Prevalence of depression and depressive symptoms among outpatients: a systematic review and meta-analysis

Abstract

Objectives

Depression and depressive symptoms are common mental disorders that have a considerable effect on patients’ health-related quality of life and satisfaction with medical care, but the prevalence of these conditions varies substantially between published studies. The aim of this study is to conduct a systematic review and meta-analysis to provide a precise estimate of the prevalence of depression or depressive symptoms among outpatients in different clinical specialties.

Design

Systematic review and meta-analysis.

Data sources and eligibility criteria

The PubMed and PsycINFO, EMBASE and Cochrane Library databases were searched to identify observational studies that contained information on the prevalence of depression and depressive symptoms in outpatients. All studies included were published before January 2016. Data characteristics were extracted independently by two investigators. The point prevalence of depression or depressive symptoms was measured using validated self-report questionnaires or structured interviews. Assessments were pooled using a random-effects model. Differences in study-level characteristics were estimated by meta-regression analysis. Heterogeneity was assessed using standard χ² tests and the I² statistic. The study protocol has been registered with PROSPERO under number CRD42017054738.

Results

Eighty-three cross-sectional studies involving 41 344 individuals were included in this study. The overall pooled prevalence of depression or depressive symptoms was 27.0% (10 943/41 344 individuals; 95% CI 24.0% to 29.0%), with significant heterogeneity between studies (p<0.0001, τ²=0.3742, I²=96.7%). Notably, a significantly higher prevalence of depression and depressive symptoms was observed in outpatients than in the healthy controls (OR 3.16, 95% CI 2.66 to 3.76, I²=72.0%, χ²=25.33). The highest depression/depressive symptom prevalence estimates occurred in studies of outpatients from otolaryngology clinics (53.0%), followed by dermatology clinics (39.0%) and neurology clinics (35.0%). Subgroup analyses showed that the prevalence of depression and depressive symptoms in different specialties varied from 17.0% to 53.0%. The prevalence of depression and depressive symptoms was higher among outpatients in developing countries than in outpatients from developed countries. Moreover, the prevalence of depression and depressive symptoms in outpatients slightly decreased from 1996 to 2010. Regarding screening instruments, the Beck Depression Inventory led to a higher estimate of the prevalence of depression and depressive symptoms (1316/4702, 36.0%, 95% CI 29.0% to 44.0%, I²=94.8%) than the Hospital Anxiety and Depression Scale (1003/2025, 22.0%, 95% CI 12.0% to 35.0%, I²=96.6%).

Conclusion

Our study provides evidence that a significant proportion of outpatients experience depression or depressive symptoms, highlighting the importance of developing effective management strategies for the early identification and treatment of these conditions among outpatients in clinical practice. The substantial heterogeneity between studies was not fully explained by the variables examined.

Strengths and limitations of this study.

• This is the first systematic review and meta-analysis to provide a comprehensive estimate of the prevalence of depression or depressive symptoms among outpatients considering different clinical specialties.

• The information provided by this study highlights the importance of developing effective management strategies for the early identification and treatment of depression among outpatients in clinical practice.

• The substantial heterogeneity between included studies was not fully explained by the variables examined.

• The paucity of longitudinal data may decrease the generalisability of the study outcomes.

Introduction

Depression is the leading cause of disability and is a major contributor to the disease burden worldwide. The global prevalence of depression and depressive symptoms has been increasing in recent decades.¹ The lifetime prevalence of depression ranges from 20% to 25% in women and 7% to 12% in men.² Depression is a significant determinant of quality of life and survival, accounting for approximately 50% of psychiatric consultations and 12% of all hospital admissions.³ Notably, the prevalence of depression or depressive symptoms is higher in patients than in the general public.^3–6 The underlying reasons include the illness itself and the heavy medical cost, unsatisfactory medical care service and poor doctor–patient relationship.^{7 8} Several informative systematic reviews on specific groups of outpatients have been published. For example, Mitchell et al estimated that the prevalence of depression in oncology and haematology patients was 9.6%–16.5%.⁵ Depression is a significant comorbidity of chronic medical disorders. The prevalence of depression in chronic medical conditions is as follows: asthma (27%),⁹ atopic dermatitis (5%),¹⁰ chronic obstructive pulmonary disease (24.6%),¹¹ gouty arthritis (20%),¹² rheumatoid arthritis (15%),¹³ systemic lupus erythematosus (22%)¹² and stroke (30%).¹⁴ Ismail et al conducted a meta-analysis of 57 studies and showed that the overall pooled prevalence of depression in patients with mild cognitive impairment was 32%.⁴ Estimates of the prevalence of depression and depressive symptoms vary substantially between published studies, particularly with respect to specialty, patient age and residence. The inconsistency across different studies may originate from the lack of a clear definition or gold standard for the diagnosis of depression and depressive symptoms. Many previous studies have focused on depression and depressive symptoms in inpatient settings; however, mental disorders in outpatients are largely underestimated.^{6 15} Depression in outpatients is associated with high indirect costs due to loss of productivity and unemployment.⁸ The combination of chronic medical illnesses and depression will lead to significant economic burden.⁸ Additionally, it is important for healthcare workers to identify mental status changes in outpatients, as mental states may affect the doctor–patient relationship and can influence patient satisfaction with medical care.¹⁶ To the best of our knowledge, no previous studies have quantitatively analysed a robust dataset with information on depression and depressive symptoms among outpatients in different clinical departments. Therefore, conducting a systematic review and meta-analysis of the depression prevalence detected during doctor visits is essential to informing efforts to prevent and treat depression and depressive symptoms among outpatients. In this study, we aimed to quantitatively summarise the prevalence of depression and depressive symptoms in different clinical departments.

Methods

Study selection

Relevant studies published before January 2016 that described the prevalence of depression or depressive symptoms in patients from different specialties were identified using the PubMed and PsycINFO, EMBASE and Cochrane Library databases (by WJH and WXH); the selected articles were then screened by title, abstract and reference lists in collaboration with study investigators using the approach recommended by the Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines.¹⁷ Potentially relevant papers were first identified through title and abstract searches. The full text of the retrieved articles was then assessed. The search strategy involved applying the ‘explode’ command to search the MeSH terms ‘depression’ and ‘depressive symptoms’ and ‘outpatient*’; the combined terms were related to ‘prevalence’, ‘cross-sectional studies’ or ‘controlled studies’ and ‘different specialties’, such as ‘internal medicine’, ‘surgical specialties’ and ‘paediatrics’, with language restrictions (articles published in English only). More search strategy details can be found in online supplementary method 1. The study inclusion criteria were the following: (1) articles that included patients diagnosed with a specific disease other than psychiatric disorders; (2) articles in peer-reviewed journals that involved only patients with a current degree of clinically relevant depression sufficient to warrant clinical intervention, regardless of the depression severity (mild, moderate or severe)¹⁸; (3) studies in which depression was confirmed by validated self-report instruments or diagnostic structured interviews¹⁹; (4) articles with study populations who were recruited from outpatient clinics only. The exclusion criteria were as follows: (1) studies that failed to report the specific prevalence of depression, (2) studies on patients whose depression predated any other physical disorder and (3) studies on patients diagnosed with more than one psychiatric disorder (in addition to depression).

Data extraction and quality assessment

Data extraction was a multistep process based on the eligibility criteria. The following information was extracted from each study independently by two investigators (JHW and XHW) independently using a standardised form: study design, research year, country, specialty category, disease, sample size, diagnostic or screening method used and reported prevalence of depression and depressive disorders. A modified version of the Newcastle-Ottawa Scale was used to evaluate the quality of non-randomised studies.²⁰ Studies were identified as having a low risk of bias (≥3 points) or a high risk of bias (<3 points). The effect of individual studies on the overall prevalence estimate was explored by serially excluding each study in a sensitivity analysis. Additionally, two reviewers (JHW and XHW) cross-checked the reference lists of all selected articles to identify other relevant studies. All discrepancies were resolved by discussion and consensus.

Statistical analysis

As considerable heterogeneity was expected because of the multiple sources of variance, a random-effects model was used to estimate the pooled prevalence of depressive symptoms.²¹ Random-effect model attempted to generalise findings beyond the included studies by assuming that the selected studies are random samples from a larger population.²² The observed heterogeneity in the depression prevalence among outpatients may be attributed to differences in the assessment methods used to detect depression, the variation in thresholds in the different validated depression measurements, the specialties examined, the study countries, study year, patient ages and other factors. Thus, subgroup analyses were performed. Binomial proportion CIs for individual studies were calculated using the Clopper-Pearson method, which allows for asymmetry.²³ Between-study heterogeneity was evaluated using standard χ² tests and the I² statistic.²¹ I² statistics were calculated to describe the percentages of total variation across studies caused by heterogeneity. A 0% value indicated no heterogeneity, and higher values represented an increase in heterogeneity. Generally, heterogeneity is categorised as 25% (low), 50% (moderate) and 75% (high).²⁴ The results of the analysis were compared in terms of descriptive characteristics (age, specialty, study year, diagnostic method and country) using subgroup analysis and meta-regression. For models with considerable heterogeneity, a meta-regression was performed to identify the moderators that might contribute to the heterogeneity of the effect sizes.²⁵ Publication bias of the studies was examined using funnel plots and Egger’s test.²⁶ All analyses were performed using R version 3.3.1 (R Foundation for Statistical Computing, Vienna, Austria) and Review Manager version 5.3 (The Cochrane Collaboration, 2015, the Nordic Cochrane Centre, Copenhagen, Denmark). Statistical tests were two-sided with a significance threshold of p<0.05. This study is registered with PROSPERO, number CRD42017054738.

Patient involvement

No patients eligible for screening were involved in the design and conduct of the study or involved in defining the research question or outcome measures. We have no intentions to disseminate our results to patients eligible for screening.

Results

Screening the titles and abstracts resulted in 3165 articles, 110 of which were duplicates, and only 207 articles were retrieved for a detailed, full-text assessment. Of these, 83 studies fulfilled the inclusion criteria; 101 studies did not meet the eligible population criteria, 8 failed to present point prevalence data and 15 used improper outcome measures and were excluded.

Eighty-three cross-sectional studies involving a total of 41 344 individuals were included in the study (figure 1). Study participants were recruited from 11 departments: 22 studies recruited patients from internal medicine clinics, 12 from primary care, 10 from neurology, 8 from dermatology, 7 from obstetrics/gynaecology, 6 from ophthalmology, 6 from oncology, 5 from infectious diseases, 4 from surgery, 3 from paediatrics and 3 from otolaryngology departments. Most (29) of the studies were conducted in Europe; 21 were performed in Asia, 19 in North America, 4 in South America, 4 in Oceania, 3 in the Middle East and 1 in Africa. Seventeen studies used the Beck Depression Inventory (BDI) to assess depression; 10, the Hospital Anxiety and Depression Scale (HADS); 7, the Patient Health Questionnaire; 6, the Hamilton Depression Scale, also called the Hamilton Depression Rating Scale; and 43, other methods. The full study characteristics are summarised in table 1. The overall prevalence estimates of depression or depressive symptoms reported by the 83 studies yielded a summary prevalence of 27.0% (10 943/41 344 individuals, 95% CI 24.0% to 29.0%), with significant between-study heterogeneity (p<0.0001, τ²=0.3742, I²=96.7%). Subgroup analyses by age, clinical department, study year, country and diagnosis method were conducted to explore the potential heterogeneity between studies. Of the 83 studies, the highest depression/depressive symptom prevalence estimates occurred in studies of outpatients from otolaryngology clinics (357/796, 53.0%, 95% CI 39.0% to 66.0%, I²=79.8%), followed by dermatology clinics (520/1558, 39.0%, 95% CI 24.0% to 56.0%, I²=96.9%) and neurology clinics (3328/9280, 35.0%, 95% CI 30.0% to 40.0%, I²=94.4%). The prevalences of depression among outpatients from other specialties are summarised in figure 2. Subgroup analysis was conducted to compare studies in developed countries versus in developing countries (7788/29 208, 24.0%, 95% CI 21.0% to 27.0%, I²=97.0%, p<0.0001 vs 3188/12 050, 33.0%, 95% CI 28.0% to 38.0%, I²=96.8%, p<0.0001). The prevalence of depression/depressive symptoms in outpatients decreased from 36.0% to 24.0% from 1990 to 2010, followed by a slight increase from 2011 to 2016. Outpatients who were younger than 30 years old showed the lowest depression prevalence, at 20.0% (170/797, 95% CI 14.0% to 28.0%, I²=81.6%, p=0.0010), whereas the highest depression prevalence was reported in outpatients older than 80 years at 34.0% (397/2128, 95% CI 15.0% to 69.0%, I²=96.8%, p<0.0001). The prevalences reported in studies stratified by year are presented in figure 3. Eight studies with healthy controls were included in a subgroup. There was a significantly higher prevalence of depression and depressive symptoms in outpatients than in healthy controls (OR 3.16, 95% CI 2.66 to 3.76, I²=72.0%, χ²=25.33) (figure 4).

Figure 1.

Flow diagram of study inclusion.

Table 1.

Selected characteristics of the 83 studies included in this systematic review and meta-analysis

Study	Specialty	Disease category	Country	Mean age (SD)	Time point (year, month)	Study design	Diagnostic method	Depression prevalence % (cases/participants)	NOS
Rohani et al53	Internal medicine	Cardiology	Iran	50.4 (NR)	April 2010–November 2010	Descriptive cross-sectional	HADS	31.0% (78/250)	5
Bokemeyer et al54	Internal medicine	Gastroenterology	Germany	43 (NR)	March 2006–July 2007	Descriptive cross-sectional	SF-36	14.9% (154/1032)	4
Tedeschini et al55	Internal medicine	Rheumatology	Italy	53.7 (12.1)	February 2005–July 2007	Descriptive cross-sectional	BDI	46.2% (36/78)	4
Yohannes et al56	Internal medicine	Pulmonary disease	UK	73 (NR)	NR	Descriptive cross-sectional	GMS, MADRS,  BASDEC	42.0% (57/137)	4
Yohannes et al57	Internal medicine	Pulmonary disease	UK	78 (5)	NR	Controlled cross-sectional	BASDEC	46.0% (44/96)	3
Noh et al58	Internal medicine	Endocrinology	Korea	52.7 (12.2)	March 2003–October 2003	Controlled cross-sectional	BDI	32.4% (66/204)	5
Janke et al59	Internal medicine	Gastroenterology	Germany	43.2 (11.0)	January 1997–December 2000	Descriptive cross-sectional	HADS	10.2% (43/429)	4
Zhang et al60	Internal medicine	Internal medicine	China	44.4 (16.5)	July 2011–June 2012	Descriptive cross-sectional	PHQ-9	9.2% (37/404)	4
He et al*61	Internal medicine	Cardiology	China	40.1 (17.3)	April 2007–July 2007	Descriptive cross-sectional	HADS, GHQ-15	26.7% (140/524)	4
He et al*61	Internal medicine	Gastroenterology	China	40.1 (17.3)	April 2007–July 2007	Descriptive cross-sectional	HADS, GHQ-15	26.9% (181/674)	4
Pontone et al62	Internal medicine	Gastroenterology	Italy	53.3 (17)	May 2009–October 2010	Descriptive cross-sectional	BDI	17.2% (22/130)	3
Urrutia et al63	Internal medicine	Pulmonary disease	Spain	44.25 (16.77)	December 2006–December 2007	Descriptive cross-sectional	HADS	1.7% (6/354)	3
Zhang et al64	Internal medicine	Endocrinology	China	55.1 (9.5)	July 2010–July 2011	Descriptive cross-sectional	PHQ-9	18.3% (107/586)	4
Birket-Smith et al65	Internal medicine	Cardiology	Denmark	67.4 (13.63)	2011	Descriptive cross-sectional	DSM-III-R, HAMD	19.7% (17/86)	4
Hajduk et al66	Internal medicine	Rheumatology	Poland	43.79 (11.66)	2009–2013	Descriptive cross-sectional	HADS-M	35.8% (19/53)	4
Qin et al67	Internal medicine	Internal medicine	China	48.1 (17.7)	November 2004–January 2006	Descriptive cross-sectional	GHQ SI	34.6% (495/1428)	4
Inagaki et al68	Internal medicine	Internal medicine	Japan	75 (NR)	July 2010	Descriptive cross-sectional	PHQ-9	9.0% (36/396)	5
Schaefert et al69	Internal medicine	Internal medicine	Germany	43.2 (14.2)	August 2009–October 2009	Descriptive cross-sectional	PHQ-15, HADS	4.6% (13/281)	3
Addolorato et al70	Internal medicine	Gastroenterology	Italy	43.9 (15.9)	1997–2015	Descriptive cross-sectional	ZUNG SDS	27.0% (442/1641)	4
Pouwer et al71	Internal medicine	Endocrinology	Netherlands	43(14)	NR	Descriptive cross-sectional	CESD	33.6% (243/724)	5
Su et al72	Internal medicine	Nephrology	China	59.8 (11.9)	NR	Descriptive cross-sectional	BDI	40.3% (129/320)	5
Xiong et al73	Internal medicine	Internal medicine	China	44.9 (16.4)	February 2011–October 2012	Descriptive cross-sectional	PHQ-9, MINI	38.3% (188/491)	5
Tsunoda et al74	Oncology	Colorectal cancer	Japan	69 (10.5)	1994–2005	Descriptive cross-sectional	HADS	36.7% (47/128)	4
Polidoro Lima and Osório75	Oncology	Oncology	Brazil	50.3 (13.9)	NR	Descriptive cross-sectional	PHQ-4	18.6% (257/1385)	4
Alcalar, N et al76	Oncology	Breast cancer	Turkey	48.32 (8.46)	September 2008–April 2009	Descriptive cross-sectional	BDI	30.9% (34/110)	3
Jehn et al77	Oncology	Metastatic breast cancer	Germany	59.9 (10.2)	NR	Descriptive cross-sectional	HADS	31.4% (22/70)	4
Qiu et al78	Oncology	Postsurgery breast cancer	China	52.02 (4.55)	January 2008–March 2009	Descriptive cross-sectional	BDI, HAMD,  MINI	18.8% (95/505)	4
Reuter et al6	Oncology	Gynaecological and  breast cancer	Germany	54 (19.81)	May 1998–June 2000	Descriptive cross-sectional	HADS	15.2% (10/66)	3
Diniz et al79	Surgery	Renal colic	Brazil	43.8 (14.4)	June 2003–October 2003	Controlled cross-sectional	BDI	59.4% (19/32)	4
Hung et al80	Surgery	Orthopaedics	China	40.7 (11.4)	NR	Descriptive cross-sectional	HADS	21.8% (49/225)	4
Jung et al81	Surgery	Chronic peritoneal dialysis	Korea	54.2 (10.24)	July 2009–October 2009	Descriptive cross-sectional	BDI	35.7% (20/56)	4
Weisbord et al82	Surgery	Chronic haemodialysis	USA	64 (NR)	2009–2011	Descriptive cross-sectional	PHQ-9	25.5% (73/286)	4
Bixo et al83	OG	Gynaecology	Sweden	43.8 (14.3)	November 1998–December 1998	Descriptive cross-sectional	PRIME-MD	27.2% (208/766)	4
Gaillard et al84	OG	Postpartum	France	31 (NR)	November 2007–November 2009	Descriptive cross-sectional	EPDS	16.7% (44/264)	4
Lorencatto et al*85	OG	Endometriosis with chronic pelvic pain	Brazil	35.3 (6.4)	NR	Descriptive cross-sectional	BDI	86.0% (43/50)	4
Lorencatto et al*85	OG	Endometriosis without chronic pelvic pain	Brazil	32.8 (7.1)	NR	Descriptive cross-sectional	BDI	38.0% (19/50)	4
Poleshuck et al86	OG	Gynaecology	USA	32.1 (NR)	March 2004–December 2004	Descriptive cross-sectional	BDI	21.8% (51/234)	4
He et al*61	OG	Gynaecology	China	40.1 (17.3)	April 2007–July 2007	Descriptive cross-sectional	HADS, GHQ-15	18.6% (103/554)	4
Wang et al87	OG	Menopause	China	NR	2004	Descriptive cross-sectional	HADS	11.1% (34/306)	2
Wojnar et al88	OG	Gynaecology	Poland	49.85 (3.09)	May 2001–October 2001	Descriptive cross-sectional	BDI, ICD-10	19.5% (442/2262)	4
Stewart et al89	Paediatrics	T1DM	USA	13.54 (1.56)	December 2001–May 2003	Descriptive cross-sectional	CESD	30.0% (62/205)	3
Zdunczyk et al90	Paediatrics	T1DM	Poland	14.2 (2.0)	October 2011–November 2012	Descriptive cross-sectional	CDI	19.4% (72/372)	4
Winter et al91	Paediatrics	Paediatrics	USA	13.89 (1.58)	NR	Descriptive cross-sectional	BDI-PC	11.0% (11/100)	3
Carson et al52	Neurology	Neurology	UK	43.4 (NR)	2003	Analytical Cohort-study	HADS, PRIME-MD	39.7% (119/300)	4
Carson et al92	Neurology	Neurology	UK	43 (16.2)	November 1997–March 1998	Descriptive cross-sectional	HADS, PRIME-MD	33.3% (100/300)	4
de Oliveira et al93	Neurology	Temporal lobe epilepsy	Brazil	40.7 (10.1)	NR	Descriptive cross-sectional	HAMD	35.4% (34/96)	4
Dickstein et al*94	Neurology	Epilepsy	USA	48 (17)	October 2007–August 2013	Descriptive cross-sectional	PHQ-9	36.3% (1003/2763)	4
Dickstein et al*94	Neurology	Multiple sclerosis	USA	51 (12)	October 2007–August 2013	Descriptive cross-sectional	PHQ-9	39.4% (1507/3823)	4
He et al*61	Neurology	Neurology	China	40.1 (17.3)	April 2007–July 2007	Descriptive cross-sectional	HADS, GHQ-15	30.7% (216/704)	4
Mao et al*95	Neurology	Parkinson’s disease	China	NR	August 2010–June 2011	Descriptive cross-sectional	HAMD	56.2% (68/121)	2
Mao et al*95	Neurology	Essential tremor	China	NR	July 2009–June 2010	Descriptive cross-sectional	HAMD	53.2% (33/62)	2
Mitsikostas and Thomas96	Neurology	Headache	Greece	41 (7)	2007	Controlled cross-sectional	HAMD	3.4% (16/470)	4
Vogel et al97	Neurology	Systemic lupus erythematosus	Denmark	41.8 (9.6)	2010	Descriptive cross-sectional	MDI	22.8% (13/57)	4
Williams et al98	Neurology	Neurology	USA	51.1 (21)	January 2001–August 2001	Descriptive cross-sectional	PHQ-9	33.3% (161/483)	5
Worku et al99	Neurology	Parkinson’s disease	USA	57.10 (10.84)	June 2013–November 2013	Descriptive cross-sectional	QIDS-C16	57.4% (58/101)	4
Attah Johnson and Mostaghimi100	Dermatology	Neurodermatitis	Papua New Guinea	NR	1992	Descriptive cross-sectional	SRQ	50.7% (67/132)	4
Balieva et al101	Dermatology	Dermatology	Norway	50.1 (17.7)	November 2011–February 2013	Controlled cross-sectional	HDRS	13.3% (77/577)	5
Hon et al102	Dermatology	Atopic eczema	China	16.0 (NR)	May 2012–October 2012	Controlled cross-sectional	DASS-42, BDI	20.8% (25/120)	4
Mattoo et al*103	Dermatology	Vitiligo	India	30.11 (12.49)	March 1998–September 1999	Controlled cross-sectional	GHQ, CPRS	22.1% (25/113)	4
Mattoo et al*103	Dermatology	Psoriasis	India	40.91 (14.26)	March 1998–September 1999	Controlled cross-sectional	GHQ, CPRS	29.1% (30/103)	4
Rasoulian et al104	Dermatology	Dermatology	Iran	NR	September 2007–December 2007	Controlled cross-sectional	HADS	70.1% (101/144)	4
Roca et al105	Dermatology	Scleroderma	USA	NR	NR	Descriptive cross-sectional	BDI	64.8% (35/54)	3
Singh et al106	Dermatology	Psoriasis	India	NR	January 2013–November 2013	Descriptive cross-sectional	PHQ-9	39.4% (41/104)	3
Tsintsadze et al107	Dermatology	Skin diseases	Ukraine	NR	NR	Descriptive cross-sectional	HDRS	56.2% (119/211)	5
Daaleman et al108	Primary care	NR	USA	57 (NR)	NR	Descriptive cross-sectional	ZUNG	2.9% (15/509)	4
Drayer et al109	Primary care	Haemodialysis	USA	61.6 (12.6)	July 2002–June 2003	Descriptive cross-sectional	PRIME-MD, PHQ-9	27.4% (17/62)	4
Hankin et al110	Primary care	NR	USA	NR	June 1993–May 1995	Descriptive cross-sectional	CESD	31.3% (676/2160)	5
Hollifield et al111	Primary care	NR	Japan	50.5 (19.3)	NR	Descriptive cross-sectional	N.I.M.H.D	23.0% (29/126)	2
Ishikawa et al112	Primary care	NR	USA	40.2 (13.4)	January 2012–June 2012	Descriptive cross-sectional	ICPC-2	2.4% (29/1194)	4
Michalski et al113	Primary care	NR	Japan	NR	NR	Descriptive cross-sectional	DSM-III-R, BDI	5.7% (40/698)	4
Okumura et al114	Primary care	Headache	France	45 (NR)	April 2005–March 2009	Descriptive cross-sectional	DSM-IV	8.4% (35/418)	5
Rondet et al115	Primary care	NR	USA	NR	September 2010–December 2010	Descriptive cross-sectional	Structure interview	56.7% (142/250)	4
Steer et al116	Primary care	NR	USA	NR	NR	Descriptive cross-sectional	BDI	24.2% (29/120)	4
Tamburrino et al117	Primary care	NR	USA	42.7 (NR)	1996–1998	Descriptive cross-sectional	PRIME-MD PQ	27.3% (478/1752)	5
Mancuso et al118	Primary care	Pulmonary disease	UK	41 (11)	NR	Descriptive cross-sectional	GDS	45.2% (104/230)	5
Lee et al119	Ophthalmology	Eye disease	USA	NR	January 2001–March 2011	Descriptive cross-sectional	Structured interview	20.0% (10/50)	4
Rovner et al120	Ophthalmology	ARMD	USA	81.2 (5.8)	NR	Descriptive cross-sectional	HDRS	23.7% (49/206)	5
Evans et al121	Ophthalmology	Eye disease	UK	81.2 (4.8)	NR	Controlled cross-sectional	GDS-15	13.5% (235/1742)	4
Brody et al122	Ophthalmology	ARMD	USA	80.11 (6.21)	January 1998–September 1999	Descriptive cross-sectional	GDS, SI	32.5% (49/151)	4
Mathew et al123	Ophthalmology	ARMD	Australia	78.0 (7.7)	December 2001–July 2005	Controlled cross-sectional	GAD	44.4% (64/145)	3
Popescu et al124	Ophthalmology	ARMD, FCD, glaucoma	Canada	80.5 (7.5)	September 2009–December 2011	Controlled cross-sectional	GDS-15	24.8% (78/315)	4
Goto et al125	E.N.T.	Otolaryngology	Japan	53.7 (18.9)	January 2006–December 2006	Descriptive cross-sectional	SDS	54.4% (49/90)	4
Asghari et al126	E.N.T.	OSAS	Iran	47.63 (11.73)	August 2008–December 2012	Descriptive cross-sectional	BDI	41.3% (293/685)	4
Lee et al127	E.N.T.	Empty nose syndrome	China	51.6 (NR)	2012–2014	Analytical cohort-study	BDI	71.4% (15/21)	4
Olley et al128	Infectious disease	HIV/AIDS	South Africa	NR	NR	Descriptive cross-sectional	MINI	34.9% (52/149)	4
Kolaric et al129	Infectious disease	HIV/AIDS	Croatia	39 (NR)	March 2003–April 2003	Controlled cross-sectional	BDI	20.0% (16/80)	3
Chan et al130	Infectious disease	HBV	China	47.2 (11.9)	October 2008–June 2009	Descriptive cross-sectional	GHQ	17.4% (26/149)	4
Wright et al131	Infectious disease	HIV/AIDS	Australia	NR	July 2005–March 2006	Descriptive cross-sectional	CES-D	36.4% (235/645)	4
Judd et al132	Infectious disease	HIV/AIDS	Australia	44.7 (9.1)	NR	Descriptive cross-sectional	BDI	34.8% (45/129)	3

*Studies that extracted different data from the same research.

ARMD, age-related macular degeneration; BASDEC, Brief Assessment Schedule Depression Cards; BDI, Beck Depression Inventory; CES-D, Centre for Epidemiologic Studies Depression Scale; CIDI, Composite International Diagnostic Interview; CPRS, Comprehensive Psychopathological Rating Scale; DASS-42, Chinese versions of the Depression, Anxiety, Stress Scales; DSM, Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition); E.N.T., otorhinolaryngology department (ear, nose, throat); EPDS, Edinburgh Postnatal Depression Scale; FCD, Fuchs Corneal Dystrophy; GADS, Goldberg Anxiety and Depression Scale; GDS, Geriatric Depression Scale; GHQ, General Health Questionnaire; HADS, Hospital Anxiety and Depression Scale; HDRS, Hamilton Depression Rating Scale; ICPC-2, International Classification of Primary Care, Second Edition; MADRS, Montgomery-Asberg Depression Rating Scale; MDI, Major Depression Inventory; MINI, MINI International Neuropsychiatric Interview; N.I.M.H.D, N.I.M.H. Diagnostic Interview Schedule; NOS, Newcastle-Ottawa Score; NR, not reported; OG, obstetrics and gynaecology; OSAS, obstructive sleep apnea syndrome; PHQ-15, 15-item Patient Health Questionnaire; PHQ-9, 9-item Patient Health Questionnaire; PRIME-MD PQ, Primary Care Evaluation of Mental Disorders Patient Questionnaire; QIDS-C16, Quick Inventory of Depressive Symptomatology; SDS, Self-Rating Depression Scale; SF-36, Short F; SRQ, Self-Rating Questionnaire; T1DM, type 1 diabetes mellitus; WHO-5, World Health Organization-5 Well-Being Index.

Figure 2.

Forest plot of the prevalence of depression or depressive symptoms among outpatients. E.N.T., ear, nose, throat.

Figure 3.

Bar graph of meta-analysis of the prevalence of depression or depressive symptoms among outpatients stratified by age and study year. (A) Prevalence of depression or depressive symptoms among outpatients stratified by year of study publication. (B) Prevalence of depression or depressive symptoms among outpatients stratified by age.

Figure 4.

Forest plot of the eight studies that included control groups.

Regarding the screening instruments used, higher prevalence estimates were found among studies that used BDI (1316/4702, 36.0%, 95% CI 29.0% to 44.0%, I²=94.8%) compared with those that used HADS (1003/2025, 22.0%, 95% CI 12.0% to 35.0%, I²=96.6%). A meta-regression analysis was conducted to evaluate the potential heterogeneity in combined subtypes to assess study year, country and specialty as sources of heterogeneity. In this analysis, country, study year and specialties represented a small amount of heterogeneity (p=0.0327, <0.05). Funnel plots and tests indicated evidence of publication bias (Egger’s test, p<0.001, figure 5). Sensitivity analyses, in which the meta-analysis was serially repeated after excluding each study, suggested that no individual study affected the overall prevalence estimate by more than 1% (online supplementary table 1).

Figure 5.

Funnel plots to test the publication bias of the 8 studies that included control groups. Each point represents a separate study on the indicated association. The points were distributed asymmetrically, indicating the existence of publication bias.

Discussion

We performed a systematic review and meta-analysis to best estimate the prevalence of depression and depressive symptoms in different clinical departments. Overall, the prevalence of depression or depressive symptoms among outpatients was 27.0%, ranging from 17.0% to 53.0% in different clinical departments. This study found that outpatients from otolaryngology clinics had the highest prevalence of depression (53.0%). Depression was found to be an important mediator for otolaryngologic conditions such as chronic tinnitus.²⁷ It was not surprising that dermatology ranked the second highest and 39.0% of outpatients from dermatology clinics suffered from depression. Atopic dermatitis was found to be associated with depression because the skin stigmata often causes embarrassment, low confidence and sadness.^{10 28} Atopic dermatitis is one of the most common dermatological disorders and was found to be associated with negative impact on the quality of life of patients, families and caregivers.^{29 30} There is psychoneuroimmunology connection between depression and medical illness.³¹ The production of pro-inflammatory cytokine (eg, IL-6) was found to be higher in patients with atopic dermatitis³² and IL-6 was found to be raised in patients with depression.³³ Raised IL-6 may cause depression in patient with atopic dermatitis. This study found that 35% of outpatients from neurology clinic suffered from depression. Genetic factors and autoantibodies play an important role in causing neuropsychiatric complications including depression.^{34 35} Stroke is a common neurological disorder and causes significant health burden.³⁶ The burden of stroke causes depression in both stroke patients and their caregivers.³⁷ Novel rehabilitation intervention targeting at motor deficit was designed to improve functional status and quality of life of patients with stroke.³⁶ This intervention might offer hope and reduce prevalence of depression in patients with stroke. Our study confirmed previous findings of the higher prevalence of depression or depressive symptoms in outpatients than in the general public.^3–6 The prevalence of depression/depressive symptoms in outpatients slightly decreased from 1990 to 2010. This decrease may be due to the potentially improved recognition of the illness and increased awareness for seeking help among the general public. However, this explanation has yet to be confirmed with population-based research. Depression or depressive symptoms are often overlooked during daily medical care by busy professionals without specific training in mental health, and our findings suggests that specialists should focus on patients’ physical problems and their mental problems. We should enhance the awareness of mental disorders during medical works and strengthen the communication between doctors and patients. Depression is expected to vary throughout the life course, as ageing is a risk factor for the development of depression and depressive symptoms. In this study, the distribution of age-related depression had two peaks and varied in different groups. Outpatients aged 30–40 years old had a similar depression prevalence as outpatients aged 80–90 years old, with rates ranging between 30.0% and 40.0%. However, previous research on the association between age and depression has shown contradictory patterns.^38–40 Klerman noted a particular emergence of childhood depression and an increase in suicide attempts and death among adolescents and young adults.³⁹ Outpatients aged 30–40 years suffering from chronic medical illnesses are at higher risk for developing depression. Depressed outpatients might develop maladaptive rumination and illness perception towards their chronic medical illness.⁴¹ Chronic medical illness may increase the risk of suicide in adult outpatients because psychosomatic complaint such as headache was found to be an important risk factor for suicide in adults.⁴² Yang showed that depression declined with age.⁴⁰ By contrast, Jorm revealed that there was no consistent pattern across studies regarding age differences in the occurrence of anxiety, depression or distress.³⁸ Our results showed that the prevalence of depression and depressive symptoms peaked among individuals aged 30–40 years and 80–90 years, consistent with the U-shaped ageing trajectory of depression reported by a previous study.⁴³ It has been suggested that depression reaches its highest level in elderly aged 80 years or older because physical dysfunction and low personal control add to personal and status losses.⁴³ Risk factors of geriatric depression include poor health, brain injury, low folate and vitamin B12 and raised plasma homocysteine levels.⁴⁴ The association between depression and chronic medical illnesses in elderly is due to accompanying poor self-reported health and functional status.⁴⁵ Further, history of depression and antidepressant treatment are important risk factors for elderly suicide.⁴⁶ The prevalence of depression and depressive symptoms was found to be higher (p<0.0001) in developing countries (24.0%) than in developed countries (33.0%), a greater difference and much higher than the 12-month prevalence estimates in developed (5.5%) and developing countries (5.9%) found in Kessler's study.⁴⁷ However, a possible limitation of this finding is that Kessler's study was restricted to a small number of countries, a narrower range of severe patients and a shorter research time.⁴⁷ More specifically, 13 studies from China were included in the present meta-analysis. The prevalence of depression or depressive symptoms among Chinese outpatients was 27.0% (1941/7194, 95% CI 22.0% to 33.0%, I²=95.4%), which fell between the prevalence observed in developing and developed countries (24.0%–33.0%) and consistent with China’s national development.

Various factors may account for the heterogeneity in this meta-analysis. First, differences in the assessments instruments and cut-offs may have affected the diagnostic sensitivity and specificity. Modified diagnostic criteria for depression and depressive symptoms have been proposed for use in different health settings, but there is no consensus regarding the optimal diagnostic approach. Whether the existing diagnostic criteria are ideal in different health settings remains to be determined. Additionally, little attention has been devoted to the ICD (International Classification of Diseases) criteria, in which a depressive episode is defined based on the number and severity of the symptoms only.⁴⁸ Second, heterogeneity between individual studies existed due to the different diagnostic methods applied in different countries. Third, the study qualities varied. For example, some studies used screening instruments with non-standard methods (eg, with cut-off scores that have not been validated) or having different thresholds in depression measurements that may increase the errors of prevalence estimates. These variations were captured in part by the modified Newcastle-Ottawa score, which assessed the risk of bias in each study.

Publication bias was assessed in this review. First, the exclusion of non-English publications likely contributed to the bias in our analysis. However, given the large number of included studies, we would not expect missed studies to significantly affect the findings. Second, because of the nature of the specialty, some studies examined the rates of depression in females only. For example, the prevalence of depression in obstetrics and gynaecology departments was 25.0%, which may have caused selection bias. Also, the ageing of the population phenomenon may have a more profound impact of depression estimates in developed counties comparing with developing countries. Third, the estimates of prevalence in some specialties were based on an inadequate number of studies, which may have affected the accuracy of the overall depression prevalence. For instance, the prevalence of depression in otorhinolaryngology departments was 53.0%, which was calculated using data from only three studies. Third, studies with fewer participants generally yielded more extreme prevalence estimates, further suggesting the presence of publication bias.⁴⁹ The study quality is also an important factor for evaluating the presence of publication bias. However, the sensitivity analysis showed that no individual study affected the overall prevalence estimate.

Limitations of this meta-analysis

Limitations should be considered when interpreting the results of this study. First, the substantial heterogeneity between studies was not fully explained by the variables examined. For example, various disease categories, the onset of depressive episode, medical expenses, medical workers’ attitudes and patients’ race and gender may contribute to the risk of depressive symptoms among outpatients. Furthermore, compared with self-report scales, interview methods commonly underestimate the prevalence of psychiatric disorders.⁵ Second, the major update from the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) to the fifth edition in 2013 may have affected the accuracy of the prevalence estimates.^{50 51} The severity indicator for depression and depressive symptoms was updated to be more precise, and this change may explain the increase in prevalence between 2010 and 2016. Third, the data were collected from studies that used different cross-sectional study designs, including different diseases and sample sizes. For example, in otorhinolaryngology departments, only three diseases were included in the meta-analysis, leading to low representativeness; however, given the available publications in the database, we would not expect this limitation to significantly affect the findings. Another limitation is the paucity of longitudinal data, which decreased the generalisability of the study outcomes. Therefore, high-quality studies that use cohort study designs to conduct follow-ups of depression might provide more precise outcomes.⁵² Fourth, using a single measure to assess depression and depressive symptom might improve the accuracy and sensitivity of the outcomes. Finally, focusing on depression alone is insufficient. Depression and depressive symptoms with other mental disorders remain an important and overlooked complication among outpatients, and this oversight calls for a more systematic approach to clinical assessment and follow-up. In conclusion, this systematic review and meta-analysis highlighted the overall high prevalence of depression and depressive symptoms, which may have long been overlooked in outpatients worldwide. Our study also provided substantial quantitative subgroup analyses that laid the foundation for researchers, clinicians and policy makers to develop effective strategies for depression management.

Conclusion

In summary, our study has several implications for clinical practice. First, we performed a systematic review and meta-analysis to estimate the prevalence of depression and depressive symptoms in different clinical departments. Second, our results suggest that more attention should be devoted to outpatient mental health, particularly in clinical departments with a high depression prevalence (eg, outpatients at otolaryngology clinics had the highest prevalence of depression (53.0%)). The inconsistency of the findings across different specialties regarding the prevalence of outpatients with depression could help modify and improve clinical guidelines for the evaluation and diagnosis of depression or depressive symptoms in different medical settings. Third, we identified that different screening instruments produce different estimates, and these findings may provide a reliable reference for developing an effective and unified measurement for diagnosing depression. Fourth, the substantial heterogeneity between studies was not fully explained by the variables examined.