aged 35–54, six of those aged 55–74 and none in the oldest age group. All but one of the people reporting a suicide attempt also reported suicidal thoughts. 3.2. Effect of socio-demographic factors Suicidal thoughts in the past year were more likely to be reported by female respondents, and those who drank more alcohol, but less likely by those who owned their own homes, lived with a partner or had more than three close personal relationships. Respondents aged 16–34 were less likely to be living with a partner or living in a home they owned than older respondents. The protective effect of these two demographic characteristics was significantly modulated by age, being relatively greater in people aged 16–34 (as demonstrated by the higher % total mediated effect in these age categories for these characteristics, shown in Table 2). Table 3 sets out the factual and counterfactual proportions for these relationships. It shows that the true proportion of people aged 16–34 who report suicidal thoughts was
Table 2 Association between patients characteristics and suicidal thoughts and percentage of total effect mediated by age group. Characteristic (reference group)
Unadjusted oddsa
% of Total effect mediatedb 16–34 (n¼1601)
35–54 (n¼2538)
55–74(n¼2304)
75þ (n¼ 946)
Demographic characteristics Female gender Living with partner Home owner Hazardous alcohol use More than 3 close family or friends
1.6 (1.2–2.1) 0.4 (0.3–0.5) 0.5 (0.4–0.6) 1.8 (1.3–2.3) 0.2 (0.2–0.3)
1 90 70 27 0.4
0 53 21 2 4
1 19 16 9 0.1
7 25 4 13 18
Child abuse Reported child sexual abuse Reported child physical abuse
3.8 (2.8–5.1) 5.0 (3.5–7.0)
9 8
18 9
4 2
17 9
Physical health Reported moderate disability Needs help with any ADLs
2.2 (1.7–2.9) 4.0 (3.1–5.1)
75 120n
16 40
16 26
54 86
Mental health Somatic symptoms Fatigue Concentration/forgetfulness Sleep Irritability Worries about physical health Depression Depressive ideas Worry Anxiety Phobias Panic Compulsions Obsessions
6.0 (4.4–8.2) 8.3 (6.2–11.2) 10.4 (7.9–13.6) 4.7 (3.6–6.1) 7.2 (5.6–9.4) 4.9 (3.6–6.6) 13.8 (10.5–18.1) 22.3 (17.0–29.2) 9.1 (7.0–11.9) 11.7 (8.9–15.3) 8.3 (6.2–11.2) 17.9 (12.6–25.5) 7.3 (5.2–10.4) 8.9 (6.7–11.8)
11 9 1 4 69 22 9 71 48 18 49 8 34 27
13 20 27 1 40 8 20 44 34 39 13 20 2 6
8 5 17 7 39 6 13 45 29 19 17 8 12 9
5 9 3 8 55 1 10 55 37 30 31 25 17 15
a b
Estimates obtained from a logistic regression without additional covariates. Estimates obtained using ‘Medeff’ Stata command; total effect mediated is a theoretical construct and is occasionally more than 100%.
46
C. Cooper et al. / Journal of Affective Disorders 177 (2015) 42–48
Table 3 Predicted and counterfactual proportionsa shown as percentages for significant predictors of suicidal thoughts. Characteristic
How 16–34 age group suicidal thoughts might change (%)
Sociodemographic …if same % Living 16–34 with partner as: 35–54 55–74 75þ Home ownerb 16–34 35–54 55–74 75þ Physical health Moderate disability
16–34 35–54 55–74 75þ Needing help with 16–34 ADLs 35–54 55–74 75þ
Mental health Irritability
Phobias
Compulsions
16–34 35–54 55–74 75þ 16–34 35–54 55–74 75þ 16–34 35–54 55–74 75þ
How 35–54 age group suicidal thoughts might change (%)
How 55–74 age group suicidal thoughts might change (%)
How 75þ age group suicidal thoughts might change (%)
5.2 3.8 3.8 5.0 5.0 4.2 4.1 4.3
7.4 5.5 5.5 7.1 6.3 5.4 5.2 5.5
3.3 2.4 2.4 3.2 3.0 2.5 2.5 2.6
2.2 1.6 1.7 2.2 2.5 2.1 2.0 2.1
5.2 5.9 7.0 9.3 5.2 5.9 7.6 10.8
4.8 5.5 6.5 8.6 4.8 5.5 7.1 10.0
1.8 2.0 2.4 3.3 1.6 1.9 2.4 3.5
1.2 1.3 1.6 2.2 1.0 1.2 1.5 2.2
5.2 5.0 3.7 3.2 5.2 4.9 4.3 3.9 5.2 4.9 4.6 4.4
5.6 5.5 4.0 3.5 5.8 5.5 4.9 4.4 5.7 5.5 5.1 5.0
3.5 3.3 2.4 2.1 3.0 2.8 2.4 2.2 2.8 2.6 2.4 2.4
3.6 3.5 2.5 2.2 3.0 2.8 2.4 2.2 2.6 2.4 2.3 2.2
a Percentage by age group reporting suicidal thoughts, and (in grey cells) the % predicted to report suicidal thoughts if the distribution of each of the characteristics studied changed to that found in a different age group. b True proportions vary due to slightly different population for which this information available.
5.2%, but if their likelihood of living with a partner equalled that of people aged 35–54 or 55–74, only 3.8% would report these thoughts, because living with a partner is protective. Similarly, the proportion of respondents aged 16–34 who reported suicidal ideation would be around one percentage point lower (4% vs 5%) if their rate of home ownership was as high as that of people in the older age groups. 3.3. Effects of child abuse Childhood sexual and physical abuse were both strong predictors of recent suicidal thoughts. Childhood sexual abuse was associated with a fourfold, and physical abuse a five fold increase in risk of experiencing suicidal ideation in unadjusted analyses (Table 2). Abuse was less frequently reported by people aged 75 þ, but the relationship between reporting these events and having recent suicidal thoughts varied very little with age. 3.4. Effects of physical health characteristics As expected, the proportion of people reporting moderate disability or needing help with ADLs rose sharply with age. These characteristics were associated with reporting more suicidal thoughts, and this relationship was stronger in the youngest (16–34) and oldest (aged 75 þ) age groups. In Table 3, the factual and counterfactual proportions demonstrate that the proportion of people aged 75 þ reporting suicidal ideation would theoretically reduce from 2.2% to 1.0% if they had the same level of need for help
with ADLs as people aged 16–34. Similarly, the proportion of the youngest people reporting suicidal ideation would theoretically be 10.8% rather than 5.2% if they had the same level of need for ADL help as people aged 75 þ. 3.5. Effects of mental health symptoms Depressive ideas and symptoms of worry were strongly associated with reporting suicidal thoughts across all age groups, as expected. These symptoms were more frequently reported by younger respondents (aged under 55 years). The strength of the relationship between these symptoms and suicidal ideation was comparable across the age groups. Symptoms of irritability, phobias and compulsions also declined with age. As well as occurring more frequently, these symptoms were all more strongly associated with reporting suicidal ideas in the youngest (16–34) than in older groups. In Table 3, the counterfactual proportions demonstrate that if people aged 16–34 had levels of irritability, phobias and compulsions comparable to those aged 75þ, then instead of 5.2% their rates of reported suicidal ideation would be 3.2%, 3.9% and 4.4% respectively.
4. Discussion Psychiatric morbidity, childhood sexual and physical abuse, physical illness and disability, hazardous alcohol use, small social networks and not living with a partner or owning one's home were
C. Cooper et al. / Journal of Affective Disorders 177 (2015) 42–48
all associated with recent suicidal thoughts across all age groups. The primary aim of our study was to clarify why suicidal thoughts decline in frequency across the life span. Our findings suggest that the lower frequency of reported depression and anxiety symptoms in people aged 55 and over, and of child abuse reported by people aged 75 and over might explain the lower rates of suicidal ideas reported by older people, given that these characteristics are strongly related to suicidal ideation across all ages in our sample. People in the youngest age group also had the highest rates of irritability, compulsions and phobias, and these symptoms were particularly strongly related to suicidal ideation in this group. Younger people (aged 16–34) also had the lowest rates of home ownership and co-habitation, both of which protected against suicidal ideation; among those younger people who did own their home or co-habit, these characteristics were relatively more protective against suicidal thoughts than in older age groups. We speculate that people in this age group who were single or renting their accommodation may have been more likely to view these situations as undesirable (and be trying to change them), whereas in older age groups these circumstances may have been life choices for some, and, where they were not, at least more accepted. This may also be a cohort effect, as home ownership levels rose steadily from 23% in 1918 to 69% in 2001, thus becoming an expected achievement for many (Office of National Statistics, 2013a). The recent decline in home ownership in Britain following the financial crisis of 2008 post-dated this survey, and might conduce to a subsequent increase in suicidal ideation. We were particularly interested in why people in older age groups report low levels of suicidal ideation. The high rates of disability in people aged 75þ are striking. Our findings that psychiatric morbidity and ADL impairments were associated with suicidal thoughts across the age groups accord with previous studies (Dennis et al., 2007; Dennis et al., 2009). Disability was an important predictor of suicidal thoughts in the youngest and the oldest age groups, so the lower rate of suicidal thoughts in people aged 75þ cannot be explained as a greater tolerance of disability and impairment. This oldest age group were also less likely than those in middle age to be living with a partner, an important protector against suicidal ideation. It has been suggested that rates of mental illness symptoms are lower in older people because of under-reporting, and our study cannot rule this out. However, if this were so, older people should be more likely to have suicidal ideation at a given level of mental symptoms than their younger counterparts, but we did not find this. The association between childhood sexual abuse and suicidal ideation has been highlighted in previous publications from this survey (Jonas et al., 2011). We found in the current paper that the association between child physical abuse and suicidal ideation was even greater than that between child sexual abuse and suicidal ideation, although the definition used for physical abuse, a severe beating, may have only detected more severe physical abuse. This is the first study to include childhood abuse in a comparison of the correlates of suicidal thinking across age groups: it appears to indicate that the negative effects of abuse on suicidality are lifelong. Many people who are abused as children end up in difficult relationships as adults (Jonas et al., 2011) and they may also have distorted ideas about their own self-worth and the way others feel about them, although we did not measure adult relationship quality or self-worth in the current study. The significance of strong and enduring relationships and their association with good mental health and well-being is well recognised in both genders (Gibb et al., 2011). Such an association must be the result of complex interactions, with choice of partner, the ability to work within a relationship, and the buffering and support provided by the relationship during times of stress all being important. The breakdown of a relationship has enormous effects on self-esteem and mental health, and is a well-recognised reason for people to state that they no longer have reasons to live.
47
4.1. Limitations The British National Psychiatric Morbidity Programme involves repeated representative household surveys which have had decreasing response rates over time; the current response rate of 57% is fairly low, although comparable with other recent national surveys (Mindell et al., 2012) (and 70% of people successfully contacted did then agree to take part). We weighted the data to correct for nonresponse on a range of sociodemographic and area characteristics. Non-response weighting had little effect on the results, showing that non-responders seem to be similar to responders (Jonas et al., 2011). Moreover, the current study explored associations, so findings are relatively less sensitive to response bias than studies of prevalence. Just 20 people aged 75þ in the sample reported suicidal thoughts, and this is a significant limitation. Great efforts were made in the survey to maintain the quality of information by using confidential data entry. The reported prevalence of childhood abuse decreased in older age. Some of this may be explained by shorter life expectancy in those experiencing adverse childhood events, by age-related increases in memory problems or by “positivity” of memories (Schlagman et al., 2006). The greater acceptance in recent years that childhood abuse, especially sexual abuse is common (Smart, 2000) may have had an effect, as abusive experiences, especially those at the less severe end of the spectrum may be more likely to be reported as abuse if they have been named as such at the time. However, even though reasons for lower levels of reported childhood abuse with increasing age remain obscure, the relationship between childhood abuse and lack of well-being remains strong in the older adult group. Around 15% (n¼52) of those reporting suicidal ideation also reported a suicide attempt in the last year. We did not distinguish those who reported acting on their thoughts from those reporting suicidal thought without associated actions. We could not analyse actual suicidal acts as the dependent variable because of the small number of participants reporting these actions, although we have previously reported that suicidal thinking represents a strong indicator of vulnerability to suicidal acts from the National Psychiatric Morbidity Surveys (Bebbington et al., 2010). Nonetheless as most people who have suicidal thoughts do not attempt suicide, knowledge about risk factors that can predict whether people with suicidal ideation progress to actual self harm are of critical importance for suicide prevention strategies, and other recent studies have completed such analyses (Glenn and Nock, 2014). 4.2. Implications Our results provide further support for vigorous opportunistic screening for suicidal ideas in both young and old, as poor mental and physical health, disability, and social isolation are associated with suicidal ideation across all age groups. The detrimental effect of childhood abuse appears to be maintained in older adults, probably operating through effects on selfesteem, and seems to increase the risk of thoughts of suicide. This indicates that clinicians should continue to ask about childhood experiences, as older people may benefit from appropriate psychological work for this, and for other relationship difficulties. Where there is insufficient family or peer support, it should be recognised that, in contradiction to current trends in mental and primary health care, there may be a need for longer term support from professional services. The alert listening ear of a regularly seen and trusted health worker may be crucial for older people. Role of the funding source No external funding was received for this work.
48
C. Cooper et al. / Journal of Affective Disorders 177 (2015) 42–48
Conflict of interest None of the authors has any financial involvement or affiliation with any organisation whose financial interests may be affected by material in this manuscript, or which potentially bias it.
Acknowledgements None.
References Alonso, J., Angermeyer, M.C., Bernert, S., Bruffaerts, R., Brugha, I.S., Bryson, H., de Girolamo, G., De Graaf, R., Demyttenaere, K., Gasquet, I., Haro, J.M., Katz, S.J., Kessler, R.C., Kovess, V., Lepine, J.R., Ormel, J., Polidori, G., Russo, L.J., Vilagut, G., Almansa, J., Arbabzadeh-Bouchez, S., Autonell, J., Bernal, M., Buist-Bouwman, M.A., Codony, M., Domingo-Salvany, A., Ferrer, M., Joo, S.S., Martinez-Alonso, M., Matschinger, H., Mazzi, F., Morgan, Z., Morosini, R., Palacin, C., Romera, B., Taub, N., Vollebergh, W.A.M., 2004. Prevalence of mental disorders in Europe: results from the european study of the epidemiology of mental disorders (ESEMeD) project. Acta Psychiatr. Scand. 109, 21–27. Australian Institute of Family Studies, 2010. Ageing, dementia and adult survivors of childhood sexual assault, Report. Bebbington, P., Brugha, T., Meltzer, H., Farrell, M., Ceresa, C., Jenkins, R., Lewis, G., 2000. Psychiatric disorder and dysfunction in the UK national survey of psychiatric morbidity. Soc. Psychiatry Psychiatr. Epidemiol. 35 (5), 191–197. Bebbington, P.E., Cooper, C., Minot, S., Brugha, T.S., Jenkins, R., Meltzer, H., Dennis, M., 2009. Suicide attempts, gender, and sexual abuse: data from the 2000 British psychiatric morbidity survey. Am. J. Psychiatry 166 (10), 1135–1140. Bebbington, P.E., Minot, S., Cooper, C., Dennis, M., Meltzer, H., Jenkins, R., Brugha, T., 2010. Suicidal ideation, self-harm and attempted suicide: results from the British psychiatric morbidity survey 2000. Eur. Psychiatry 25 (7), 427–431. Berntsen, D., Rubin, D.C., 2002. Emotionally charged autobiographical memories across the life span: the recall of happy, sad, traumatic, and involuntary memories. Psychol. Aging 17 (4), 636–652. Brugha, T.S., Morgan, Z., Bebbington, P., Jenkins, R., Lewis, G., Farrell, M., Meltzer, H., 2003. Social support networks and type of neurotic symptom among adults in British households. Psychol. Med. 33 (2), 307–318. Buis, M., 2014. LDECOMP: Stata module decomposing the total effects in a logistic regression into direct and indirect effects. Stat. Softw. Compon.. Coffey, P., Leitenberg, H., Henning, K., Turner, T., Bennett, R.T., 1996. Mediators of the long-term impact of child sexual abuse: perceived stigma, betrayal, powerlessness, and self-blame. Child Abuse Negl. 20 (5), 447–455. Conwell, Y., Duberstein, P.R., Cox, C., Herrmann, J., Forbes, N., Caine, E.D., 1998. Age differences in behaviors leading to completed suicide. Am. J. Geriatr. Psychiatry 6 (2), 122–126. Dennis, M., Baillon, S., Brugha, T., Lindesay, J., Stewart, R., Meltzer, H., 2009. The influence of limitation in activity of daily living and physical health on suicidal ideation: results from a population survey of Great Britain. Soc. Psychiatry Psychiatr. Epidemiol. 44 (8), 608–613. Dennis, M., Baillon, S., Brugha, T., Lindesay, J., Stewarvand, R., Meltzer, H., 2007. The spectrum of suicidal ideation in Great Britain: comparisons across a 16–74 years age range. Psychol. Med. 37 (6), 795–805. Ferri, C.P., Prince, M., Brayne, C., Brodaty, H., Fratiglioni, L., Ganguli, M., Hall, K., Hasegawa, K., Hendrie, H., Huang, Y.Q., Jorm, A., Mathers, C., Menezes, P.R., Rimmer, E., Scazufca, M., 2005. Global prevalence of dementia: a Delphi consensus study. Lancet 366 (9503), 2112–2117. Gibb, S.J., Fergusson, D.M., Norwood, L.J., 2011. Relationship duration and mental health outcomes: findings from a 30-year longitudinal study. Br. J. Psychiatry 198 (1), 24–30. Glenn, C.R., Nock, M.K., 2014. Improving the short-term prediction of suicidal behavior. Am. J. Prev. Med. 47 (3S2), 176–180.
Gunnell, D., Lewis, G., 2005. Studying suicide from the life course perspective: implications for prevention. Br. J. Psychiatry 187, 206–208. Handley, T.E., Hiles, S.A., Inder, K.J., Kay-Lambkin, F.J., Kelly, B.J., Lewin, T.J., McEvoy, M., Peel, R., Attia, J.R., 2014. Predictors of suicidal ideation in older people: a decision tree analysis. Am. J. Geriatr. Psychiatry 22 (11), 1325–1335. Hawton, K., Harriss, L., 2008. How often does deliberate self-harm occur relative to each suicide? A study of variations by gender and age. Suicide Life-Threat. Behav. 38 (6), 650–660. Hawton, K., van Heeringen, K., 2009. Suicide. Lancet 373 (9672), 1372–1381. Hicks, R., Tingley, D., 2011. Causal mediation analysis. Stata J. 11, 1–15. Hicks, R., Tingley, D., 2012. Mediation: Stata Module for Causal Mediation Analysis and Sensitivity Analysis, Computer Program. Holland, A.C., Kensinger, E.A., 2010. Emotion and autobiographical memory. Phys. Life Rev. 7 (1), 88–131. Jenkins, R., Meltzer, H., Bebbington, P., Brugha, T., Farrell, M., McManus, S., Singleton, N., 2009. The British mental health survey programme: achievements and latest findings. Soc. Psychiatry Psychiatr. Epidemiol. 44 (11), 899–904. Jonas, S., Bebbington, P., McManus, S., Meltzer, H., Jenkins, R., Kuipers, E., Cooper, C., King, M., Brugha, T., 2011. Sexual abuse and psychiatric disorder in England: results from the 2007 adult psychiatric morbidity survey. Psychol. Med. 41 (4), 709–719. Kish, L., 1965. Survey Sampling. Wiley and Sons, London. Lewis, G., Pelosi, A.J., Araya, R., Dunn, G., 1992. Measuring psychiatric disorder in the community: a standardized assessment for use by lay interviewers. Psychol. Med. 22 (2), 465–486. Martindale, B., 2007. Resilience and vulnerability in later life. Br. J. Psychother. 23, 205–216. McBride, O., Bebbington, P., Cooper, C., 2013. Could the lower prevalence of affective disorder in older people be due to measurement error? Reliability of the revised clinical interview schedule in younger and older adults. J. Affect. Disord. 148 (2–3), 310–315. McManus, S., Meltzer, H., Brugha, T.S., Bebbington, P.E., Jenkins, R., 2009. Adult Psychiatric Morbidity in England, 2007: Results of a Household Survey, HSCIC, Report. Mindell, J., Biddulph, J.P., Hirani, V., Stamatakis, E., Craig, R., Nunn, S., Shelton, N., 2012. Cohort profile: the health survey for England. Int. J. Epidemiol. 41 (6), 1585–1593. Nock, M.K., Borges, G., Bromet, E.J., Alonso, J., Angermeyer, M., Beautrais, A., Bruffaerts, R., Chiu, W.T., de Girolamo, G., Gluzman, S., De Graaf, R., Gureje, O., Haro, J.M., Huang, Y., Karam, E., Kessler, R.C., Lepine, J.P., Levinson, D., Medina-Mora, M.E., Ono, Y., Posada-Villa, J., Williams, D., 2008. Cross-national prevalence and risk factors for suicidal ideation, plans and attempts. Br. J. Psychiatry 192 (2), 98–105. Office of National Statistics, 2013a. A Century Home Ownersh, Online Source. Office of National Statistics, 2013b. Suicides United Kingdom, 2011, Report. Russell, D., Turner, R.J., Joiner, T.E., 2009. Physical disability and suicidal ideation: a community-based study of risk/protective factors for suicidal thoughts. Suicide Life-Threat. Behav. 39 (4), 440–451. Salib, E., Tadros, G., Cawley, S., 2001. Elderly suicide and attempted suicide: one syndrome. Med. Sci. Law 41 (3), 250–255. Saunders, J.B., Aasland, O.G., Babor, T.F., Delafuente, J.R., Grant, M., 1993. Development of the alcohol-use disorders identification test (audit) – who collaborative project on early detection of persons with harmful alcohol-consumption 2. Addiction 88 (6), 791–804. Schlagman, S., Schulz, J., Kvavilashvili, L., 2006. A content analysis of involuntary autobiographical memories: examining the positivity effect in old age. Memory 14 (2), 161–175. Smart, C., 2000. Reconsidering the recent history of child sexual abuse, 1910–1960. J. Soc. Policy 29, 55–71. Statacorp, 2013. Stata Statistical Software: Release 13. College Station, TX, Statacorp, Computer Program. Ware, J.E., Kosinski, M., Keller, S.D., 1996. A 12-item short-form health survey – construction of scales and preliminary tests of reliability and validity. Med. Care 34 (3), 220–233.