PsychologicalMedicine

http://journals.cambridge.org/PSM
AdditionalservicesforPsychologicalMedicine:
Emailalerts:Clickhere
Subscriptions:Clickhere
Commercialreprints:Clickhere
Termsofuse:Clickhere

Spiritualandreligiousbeliefsasriskfactorsfortheonsetofmajor
depression:aninternationalcohortstudy
B.Leurent,I.Nazareth,J.BellnSaameo,M.I.Geerlings,H.Maaroos,S.Saldivia,I.vab,F.TorresGonzlez,M.Xavier
andM.King
PsychologicalMedicine/FirstViewArticle/January2013,pp112
DOI:10.1017/S0033291712003066,Publishedonline:

Linktothisarticle:http://journals.cambridge.org/abstract_S0033291712003066
Howtocitethisarticle:
B.Leurent,I.Nazareth,J.BellnSaameo,M.I.Geerlings,H.Maaroos,S.Saldivia,I.vab,F.TorresGonzlez,M.Xavier
andM.KingSpiritualandreligiousbeliefsasriskfactorsfortheonsetofmajordepression:aninternationalcohortstudy.
PsychologicalMedicine,AvailableonCJOdoi:10.1017/S0033291712003066
RequestPermissions:Clickhere

Downloadedfromhttp://journals.cambridge.org/PSM,IPaddress:128.40.240.4on30Jan2013

Psychological Medicine, Page 1 of 12.

doi:10.1017/S0033291712003066

f Cambridge University Press 2013

O R I G I N A L AR T I C LE

Spiritual and religious beliefs as risk factors for

the onset of major depression: an international
cohort study
B. Leurent1,2, I. Nazareth2, J. Bellon-Saameno3, M.-I. Geerlings4, H. Maaroos5, S. Saldivia6, I. Svab7,
F. Torres-Gonzalez8, M. Xavier9 and M. King1*
1

Mental Health Sciences Unit, Faculty of Brain Sciences, University College London Medical School, UK
Research Department of Primary Care and Population Health, University College London Medical School, UK
3
Department of Preventive Medicine, El Palo Health Centre, Malaga, Spain
4
University Medical Centre, Utrecht, The Netherlands
5
Faculty of Medicine, University of Tartu, Estonia
6
Departamento de Psiquiatra y Salud Mental, Universidad de Concepcion, Chile
7
Department of Family Medicine, University of Ljubljana, Slovenia
8
CIBERSAM-Granada University, Granada, Spain
9
Department of Mental Health, Faculdade Ciencias Medicas, CEDOC, Lisboa, Portugal
2

Background. Several studies have reported weak associations between religious or spiritual belief and psychological
health. However, most have been cross-sectional surveys in the USA, limiting inference about generalizability. An
international longitudinal study of incidence of major depression gave us the opportunity to investigate this
relationship further.
Method. Data were collected in a prospective cohort study of adult general practice attendees across seven countries.
Participants were followed at 6 and 12 months. Spiritual and religious beliefs were assessed using a standardized
questionnaire, and DSM-IV diagnosis of major depression was made using the Composite International Diagnostic
Interview (CIDI). Logistic regression was used to estimate incidence rates and odds ratios (ORs), after multiple
imputation of missing data.
Results. The analyses included 8318 attendees. Of participants reporting a spiritual understanding of life at
baseline, 10.5 % had an episode of depression in the following year compared to 10.3 % of religious participants
and 7.0 % of the secular group (p<0.001). However, the ndings varied signicantly across countries, with
the dierence being signicant only in the UK, where spiritual participants were nearly three times more likely
to experience an episode of depression than the secular group [OR 2.73, 95 % condence interval (CI) 1.594.68].
The strength of belief also had an eect, with participants with strong belief having twice the risk of
participants with weak belief. There was no evidence of religion acting as a buer to prevent depression after
a serious life event.
Conclusions. These results do not support the notion that religious and spiritual life views enhance psychological
well-being.
Received 13 July 2012 ; Revised 19 November 2012 ; Accepted 6 December 2012
Key words : General practice, longitudinal, major depression, religion, spirituality.

Background
Research ndings, most originating from the USA,
have generally reported a positive, albeit weak, relationship between higher levels of religious involvement and better health once other inuences, such as

* Address for correspondence : Professor M. King, Mental Health

Sciences Unit, University College London, Charles Bell House, 6773
Riding House Street, London W1W 7EJ, UK.
(Email : michael.king@ucl.ac.uk)

age, sex and social support, have been taken into account (Koenig et al. 1998, 2001 ; McCullough & Larson,
1999 ; Johnstone et al. 2008). Nevertheless, many other
studies do not nd such an association (Payne et al.
1991 ; Schaefer, 1997 ; Lewis et al. 2000 ; Gartner et al.
2012). There have been at least two meta-analyses of
relevant studies. The rst examined the relationship
between religiosity and psychological adjustment
(Hackney & Sanders, 2003), while excluding studies
measuring a wider concept of spirituality or those
examining the relationship with mental disorders.

B. Leurent et al.

A small positive correlation between religiosity and

psychological status was found (r=0.10, p<0.0001)
when combining all eect sizes from 35 cross-sectional
studies. Seventy-eight negative relationships were
found in the data set of 264 eect sizes. Greater eects
were seen with institutional religiosity than personal
devotion, and with psychological distress rather
than of life satisfaction. The authors concluded that
greater internality of religious belief was associated
with more positive psychological outcomes. A second
meta-analysis of 147 studies (of which 15 were longitudinal) examined associations between religiosity
and/or spirituality and depressive symptoms and/or
depressive disorder (Smith et al. 2003). Religious/
spiritual belief seemed to have a small negative
(x0.096) correlation with depressive symptoms but
the protective eect of belief was stronger on risk
of major depression after signicant life events.
Unfortunately, the authors did not distinguish evidence from longitudinal versus cross-sectional studies
in their analysis.
Numerous cross-sectional studies have been conducted since 2003, including some outside the USA. In
one very large Canadian community health study,
greater participation in worship was associated with
lower odds of psychiatric disorders but people who
placed greater importance on spiritual values had
higher odds of most psychiatric disorders (Baetz et al.
2006). A further cross-sectional study of more than
6000 people in Korea also reported that strong spiritual values were associated with increased rates of
current depressive disorder (Park et al. 2012).
Most of the studies in the meta-analyses described
above (Hackney & Sanders, 2003 ; Smith et al. 2003)
were cross-sectional in nature and thus we need more
prospective research in a variety of cultures and societies. Recent prospective studies have been small
scale but generally positive in their ndings. Kasen
et al. (2012) and Miller et al. (2012) reported data on 114
adults who were the grown-up children of a cohort of
people at high risk of depression, and matched controls, enrolled in a multi-generational 10-year longitudinal study. Participants for whom religion or
spirituality was highly important seemed to be protected from major depression, especially relapse in
those with a history of depression (Miller et al. 2012).
Religion was also more protective in participants
exposed to negative life events (Kasen et al. 2012).
However, there were few participants, the sample
was very specic (catholic or protestant Caucasians,
who were ospring of depressed parents for the exposed cohort) and ndings were of borderline signicance.
In earlier cross-sectional work in a large sample of
people from a range of ethnic groups in England and

Wales, we reported that holding a spiritual life view

without religious aliation was associated with a
higher prevalence of anxiety and depression (King
et al. 2006a). However, the direction of this association
was unclear. Lack of religion may lead to common
mental disorders in vulnerable people who seek
meaning in their lives. Conversely, people developing
a common mental disorder who are not aliated
to any religious group may become involved in a
search for meaning for relief from symptoms. The
New Age movements and other non-traditional faiths
in Western Europe may reect a search for meaning in
societies such as the UK, where religious practice has
declined steeply in recent decades. We previously
undertook a prospective study to develop a risk prediction algorithm for the onset of major depression in
general practice attendees in seven countries : six
European and one Latin American (King et al. 2008).
These data provided an opportunity to examine the
impact of spiritual and religious beliefs on the development of depression. Our aims were to : (1) examine
the impact of a religious or spiritual life view on
onset of major depression over 12 months ; (2) assess
whether this impact varied for rst or recurrent
episodes of depression ; (3) examine how the impact
varied by religious denomination, change in life
view and change in strength of belief over the
12 months ; and (4) determine whether the form of life
view mediated the impact of signicant life events on
onset of major depression. Our principal hypothesis
was that people expressing a spiritual life view in the
absence of religious aliation or practice are more
likely to develop DSM-IV major depression than those
who have a religious life view or are secular in outlook.

Method
Study setting and design
The study, described in detail elsewhere (King et al.
2006b, 2008), was approved by research ethics committees in each country. It was a prospective cohort
study conducted in (1) 25 general practices in the
Medical Research Council (MRC) General Practice
Research Framework (GPRF) in the UK ; (2) nine large
primary care centres in Andaluca, Spain ; (3) 74 general practices nationwide in Slovenia ; (4) 23 general
practices nationwide in Estonia ; (5) seven large general practice centres near Utrecht, The Netherlands ;
(6) two large primary care centres in the Lisbon area of
Portugal ; and (7) 78 general practices in Concepcion
and Talcahuano in the Eighth Region of Chile. General
practices covered urban and rural populations with
considerable socio-economic variation.

Spiritual and religious beliefs and psychological well-being

Study participants
Consecutive attendees aged 1875 years were recruited in Europe between April 2003 and September
2004 and in Chile between October 2003 and February
2005. Exclusion criteria were inability to understand
the countrys main language, psychosis, dementia and
incapacitating physical illness. Recruitment varied
slightly because of local service dierences. In the UK
and The Netherlands, researchers spoke to patients
while they waited in the practices. In the remaining
European countries, general practitioners (GPs) introduced the study before contact with researchers. In
Chile, attendees were stratied on age and gender on
the basis of local gures and participants were selected
randomly within each stratum. Participants gave informed consent and undertook a research evaluation
within 2 weeks. Only attenders without a DSM-IV diagnosis of major depressive disorder at baseline were
part of this analysis.
Assessments at baseline
Each instrument or question not available in the relevant languages was translated from English and
back-translated by professional translators (King et al.
2006b).
Demography
We collected standard information on participants
sex, age, education, marital status, employment and
ethnicity.
Religious and spiritual beliefs
The self-report version of the Royal Free Interview
for Spiritual and Religious Beliefs (King et al. 2001)
examines religious aliation and practice, and spiritual beliefs whether or not in the context of religion. In
this study only the rst three items of the questionnaire were used. Before the questions are posed the
respondent reads an introductory statement : In using
the word religion, we mean the actual practice of a
faith, e.g. going to a temple, mosque, church or synagogue. Some people do not follow a religion but do
have spiritual beliefs or experiences. For example, they
believe that there is some power or force other than
themselves, which might inuence their life. Some
people think of this as God or gods, others do not.
Some people make sense of their lives without any
religious or spiritual belief . On the basis of this statement, respondents were asked to indicate whether
their understanding of life was primarily religious,
spiritual, or neither religious nor spiritual (this latter
category will be referred to as secular ). If religious or

spiritual they were then asked to indicate whether

they regarded themselves as having a specic religion.
People with a spiritual life view may identify themselves with a religion, even if they do not practice it.
Finally, religious and spiritual participants were asked
to indicate on scale from 1 to 6 how strongly they held
their life view.
Diagnosis of depression
A DSM-IV diagnosis of major depression in the preceding 6 months was made using the Depression
Section of the Composite International Diagnostic
Interview (CIDI ; Robins et al. 1988 ; WHO, 1997).
Screen for lifetime history of depression
Lifetime depression was considered possible if the respondent answered armatively to both of the rst
two questions of the CIDI Depression Section (WHO,
1997).
Life events
The List of Threatening Experiences questionnaire
(Brugha et al. 1985) enquired about major life events in
the preceding 6 months.
Social support
Adequacy of support from family and friends was
measured using brief standardized questions (Blaxter,
1990).
Follow-up assessments at 6 and 12 months
All participants were re-evaluated for DSM-IV major
depression after 6 and 12 months. At 6 months they
also completed the Royal Free Interview for Spiritual
and Religious Beliefs and the questions on life events.
Statistical analysis
To manage missing data we undertook multiple imputation by chained equation, using the ice procedure
in Stata (Royston, 2005). We generated two imputed
databases each containing 30 imputed versions using
all relevant variables predicting depression or missingness ; one database was used for the analysis of the
understanding of life and the other for the analysis of
religious denomination and strength of belief in religious or spiritual participants. Regression results
were combined using Rubins rule (mim command
in Stata). We explored the pattern of missing data
and performed sensitivity analyses on complete cases.
t tests and x2 tests were used to compare groups at
baseline. Incidence rates of major depression were

B. Leurent et al.

obtained on imputed data by univariable logistic regression and presented graphically. Odds ratios (ORs)
unadjusted and adjusted for age, sex, education, employment, social support, past history of depression
and country were computed using logistic regression.
Interactions between the predictor and each covariate
were tested and the model was stratied if an interaction was found. In view of multiple testing, we
considered an interaction signicant at the level of
p<0.01. All other p values reported are two-sided, and
considered signicant at the level of p<0.05. Incidence
rates and ORs are reported along with their 95 % condence intervals (CIs). Standard errors are based on
robust sandwich estimates to account for the clustering eect of each general practice (Huber, 1967).
For the analyses on imputed data, the exact number
of participants in each category of the exposure
variables cannot be reported as it varied slightly
with each imputation. In the presence of missing
data, CIs are a more reliable guide to the precision
around estimates than estimated frequencies. Statistical analyses were performed using Stata Release 11
(Stata Corp, 2009).

Results
Response rates and prevalence of DSM-IV major
depression at recruitment
Response to recruitment was high in Portugal (76 %),
Estonia (80 %), Slovenia (80 %) and Chile (97 %) but
lower in the UK (44 %) and The Netherlands (45 %).
Ethical constraints did not allow collection of data on
non-responders at baseline. Across all countries the
response to follow-up at 6 months was at 89.5 %. A
total of 10045 people took part : 219 were excluded on
grounds of age, 143 had a missing CIDI diagnosis
and ve a missing practice identication at baseline.
Of the remaining 9678, 8318 without DSM-IV major
depression (86 %) at baseline were analysed. The median age was 49 years, two-thirds were women,
and 75 % of participants held a religious or spiritual
understanding of life. Women were more likely
than men to have a religious or spiritual understanding of life, and a past history of depression was least
common in people with a spiritual understanding
(Table 1). The characteristics of participants by country are reported in Supplementary Table S1 (available
online). Eleven per cent of participants did not complete the 6-month follow-up and 16 % did not complete the 12-month follow-up. Non-completers tended
to be younger and less educated and had experienced
more serious life events in the 6 months before baseline.

Understanding of life and onset of DSM-IV major

depression over 12 months
Of participants reporting a religious understanding of
life, 10.3 % experienced an episode of major depression
over the subsequent 12 months, compared to 10.5 % of
participants with a spiritual life view and 7.0 % of the
secular group (p<0.001). This nding was examined
more closely in a logistic regression with secular
participants as the reference group. The results are
reported unadjusted and adjusted for age, sex, education, employment, social support, past history of
depression and country. Participants with a spiritual
understanding of life had a greater risk of major depression at 6 or 12 months than participants with
neither a spiritual nor a religious life view (Table 2).
Participants holding a religious understanding of life
were also more at risk than secular participants, but
this nding lost statistical signicance after adjustment. When stratied by country, however, our nding that a spiritual life view predisposed people to
major depression was signicant only in the UK,
where spiritual participants were nearly three times
more likely to experience an episode of depression
than the secular group (OR 2.73, 95 % CI 1.594.68)
(Table 2). In a post-hoc analysis, using international
surveys (European Values Study, 2012 ; World Values
Survey, 2012) we ranked the countries on the proportion of people considering themselves religious
(from least to most : the UK, Estonia, The Netherlands,
Spain, Chile, Slovenia, Portugal). The variation in results between countries was not explained by the importance of religion in each country (results available
from the authors on request).
Understanding of life and onset of rst episode or
recurrence of major depression over 12 months
To investigate the eect of religion on rst episode
versus recurrence of depression, we stratied the
analysis by lifetime history of depression. The eect of
a religious versus a secular understanding of life
was similar in predicting new onset (OR 1.60, 95 % CI
1.052.42) and recurrence (OR 1.44, 95 % CI 1.071.93).
A spiritual view of life did not predict onset of depression in participants with no history of depression
(OR 1.04, 95 % CI 0.631.73) but was related to new
occurrence for participants with a past history of depression (OR 1.61, 95 % CI 1.192.17). However, this
interaction was not signicant (p=0.15).
Religious denomination and onset of major
depression
In the 6094 participants with a spiritual or religious
understanding of life, the incidence of major

Spiritual and religious beliefs and psychological well-being

Table 1. Demographic characteristics and understanding of life

Understanding of life
Religious
(n=4348, 52 %)
Country, n ( %)
UK
Spain
Slovenia
Estonia
The Netherlands
Portugal
Chile
Gender, n ( %)
Female
Male
Age (years), n ( %)
1829
3039
4049
5059
6069
7076
Married/living with partner, n ( %)
No
Yes
Missing
Education, n ( %)
Above school
Secondary
Primary/no education
Trade/other
Missing
Employment, n ( %)
Employed/student
Retired
Unemployed/other
Missing
European ethnicity, n ( %)
No
Yes
Missing
Past history of depression, n ( %)
No
Yes
Missing
Family and friends support
Mean (S.D.)
Missing
Religious denomination, n ( %)
Catholic
Protestant
Other religion
No specic religion
Missing
Strength of belief
Mean (S.D.)
Missing
S.D.,
a

Spiritual
(n=1746, 21 %)

Neither
(n=2087, 25 %)

Totala
(n=8318)

p valueb

462 (10.6)
714 (16.4)
322 (7.4)
189 (4.3)
358 (8.2)
833 (19.2)
1470 (33.8)

300 (17.2)
136 (7.8)
328 (18.8)
227 (13.0)
143 (8.2)
75 (4.3)
537 (30.8)

354 (17.0)
153 (7.3)
387 (18.5)
506 (24.2)
469 (22.5)
97 (4.6)
121 (5.8)

1131 (13.6)
1006 (12.1)
1048 (12.6)
923 (11.1)
1077 (12.9)
1005 (12.1)
2128 (25.6)

<0.001

3045 (70.0)
1303 (30.0)

1235 (70.7)
511 (29.3)

1235 (59.2)
852 (40.8)

5599 (67.3)
2719 (32.7)

<0.001

498 (11.5)
585 (13.5)
764 (17.6)
946 (21.8)
1022 (23.5)
533 (12.3)

332 (19.0)
356 (20.4)
356 (20.4)
333 (19.1)
266 (15.2)
103 (5.9)

406 (19.5)
429 (20.6)
398 (19.1)
435 (20.8)
293 (14.0)
126 (6.0)

1244 (15.0)
1380 (16.6)
1540 (18.5)
1745 (21.0)
1619 (19.5)
790 (9.5)

<0.001

1333 (30.7)
3005 (69.3)
10

615 (35.3)
1129 (64.7)
2

634 (30.5)
1445 (69.5)
8

2614 (31.5)
5681 (68.5)
23

0.001

656 (15.1)
1324 (30.5)
2011 (46.3)
349 (8.0)
8

533 (30.5)
627 (35.9)
388 (22.2)
197 (11.3)
1

749 (36.0)
820 (39.4)
343 (16.5)
167 (8.0)
8

1965 (23.7)
2844 (34.3)
2769 (33.4)
718 (8.7)
22

<0.001

1663 (38.4)
1017 (23.5)
1654 (38.2)
14

975 (56.1)
276 (15.9)
487 (28.0)
8

1313 (63.5)
336 (16.2)
420 (20.3)
18

4003 (48.4)
1662 (20.1)
2610 (31.5)
43

<0.001

1535 (35.4)
2806 (64.6)
7

563 (32.3)
1180 (67.7)
3

162 (7.8)
1921 (92.2)
4

2265 (27.3)
6039 (72.7)
14

<0.001

2021 (46.5)
2322 (53.5)
5

728 (41.8)
1015 (58.2)
3

1039 (49.9)
1046 (50.2)
2

3863 (46.5)
4445 (53.5)
10

<0.001

12.7 (2.4)
34

12.3 (2.7)
10

12 (2.7)
19

12.4 (2.5)
77

<0.001

2680 (62.2)
1248 (29.0)
243 (5.6)
137 (3.2)
40

563 (33.0)
317 (18.6)
118 (6.9)
708 (41.5)
40

3243 (53.9)
1565 (26.0)
361 (6.0)
845 (14.1)
80

<0.001

4.3 (1.5)
90

3.7 (1.6)
30

4.1(1.6)
120

<0.001

N.A.

N.A.

Standard deviation ; N.A., not applicable.

Including 137 missing understanding of life.
b
p value for dierence between understanding of life, from t tests or x2 tests, as appropriate.

B. Leurent et al.

Table 2. Odds ratios (ORs) for onset of major depression over 12 months by understanding
of life, compared to neither religious nor spiritual
Unadjusted OR (95 % CI)

Adjusted ORa (95 % CI)

Overall
Religious
Spiritual

1.52 (1.191.93)**
1.56 (1.212.02)**

1.14 (0.871.50)
1.32 (1.021.70)*

UK
Religious
Spiritual

1.35 (0.682.68)
2.73 (1.594.68)***

1.86 (0.883.92)
2.68 (1.524.71)**

Spain
Religious
Spiritual

1.41 (0.782.54)
1.53 (0.733.20)

1.34 (0.732.46)
1.50 (0.733.07)

Slovenia
Religious
Spiritual

0.62 (0.261.46)
1.06 (0.522.16)

0.69 (0.271.73)
1.10 (0.522.34)

Estonia
Religious
Spiritual

1.16 (0.632.13)
1.08 (0.591.98)

1.06 (0.512.22)
1.14 (0.622.12)

The Netherlands
Religious
Spiritual

0.63 (0.311.28)
1.15 (0.622.13)

0.69 (0.351.37)
1.07 (0.572.02)

Portugal
Religious
Spiritual

2.40 (0.5510.43)
1.96 (0.3810.05)

1.78 (0.398.08)
1.52 (0.278.48)

Chile
Religious
Spiritual

1.12 (0.582.17)
1.03 (0.532.00)

1.08 (0.542.14)
1.04 (0.561.96)

CI, Condence interval.

Values based on imputed data.
a
Adjusted for age, sex, education, employment, social support, past history of
depression, and country.
* p <0.05, ** p<0.01, *** p<0.001.

depression over the subsequent 12 months was similar

across the dierent religious denominations (Catholic
9.8 %, Protestant 10.9 %, other religion 11.5 %, no specic religion 10.8 %, p=0.65). In a post-hoc analysis
in the 1746 participants who reported a spiritual
understanding of life, a similar incidence of major depression over the subsequent 12 months was found
between those who were able to nominate a religious
aliation and those who were not (10.7 % v. 11.1 %,
p=0.41).

Strength of belief at baseline and onset of major

depression
Higher strength of belief in the 6094 participants
who reported a spiritual or religious life view was associated with a greater likelihood of DSM-IV major
depression over the subsequent 12 months after adjustment for age, sex, education, employment, social

support, past history of depression, and country

(unadjusted OR per unit increase on the scale of
strength of belief 1.14, 95 % CI 1.071.21 ; adjusted OR
1.08, 95 % CI 1.001.15) (Fig. 1). There was no interaction between country and strength of belief (p=0.16).
Those with a strongly held belief were twice as likely
to experience major depression in the subsequent
12 months as those with a weakly held belief.

Change in strength of belief and subsequent onset

of major depression
The incidence rate of depression between 6 and
12 months was 5.5 %. We examined whether change
over the rst 6-month follow-up in the strength of
spiritual or religious belief (strength score at 6 months
minus score at baseline) was associated with onset
of major depression between 6 and 12 months, after
adjustment for strength of belief at baseline and other

Incidence of depression over 12 months (%)

Spiritual and religious beliefs and psychological well-being

16

Incidence

95% CI

14
12.5

12

11.7
10.1

10

9.2
8

7.4
6.2

6
4
2
n*= 520

443

1075

1255

947

1734

Weakly

Strongly

How strongly do you hold your religious or spiritual view of life?

OR and 95%CI for depression at 12 months

Fig. 1. Incidence of major depression versus strength of spiritual or religious belief. Based on imputed data. * Frequencies based
on observed data, not numbers included in analysis. CI, Condence interval.

2.21
2
1.47
1.25
1.00

0.94

0.93

0.92
0.77

0.79

0.5

0.25

n*= 51
5/4

140

328

844

2193

806

405

171

113

0 (Ref.)

+1

+2

+3

+4/+5

Difference in strength of belief between baseline and 6 months

Fig. 2. Adjusted odds ratio (OR) for onset of major depression between 6 and 12 months after a change in strength of belief
between baseline and 6 months. OR adjusted for strength of belief at baseline, age, sex, education, employment, social support,
past history of depression, and country. Values based on imputed data. * Frequencies based on observed data, not numbers
included in analysis. CI, Condence interval.

covariates. Those whose belief decreased seemed

to be at greater risk of depression whereas those
whose belief increased had slightly less risk (Fig. 2).
However, CIs were wide, especially for the larger
degrees of change.
Change in the nature of life view and subsequent
onset of major depression
Although an understanding of life (religious, spiritual
or secular) is relatively stable in most people

(King et al. 1999), 27.1 % of participants in this

study changed their life view between baseline and
the 6-month interview. Thus, we examined whether
change in the nature of the life view between
baseline and 6 months had any association with onset
major depression between 6 and 12 months. Point
estimates were in the direction of higher risk of depression for change in a religious direction and
lower risk for change in a secular direction, but the
CIs were wide and no signicant dierence was found
(Table 3).

B. Leurent et al.

Table 3. Odds ratios (ORs) for onset of major depression between 6 and 12 months, after a
change in understanding of life between baseline and 6 months
Understanding of life
At baseline

At 6 months

Frequencya

Unadjusted OR
(95 % CI)

Adjusted ORb
(95 % CI)

Religious

Religious
Spiritual
Neither

3004
658
175

1.00 (Reference)
0.98 (0.621.53)
0.64 (0.261.56)

1.00 (Reference)
0.97 (0.611.55)
0.87 (0.352.16)

Spiritual

Religious
Spiritual
Neither

436
887
253

1.40 (0.842.32)
1.00 (Reference)
1.05 (0.542.03)

1.51 (0.872.61)
1.00 (Reference)
1.09 (0.562.14)

Neither

Religious
Spiritual
Neither

196
265
1413

1.53 (0.703.36)
1.46 (0.722.96)
1.00 (Ref)

1.31 (0.533.19)
1.35 (0.662.78)
1.00 (Ref)

Incidence of depression over 12 months (%)

CI, Condence interval.

Values based on imputed data.
a
Frequencies based on observed data, not numbers included in analysis.
b
Adjusted for age, sex, education, employment, social support, past history of
depression, and country.

20
Religious

Spiritual

Neither
15

10

n* = 3474
0

1749

2235
1

>1

Life events between baseline and 6 months

Fig. 3. Modifying eect of life understanding on major depression after serious life events. Values based on imputed data.
* Frequencies based on observed data, not numbers included in analysis.

Impact of belief on the relationship between serious

life events and onset of major depression

Sensitivity of the analysis to imputation of missing

data

Finally, we examined whether a religious or spiritual

understanding of life modied the risk of major depression following a serious life event. As expected,
the incidence of major depression between baseline
and 12 months was higher for patients who experienced serious life events (Fig. 3). Understanding of
life was not a signicant modifying factor of the
eect of life events on depression (unadjusted p=0.57,
adjusted p=0.48).

When imputed data were compared to the observed

data we found a similar distribution between understanding of life categories, but the incidence of major
depression at 6 or 12 months was higher in the imputed data (8.7 % v. 9.5 %), reecting the higher likelihood of dropping out for patients at risk of depression
at baseline. No major discrepancy was found between
imputed data and complete-case analyses. For example, the OR of developing major depression by 6 or

Spiritual and religious beliefs and psychological well-being

12 months was 1.64 (95 % CI 1.262.15) for people with
a religious understanding and 1.74 (95 % CI 1.342.26)
for people with a spiritual understanding of life,
compared to 1.52 and 1.56 respectively on imputed
data (Table 2).

Discussion
Main ndings
We found that people who held a religious or spiritual
understanding of life had a higher incidence of depression than those with a secular life view. However,
this nding varied by country ; in particular, people in
the UK who had a spiritual understanding of life were
the most vulnerable to the onset of major depression.
Regardless of country, the stronger the spiritual or religious belief at baseline, the higher the risk of onset of
depression. Although our main nding of an association between religious life understanding and onset
of depression varied by country, we found no
evidence that spirituality may protect people, and
only weak evidence that a religious life view was
possibly protective in two countries (Slovenia and The
Netherlands). Finally, there was no moderating eect
of religious and spiritual understanding of life on the
impact of life events on onset of major depression.
Strengths and limitations
The main strengths of our study are its prospective
cohort design, the involvement of several countries,
the large sample size and the use of standardized assessments of religious/spiritual belief, other risk variables and major depression. However, combining
people from dierent cultures creates heterogeneity
and runs the risk of missing real dierences within
countries. Despite a large sample size, we lacked
power for some of the analyses, particularly the assessment of statistical interactions. However, to our
knowledge there are very few data sets available containing variables on spiritual and religious beliefs that
are large enough to undertake interaction analyses.
Thus we consider it is important to show the results to
avoid a reporting bias that favours statistically signicant results (Dwan et al. 2008). Another strength
of this study is the use of multiple imputation to take
account of participants who dropped out. Assuming
predictors of missingness have been examined at
baseline, this should give unbiased ndings, as opposed to restricting the analysis to participants who
completed both follow-ups only, which may represent
a sample at lesser risk of depression than the original
one. However, participations rates were low in some
countries, and we cannot assume that ndings in

general practice attendees can be generalized to the

whole population. Our study sample is likely to have
a more complex medical history ; for example, 53 %
reported a lifetime history of depression although
other evidence would suggest this gure is usually in
the range 3040 % (Kruijshaar et al. 2005). However,
our population may be more representative of life
views than those who might participate in a study of
religion. The original aim of this study was to develop
a risk prediction algorithm for depression (PredictD ;
King et al. 2008), and thus our analysis is limited to the
religiosity data available. Religion did not appear as a
variable in the nal prediction algorithm PredictD,
which included only the most parsimonious combination of risk factors of depression, but this does
not imply the absence of a relationship between religiosity and depression. Although our questions were
limited to peoples overall view of religion and spirituality, rather than the detail of any specic belief,
they applied to all people and not just those with a
Christian background, which is often the case in North
American research. Another limitation is the diculty
distinguishing religion and spirituality and the trouble
respondents may have in placing themselves in one of
the two categories. This may account for a proportion
of the 27 % of participants who changed their life view
between baseline and 6 months. However, the questions have high repeatability (King et al. 2001) and
clear denitions were given in a short note preceding
the question, ensuring a universal understanding of
the term across the dierent countries and cultures.
Furthermore, they avoided the common pitfall in religion studies of conating questions on religion and
spiritual belief with those on psychological well-being
(Koenig, 2008). Finally, the relationship between
religion, spirituality and mental symptoms should
not necessarily be interpreted as causal in nature.
Although the longitudinal design removes the possibility of reverse causation, unmeasured confounders
in the complex relationship between religiosity and
depression are likely to remain, even after adjustment
for main risk factors of depression.
Relationship to other ndings
Our work adds to a growing body of evidence that
spiritual beliefs in the absence of a clear religious
aliation or practice increase vulnerability to depression (Baetz et al. 2006 ; Braam et al. 2007 ; Park et al.
2012) ; it also contrasts with many studies where religious and spiritual beliefs and practice have been
found to be associated with better mental health.
An explanation for this disparity could be the complex
relationship between the concepts of religiosity
and well-being, and that ndings in any specic

10

B. Leurent et al.

population may not generalize to another. Research in

this eld has been dominated by North American
studies, whereas in the more secular cultures of
Europe, religious people may feel less supported in
their faith. Alternatively, it may relate to the ways in
which such beliefs are measured in research. In their
meta-analysis, Hackney & Sanders (2003) criticized
many studies for measuring religion and spirituality
with insucient depth and suggested that a multifaceted concept such as religious belief and practice
may have complex associations with mental health.
In their meta-analysis of 147 studies on religious
belief and depression, Smith et al. (2003) found that
extrinsic religious orientation and so-called negative
religious coping were associated with higher levels of
depressive symptoms. However, they also reported
that the negative correlation between religiousness
and depressive symptoms was at its greatest in the
presence of life stress and suggested that religion may
have a buering eect on the impact of life events. We
did not nd evidence of such buering. They speculated that social desirability of response (exaggerating
religiousness and downrating depressive symptoms),
or the possibility that religious people might be better
at expressing emotion and thus coping with stress,
might explain some of the correlation in their studies
that were mainly cross-sectional in nature.
Recent longitudinal research suggests that attaching
a high importance to religion was associated with
lower risk of recurrence of depression in the subsequent
10 years (Miller et al. 2012) and had a protective eect
after negative life events (Kasen et al. 2012). We have
not been able to replicate either of these ndings in
this cohort. Our study was, however, consistent with
their nding of no clear relationship between religious
denomination and depression.
The disparity in the ndings suggests that, if there
is an association between religion/spirituality and
psychological well-being, it is likely to be weak. If religious belief has a powerful positive eect on mental
health, we would expect to detect it in most studies.

Implications
Why a religious or spiritual life view might place
people at risk of depression remains unclear. One explanation is that people predisposed to depression at
baseline may seek meaning in spiritual or religious
sources. The possibility that people predisposed to
depression increase their search for existential meaning in religion and spirituality is supported by our
nding that change in belief over time towards greater
religiosity did seem to be related to greater risk of depression.

In conclusion, we found that holding a religious or

spiritual life view, in contrast to a secular outlook,
predisposed people to the onset of major depression
and that such beliefs and practice did not act as a
buer to adverse life events. Our ndings highlight
the complexity of the relationship between religion,
spirituality and mental health and oer a challenge to
an increasing tendency to regard religion and spirituality as being good for mental well-being (Schumann
& Meador, 2003).

Supplementary material
For supplementary material accompanying this paper
visit http://dx.doi.org/10.1017/S0033291713003066.

Acknowledgements
The study in Europe was funded by a European
Commission Vth Framework grant (PREDICT-QL4CT2002-00683). Funding in Chile was provided by
project FONDEF DO2I-1140. We are grateful for part
support in Europe from : the Estonian Scientic
Foundation (grant 5696) ; the Slovenian Ministry for
Research (grant 4369-1027) ; the Spanish Ministry of
Health (FIS references : PI041980, PI041771, PI042450)
and the Spanish Network of Primary Care Research,
redIAPP (ISCIII-RETIC RD06/0018) and SAMSERAP
group ; and the UK NHS Research and Development
oce for service support costs in the UK. The funders
had no direct role in the design or conduct of the
study, interpretation of the data or review of the
manuscript.
M. King had full access to the data and takes responsibility for their integrity and the accuracy of
the data analysis. We thank all patients and general
practice sta who took part ; the European Oce at
University College London for their administrative
assistance at the coordinating centre ; K. McCarthy,
the projects scientic ocer in the European
Commission, Brussels, for his helpful support and
guidance ; the UK MRC GPRF ; L. Letley from the MRC
GPRF ; the GPs of the Utrecht General Practitioners
Network ; and the Camden and Islington Mental
Health and Social Care Trust. We also acknowledge
the Maristan network, through which the collaboration in Spain, Portugal, the UK and Chile rst developed.

Declaration of Interest
None.

Spiritual and religious beliefs and psychological well-being

References
Baetz M, Bowen R, Jones G, Koru-Sengul T (2006). How
spiritual values and worship attendance relate to
psychiatric disorders in the Canadian population. Canadian
Journal of Psychiatry 51, 654661.
Blaxter M (1990). Health and Lifestyles. Routledge : London.
Braam AW, Deeg DJ, Poppelaars JL, Beekman AT,
van Tilburg W (2007). Prayer and depressive symptoms
in a period of secularization : patterns among older adults
in the Netherlands. American Journal of Geriatric Psychiatry
15, 273281.
Brugha T, Bebbington P, Tennant C, Hurry J (1985). The
List of Threatening Experiences : a subset of 12 life event
categories with considerable long-term contextual threat.
Psychological Medicine 15, 189194.
Dwan K, Altman DG, Arnaiz JA, Bloom J, Chan AW,
Cronin E, Decullier E, Easterbrook PJ, Von Elm E,
Gamble C, Ghersi D, Ioannidis JP, Simes J,
Williamson PR (2008). Systematic review of the empirical
evidence of study publication bias and outcome reporting
bias. PLoS One 3, e3081.
European Values Study (2012). Research theme : religion
(www.europeanvaluesstudy.eu/evs/research/themes/
religion/). Accessed 6 July 2012.
Gartner J, Larson DB, Allen DG (2012). Religious
commitment and mental health : a review of the empirical
literature. Journal of Psychology and Theology 19, 625.
Hackney CH, Sanders GS (2003). Religiosity and mental
health : a meta-analysis of recent studies. Journal for the
Scientic Study of Religion 42, 4355.
Huber PJ (1967). The behavior of maximum likelihood
estimates under non-standard conditions. Proceedings of the
Fifth Berkeley Symposium on Mathematical Statistics and
Probability 1, 221233.
Johnstone B, Franklin KL, Yoon DP, Burris J, Shigaki C
(2008). Relationships among religiousness, spirituality, and
health for individuals with stroke. Journal of Clinical
Psychology in Medical Settings 15, 308313.
Kasen S, Wickramaratne P, Gamero MJ, Weissman MM
(2012). Religiosity and resilience in persons at high
risk for major depression. Psychological Medicine 42,
509519.
King M, Speck P, Thomas A (1999). The eect of spiritual
beliefs on outcome from illness. Social Science and Medicine
48, 12911299.
King M, Speck P, Thomas A (2001). The Royal Free
Interview for Spiritual and Religious Beliefs : development
and validation of a self-report version. Psychological
Medicine 31, 10151023.
King M, Walker C, Levy G, Bottomley C, Royston P,
Weich S, Bellon-Saameno J, Moreno B, Svab I, Rotar D,
Rifel J, Maaroos H, Aluoja A, Kalda R, Neeleman J,
Geerlings MI, Xavier M, Carraca I, Goncalves-Pereira M,
Vicente B, Saldivia S, Melipillan R, Torres-Gonzalez F,
Nazareth I (2008). Development and validation of
an international risk prediction algorithm for
episodes of major depression in general practice
attendees : the PredictD study. Archives of General
Psychiatry 65, 13681376.

11

King M, Weich S, Nazroo J, Blizard R (2006a).

Religion, mental health and ethnicity. EMPIRIC a
national survey of England. Journal of Mental Health 15,
153162.
King M, Weich S, Torres-Gonzalez F, Svab I, Maaroos HI,
Neeleman J, Xavier M, Morris R, Walker C,
Bellon-Saameno JA, Moreno-Kustner B, Rotar D, Rifel J,
Aluoja A, Kalda R, Geerlings MI, Carraca I,
de Almeida MC, Vicente B, Saldivia S, Rioseco P,
Nazareth I (2006 b). Prediction of depression in European
general practice attendees : the PREDICT study. BMC
Public Health 6, 6.
Koenig HG (2008). Concerns about measuring spirituality
in research. Journal of Nervous and Mental Disease 196,
349355.
Koenig HG, George LK, Peterson BL (1998). Religiosity and
remission of depression in medically ill older patients.
American Journal of Psychiatry 155, 536542.
Koenig HK, McCullough ME, Larson DB (2001).
Handbook of Religion and Health. Oxford University Press :
Oxford.
Kruijshaar ME, Barendregt J, Vos T, de Graaf R, Spijker J,
Andrews G (2005). Lifetime prevalence estimates of major
depression : an indirect estimation method and a
quantication of recall bias. European Journal of
Epidemiology 20, 103111.
Lewis CA, Maltby J, Burkinshaw S (2000). Religion and
happiness : still no association. Journal of Beliefs and Values
21, 233236.
McCullough ME, Larson DB (1999). Religion and
depression : a review of the literature. Twin Research 2,
126136.
Miller L, Wickramaratne P, Gamero MJ, Sage M,
Tenke CE, Weissman MM (2012). Religiosity and
major depression in adults at high risk : a tenyear prospective study. American Journal of Psychiatry
169, 8994.
Park JI, Hong JP, Park S, Cho MJ (2012). The relationship
between religion and mental disorders in a Korean
population. Psychiatry Investigation 9, 2935.
Payne IR, Bergin AE, Bielema KA, Jenkins PH (1991).
Review of religion and mental health : prevention and the
enhancement of psychosocial functioning. Prevention in
Human Services 9, 1149.
Robins LN, Wing J, Wittchen HU, Helzer JE,
Babor TF, Burke J, Farmer A, Jablenski A, Pickens R,
Regier DA, Sartorius N, Towle LH (1988). The
Composite International Diagnostic Interview.
An epidemiologic instrument suitable for use in
conjunction with dierent diagnostic systems and
in dierent cultures. Archives of General Psychiatry 45,
10691077.
Royston P (2005). Multiple imputation of missing values :
update of ice. Stata Journal 5, 527536.
Schaefer WE (1997). Religiosity, spirituality, and personal
distress among college students. Journal of College Student
Development 38, 633644.
Schumann JJ, Meador KG (2003). Heal Thyself : Spirituality,
Medicine, and the Distortion of Christianity. Oxford
University Press : New York.

12

B. Leurent et al.

Smith TB, McCullough ME, Poll J (2003). Religiousness and

depression : evidence for a main eect and the moderating
inuence of stressful life events. Psychological Bulletin 129,
614636.
StataCorp (2009). Stata Statistical Software : Release 11. Stata
Corporation : College Station, TX.

WHO (1997). Composite International Diagnostic

Interview (CIDI). Version 2.1. World Health Organization :
Geneva.
World Values Survey (2012). Online Data Analysis, WVS
20052008 (www.wvsevsdb.com/wvs/
WVSAnalizeStudy.jsp). Accessed 6 July 2012.