Prevalence and sociodemographic correlates of mental and behavioural disorders in Pakistan: findings from the first nationally representative cross-sectional survey

Abstract

Introduction

Comprehensive population-based nationally representative data on the prevalence of mental and behavioural disorders in Pakistan (population 247 million) have been lacking—hindering the development of prevention policies and responsive mental health services. We conducted the National Psychiatric Morbidity Survey to ascertain the prevalence and sociodemographic correlates of these disorders to inform the development of mental healthcare policies to reduce stigma, disability and morbidity associated with these disorders.

Methods

We conducted a cross-sectional household survey using a stratified multistage cluster sampling design across the four provinces of Pakistan between 2019 and 2022. A total of 17 773 adults aged ≥18 years were interviewed in-person by trained researchers using the Mini International Neuropsychiatric Interview (V.7.0.2). Current (point) and lifetime precise and weighted prevalence was determined according to the International Classification of Diseases ICD-10 classification of mental and behavioural disorders. Functional impairment was measured using the Sheehan Disability Scale. Weighted prevalence estimates and multivariable logistic regression were used to assess associations with sociodemographic variables. National Bioethics Committee of Pakistan granted ethical approval for the survey.

Results

The overall current and lifetime weighted prevalence of all mental and behavioural disorders (ICD-10 codes, F10–F60) was 32.3% (95% CI 31.6 to 32.9) and 37.9% (95% CI 37.2 to 38.6), respectively. The highest current prevalence (24.5%) was observed for neurotic/stress-related disorders (F40–F48), followed by post-traumatic stress disorder (PTSD) (F43.1, 19.9%) and mood disorders (F30–F39, 11.6%)—which was essentially due to the high prevalence of episodic/recurrent depressive disorders (F32–F33, 10.6%). About 6% of the participants reported suicidality in the past month and 1% had made a lifetime suicide attempt. The highest prevalence was observed in people who were aged 40+ years, females, lived in urban areas, separated, divorced or widowed, had no formal education and were retired or categorised as homemakers. In the multivariable logistic regression analyses, the likelihood of psychiatric morbidity significantly increased with increasing age (p value for trend <0.001), being separated, divorced or widowed (p value for trend <0.01) and being unemployed, homemaker, student or retired (p value for trend <0.01) and decreased with increasing level of education and income (p value for trend <0.01) and residence in rural area.

Conclusion

This first nationally representative survey of psychiatric morbidity in Pakistan showed that one in three adults in the country is living with a diagnosable mental and behavioural disorder. The relatively high prevalence of PTSD (1 in 5) may be associated with broader contextual factors, although the cross-sectional design does not permit attribution to specific events or exposures. These findings inform the development of national mental health prevention strategy, healthcare services and community support programmes to reduce the burden of psychiatric morbidity in Pakistan.

WHAT IS ALREADY KNOWN ON THIS TOPIC.

WHAT THIS STUDY ADDS

• This study represents the first and largest nationally representative, population-based survey of psychiatric morbidity in Pakistan, providing comprehensive and contemporary prevalence estimates across the full spectrum of mental and behavioural disorders (International Classification of Diseases ICD-10 codes F10–F60), stratified by age, sex and sociodemographic characteristics and interpreted within the Pakistani sociocultural context.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• Provides the first nationally representative evidence base for mental health in Pakistan, informing the development of clinical guidelines, capacity building, and integrated models of primary prevention and care to improve the health and well-being of people with mental and behavioural disorders, while enabling psychiatrists, policymakers, Non-Governmental Organisations (NGOs), and government agencies to design culturally sensitive strategies to reduce stigma, increase awareness, and reduce psychiatric morbidity.

• Highlights the need for sustainable, culturally relevant mental healthcare services, resource allocation and community-based support, particularly in underserved areas, and offers a model for other low- and middle-income countries to implement nationally representative surveys and evidence-based mental health practice and policies aligned with the WHO’s Comprehensive Mental Health Action Plan 2013–2030.

Introduction

Mental and behavioural disorders are a leading cause of disability worldwide, contributing significantly to the global burden of disease. According to the Global Burden of Diseases, Injuries and Risk Factors Study (GBD, 2019), the prevalence of mental disorders increased by 48.1% between 1990 and 2019 (from 654.8 million to 970.1 million estimated cases), disability-adjusted life-years (DALYs) due to mental disorders increased from 80.8 million to 125.3 million and about 14.6% of years lived with disability in 2019 were attributable to mental disorders.¹ Depressive and anxiety disorders are among the leading causes of DALYs worldwide. The WHO has estimated that in 2021, one in seven persons around the world was living with a mental disorder, with depressive and anxiety disorders being the most common—in 2019, 280 million individuals globally were living with depression and, in 2021, 359 million were living with anxiety disorder.² These conditions and other mental disorders frequently co-occur and often go unrecognised, untreated or poorly managed, particularly in low- and middle-income countries (LMICs) like Pakistan. LMICs face a high burden of mental disorders and mental health needs remain largely unmet (eg, >80% of affected individuals receive no adequate care) due to systemic barriers such as under-resourced health systems (<1% of health budget), lack of trained mental health professionals, poor integration into primary care, stigma and limited population-level data to guide healthcare planning and resource allocation.³ Pakistan, the fifth most populous country in the world, exemplifies these challenges. Despite having a population of over 247 million (2023 estimate; https://data.who.int/countries/586), the country has one of the lowest psychiatrist-to-population ratios globally—that is, 0.19 certified psychiatrists per 100 000 population, which equates to about 470 psychiatrists, and most of these specialists are concentrated in urban centres.^4,5 Mental health spending in Pakistan is <0.5% of the total healthcare budget, and specialised services are limited, fragmented and inaccessible to the majority of the population.⁶

The reported prevalence of (selected) mental disorders in Pakistan ranges from 10% to 72%, based on urban/rural residence and gender differences.⁷ The various estimates reported by WHO and other agencies are extrapolated from global or regional datasets and are not grounded in empirical, nationally representative evidence. Local studies, while valuable, have mostly focused on the prevalence of anxiety and depressive disorders and have been primarily small-scale, geographically restricted and methodologically inconsistent. Most have relied on screening questionnaires like the General Health Questionnaire rather than standardised diagnostic questionnaires, resulting in wide-ranging and often conflicting prevalence estimates—ranging from 10% to 33% in males and 29% to 66% in females.⁸ These studies also vary in how they define mental disorders, their sampling strategies and their generalisability. For example, earlier surveys focused on specific cities or rural districts, limiting their relevance to broader health system planning.

Moreover, existing research often overlooks the complex sociocultural and geopolitical context of Pakistan. Factors such as chronic poverty, gender disparities (particularly marginalisation and mistreatment of women with mental health issues), religious and ethnic marginalisation, large-scale displacement due to conflict and natural disasters and economic instability can significantly influence the onset, reporting and outcomes of mental disorders.^3,9 Additionally, stigma, traditional beliefs and lack of mental health awareness/literacy can shape how individuals understand and respond to psychological distress, often delaying or preventing access to care.¹⁰

Another major gap in the existing literature is the lack of data on the functional impact of psychiatric disorders. While the prevalence data provide a snapshot of the burden of psychiatric disorders, understanding how these disorders impair daily functioning across work, family and social domains is essential for estimating their true economic and societal costs. In Pakistan, few studies have examined the relationship between psychiatric morbidity and disability and even fewer have used validated questionnaires like the Sheehan Disability Scale for such assessments.

In this context, the national psychiatric morbidity (cross-sectional) survey (NPMS) was designed to address these critical evidence gaps. This is the first nationally representative household survey in Pakistan using a structured diagnostic tool to estimate the current (point) and lifetime prevalence of the full spectrum of the mental and behavioural disorders (according to the International Classification of Diseases ICD-10 categories), evaluate sociodemographic correlates, assess functional impairment and explore suicidal ideation and behaviours. Findings from the NPMS are also essential for mental health advocacy, programme development, service integration and workforce training. By aligning with global standards and focusing on the unique sociocultural landscape of Pakistan, this survey offers a robust evidence base to inform public health prevention policy and promote mental health equity.

The objectives of the NPMS were to: (1) assess the prevalence of common psychiatric disorders in a nationally representative adult sample in Pakistan, (2) examine the association between psychiatric morbidity and different demographic variables (age, sex, district of residence, marital status, level of education, employment and income), (3) evaluate the prevalence of suicidal ideation and behaviours and (4) assess the level of disability associated with psychiatric disorders.

Methods

Study design and participants

An epidemiological cross-sectional survey design was used to estimate prevalence of psychiatric morbidity, according to the ICD-10 classification of mental and behavioural disorders (ICD codes, F10–F60). We used a three-stage, stratified random cluster sampling technique to ensure national representativeness. In the first stage, the Pakistan Bureau of Statistics provided technical expertise in stratifying the country by province and urban/rural areas and selected enumeration blocks containing 200–300 households using probability proportional to size. In the second stage, a systematic random sample of households was drawn from each selected block after listing all the households. In the third stage, one eligible adult (aged 18+ years) from each selected household was randomly chosen using the Kish grid method, ensuring each adult had an equal chance of selection. This approach allowed for geographically diverse and demographically representative data collection across the country. The survey was implemented on a de facto basis (ie, individuals present at the time of the survey).

The survey was conducted between 2019 and 2022 in the four provinces of Pakistan—Punjab, Sindh, Khyber Pakhtunkhwa (KPK) and Balochistan. Data collection was temporarily paused during the COVID-19 lockdowns in 2020–2021 and resumed with strict safety protocols (including use of personal protective equipment, physical distancing and health screening of field teams to ensure safety of both participants and enumerators). Eligible participants were adults aged 18+ years, residing in the selected households for at least 6 months and sharing the same kitchen, who were able to comprehend the purpose of the study and willing to provide informed consent. Individuals with severe cognitive and/or physical disabilities that prevented them from understanding the interview were excluded. Visitors and family members living elsewhere were also excluded. If no eligible individual was present in the selected household, the interviewers proceeded to the next household on the serial list.

Data collection

Data were collected by using the Mini International Neuropsychiatric Interview (MINI), for Diagnostic and Statistical Manual of Mental Disorders DSM-5 (V.7.0.2), which was adapted and translated into Urdu (the national language of Pakistan) using forward-backward translation and reviewed by bilingual experts for cultural relevance. The MINI was chosen because it is comprehensive, reliable and simple to use. It has built-in diagnostic algorithms that are in alignment with DSM-5 and ICD-10. There is also good concordance between the MINI and both Structured Clinical Interview for DSM-5 Disorders (SCID) and Composite International Diagnostic Interview, and it has been validated against SCID. A number of studies have positively tested the validity of the MINI and its inter-rater and test-retest reliability. Its inbuilt structure circumvents the problem of two stage interviews required in population based psychiatric epidemiological surveys. Functional impairment was assessed using the SDS.^11,12 The SDS was administered to the participants who met the criteria for mental and behavioural disorders. Sociodemographic information—including age, sex, province and district of residence, religion, type of family, marital status, level of education, employment and income—was collected using a structured questionnaire. Data were collected (using electronically filled forms) and recorded electronically via tablet computers and uploaded to a centralised cloud-based server with regular supervisory checks for quality assurance and completeness.

Interviews were conducted face-to-face at the participants’ home by trained pairs (at least one male and one female) of interviewers. The survey was overseen by an executive committee, which included members of the Pakistan Psychiatric Society, epidemiologists, subject experts in mental health, demographers and biostatisticians. To maintain uniformity in the data collection/fieldwork process, the committee developed a master protocol and operational instructions manual. Enumerators, primarily psychology and health science graduates, underwent 5 days of intensive training on survey methods/questionnaires, consent procedures and interview techniques. Competency was ensured by pre-training and post-training assessments. A three-tiered monitoring system was used, including spot checks, supervisory field visits and weekly quality review meetings. A pilot study was conducted in peri-urban area—this informed the refinement of protocol/operational instructions manual, including the necessity of mixed-gender interview teams to improve community acceptability.

The Pakistan Bureau of Statistics calculated a minimum required sample size of 11 180 adults (aged 18+ years) to ensure national representativeness. This calculation was based on an assumed prevalence of overall psychiatric morbidity of 34%, derived from previous subnational studies and a systematic review.⁸ These studies had reported a wide range of prevalence estimates using general (non-specific) questionnaires like the General Health Questionnaire. Given this variability, 34% was assumed as a conservative midpoint to ensure adequate statistical power. The sample size also incorporated a design effect of four to account for the complex cluster sampling, a 4% margin of error, a 95% confidence level and an anticipated non-response rate of 3%. To enhance statistical power and enable robust subgroup analyses, the sample size was intentionally doubled in Punjab and Sindh—the two most populous provinces. This approach allowed for more precise provincial estimates while maintaining national representativeness through appropriate sampling weights. Selection bias was minimised through rigorous random sampling and substitution protocols for non-eligible households. Interviewer bias was mitigated through standardised training. Social desirability bias was addressed by ensuring privacy and employing same-gender interviewers where culturally appropriate.

The study was approved by the National Bioethics Committee of Pakistan (Reference no. 4–87/NBC-268/17/1370). Informed consent was obtained from all participants and privacy rights observed. Permission from the head of the household was sought in situations where it was deemed culturally appropriate or expected—particularly in conservative or rural settings where decision-making is often guided by the household hierarchy. This approach respected the cultural context, sensitivity and etiquette of the household and helped build trust, enhanced community acceptance and ensured ethical engagement with the participants. All participants were assured of confidentiality and were offered referral information for their district/local mental healthcare service if needed. Prior arrangements were made with the district/local government mental healthcare services, and all participants with suicidal plan or intention and acute disorders were referred to these services for clinical assessment and management.

Statistical analysis

Variables, including demographic characteristics and psychiatric morbidity, were summarised using frequencies and percentages. Weighted prevalence (%) estimates and corresponding 95% CIs were calculated using the normal approximation (the Wald method) to the binomial distribution, based on the SE of the proportion. The levels of disability due to psychiatric morbidity, according to SDS (1–12 mild disability, 13–24 moderate disability, 25–36 severe disability, 37–48 extreme disability) are reported as mean and SD.

Associations between psychiatric morbidity and sociodemographic variables were initially explored using cross-tabulations. Due to the unequal sampling probability—particularly oversampling in Punjab and Sindh—and differential response rates, all prevalence estimates were weighted using normalised sampling weights to ensure national representativeness. The dataset was derived from a multistage stratified sampling design, with province of residence serving as the primary stratification variable. Sampling weights, reflecting provincial representation, were applied using the ‘Weight Cases’ procedure in IBM SPSS Statistics (V.25) to account for unequal probabilities of selection and to improve the representativeness of prevalence estimates and logistic regression analyses. Weighted prevalence estimates and ORs were obtained after application of these sampling weights. However, variables required to fully specify the complex survey design—such as primary sampling units and clustering identifiers—were not available in the dataset provided by the Pakistan Bureau of Statistics. Consequently, the multistage sampling structure could not be explicitly incorporated into variance estimation. CIs and p values for trend were therefore derived using conventional logistic regression methods. Overall psychiatric morbidity was examined as the dependent variable. Binary and multivariable logistic regression models were employed separately to assess associations with key sociodemographic variables. All categorical variables were dummy coded for inclusion in the regression models.

Potential confounders were identified based on existing literature and conceptual relevance—these included age, sex, district of residence (urban/rural), marital status, level of education, employment and income. All these variables were included as covariates in the multivariable models to adjust for potential confounding. Adjusted OR and 95% CI were calculated for each factor while incorporating sampling weights. Missing data were minimal (<5%) for most variables, except for income, where 13.7% of participants did not respond or refused to answer. In multivariable regression analyses, listwise deletion was applied, excluding participants with missing covariate data from the respective models. Descriptive analyses were conducted using all available cases. No imputation was undertaken as the overall level of missingness was low and unlikely to materially affect the findings. To assess the potential impact of missing income data, a sensitivity analysis was conducted by re-running the multivariable regression models excluding the income variable. The estimated associations were materially unchanged—with no meaningful differences in the direction, magnitude or statistical significance of the associations compared with the primary analysis—indicating that the findings were robust to the exclusion of income.

Patient and public involvement

Patients and the public were not involved in the design, conduct, reporting or dissemination planning of this study.

Results

A total of 20 833 individuals were approached for participation in the study; of them, 17 773 (49.2% males, 50.8% females) were successfully interviewed (response rate=86.2%). Table 1 presents sociodemographic characteristics of the study population (n=17 773; median age (IQR)=35 years (26–43)). The large majority of participants were aged 18–29 years (33.6%), lived in rural areas (63.8%), lived as nuclear family (66.4%), married (81.1%), had no formal education (33.7%) and employed (39.6%).

Table 1. Sociodemographic characteristics of the study population (n=17 773).

Characteristic	N	%
Age (years)
18–29	5970	33.6
30–39	5500	30.9
40–49	3590	20.2
50–59	1676	9.4
60+	1037	5.8
Sex
Male	8750	49.2
Female	9023	50.8
Province of residence
Punjab	7680	43.2
Sindh	5680	32
KPK	2524	14.2
Baluchistan	1889	10.6
District of residence
Rural	11 338	63.8
Urban	6435	36.2
Religion
Islam	17 314	97.4
Hindu	255	1.4
Christian	204	1.1
Type of family
Nuclear	11 800	66.4
Extended	5973	33.6
Marital status
Unmarried	2844	16
Married	14 419	81.1
Separated	73	0.4
Divorced	47	0.3
Widowed	390	2.2
Level of education
No formal education	5991	33.7
Primary/middle (grades 1–8)	4496	25.3
Secondary (grades 9–10)	3347	18.8
Intermediate (grades 11–12)	1901	10.7
University graduate	1552	8.7
Other	486	2.7
Employment status
Employed/self-employed	7046	39.6
Unemployed	4236	23.8
Homemaker	4070	22.9
Student	379	2.1
Retired	143	0.8
Other	1899	10.7
Income (Pakistani rupees/month)
≤25 000	11 046	62.2
25 001–45 000	3026	17
45001+	1273	7.2
Did not respond/refused	2428	13.7

KPK, Khyber Pakhtunkhwa.

Table 2 shows the weighted current (point) and lifetime prevalence (%) of mental and behavioural disorders according to ICD-10 Classification of Mental and Behavioural Disorders: Diagnostic Criteria for Research (DCR) codes. The current prevalence of all disorders combined (F10–F60) was 32.3% (95% CI 31.6 to 32.9) and lifetime prevalence was 37.9% (95% CI 37.2 to 38.6). Among the categories of disorders, the highest current prevalence (24.5%) was observed for neurotic/stress-related disorders (F40–F48), followed by post-traumatic stress disorder (PTSD) (19.9%) and mood disorders (F30–F39, 11.6%)—which was essentially due to the high prevalence of episodic/recurrent depressive disorders (F32–F33, 10.6%).

Table 2. Prevalence (%) of mental and behavioural disorders according to the ICD-10 Classification of Mental and Behavioural Disorders: Diagnostic Criteria for Research(DCR) codes in the study population.

ICD-10 DCR code*	Disorder	Current (point) prevalence %† (95% CI)	Lifetime prevalence %† (95% CI)
F10–F60	All disorders combined	32.3 (31.6 to 32.9)	37.9 (37.2 to 38.6)
F10–F19	All disorders due to psychoactive substance use	0.85 (0.7 to 1.0)	–
F10	Disorders due to alcohol use	0.5 (0.4 to 0.6)	–
F11–F19	Other disorders due to psychoactive substance use (except F17, tobacco use)	0.4 (0.3 to 0.5)	–
F20–F29	All schizophrenia/schizotypal/delusional disorders	2.7 (2.4 to 2.9)	4.0 (3.8 to 4.3)
F30–F39	Mood (affective) disorders	11.6 (11.1 to 12.0)	19.6 (19.1 to 20.2)
F30–F31	Manic episode/bipolar affective disorders	2.4 (2.2 to 2.6)	4.5 (4.2 to 4.7)
F32–F33	Episode/recurrent depressive disorders	10.6 (10.1 to 11.0)	17.8 (17.3 to 18.3)
F40–F48	Neurotic/stress-related/somatoform disorders	24.5 (23.9 to 25.1)	24.8 (24.2 to 25.4)
F32–F33, F40–F48	Combined depressive and neurotic/stress-related/somatoform disorders	30.7 (30.0 to 31.3)	35.7 (35.0 to 36.4)
F40	Phobic anxiety disorders	1.4 (1.3 to 1.6)	–
F40.0	Agoraphobia	0.8 (0.6 to 0.9)	–
F40.1	Social phobias	0.9 (0.8 to 1.1)	–
F41	Other anxiety disorders	5.3 (5.0 to 5.6)	5.8 (5.4 to 6.1)
F41.0	Panic disorder (episodic paroxysmal anxiety)	1.2 (1.0 to 1.3)	1.7 (1.5 to 1.9)
F41.1	Generalised anxiety disorder	4.5 (4.2 to 4.8)	–
F42	Obsessive-compulsive disorder	1.3 (1.1 to 1.5)	–
F43.1	Post-traumatic stress disorder	19.9 (19.4 to 20.5)	–
F60.2	Dissocial personality disorder	–	0.9 (0.8 to 1.0)

^*The ICD-10 Classification of Mental and Behavioural Disorders: Diagnostic Criteria for Research (ICD-10 DCR).

^†Prevalence was estimated as a weighted proportion using survey weights. The corresponding 95% CI was derived via the normal approximation to the binomial distribution, based on the standard error of the weighted proportion.

Table 3 shows the period (ie, past 1 month) prevalence (%) of suicidality according to the MINI (B-module, B1–B18). Suicidality was reported by 6.2% of the participants and 1.1% had made a suicide attempt. Passive suicidal ideation was reported by 4.1%, while active suicidal ideation was reported by 3.2%.

Table 3. Period (past 1 month) prevalence (%) of suicidality^* according to MINI (B-module, B1–B19) in the study population.

Suicidality in the past 1 month	N	Period prevalence %†
		(95% CI)
B2–B19: suicidality	1301	7.3 (6.9 to 7.7)
B1b, B2-B12, B14, B16: suicidality excluding suicidal attempt	1097	6.2 (5.8 to 6.5)
B1a: plan/intent to self-harm in an accident	177	1.0 (0.9 to 1.1)
B2: passive suicidal ideation	735	4.1 (3.8 to 4.4)
B3: active suicidal ideation	570	3.2 (3.0 to 3.5)
B4a: psychotic suicidality (voice(s) telling you to kill yourself)	109	0.6 (0.5 to 0.7)
B4b: suicidal dreams (dreams with suicidal content)	88	0.5 (0.4 to 0.6)
B5 and/or B6 and/or B7 and/or B8 and/or B9: suicidal plan/intent (method, means, place, date, suicide note)	325	1.8 (1.6 to 2.0)
B9 or B14: suicidal preparatory behaviour	232	1.3 (1.1 to 1.5)
B10 or B11: suicidal intent	365	2.1 (1.9 to 2.3)
B12: impulsive suicidality	179	1.0 (0.9 to 1.2)
B15: non-suicidal self-injury/harm	151	0.9 (0.7 to 1.0)
B16: factors that interrupted suicide attempt
B16a: started a suicide attempt, but decided to stop	104	0.6 (0.5 to 0.6)
B16b: started a suicide attempt, but was interrupted	95	0.5 (0.4 to 0.6)
B16c: went through with a suicide attempt, but was rescued	119	0.7 (0.5 to 0.8)
B17: usual time spent/day with any suicidal impulses, thoughts or actions
≤120 min	448	2.5 (2.3 to 2.8)
120–360 min	379	2.1 (1.9 to 2.4)
360+ minutes	16	0.09 (0.0 to 0.1)
B18: state of the last suicide attempt (ever)	186	1.1 (0.9 to 1.2)
Current: within the past 12 months	101	0.6 (0.5 to 0.7)
In early remission: between 12 months and 24 months ago	29	0.2 (0.1 to 0.2)
In remission: more than 24 months ago	56	0.3 (0.2 to 0.4)

^*Based on at least one ‘Yes’ response to B1–B19 questions in MINI B-module.

^†Prevalence was estimated as a proportion. The corresponding 95% CI was derived via the normal approximation to the binomial distribution, based on the SE of the proportion.

MINI, Mini International Neuropsychiatric Interview .

Table 4 presents the current (point) and lifetime prevalence (%) and sociodemographic correlates (crude and adjusted ORs and 95% CI) of psychiatric morbidity. The highest current prevalence of psychiatric morbidity was observed in people who were aged 40+ years, females, lived in urban areas, separated, divorced or widowed, had no formal education and retired and categorised as homemaker (ie, women, most often wife/mother, whose primary role is managing the household/family life rather than engaging in employment outside the home). In the multivariable logistic regression analyses, there was a significant positive association (or the likelihood of psychiatric morbidity increased) with increasing age (p value for trend <0.001), being separated, divorced or widowed (p value for trend <0.01) and being unemployed, homemaker, student or retired (p value for trend <0.01). There was an inverse (or protective) association (or the likelihood of psychiatric morbidity decreased) with increasing level of education and income (p value for trend <0.01) and residence in rural area.

Table 4. Prevalence (%) and sociodemographic correlates of psychiatric morbidity in Pakistan.

	Current (point) prevalence %* (95% CI)	Lifetime prevalence %1 (95% CI)	Crude OR† (95% CI)	Adjusted OR (95% CI)
Age (years)
18–29	15.8 (14.9 to 16.6)	23.3 (22.2 to 24.3)	1.0	1.0
30–39	22.7 (21.6 to 23.7)	29.3 (28.2 to 30.5)	1.6 (1.4 to 1.7)	1.6 (1.4 to 1.7)
40–49	57.0 (55.5 to 58.6)	59.8 (58.3 to 61.4)	7.1 (6.5 to 7.8)	7.3 (6.6 to 8.1)
50–59	60.7 (58.4 to 63.0)	63.7 (61.4 to 65.9)	8.2 (7.3 to 9.3)	8.3 (7.3 to 9.5)
60+	56.7 (53.8 to 59.7)	59.2 (56.2 to 62.1)	7.0 (6.1 to 8.0)	6.5 (5.6 to 7.6)
P value for trend			<0.001
Sex
Male	31.1 (30.1 to 32.0)	37.3 (36.3 to 38.3)	1.0	1.0
Female	33.5 (32.5 to 34.4)	38.5 (37.5 to 39.5)	1.1 (1.1 to 1.2)	0.8 (0.7 to 0.8)
District of residence
Urban	33.8 (32.7 to 35.0)	39.3 (38.1 to 40.5)	1.0	1.0
Rural	31.5 (30.7 to 32.3)	37.2 (36.4 to 38.1)	0.9 (0.8 to 1.0)	0.8 (0.8 to 0.9)
Marital status
Unmarried	18.6 (17.3 to 20.2)	26.2 (24.6 to 27.7)	1.0	1.0
Married	34.1 (33.4 to 34.9)	39.4 (38.6 to 40.1)	2.3 (2.1 to 2.5)	1.0 (0.9 to 1.1)
Separated	40.1 (30.5 to 49.7)	51.9 (42.1 to 61.6)	2.9 (2.0 to 4.4)	1.6 (1.0 to 2.5)
Divorced	63.8 (49.5 to 78.1)	71.4 (57.9 to 84.8)	7.7 (4.2 to 14.1)	4.3 (2.2 to 8.4)
Widowed	60.9 (55.9 to 65.9)	66.4 (61.5 to 71.2)	6.8 (5.4 to 8.5)	1.4 (1.1 to 1.9)
P value for trend			<0.01
Level of education
No formal education	39.1 (37.9 to 40.2)	42.8 (41.6 to 44.1)	1.0	1.0
Primary/middle (grade 1–8)	35.1 (33.8 to 36.5)	40.9 (39.5 to 42.3)	0.8 (0.8 to 0.9)	1.1 (1.0 to 1.2)
Secondary (grade 9–10)	28.1 (26.6 to 29.5)	33.6 (32.0 to 35.1)	0.6 (0.6 to 0.7)	0.9 (0.8 to 1.0)
Intermediate (grade 11–12)	18.9 (17.2 to 20.5)	25.5 (23.6 to 27.3)	0.4 (0.3 to 0.4)	0.7 (0.6 to 0.8)
University graduate	24.6 (22.5 to 26.7)	32.8 (30.5 to 35.1)	0.5 (0.5 to 0.6)	0.9 (0.8 to 1.0)
Other	30.4 (26.8 to 34.0)	44.8 (40.9 to 48.6	0.7 (0.6 to 0.8)	0.9 (0.7 to 1.1)
P value for trend			<0.01
Employment
Employed/self-employed	27.4 (26.4 to 28.4)	32.8 (31.7 to 33.9)	1.0	1.0
Unemployed	31.1 (29.7 to 32.5)	36.4 (35.0 to 37.9)	1.2 (1.1 to 1.3)	1.4 (1.2 to 1.5)
Homemaker	39.0 (37.6 to 40.4)	43.8 (42.3 to 45.2)	1.7 (1.6 to 1.8)	2.5 (2.3 to 2.8)
Student	23.9 (19.8 to 27.9)	33.3 (28.8 to 37.8)	0.8 (0.7 to 1.0)	2.4 (1.8 to 3.0)
Retired	62.7 (54.8 to 70.6)	66.2 (58.4 to 73.9)	4.5 (3.2 to 6.2)	2.9 (2.0 to 4.2)
Other	35.9 (34.0 to 37.8)	43.8 (41.8 to 45.8)	1.5 (1.3 to 1.6)	1.9 (1.7 to 2.1)
P value for trend			<0.01
Income (Pakistani rupees/month)
≤25 000	34.1 (33.2 to 35.0)	39.1 (38.2 to 39.8)	1.0	1.0
25 001–45 000	29.3 (27.8 to 30.8)	35.8 (34.2 to 37.4)	0.8 (0.7 to 0.9)	0.9 (0.8 to 0.9)
45001+	26.6 (24.2 to 29.0)	32.5 (29.9 to 35.1)	0.7 (0.6 to 0.8)	0.8 (0.7 to 0.9)
Did not respond/refused	31.2 (29.5 to 32.8)	38.0 (36.3 to 39.7)	0.9 (0.8 to 1.0)	0.8 (0.7 to 0.9)
P value for trend			<0.01

^*Prevalence was estimated as a weighted proportion using survey weights. The corresponding 95% CI was derived via the normal approximation to the binomial distribution, based on the SE of the weighted proportion.

^†ORs are based on current prevalence. Adjusted OR and their corresponding 95% CI are based on multivariate logistic regression analysis.

Table 5 presents mean (SD) levels of current and lifetime disability due to mental and behavioural disorders according to SDS. All disorders were associated with functional impairment across domains of work, family and social life. Mean current disability scores ranged from 18.3 to 23.7, with the highest scores observed among individuals with disorders due to alcohol use (23.7), manic episodes/bipolar disorder (23.5), schizophrenia (23.3), generalised anxiety disorder (21.5), PTSD (20.9) and neurotic and stress-related disorders (20.2).

Table 5. Mean (SD) level of functional disability due to mental and behavioural disorders according to SDS^* in the study population.

ICD-10 DCR code†	Disorder	Current (point) SDS Mean (SD)	Lifetime SDS Mean (SD)
F10–F60	All disorders combined	19.2 (13.0)	19.0 (13.1)
F10–F19	All disorders due to psychoactive substance use	20.2 (11.4)	–
F10	Disorders due to alcohol use	23.7 (11.4)	–
F11–F19	Other disorders due to psychoactive substance use (except F17, tobacco use)	16.6 (10.3)	–
F20–F29	All schizophrenia/schizotypal/delusional disorders	23.3 (11.7)	22.6 (11.6)
F30–F39	Mood (affective) disorders	18.8 (12.3)	19.2 (12.8)
F30–F31	Manic episode/bipolar affective disorders	23.5 (10.7)	23.8 (11.1)
F32–F33	Episode/recurrent depressive disorders	18.3 (12.2)	18.8 (12.9)
F40–F48	Neurotic and stress-related disorders	20.2 (13.0)	20.1 (13.0)
F32–F33, F40–F48	Combined depressive and neurotic/stress-related/somatoform disorders	19.3 (13.0)	19.1 (13.2)
F40	Phobic anxiety disorders	15.5 (10.1)	–
F40.0	Agoraphobia	17.0 (10.9)	–
F40.1	Social phobias	14.9 (9.0)	–
F41	Other anxiety disorders	20.6 (11.8)	20.3 (11.9)
F41.0	Panic disorder (episodic paroxysmal anxiety)	15.5 (14.4)	18.2 (12.4)
F41.1	Generalised anxiety disorder	21.5 (11.3)	–
F42	Obsessive-compulsive disorder	17.2 (10.6)	–
F43.1	Post-traumatic stress disorder	20.9 (13.2)	–
F60.2	Antisocial personality disorder	–	20.7 (10.1)

^*SDS: 1–12 mild disability, 13–24 moderate disability, 25–36 severe disability, 37–48 extreme disability.

^†The ICD-10 Classification of Mental and Behavioural Disorders: Diagnostic Criteria for research. https://iris.who.int/bitstream/handle/10665/37958/9241544228_eng.pdf

.SDS, Sheehan Disability Scale.

Discussion

Main findings

This study provides the first nationally representative population-based estimates of the prevalence and functional burden of mental and behavioural disorders in Pakistan, offering critical insight into the mental health status of a country with a population of over 247 million people. The study revealed that about one-third (ie, one in three) of adults are currently experiencing a diagnosable mental health disorder, with neurotic/stress-related disorders being most prevalent (24.5%), followed by PTSD (19.9%) and mood disorders (11.6%), which was essentially due to the high prevalence of episodic/recurrent depressive disorders (10.6%).

Contextualisation with Pakistani culture and comparison with other studies

In Pakistan, mental disorders account for >4% of the total disease burden—it has been estimated by WHO that 24 million people in the country need psychiatric assistance.⁵ A systematic review found that the prevalence of anxiety and depressive disorders, across smaller regional studies, ranged from 10% to 33% in males and 29% to 66% in females, highlighting inconsistencies due to the use of non-diagnostic screening questionnaires and varied sampling methods.⁸

The lifetime prevalence of all mental and behavioural disorders combined (37.9%) in our study is within the global range of 12–47.4%, reported in the WHO’s World Mental Health Surveys.¹³ A direct comparison of our findings with studies from other countries is limited due to differences in culture, values, diagnostic criteria, age groups included in studies and questionnaires used for data collection.¹⁴ In a nationally representative survey of psychiatric disorders in Iran, the overall 12-month period prevalence of any disorder was 23.6% (26.5% in women and 20.8% in men). The most common disorder was any anxiety disorder (15.6%), followed by major depressive disorder (12.7%), generalised anxiety disorder (5.2%) and obsessive-compulsive disorder (5.1%).¹⁵ In a recent meta-analysis of 25 studies of mental disorders in South Asia, the pooled prevalence was 16% for depression and 12% for anxiety.¹⁶ In a similar survey in Malaysia (using the Diagnostic Interview Schedule), the overall prevalence of mental disorders among adults was 29%—the highest prevalence was observed in the rural region of East Malasia (43%), followed by the capital Kuala Lumpur (40%).¹⁷ In the National Mental Health Survey of India, conducted across 12 states in 2015–2016, the current prevalence of any mental disorder was 10.6% and lifetime prevalence was 13.7%—the most common mental and behavioural disorders were due to use of psychoactive substance (22.4%), mood disorders (5.6%) and neurotic and stress-related disorders (3.7%).^18,19 The relatively higher burden of mental disorders in Pakistan may be attributed to chronic sociopolitical stressors, economic challenges/instability, cultural stigma, gaps in mental health service infrastructure, shortages of qualified mental health professionals and insufficient mental healthcare funding.²⁰

The sociodemographic correlates identified in our study mirror findings from other LMICs. In our study, the likelihood of psychiatric morbidity increased with increasing age, being separated, divorced or widowed and being unemployed, homemaker, student or retired and decreased with increasing level of education and income and residence in rural area. The National Mental Health Survey of India also found similar associations between mental disorders and poverty, unemployment and urban stress.^18,19 People living in urban areas may be more prone to factors such as overcrowding, pollution, poor housing and social fragmentation, which may be associated with psychiatric morbidity.²¹ Multivariable analysis revealed lower adjusted odds of psychiatric morbidity in females compared with males. This paradox (ie, reversal of association) may be partly due to adjustment for the confounding variables (ie, age, marital status, level of education, employment) and may also reflect underreporting among men due to stigma and societal expectations around masculinity.²² Income data were missing for 13.7% of participants, and the non-response may not have been random. While the sensitivity analyses, excluding income from the regression models, yielded similar results, we cannot rule out the selection bias arising from participants intentionally not disclosing their income due to reasons best known to them.

Our study did not directly assess gender-based violence or sociocultural stressors that disproportionately affect women, such as restricted mobility and caregiving burdens—factors previously identified in regional studies as major contributors to mental distress among South Asian women.²³ Future research should address these dimensions more explicitly.

In our study, there was a relatively high current prevalence (19.9%) of PTSD. The Pakistan population has endured decades of sociopolitical strife including armed conflicts, displacement and natural disasters (ie, floods, earthquakes, droughts, heatwaves). In a recent systematic review, the prevalence of PTSD and depressive symptoms among civilian residing in armed conflict-affected regions was 23.7% and 25.6%, respectively.²⁴ In a study of flood affected survivors, the prevalence of PTSD symptoms was 34% in Punjab and 42% in the KPK province.²⁵ A systematic review/meta-analysis of PTSD prevalence following earthquake in Iran and Pakistan showed an overall polled prevalence of 55.6% (60.2% in Iranian survivors and 49.2% in Pakistani survivors).²⁶ There is also evidence of the impact of COVID-19 (and other infectious disease pandemics of the 21st century) on stress-related disorders/PTSD. In a recent meta-analysis and systematic review, the pooled overall prevalence of post-pandemic PTSD in the general population was 19.3% (95% CI 15.3% to 23.2%)—it was 24.2% in high-income countries and 17.0% in LMICs.²⁷ In another systematic review and meta-analysis of studies conducted in sub-Saharan Africa, the overall pooled prevalence of PTSD was 22% (95% CI 13% to 32%) and the current prevalence was 25% (95% CI 16% to 36%). The prevalence of PTSD in conflict-exposed regions was remarkably higher at 30% (95% CI 21% to 40%).²⁸ In a similar recent study, the point prevalence of PTSD in adult war/conflict-survivor sample was 26.5%.²⁹

The relatively high prevalence of PTSD observed in this study may be related to broader contextual factors, including the COVID-19 pandemic, natural disasters, economic instability and sociopolitical strife. However, given the cross-sectional design and the absence of temporal analyses or direct measures of exposure to these events, the study does not permit attribution of the observed prevalence to any specific contextual factor. Consequently, these explanations remain speculative and should be interpreted with caution. The prevalence of PTSD in our study was higher than that reported in many nationally representative surveys from other settings. This difference may reflect variation in diagnostic instruments, assessment methods, sociocultural factors, patterns of symptom reporting or population characteristics across studies. Accordingly, direct comparisons should be made cautiously. Notwithstanding these considerations, the findings highlight the substantial burden of PTSD symptoms in the population and underscore the need for trauma-informed mental healthcare and psychosocial support services.

We have also documented the burden and issues around suicidality in Pakistan (6.2% of the participants reported suicidality in the past month and 1.1% reported a lifetime suicide attempt). In cross-sectional studies of suicidality in five LMICs, reported lifetime prevalence of suicidal ideation ranged from 3.5% to 11.1% in community samples.³⁰ In our study, the prevalence of active/passive suicidal ideation and attempts was much lower relative to overall point/lifetime prevalence of mental disorders. Under-reporting (or concealment) of suicidal behaviour in Pakistan may be influenced by legal, cultural, religious and systemic factors. This may be due to the following factors: (1) until December 2022, reported attempted suicide was a criminal offence punishable by law (imprisonment/fine)—this may have discouraged individuals and families from reporting suicide attempts, (2) at the societal level, suicide is often perceived as a moral failing, leading to significant stigma—families may therefore conceal suicide incidents to avoid social ostracisation and protect family honour, (3) suicidality is prohibited in Islam—and the religion is considered a protective factor against suicide. Religious scholars and communities emphasise compassion, acknowledging that suicide often results from severe mental anguish and there is increasing recognition of the need for support (not stigma) when addressing mental health issues within Muslim communities. In Pakistan, suicide is highly stigmatised, which likely contributes to under-reporting.^31,32 The findings of our study contextualise Pakistan’s religio-legal/sociocultural norms and dynamics and emphasise the need for increasing awareness and public health led suicide prevention programmes and improving access to supportive mental health services.³³ In our study, there was a substantial functional impairment associated with psychiatric morbidity across domains of work, family and social life. These findings are consistent with those from the WHO World Mental Health Report, where mental disorders were associated with significant disability and reduced productivity, particularly in LMICs.^34,35 The indirect economic burden of mental disorders in Pakistan is likely immense, though poorly quantified.^6,36

The mental healthcare in Pakistan remains a significant concern. According to WHO, Pakistan has only 0.19 psychiatrists per 100 000 population, which equates to about 470 psychiatrists for a population of over 247 million.^5,37 This shortage is compounded by limited integration of mental healthcare into primary and preventive services.³⁸ Cultural beliefs, lack of awareness, misinformation and reliance on traditional and spiritual healers further delay or divert care-seeking from certified mental healthcare professionals. In a study of 87 adult psychiatric patients treated at public hospitals in Lahore, Pakistan, the patient reported seeing diverse traditional/spiritual healers, including homeopaths, naturopaths (Tibb), Islamic faith healers and sorcerers for their psychiatric disorders prior to their current treatment from certified psychiatrists, with the majority indicating that they had consulted more than one of these healers.³⁹ Studies from LMICs suggest that integrating mental healthcare into community health platforms and improving mental health literacy can promote access to services, reduce stigma and increase compliance with treatment.⁴⁰

Strengths and limitations

Our study has several key strengths. It is the first nationally representative household survey in Pakistan to assess the current and lifetime prevalence of the full spectrum of psychiatric morbidity using a structured diagnostic questionnaire (MINI), ensuring standardised and internationally comparable measures of mental health and behavioural disorders. The large sample size, rigorous multistage sampling strategy and geographic coverage across the four provinces enhanced generalisability of the findings to the adult population of Pakistan. The use of electronic data capture, quality assurance protocols and trained interviewers enabled collection of high-quality data and high response rate. Other strengths included: (1) language adaptation and recruitment of interviewers from the same province as the participants; (2) pilot testing of the questionnaires (for comprehension and cultural sensitivity); (3) culturally competent training/supervision of the interviewers and (4) community engagement and stakeholders’ involvement. Another key strength is the evaluation of suicidality and functional disability.

The study also has a number of limitations. Stigma and social desirability bias may have led to under-reporting (or concealment) of sensitive issues, particularly suicidal thoughts and behaviours and substance use. Although gender-matched interviewers and private setting/location were prioritised to improve disclosure, these conditions could not be consistently guaranteed in all locations—especially in more conservative or rural communities. As the information on clustering and primary sampling units from the complex survey design was not available, these variables therefore could not be incorporated into the logistic regression analysis. Consequently, the CIs and p values for trend estimated in the logistic regression may be underestimated, resulting in overly narrow CIs and inflated statistical significance.

Conclusions and clinical and public health policy implications

In summary, this first nationally representative cross-sectional survey of psychiatric morbidity showed that one in three adults in Pakistan is living with a mental health and behavioural disorder. The relatively high prevalence of PTSD (one in five) may reflect, at least in part, the influence of broader contextual factors such as the COVID-19 pandemic, natural disasters, economic instability and sociopolitical strife; however, these hypotheses warrant further investigation. Considering that <0.5% of the total health budget is spent on mental healthcare and there are just about 470 certified psychiatrists nationwide, these findings quantify the burden of mental health disorders in Pakistan and inform the development of mental healthcare policy, psychotherapy, digital health platforms and possible integration of mental healthcare into primary/preventive/public health services. There is a need to strengthen community-based services, train frontline health workers in mental health identification and support and invest in culturally adapted, scalable interventions. Importantly, mental health policies should be informed by local realities, including poverty, inequality, displacement and stigma. Intersectoral collaboration—with education, social welfare and justice systems—is essential to address structural and systemic determinants of poor mental health. Future research should focus on longitudinal data, adolescent mental health, minority and marginalised populations, the mental health of displaced populations and health system readiness to respond to this burden.

Data availability statement

Data are available upon reasonable request.