The prevalence and the correlates of mental disorders among the elderly population: results from China Mental Health Survey

Abstract

Background

Mental disorders among the elderly are a growing public health concern contributing significantly to disease burden, disability, and mortality. However, there is a lack of nationally representative studies examining the prevalence and correlates of mental disorders among older adults in community settings. Considering 55 years old is the beginning of “young old”, our study targets the population of adults aged 55 years old and above. Using data from the China Mental Health Survey (CMHS), we aim to estimate the prevalence and distribution of mental disorders and to investigate the correlates of mental disorders.

Methods

Data of study was derived from the CMHS, a nationally representative community-based epidemiological survey. CMHS employed Composite International Diagnostic Interview (CIDI), a structured diagnostic tool, to collect relevant data. A total of 12,667 adults aged 55 and above were included in this survey. Weighted prevalence estimates were calculated, and design-corrected Rao-Scott χ² test, along with logistic regression model were used to identify correlates of mental disorders.

Results

A total of 10,840 participants (85.6%) completed the CIDI. The lifetime and 12-month prevalence of mental disorders among the Chinese elderly population were 19.16% and 10.62%, respectively. Anxiety disorders were the most prevalent mental disorders, with a lifetime prevalence and 12-month prevalence was 9.07% and 5.97%, respectively. The corresponding data for mood disorders were 8.19% and 4.36%, and for substance use disorders were 4.16% and 0.89%, respectively. Having ≥ 3 physical diseases (OR = 3.22, 95% CI: 2.35–4.40), experiencing chronic pain (OR = 2.94, 95% CI: 1.77–4.90), and having sleep disturbances (OR = 4.02, 95% CI: 3.14–5.13) were all significantly associated with higher odds of mental disorders. Conversely, individuals aged 70 years and older had significantly lower odds of mental disorders (OR = 0.42, 95% CI: 0.29–0.62). All associations were statistically significant (p < 0.05).

Conclusion

Mental disorders are highly prevalent among Chinese adults aged 55 years and above, with anxiety disorders, mood disorders, and substance use disorders being the most common. Chronic disease, chronic pain and sleep disturbances played an important role in the risk of mental disorders mental disorders.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-23993-6.

Introduction

The world is undergoing a significant demographic shift, characterized by a growing proportion of elderly individuals in the population. By 2050, it is projected that the global population will officially transition into an"ageing society"[1]. The elderly population is predicted to surge from 1.05 billion in 2020 to 2.1 billion in 2050, globally outnumbering the younger population aged 10–24 years (2 billion) [2]. China has already entered an aging society, with the elderly accounting for 18.7% of the total population in 2020 [3]. The rapidly growing elderly population in China poses significant public health challenges in addressing their physical and mental health needs [4, 5].

On the mental health front, the elderly population often faces mood disorders, anxiety disorders and substance use disorders [6]. Previous studies have shown that the global prevalence of anxiety and depressive disorders ranges from 2 to 15% in community-dwelling elderly adults [7] and the prevalence of substance use disorders increased from 26% in 1990 to 47% in 2016 [8]. Furthermore, mental disorders contribute significantly to the global disease burden [9] but also the disability and mortality [10]. Depressive disorders rank as the third leading cause of global disease burden [11]. Given these findings, studies focusing on the prevalence of mental disorders among the elderly population and their correlates are essential to mitigating their adverse effects. Previous studies have shown that several socio-demographic [12], psychological and physical factors [13] are related to mental disorders among elderly population. For example, education level, female gender, older age, low income, chronic diseases, and sleep disorders have been linked to depressive disorders [14], anxiety disorders [15] and substance use disorders [16]. However, the findings across these studies are inconsistent, necessitating further investigation, especially in China, where no nationally representative data has been reported in this field.

Entering the twenty-first century, rapid economic growth in terms of higher GDP per capita has been associated with a deterioration in mental health outcomes [17]. Nonetheless, epidemiological investigations of mental disorders in China have predominantly focused on specific symptoms [18] or have been largely localized. For instance, a survey conducted among the elderly population in Tianjin identified gender, age, marital status, education level, and economic status as significant influencing factors of mental disorders [12]. The China Health and Retirement Longitudinal Study (CHARLS) primarily explored the risk factors of depressive symptoms rather than mental disorders among the elderly population [19]. As a result, there is an urgent need for a population-based study focusing on mental disorders among the elderly to provide a comprehensive understanding of their prevalence and correlates in China. China Mental Health Survey (CMHS) [20] is a nationally representative survey, which depicted the prevalence of mental disorders and their related factors among adults using the Composite International Diagnostic Interview (CIDI) 3.0. In this article, we focus on the elderly population from CMHS to investigate the mental health status of adults aged 55 and above in China, providing a theoretical basis for policy formulation and intervention strategies.

Studies have established the age of 55 as the onset of “young old” [21], identifying this age group as particularly vulnerable to mental disorders, with the prevalence of such disorders increasing after this age. For example, a recent analysis revealed that the prevalence of anxiety disorders in Chinese men steadily rises from ages 20 to 60, peaking at 55–59 years [22]. Similarly, an epidemiological survey conducted across four Chinese provinces found higher rates of mood and anxiety disorders among adults aged 55 and above compared to younger individuals [23]. Moreover, China’s retirement policy, which mandates retirement at ages 50–55 for women and 60 for men, exacerbates mental health risks [24]. Physical health decline, which often accompanies aging, is another major contributor to mental health challenges [25]. Chronic illnesses increase sharply after age 55, further compounding these risks. A 10-year longitudinal study revealed that women aged 55–65 experience significant declines in physical health and functional capacity, which are closely linked to depressive and anxiety symptoms [26]. Therefore, guided by the World Health Organization’s Decade of Healthy Aging (2021–2030) [27], which emphasizes addressing early-onset aging risks in vulnerable groups, and informed by Asian research precedents that utilize 55 + as a threshold for aging-related mental health trends [28], our study focuses on adults aged 55 and above in China. This study aims to map the prevalence of mental disorders in this population and identify associated risk factors to inform evidence-based interventions.

Methods

Sample and procedures

Data for this study were derived from the CMHS, the first nationally representative community-based cross-sectional epidemiological survey in China. Conducted between July 2013 and March 2015, it comprehensively documented the prevalence and distribution of mental disorders nationwide. CMHS was conducted among adults aged 18 and above from 31 provinces, autonomous regions and municipalities across China (excluding Hong Kong, Macao and Taiwan). The sampling design adopted the national disease surveillance points (DSPs) framework, employing a multi-stage disproportionate stratified approach. The seven-stage process included: (1) selecting DSPs, (2) selecting streets/towns from DSPs using Probability Proportionate to Size Sampling (PPS), (3) selecting communities/villages from selected streets/towns using PPS, (4) selecting CMHS communities/villages from DSP communities via simple random sampling, (5) selecting DSP households randomly, (6) selecting CMHS households from residential groups, and (7) selecting final respondents using Kish table sampling. After seven stages of sampling design, CMHS selected 1256 communities or villages from 628 streets or towns in 157 districts or counties. Specific sampling methods and procedure had been described elsewhere [20, 29, 30].

Ultimately, a total of 12,667 individuals aged 55 and older were enrolled in the study (Fig. 1). CMHS trained lay interviewers to administer the Composite International Diagnostic Interview (CIDI) for diagnosing non-psychotic disorders, utilizing face-to-face computer-assisted personal interviews (CAPI) with all respondents. The research protocol of CMHS was approved by the Institutional Review Board (IRB) of the Sixth Hospital of Peking University, Beijing, China (Approval No. IMH-IRB-2013–13-1). All participants provided written informed consent before their enrollment in the study.

Fig.1.

Selection process of the sample. DSP = Disease Surveillance Points. CIDI = Composite International Diagnostic Interview 3.0

Instruments

The Composite International Diagnostic Interview (CIDI) 3.0 [28]

CIDI 3.0 is a fully structured diagnostic instrument consisting of two parts. Part one (completed by all participants) was used to assess most of the non-psychotic disorders (including mood disorders, anxiety disorders excluding post-traumatic stress disorder and anxiety disorder not otherwise specified, substance-use disorders, impulse-control disorders, and eating disorders) [21]. Part two was administered to respondents who met lifetime criteria or sub-threshold criteria for disorders identified in part one, as well as to a randomly selected 25% of participants who screened negative. Details regarding parts one and two are provided in Additional File 1. CMHS utilized the Chinese version of the CIDI, with minor modifications to the interview schedule to enhance respondent comfort and improve communication. Lifetime and 12-month prevalence rates for non-psychotic disorders were generated using computerized algorithms based on DSM-IV criteria. Part one also collected data on the treatment of mental disorders and suicidal behaviors, while part two gathered pharmacological and socio-demographic information, as well as data on other related factors.

Risk factors or correlates

Socio-demographic factors including age, gender, education, marital status, income, residential area, region was collected in CIDI. Information on the lifetime presence of 12 chronic diseases was obtained by asking participants whether they had any of the following conditions: arthritis or rheumatism, seasonal allergies, stroke, heart disease, high blood pressure, asthma, tuberculosis, chronic lung disease, diabetes, ulcers, HIV or AIDS, epilepsy, or cancer. In addition, participants were assessed for accidents, injuries, or poisoning incidents within the past 12 months, and the impacts of these events were also recorded. Chronic pain was defined as the presence of frequent and severe headaches or other forms of chronic pain within the past 12 months. Finally, participants were asked about the presence of sleep disturbances over the past 12 months and the effects of these disturbances.

Quality control and consent

Each step of the study underwent strict quality control, including subjects’ selections, interviewers training, data collection, and interviews. During data collection, quality control was ensured through data verification, audio verification, telephone verification, and field verification. The quality control personnels had received rigorous training to maintain high data integrity [20, 29, 30]. All participants provided oral informed consent prior to their participation, with approval granted by the ethics committees of the Sixth Hospital of Peking University and Tianjin Anding Hospital.

Statistical analysis

This study used SAS9.4 to estimate the weighted lifetime and 12-month prevalence of mental disorders. Sampling weights were calculated as the reciprocal of the selection probability at each level of the multi-stage sampling process. The combined sampling weight was the product of weights across seven levels: county or district, town or street, primary village or community, secondary village or community, residential group, household, and individual. Non-response weights were determined as the reciprocal of the non-response rate, estimated using logistic regression models to adjust for factors associated with non-participation. Post-stratification weights were applied to align the sample distribution with the national population of individuals aged 55 years and older. These weights were based on gender (male or female), age groups (55–59 years, 60–69 years, 70–79 years, and 80 years or older), and urbanization level (rural or urban), using data from the sixth national population census of 2010 as the reference standard. For questionnaire items with low missing rates, missing data were imputed using median values observed in the dataset. Extreme post-stratification weights were trimmed to the 0.01 and 0.99 quantiles to minimize variance. The final weight for each participant was computed as the product of sampling weights, non-response weights, post-stratification weights, and adjustments for extreme weights. The weighted methods used in this study have been previously described [30]. The frequency and percentage of categorical variables were calculated. Design-corrected Rao-Scott χ2 test was used to examine the relationship between gender, age or area and mental disorders. To identify correlates of mental disorders in the elderly population, we first conducted univariable logistic regression analyses to examine bivariate associations between candidate variables and mental disorders. Variables were selected based on established theoretical frameworks in mental health epidemiology, including age, gender, education level, urban–rural residence, geographic region, marital status, income, sleep disturbances, chronic pain, and chronic physical diseases. All variables assessed in univariable analyses were subsequently included in a multivariable logistic regression model to control for potential confounding. Bonferroni correction was used to adjust for multiple testing. Odds ratios (ORs) and 95% confidence intervals (CIs) were reported, with statistical significance at P < 0.05. Only mental disorders evaluated in Part One of CIDI were included in the logistic regression models. This decision was necessitated by the fact that the majority of participants did not undergo CIDI part two assessments, and applying uniform weighting would have significantly diminished the sample size. A P-value < 0.05 was considered statistically significant.

Results

Demographic characteristics of participants

A total of 12,667 adults aged 55 years and above were initially selected for this survey. Of these, 10,840 participants (85.58%) completed the Composite International Diagnostic Interview (CIDI) and were included in the analysis. The demographic characteristics of these 10,840 participants are presented in Table 1.

Table 1.

Sociodemographic characteristics of the sample (N = 10,840)

Characteristics	Frequency	Proportion (%)
Age
55–59	3386	31.24
60–64	3030	27.95
65–69	1992	18.38
70 and over	2432	22.44
Gender
Female	5591	51.58
Male	5249	48.42
Resident area
Rural	5796	53.47
Urban	5044	46.53
Region
Eastern	3877	35.77
Central	3818	35.22
Western	3145	29.01
Education levela
Illiterate/below primary school	5062	46.76
Primary school	2392	22.09
Junior high school	2088	19.29
Senior high school and above	1283	11.85
Marital status
Married or cohabitating	8944	82.51
Othersb	1896	17.49
Household incomec
Low	4400	40.59
Middle	3598	33.19
High	2842	26.22

^aData missing from 11 respondents

^bOthers included never married, separated, widowed, and divorced

^cThe household income was categorized into three levels based on yearly income percentiles: low (< 10,500 yuan), middle (10,500–42,500 yuan), and high (42,500 yuan and above)

The lifetime and 12-month prevalence of mental disorders

The lifetime and 12-month prevalence of any disorders are shown in Table 2. Overall, the weighted lifetime prevalence of any disorders was 19.16% (95%CI: 16.51-0.21.82). Anxiety disorders were the most prevalent (9.07%), followed by mood disorders (8.19%) and substance use disorders (4.16%). The weighted 12-month prevalence of any disorders was 10.62% (95%CI: 8.75–12.50). The 12-month prevalence rates were 5.97% for anxiety disorders, 4.36% for mood disorders, and 0.89% for substance use disorders.

Table 2.

Unweighted and weighted lifetime and 12-month prevalence of mental disorders (n = 10,840)

Mental disorders	Lifetime prevalence			12-month prevalence
	Frequency, n	Unweighted%(95% CI)	Weighted % (95% CI)	Frequency, n	Unweighted%(95% CI)	Weighted % (95% CI)
Mood disorders	864	7.97 (7.46–8.48)	8.19 (6.82–9.57)	448	4.13 (3.76–4.51)	4.36 (3.57–5.16)
Anxiety disordersa	720	16.04 (14.87–17.22)	9.07 (7.47–10.66)	500	11.50 (10.47–12.52)	5.97 (4.88–7.07)
Substance use disorders	428	3.95 (3.58–4.32)	4.16 (3.47–4.86)	100	0.92 (0.74–1.10)	0.89 (0.57–1.20)
Impulse-control disorders	122	1.13 (0.93–1.32)	1.17 (0.73–1.61)	81	0.75 (0.6–0.91)	0.79 (0.45–1.14)
Eating disorders	5	0.05 (0.01–0.09)	0.02 (0.00–0.04)	2	0.02 (0.00–0.04)	0.01 (0.00–0.02)
Any mental disordersa	1658	33.21 (30.22–36.19)	19.16 (16.51–21.82)	931	20.80 (18.69–22.91)	10.62 (8.75–12.50)

^aAs this disorder involved CIDI part one and part two, unweighted prevalence was calculated using probability of selectively entering part two

12-month prevalence in different genders, ages and areas

Table 3 displays the 12-month prevalence in different genders. The overall 12-month prevalence of any disorders was 9.55% in males and 11.71% in females, with no statistically significant difference observed between genders. However, males were more likely to have substance use disorders than females (P < 0.001), with the prevalence of 1.51% and 0.26%, respectively.

Table 3.

Weighted 12-month prevalence of mental disorders by gender (n = 10,840)

Mental Disorders	MALE		FEMALE		Pa
	Frequency, n	Prevalence% (95% CI)	Frequency, n	Prevalence%(95% CI)
Mood disorders	169	3.47 (2.57–4.37)	279	5.27 (3.92–6.62)	0.175
Anxiety disordersb	221	5.27 (4.08–6.46)	279	6.69 (5.17–8.21)	0.525
Substance use disorders	81	1.51 (0.93–2.08)	19	0.26 (0.07–0.45)	< 0.001
Impulse-control disorders	53	1.07 (0.45–1.69)	28	0.51 (0.12–0.90)	0.973
Eating disorders	2	0.01(0.00–0.03)	0	< 0.001c	NA
Any mental disordersb	427	9.55 (7.23–11.87)	504	11.71 (9.05–14.36)	0.192

^aP values are Bonferroni-corrected, P > 1 is shown as P = 1

^bAs this disorder involved CIDI part one and part two, unweighted prevalence was calculated using probability of selectively entering part two

^cPrevalence 95% CI could not be calculated when the frequency was equal to 0. NA Not applicable

As shown in Table 4, the weighted 12-month prevalence of any disorders varied significantly by age group. The prevalence for individuals aged 55–59 years, 60–64 years, 65–69 years, and ≥ 70 years were 11.83%, 12.74%, 10.31%, and 7.75%, respectively. The prevalence of anxiety disorders was 5.92%, 8.56%, 5.45% and 4.27% in the age groups of 55–59 years, 60–64 years, 65–69 years and ≥ 70 years, respectively (P = 0.070). Additionally, Substance use disorders were most prevalent among individuals aged 55–59 years (1.57%) and decreased progressively with age, with prevalence rates of 1.03%, 0.44%, and 0.28% for those aged 60–64 years, 65–69 years, and ≥ 70 years, respectively (P < 0.001).

Table 4.

Weighted 12-month prevalence of mental disorders by age (N = 10,840)

Mental Disorders	55–59 YEARS		60–64 YEARS		65–69 YEARS		≥ 70 YEARS		Pa
	Frequency n	prevalence%(95% CI)	Frequency n	prevalence%(95% CI)	Frequency n	prevalence%(95% CI)	Frequency n	prevalence%(95% CI)
Mood disorder	131	4.50 (3.31–5.69)	146	5.06 (3.86–6.25)	93	4.48 (3.21–5.75)	78	3.61 (2.44–4.78)	1
Anxiety disorderb	172	5.92 (4.17–7.67)	151	8.56 (6.10–11.02)	94	5.45 (3.63–7.26)	83	4.27 (2.57–5.98)	0.070
Substance use disorders	43	1.57 (0.84–2.29)	32	1.03 (0.42–1.65)	13	0.44 (0.10–0.66)	12	0.28 (0.04–0.52)	< 0.001
Impulse-control disorders	36	1.60 (0.62–2.58)	30	0.90 (0.47–1.32)	11	0.29 (0.08–0.52)	4	0.09 (0.00–0.23)	NA
Eating disorders	0	< 0.001c	1	0.01 (0.00–0.02)	0	< 0.001 c	1	0.02 (0.00–0.05)	NA
Any mental disordersb	317	11.83 (8.76–14.91)	283	12.74 (10.20–15.28)	178	10.31(7.65–12.96)	153	7.75 (5.68–9.83)	0.004

^aP values are Bonferroni-corrected, P > 1 is shown as P = 1

^bAs this disorder involved CIDI part one and part two, unweighted prevalence was calculated using probability of selectively entering part two

^cPrevalence 95% CI could not be calculated when the frequency was equal to 0. NA Not applicable

The 12-month prevalence in different areas is shown in Table 5. There was no statistically significant difference in the prevalence of any disorders between different regions, with a prevalence of 11.21% in urban areas and 10.14% in rural areas. Similarly, there was no statistically significant difference in the prevalence of other mental disorders by region.

Table 5.

Weighted 12-month prevalence of mental disorders by region (N = 10,840)

Mental Disorders	URBAN		RURAL		Pa
	Frequency, n	Prevalence%(95% CI)	Frequency, n	Prevalence%(95% CI)
Mood disorders	183	3.95 (3.05–4.86)	265	4.71 (3.62–5.80)	1
Anxiety disordersb	205	6.06 (4.48–7.63)	295	5.91 (4.50–7.32)	1
Substance use disorders	42	0.82 (0.36–1.28)	58	0.94 (0.58–1.31)	1
Impulse-control disorders	35	0.75 (0.39–1.10)	46	0.83 (0.34–1.33)	1
Eating disorders	1	0.00 (0.00–0.00)	1	0.01 (0.00–0.02)	1
Any disordersb	385	11.21 (8.65–13.76)	546	10.14 (7.96–12.31)	0.455

^aP values are Bonferroni-corrected, P > 1 is shown as P = 1

^bAs this disorder involved CIDI part one and part two, unweighted prevalence was calculated using probability of selectively entering part two

The correlates of 12-month prevalence of mental disorders

Table 6 presents the correlates of the 12-month prevalence of mental disorders. In the multivariable analysis, having ≥ 3 physical diseases (OR = 3.22, 95% CI: 2.35–4.40), experiencing chronic pain (OR = 2.94, 95% CI: 1.77–4.90), and having sleep disturbances (OR = 4.02, 95% CI: 3.14–5.13) were associated with higher odds of mental disorders. Conversely, individuals aged 70 years and older were found to have lower odds of mental disorders (OR = 0.42, 95% CI: 0.29–0.62).

Table 6.

Association between demographic, health status and 12-month prevalence of mental disorders^a

Characteristics	Univariable,		Multivariable
	OR (95% CI)	P	OR (95% CI)	P
Age
55–59	1		1
60–64	1.025(0.79–1.34)	0.8532	0.89(0.67–1.18)	0.421
65–69	0.82(0.60–1.10)	0.1834	0.66(0.47–0.92)	0.015
≥ 70	0.58(0.42–0.79)	0.0010	0.42(0.29–0.62)	< 0.001
Gender
Female	1		1
Male	0.88(0.70–1.11)	0.2749	1.01(0.79–1.30)	0.945
Region
Eastern	1		1
Central	1.21(0.79–1.85)	0.3681	1.12(0.75–1.67)	0.577
Western	0.95(0.64–1.41)	0.8016	0.82(0.58–1.16)	0.269
Resident area
Rural	1		1
Urban	1.14(0.90–1.44)	0.2865	1.20(0.94–1.52)	0.144
Education level
Illiterate/below primary school	1		1
Primary school	1.10(0.88–1.38)	0.3971	1.19(0.93–1.52)	0.174
Junior high school	1.25(0.92–1.70)	0.1452	1.38(0.97–1.98)	0.077
Senior high school and above	1.24(0.85–1.83)	0.2647	1.36(0.89–2.06)	0.153
Marital status
Married/Cohabitating	1		1
Others	1.00(0.74–1.35)	0.9823	1.13(0.81–1.58)	0.483
Household incomeb
Low	1		1
Middle	0.91(0.73–1.14)	0.4083	0.92(0.71–1.18)	0.492
High	0.93(0.68–1.28)	0.6595	1.04(0.77–1.41)	0.810
The number of chronic disease
None	1		1
1–2	1.93(1.42–2.61)	<.0001	1.39(0.98–1.97)	0.065
≥ 3	5.87(4.33–7.95)	<.0001	3.22(2.35–4.40)	< 0.001
Chronic pain
No	1		1
Yes	7.26(4.42–11.93)	<.0001	2.94(1.77–4.90)	< 0.001
Sleep disturbances
No	1		1
Yes	5.43(4.29–6.86)	<.0001	4.02(3.14–5.13)	< 0.001

^aMental disorders only included disorders involved in CIDI part one (including mood disorders, anxiety disorders excluding post-traumatic stress disorder and anxiety disorder not otherwise specified, substance-use disorders, impulse-control disorders, and eating disorders)

^bThe household income was categorized into three levels based on yearly income percentiles: low (< 10,500 yuan), middle (10,500–42,500 yuan), and high (42,500 yuan and above)

Discussion

According to our findings, the lifetime prevalence of any mental disorders was 19.16%, while the 12-month prevalence was 10.62%. Anxiety disorders were the most prevalent mental disorders, followed by mood disorders and substance use disorders. Furthermore, the distribution of mood, anxiety, and substance use disorders varied across different genders and age groups. Age, the number of chronic diseases, chronic pain, and sleep disturbances were significantly associated with the 12-month prevalence of any mental disorders.

After entering the twenty-first century, studies that comprehensively investigate the prevalence of mental disorders among the elderly in China remain remarkably scarce. Studies examining any mental disorders in different provinces revealed a lifetime prevalence of 24.20% in Tianjin [12] and 21.9% in Hebei [6], which both had a higher prevalence than our study. The mental disorders in Hebei survey mainly included anxiety disorders, mood disorders, substance use disorders and psychotic disorders. However, the Tianjin survey included dementia and mental retardation, which were excluded from our analysis due to methodological differences. Moreover, the lifetime prevalence rates of anxiety disorders identified in the Tianjin and Hebei surveys were 3.71% and 10.13%, respectively. Similarly, the lifetime prevalence of mood disorders and substance use disorders were 9.75% and 5.58% in Tianjin, and 7.73% and 7.48% in Hebei, respectively. These findings are difficult to compare directly with our results due to these differences. Therefore, broader investigations are needed to provide a more comprehensive understanding of the prevalence of mental disorders among the elderly population in China. When compared with other countries, the prevalence of mental disorders among Chinese elderly population aligns with that of other Asian nations but remains notably lower than in Western counterparts. For instance, in Singapore, the lifetime and 12-month prevalence of mental disorders (including major depressive, bipolar, generalized anxiety, obsessive–compulsive and alcohol use disorders) [31] among adults aged 50 years and above were 7.9% and 3.1%, respectively, with a downward trend observed in older age groups. In contrast, among European elderly aged 65 to 84 years, the lifetime and 12-month prevalence of mental disorders (primarily anxiety, mood and substance use disorders) were 47.0% and 35.2%, respectively [32]. The 12-month prevalence of anxiety, mood and substance use disorders were 25.6%, 14.3% and 18.2% among the European elderly [32], compared to 11.39%, 6.77%and 3.75% among American adults aged 55 years old and above [21].

This finding aligns with previous study showing that the prevalence of mental disorders in Asia is generally lower than in Western countries [33]. Several factors may explain this difference. First, the multi-ethnic nature of Asia may lead to genetic polymorphism, and the genetic susceptibility of the Asian population to mental disorders may differ from Western populations [34]. Second, Asian cultures typically emphasize family and community values, which may provide stronger social and psychological support for individuals, thereby reducing the incidence of mental disorders. At the same time, Asian cultural values, especially Chinese cultural values, are also associated with greater stigma surrounding mental disorders. This stigma often leads to reluctance to acknowledge mental health issues or seek treatment [33].

In contrast to European findings, our study reported a notably lower prevalence of mental disorders among older adults. Several factors elucidate these cross-regional discrepancies. European research highlights that historical underestimation of geriatric mental disorder prevalence stemmed from non-age-adapted assessment tools [33]. The CIDI65 + interview used in Europe, featuring simplified language, improved response validity and uncovered higher prevalence. By contrast, the standard CIDI used in our study lacks age-specific adaptations, which may have caused partial misinterpretation of questions by older adults and potentially underestimated prevalence. The challenge of non-response bias complicates prevalence data interpretation. Global studies on non-response and mental disorder prevalence are inconsistent: some show higher rates among non-responders [35], while others report no significant link [36], suggesting estimates may be conservative if bias exists. Moreover, a cross-sectional study in China [37] revealed that a significant proportion of Chinese individuals hold negative attitudes toward people with mental illnesses, highlighting that cultural stigma around mental health is a critical factor contributing to underreporting. These multifaceted influences highlight the necessity of adopting culturally sensitive and methodologically robust approaches in future mental health research.

According to our findings, the 12-month prevalence of different mental disorders among the elderly population varied significantly by age and gender. First, it was notable that anxiety disorders were more prevalent among individuals aged 55–64 years than among those aged 65 years and older. The finding was partly consistent with previous studies, which have shown that the prevalence of anxiety disorders decreases with increasing age, particularly with a significant drop observed in individuals aged 75 years and older [38]. As is well known, individuals entering the"young old"age group begin to experience neurobiological changes, along with an increase in physical illnesses and cognitive decline, all of which are associated with anxiety disorders in aging populations [39]. Their psychological adjustment abilities may not yet be well-developed, which could explain the higher prevalence of anxiety disorders in the 55–64 age group. Unlike previous studies that typically set 60 or 65 as the starting point of old age, our study defined the beginning of old age as 55. Consequently, we found that the prevalence among individuals aged 55–59 was as high as those aged 60–64, suggesting that interventions for anxiety disorders should target the'young old.

Our results also indicate that age plays a significant role in the prevalence of any mental disorders, with increasing age (≥ 70 years) associated with a reduced risk of these disorders. This finding is partly in line with some studies in this area [32], which have similarly observed a decline in the prevalence of mental disorders in individuals aged 75 years and older. Although the ageing process reduces the capacity of the elderly to cope with stressful life events, the oldest-old tend to have higher resilience and more optimistic than the young-old [40]. In contrast, two studies from India and Brazil have shown that the prevalence of mental disorder was significant higher in older people(> 80 years old) [41, 42]. This discrepancy may be attributed to socio-economic factors, such as reduced functional ability in older adults due to limited economic development and inadequacies in pension systems.

Secondly, mood disorders were more common in female than male, which was in line with a handful of comparable studies in this area [43]. This difference may be attributed to greater emotional sensitivity, negative emotional experiences associated with childbirth, menopause, and other unique life stages in women compared to men [44].

Thirdly, substance use disorders were more prevalent in male and decreased with increasing age, which may be related to Chinese “wine culture”. Notably, substance use disorders in the Chinese population are primarily associated with alcohol use [20]. And a large proportion of studies had shown that alcohol use decreased as individuals age [30]. This decrease may be attributed to the adverse outcomes of drinking becoming more pronounced with age, leading to a decline in health status among older adults. Moreover, the liver’s ability to metabolize alcohol diminishes with age, resulting in reduced alcohol consumption.

Notably, our study demonstrated scant significant variations in the prevalence of mental disorders between elderly individuals residing in urban and rural areas. This finding aligns with previous analyses of the CMHS dataset, which reported no significant urban–rural disparities in the prevalence of most mental disorders (including mood, anxiety, and substance use disorders) [21]. Our results also resonate with a Korean study, which found no significant differences in the prevalence of depression between urban and rural samples [45]. In contrast, a 2001–2005 Chinese study that utilized the SCID to diagnose all mental disorders reported higher prevalence of depressive disorders and alcohol dependence among rural residents [23]. Several factors may account for this discrepancy. First, rapid socioeconomic development in rural areas has improved healthcare access and mitigated historical stressors [46]. Methodologically, the SCID used in the prior study [23] likely detects symptoms of certain disorders, such as bipolar disorders and anorexia nervosa, more sensitively than the CIDI [47]. Additionally, lower literacy among rural older adults may have caused misinterpretations of CIDI questions, leading to underreporting.

The distribution of the 12-month prevalence of different mental disorders provides a more accurate depiction of the mental health status of the elderly. Furthermore, our results emphasize the importance of addressing chronic diseases, pain, and sleep disturbances as part of strategies to prevent mental disorders in this population. Contrary to some prior studies linking low income to mental disorders [48, 49], our findings did not identify income as a significant correlate of mental disorders among the elderly in China. This divergence may stem from the relatively homogeneous socio-economic status within the China Mental Health Survey (CMHS) sample or from cultural factors, such as stigma, that influence the reporting and perception of mental health issues in this population.

Our results were concordant with previous studies showing that having diagnosis of chronic non communicable disease is associated with a higher risk of mental disorders, with the risk increasing further for individuals with three or more chronic conditions [50]. At the same time, our study found that pain was associated with higher odds of mental disorders, consistent with previous findings that highlight the relationship between pain and mental health [51]. Comorbidity and pain can amplify the painful experience of the individual, increasing the risk of mental illness. Furthermore, comorbid conditions and pain reduce physical activity, which in turn diminishes executive functioning and exacerbates mental disorders [52].

Similarly, our findings align with numerous studies confirming that sleep problem is a significant predictor for the onset of mental disorders, particularly depression, followed by anxiety and substance use disorders [42]. Studies have also shown that greater insomnia severity predicts an increased likelihood of developing depressive and anxiety disorders. Importantly, adults aged 55 and older account for 80% of all individuals with insomnia [53], emphasizing the critical need to investigate the relationship between insomnia and mental disorders and to address sleep issues among the older population. Nonetheless, the mechanisms linking insomnia to mental disorders remain incompletely explained. Some potential explanations are as follows On a physiological level, wake-sleep regulation and mental disorders share specific neuronal pathways, neurotransmitters, and receptors [54]. On a genetic and environmental level, depression and insomnia overlap in terms of genetic predispositions and environmental influences [55]. On a psychological level, stress and adversity are strongly associated with the co-occurrence of insomnia and depression [55].

In summary, it is essential to integrate mental disorder screening into aging-related disease assessments, particularly setting 55 years old as the starting age, which can be highly meaningful for public health.

Limitations and strength

There are several limitations deserve to be concerned. First, as with other cross-sectional studies, retrospective reporting may cause recall bias or influenced by current mental state when diagnose lifetime mental disorders. Cognitive decline in the elderly may increase the influence, though interviewers recorded mental state during interviews to mitigate its impact. Second, our study was based on community population, which excluded institutionalized elderly individuals. This exclusion may have limited the generalizability of our findings. Third, Chinese cultural values, which are associated with greater stigma surrounding mental disorders, may have influenced the reported prevalence rates. Additionally, the data were collected ten years ago. Since that time, China has experienced significant economic growth as well as advances in medicine and technology, which may impact the current relevance and generalizability of these findings. However, the CHMS remains the most recent nationally representative survey available and still provides important insights into the phenomenon studied. Lastly, the survey's limited generalizability resulted from the exclusion of mental disorders such as dementia in our CIDI assessment. Finally, the family history of mental disorders, a known risk factor, was not collected in this study.

Despite its limitations, our study presents several significant strengths. First, in contrast to previous research that predominantly centered on mental symptoms, this study employed standardized diagnostic tools to investigate the prevalence of mental disorders among Chinese elderly, ensuring a more accurate and clinically relevant assessment. Second, by comprehensively collecting a wide range of variables, including demographic information, physical diseases, and other related factors, our study offers a holistic understanding of the determinants contributing to mental health issues in the elderly. Additionally, defining the"young old"as individuals aged 55 and above offers a precise basis for targeted prevention strategies, enabling early interventions attuned to this vulnerable group's mental health needs.

Conclusion

The prevalence of mental disorders was high among individuals aged 55 years and older, with anxiety disorders, mood disorders, and substance use disorders being the most common. While different mental disorders were associated with specific demographic factors, chronic diseases, pain, and sleep disturbances played a important role in the overall prevalence of mental disorders.

Abbreviations

CMHS

China Mental Health Survey

CIDI

Composite International Diagnostic Interview 3.0

DSP

Disease surveillance points

CAPI

Computer-assisted personal interview

Authors’ contributions

Yueqin Huang and Guangming Xu made a contribution to the concept or design of the study. Jingjuan Pang wrote the first draft of the manuscript. Zhaorui Liu and Huifang Yin revised the manuscript. Tingting Zhang, Minghui Li, Xiaofei Hou and Guoli Yan managed the statistical analysis and interpretation. Xiangdong Xu, Limin Wang, Yongping Yan, Shuiyuan Xiao, Lingjiang Li, Jie Yan, Yaqin Yu, Xiufeng Xu, Zhizhong Wang, Yifeng Xu and Tao Li participated in data acquisition. All authors contributed to and have approved the final manuscript.

Funding

This work was supported by Capital’s Funds for Health Improvement and Research (2024-2G-4113), the National Twelfth Five-Year Plan for Science and Technology Support from the Chinese Ministry of Science and Technology (2012BAI01B01), the Special Research Project for Non-Profit Public Service of the Chinese Ministry of Health (201202022), the Tianjin Traditional Chinese Medicine Research Project (2023225) and the Tianjin Municipal Education Committee fund (2022KJ266, 2023KJ046).

Data availability

The data that support the findings of this study is available from the corresponding authors.

Declarations

Ethics approval and consent to participate

The Ethical Committee of the Sixth Hospital of Peking University and Tianjin Anding Hospital approved. Participants signed the relevant informed consent. All procedures were in accordance with the ethical standards of national research committee or the institutional, and with the 1964 Helsinki declaration relevant ethical standards.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.