Depressive symptoms and psychometric evaluation of the PHQ-9 among caregivers of breast cancer patients in Pakistan: a cross-sectional study

Abstract

Introduction

Caregivers of breast cancer patients experience a significant psychological burden, which may lead to moderate-to-severe depressive symptoms. However, limited research has explored the prevalence among caregivers in the Pakistani population. This study aimed to assess the prevalence of moderate-to-severe depressive symptoms among these caregivers.

Methodology

We conducted a multi-step psychometric analysis of the Patient Health Questionnaire-9 (PHQ-9) among caregivers of breast cancer patients recruited from tertiary hospitals across Pakistan. Exploratory Factor Analysis (EFA) employed principal axis factoring with Promax rotation. Assumptions were verified using Bartlett’s Test of Sphericity and Kaiser-Meyer-Olkin sampling adequacy. Confirmatory Factor Analysis (CFA) tested the EFA-derived structure, with model fit evaluated using RMSEA, TLI, and χ². Socio-demographic and health-related data were collected through structured questionnaires. Statistical analysis was performed using JAMOVI software, incorporating descriptive statistics, chi-square tests, t-tests/ANOVA, and binomial logistic regression to identify factors associated with moderate-to-severe depressive symptoms among the caregivers of the breast cancer patients.

Results

Among 402 caregivers, 186 (46.3%) had moderate-to-severe depressive symptoms, while 216 (53.7%) had no or mild depressive symptoms based on PHQ-9 scores. Overall, 27.1% had mild symptoms, 24.1% had moderate symptoms, 16.4% had moderately severe symptoms, and 5.7% had severe symptoms. EFA revealed a 3-factor solution explaining 46.75% of variance, all primary loadings > 0.52. The CFA confirmed this structure (RMSEA = 0.064; CFI = 0.96; TLI = 0.94), though χ² was significant (p < 0.05). Among caregivers, low monthly income (p = 0.033), a sedentary lifestyle (p = 0.012), and urban residence (p = 0.044) were associated with a higher risk of depressive symptoms. In a multivariate logistic regression analysis, lower education levels were found to be a significant predictor of moderate-to-severe depressive symptoms (p = 0.007).

Conclusion

A substantial proportion of caregivers of patients with breast cancer in Pakistan experience moderate-to-severe depressive symptoms, indicating a significant and underrecognized mental health burden. The PHQ-9 demonstrated acceptable psychometric performance with a multidimensional structure in this population. Secondary education, higher physical activity, and higher income were associated with increased odds of depressive symptoms. These findings should be interpreted in light of the cross-sectional design and use of self-reported measures.

Clinical trial number

Not applicable.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12888-026-08180-0.

Introduction

Breast cancer is the most frequently diagnosed malignancy among women worldwide. In 2020, approximately 2.3 million new cases and 685,000 deaths were reported globally, making breast cancer a leading contributor to cancer-related morbidity and mortality [1]. The burden is particularly pronounced in Pakistan, where breast cancer is the most common cancer among women, affecting nearly one in four females [2]. In 2020, breast cancer accounted for 28.7% of all newly diagnosed cancers and 11.7% of cancer-related deaths in the country [3]. These figures reflect not only high disease incidence but also systemic challenges such as limited screening, delayed diagnosis, and constrained oncology care.

The growing burden of breast cancer extends beyond patients to informal caregivers, who play a central role throughout the disease trajectory. Caregivers provide emotional support, assist with daily activities, accompany patients to medical appointments, and participate in treatment-related decision-making [4]. In Pakistan, caregiving responsibilities are predominantly undertaken by family members within a healthcare system that remains largely patient-centered, with limited formal attention to caregiver well-being. Structural constraints, financial pressures, and inadequate psychosocial support services further intensify the caregiving role, particularly during prolonged and complex cancer treatment [5, 6].

Sustained caregiving in oncology settings is associated with substantial psychological morbidity. Caregivers are exposed to chronic stress, emotional exhaustion, role strain, and uncertainty related to disease prognosis, all of which increase vulnerability to mental health disorders [7]. Among these, major depressive disorder (MDD) is one of the most clinically significant outcomes, adversely affecting caregivers’ quality of life and their capacity to provide effective care. A systematic review and meta-analysis reported that approximately 42% of caregivers of cancer patients experience depressive symptoms, with a higher prevalence observed among female caregivers [8]. Studies from regional settings have reported similarly high levels of depressive symptoms among caregivers of patients with breast cancer, including 52.5% in India and 27.9% in Jordan [9, 10]. The severity of depressive symptoms has been linked to factors such as younger age, spousal caregiving, lower socioeconomic status, and caregiving for patients with advanced-stage disease [9, 11].

Cultural context further shapes the recognition and expression of depressive symptoms. In Pakistan, mental health conditions remain highly stigmatized, and psychological distress is frequently somatized or underreported. This contributes to underrecognition and delayed identification of depression among caregivers, limiting opportunities for timely intervention and support [12]. Despite the high burden of breast cancer and the central role of caregivers within the Pakistani healthcare system, research focusing on caregivers’ mental health remains limited, with most studies prioritizing patient outcomes.

Accurate identification of depressive symptoms in caregivers requires screening instruments that are psychometrically valid within the target population. Although the Patient Health Questionnaire-9 (PHQ-9) is widely used to screen for major depressive disorder, its psychometric properties may vary across cultural contexts and caregiving populations due to differences in symptom expression, caregiving stressors, and sociocultural norms. Caregivers of patients with breast cancer represent a distinct population exposed to prolonged emotional and psychological stress, which may influence the dimensional structure and performance of PHQ-9 items. However, we could not find evidence supporting the validity and factor structure of the PHQ-9 among caregivers of patients with breast cancer in Pakistan. Therefore, this study aims to assess the prevalence of moderate-to-severe depressive symptoms and to evaluate the psychometric properties of the PHQ-9 among caregivers of patients with breast cancer in Pakistan, to support accurate screening and inform caregiver-focused mental health interventions.

Methodology

Study design and study setting

This cross-sectional study was conducted over a period of three months across tertiary care hospitals in Pakistan, following ethical approval from the institutional review board of Rahmah Health Foundation, Islamabad, Pakistan. The study setting included hospitals with multidisciplinary cancer care teams comprising medical, surgical, radiation, and oncology specialists. However, at the time of the study, no dedicated caregiver support services were in place.

Sampling strategy

A hospital-based consecutive sampling approach was used. Caregivers of patients with breast cancer were recruited from oncology outpatient and inpatient services during the data collection period. Eligible caregivers accompanying patients for treatment or follow-up visits were approached sequentially and invited to participate until the target sample size was achieved.

Inclusion and exclusion criteria

Caregivers were eligible if they were aged ≥ 18 years and identified as the primary informal caregiver of a patient with a confirmed diagnosis of breast cancer. A primary caregiver was defined as a family member or close relative providing the majority of unpaid physical, emotional, or logistical support for at least three months prior to enrollment. Participants were required to be actively involved in daily caregiving and able to complete the study questionnaire.

Caregivers were excluded if they had a previously diagnosed psychiatric disorder under active treatment, severe cognitive impairment, or were unable to provide informed consent. Professional or paid caregivers were also excluded.

Sample size and eligibility criteria

A sample size of 384 caregivers was determined using the OpenEpi sample size calculator, with a 95% confidence interval and a 5% margin of error. However, we included 402 participants; the oversampling was intentional to account for potential incomplete responses or data loss.

Data collection

Data collection was carried out through face-to-face interviews conducted by a trained researcher. Data collection involved a structured questionnaire to gather information on socio-demographic factors, caregiver burden, and health-related variables.

The study used a previously validated Urdu translation of the PHQ-9 scale, originally developed and validated by Ahmed S et al. [13]. This validated instrument has been widely used for assessing depressive symptoms in Urdu-speaking populations.

According to PHQ-9 scoring guidelines, depression symptom severity was categorized as minimal (0–4), mild (5–9), moderate (10–14), moderately severe (15–19), and severe (20–27). For analysis, caregivers were grouped based on the presence or absence of moderate to severe depressive symptoms. Individuals with a PHQ-9 score of less than 10 (i.e., minimal or mild depression symptoms) were classified as having no moderate to severe depressive symptoms, whereas those with a PHQ-9 score of 10 or higher were classified as having moderate to severe depression symptoms. The data collection process was supervised to ensure accuracy, and participants were assured of the confidentiality and anonymity of their responses.

Study variables

The study analyzed various caregiver-related variables, including demographics (age, gender, education, family structure, income, and residential area) and health factors (BMI, chronic conditions, and smoking status). Additionally, caregiver burden, including the number of caregiving hours, financial support such as contributing significantly to medical bills, rent or mortgage payments, daily living expenses, transportation costs, or other essential needs, transport facilities such as daily commuting, frequent medical visits, therapy sessions, or other essential destinations, and domestic support such as cooking, cleaning, laundry, or grocery shopping was assessed. Data quality control was maintained through close supervision, daily monitoring, and a double-entry validation process to eliminate errors.

Statistical analysis

The statistical methods were selected a priori to align with the dual objectives of this study: (1) to evaluate the psychometric properties of the PHQ-9 among caregivers of patients with breast cancer, and (2) to examine the prevalence and correlates of moderate-to-severe depressive symptoms in this population. Psychometric analyses were conducted to assess internal consistency, latent factor structure, and model fit, given the absence of prior validation studies of the PHQ-9 in this specific caregiving context. Subsequently, inferential statistical methods were applied to compare caregiver groups and identify predictors of depression, with analytical choices guided by data distribution, measurement scale, and study design considerations.

The distribution of continuous variables was assessed using the Shapiro–Wilk test and visual inspection of histograms and Q–Q plots. Most variables demonstrated non-normal distributions; therefore, appropriate parametric or non-parametric tests were applied as indicated.

Multivariable regression analyses were performed to identify factors independently associated with moderate-to-severe depressive symptoms.

Reliability analysis

To assess the internal consistency of the items, we conducted a reliability analysis using both Cronbach’s alpha (α) and McDonald’s omega (ω). These coefficients were computed for the full scale and each item, if deleted, providing insights into the contribution of individual items to overall reliability.

Exploratory factor analysis (EFA)

We performed an EFA to examine the underlying factor structure of the scale. The analysis used principal axis factoring (PAF) as the extraction method, which is robust to non-normality, and a Promax rotation, allowing factors to correlate. The number of factors was determined using parallel analysis, which compares observed eigenvalues to those derived from simulated random data. Factor retention decisions were further supported by scree plot evaluation and eigenvalues > 1.0.

Assumptions for factor analysis were verified using Bartlett’s test of sphericity (to confirm sufficient item correlations) and the Kaiser-Meyer-Olkin (KMO) Measure of Sampling Adequacy (to ensure the data were suitable for factor analysis). Factor loadings below 0.3 were suppressed for clarity. Additional output included factor summaries (e.g., variance explained), inter-factor correlations, model fit indices, and a scree plot to support factor retention decisions. Model fit was evaluated using multiple indices based on established thresholds. A relative chi-square (χ²/df) value of 3.0 or less was considered indicative of a good fit. The Root Mean Square Error of Approximation (RMSEA) was interpreted as acceptable when values were 0.08 or below, with lower values indicating better fit. The Comparative Fit Index (CFI) was deemed acceptable at values of 0.90 or higher, reflecting adequate model performance compared to a null model. For the Standardized Root Mean Square Residual (SRMR), values of 0.08 or less were considered indicative of a good fit. Lastly, Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) were used for model comparison, with preference given to models showing lower AIC and BIC values, reflecting better overall model parsimony and explanatory power.

Confirmatory factor analysis (CFA)

Following the EFA, a CFA was conducted to validate the identified factor structure. Although χ² was statistically significant (p < 0.05), this is expected given the large sample size, as χ² is sensitive to N. The model was estimated using maximum likelihood estimation, and fit indices, including the comparative fit index (CFI), Tucker-Lewis Index (TLI), root mean square error of approximation (RMSEA), and Bayesian information criterion (BIC), were examined to evaluate model adequacy. Furthermore, AIC and BIC were used to evaluate model parsimony and support model retention. Factor covariances and standardized loadings were also reported to assess the relationships between items and their respective latent factors.

Post-factor analysis: group comparisons and predictors of depression

After the factor analysis of the questionnaire, the statistical analysis was conducted to compare caregivers with moderate-to-severe depression and those without depression across demographic, socioeconomic, health-related, and caregiving-related variables. Descriptive statistics were reported as means and standard deviations for normally distributed continuous variables, medians and interquartile ranges for non-normally distributed data, and frequencies and percentages for categorical variables. For inferential analysis, independent samples t-tests were used for normally distributed continuous variables, while the Wilcoxon rank-sum test was applied for non-normally distributed data. Categorical variables were analyzed using Pearson’s chi-square test, while ordinal variables with multiple categories were assessed using the Kruskal-Wallis test. A p-value of < 0.05 was considered statistically significant. The missing data were handled by excluding cases with incomplete information for the specific analysis. Assumptions for each test, including normality and homogeneity of variance, were verified before analysis. Effect sizes were reported for significant findings to assess the magnitude of observed differences.

A binomial logistic regression analysis was performed to assess the relationship between caregiver depression status (binary outcome: no depression vs. moderate-severe depression) and various predictor variables, including demographic, socioeconomic, health-related, and caregiving-related factors. The model was fitted using maximum likelihood estimation to estimate the log-odds of depression status.

Model specification and evaluation

The overall model fit was evaluated using the likelihood ratio test, McFadden’s pseudo R² (R² McF), Akaike Information Criterion (AIC), and deviance statistics. Predictor significance was assessed via Wald chi-square tests, with odds ratios (ORs) and 95% confidence intervals (CIs) reported for interpretation.

All continuous predictors were checked for linearity in the log-odds assumption, while categorical variables were included as factors. Multicollinearity was examined using the variance inflation factor (VIF) and tolerance statistics, with VIF < 5 and tolerance > 0.2 considered acceptable. This approach ensured a robust comparison between caregiver groups while addressing potential confounding factors. All the data analysis was performed using JAMOVI 2.6.26.

Ethics approval and consent to participate

This study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki. Written informed consent was obtained from all participants prior to inclusion in the study.

Results

Reliability analysis

The scale demonstrated strong internal consistency, with Cronbach’s α = 0.82 and McDonald’s ω = 0.82. Item-level analysis revealed that removing any single item did not substantially improve reliability (α and ω ranged from 0.79 to 0.83 when items were deleted), supporting the retention of all items in the initial analysis. The loading of factors is illustrated in Table 1:

Table 1.

Promax-rotated factor loadings from exploratory factor analysis (n = 402)

	Factor
	1	2	3	Uniqueness
PHQ_1	0.7801			0.5199
PHQ_2	0.6997			0.4839
PHQ_3		0.7121		0.4342
PHQ_4	0.4365	0.3681		0.5015
PHQ_5		0.7240		0.5292
PHQ_6			0.5830	0.5537
PHQ_7	0.5834			0.5576
PHQ_8	0.7230			0.4609
PHQ_9			0.5214	0.7514

Note. Loadings < |0.3| omitted for clarity. Bolded values indicate primary factor assignments. Extraction method: Principal axis factoring. Rotation: Promax

Exploratory factor analysis

The KMO measure (0.85) and significant Bartlett’s Test (χ² = 1023, p < 0.001) confirmed the suitability of the data for factor analysis. Parallel analysis suggested retaining three factors, collectively explaining 46.75% of the total variance. Factor loadings revealed a clear structure:

• Factor 1 (25.49% variance) included PHQ-1, PHQ-2, PHQ-4, PHQ-7, and PHQ-8 (core affective-cognitive symptoms).

• Factor 2 (14.00% variance) comprised PHQ-3 and PHQ-5 (somatic symptoms).

• Factor 3 (7.26% variance) included PHQ-6 and PHQ-9 (suicidality/ and self-worth).

The exploratory factor analysis identified a multidimensional structure of depressive symptoms among caregivers of patients with breast cancer. The first factor comprised items primarily reflecting affective and cognitive symptoms, including depressed mood, anhedonia, feelings of worthlessness, and concentration difficulties. The second factor captured somatic symptoms, such as fatigue, sleep disturbance, and appetite changes, which may reflect the physical and emotional demands of prolonged caregiving. The third factor included items related to psychomotor changes and suicidal ideation, representing more severe manifestations of depressive symptomatology. Together, these factors suggest that depressive symptoms in caregivers may be expressed across distinct but interrelated domains.

Notably, PHQ-4 cross-loaded on Factors 1 and 2, and PHQ-9 had a weak loading (0.52) and high uniqueness (0.75), suggesting potential misfit. The loading of factors is illustrated in Table 1. The Promax-rotated solution showed moderate correlations between factors (r = 0.35–0.68). The inter-factor correlations are illustrated in the supplementary file. Model fit was acceptable (RMSEA = 0.064, TLI = 0.94).

The three-factor solution was retained based on multiple complementary criteria. Parallel analysis indicated retention of three factors with observed eigenvalues exceeding those generated from random data, and scree plot inspection demonstrated a clear inflection after the third factor. This solution accounted for a substantial proportion of the total variance and showed a stable pattern of salient loadings with minimal cross-loading. Moreover, the three-factor model demonstrated conceptual coherence, with each factor representing a distinct domain of depressive symptomatology, supporting its interpretability beyond purely statistical considerations.

Confirmatory factor analysis

The CFA supported the three-factor structure, with good fit indices (CFI = 0.96, TLI = 0.94, RMSEA = 0.064), as reported in Table 2:

Table 2.

Model fit indices for exploratory and confirmatory factor analysis

Fit Index	EFA	CFA	Threshold for Good Fit
RMSEA	0.064	0.064	< 0.08
90% CI	[0.037, 0.092]	[0.045, 0.083]	–
TLI	0.94	0.94	≥ 0.90
CFI	–	0.96	≥ 0.90
χ² (df)	31.72 (12)	63.37 (24)	χ²/df = ≤ 3.0
BIC	−40.24	9664	Lower = Better

The standardized factor loadings from the confirmatory factor analysis, presented in Table 3, demonstrate strong and statistically significant associations between all items and their respective latent factors (all p < 0.001), except for PHQ-9, which showed a weaker but still significant loading.

Table 3.

Standardized factor loadings from confirmatory factor analysis with standard errors and significance tests

Factor	Indicator	Estimate	SE	Z	p
Factor 1	PHQ-1	0.7355	0.05518	13.328	< 0.001
	PHQ-2	0.7945	0.05302	14.985	< 0.001
	PHQ-4	0.7890	0.05499	14.347	< 0.001
	PHQ-7	0.7579	0.05343	14.185	< 0.001
	PHQ-8	0.8179	0.05260	15.549	< 0.001
Factor 2	PHQ-3	0.9188	0.06371	14.421	< 0.001
	PHQ-5	0.7150	0.05896	12.127	< 0.001
Factor 3	PHQ-6	0.8725	0.17286	5.047	< 0.001
	PHQ-9	0.1840	0.04635	3.969	< 0.001

Note. All factor loadings were statistically significant (*p* < 0.001). PHQ = Patient Health Questionnaire

The path diagram of the three-factor solution is illustrated in Fig. 1:

Fig. 1.

Path diagram of the three-factor model of caregiver depression symptoms

Post-factor analysis: group comparisons and predictors of moderate-to-severe depressive symptoms

Having established the robust psychometric properties of the three-factor model, we next examined its clinical utility in characterizing depression symptoms among caregivers. Psychological assessment using the Patient Health Questionnaire-9 (PHQ-9) revealed a mean score of 9.25 (SD = 6.13). Based on PHQ-9 scoring criteria, 73.4% of caregivers were categorized as having depressive symptoms. Specifically, 27.1% had mild, 24.1% moderate, 16.4% moderately severe, and 5.7% severe depressive symptoms. While 26.6% exhibited no depressive symptoms. The mean age of participants was 37.84 years (SD = 12.84), with 37.1% identifying as female. Most caregivers were married (76.4%) and resided in urban areas (53.5%). Educational attainment varied, with 34.1% having completed secondary education and 27.1% possessing a diploma, bachelor’s degree, or higher. In terms of employment, 34.6% were employed full-time, 25.6% part-time, while 37.3% were unemployed. Regarding the caregiver–patient relationship, 42.8% were spouses, 28.6% were offspring, 8.7% were parents, and 10.2% reported other relationships. Most caregivers (93.5%) did not reside with the patient. Nearly half (48.8%) lived in joint family households. The mean number of siblings reported by caregivers was 5.21 (SD = 2.19).

Health-related characteristics indicated that 47.8% of caregivers reported good health and 36.1% reported excellent health. A high prevalence of chronic medical conditions was observed, with hypertension being the most common (78.6%), followed by diabetes (8.0%), cardiac conditions (2.5%), asthma (2.0%), and hepatitis (1.7%). The mean body mass index (BMI) was 24.43 kg/m² (SD = 4.42). A small proportion reported smoking tobacco (15.4%) or using sleep-inducing medications (7.5%). The average sleep duration was 7.56 h per night (SD = 2.47), and the mean duration of physical activity was 0.76 h per week (SD = 0.91). A history of depression was reported in 9.2% of caregivers.

The median monthly income of caregivers was 25,500. Pakistani rupees (PKR) (IQR = 30500 Over half (54.2%) reported providing significant financial support to the patient. The reported median number of caregiving hours per day was 5 (IQR = 5.37). The median distance from caregivers’ residences to the hospital was 40 km (IQR = 80), and patients visited the hospital an average of 1.65 times per week (SD = 1.42). Table 4 presents descriptive statistics comparing caregivers with major depressive disorders (MDD) and without depression across sociodemographic, economic, and health-related variables.

Table 4.

Comparison of demographic, socioeconomic, health-related, and caregiving-related variables between caregivers with and without moderate-to-severe depressive symptoms

Variables	N	Moderate-severe depressive symptoms	No/Mild depressive symptoms	Test Statistic
		(N = 186)	(N = 216)
Age of caregiver in years	402	Q1:27.0 Q2 = 38.0 Q3 = 48.0	Q1 = 27.4 Q2 = 37.0 Q3 = 45.0)	F (1, 400) = 0.81, P = 0.373
Age of caregiver in years	402			χ² (3, N = 402) = 3.09, p = 0.382
Up to 20		15/186	22/216
21–40		91/186	117/216
41–60		72/186	66/216
Greater than 60		8/186	11/216
Gender of the caregiver: female	402	68/186	81/216	χ² (1, N = 402) = 0.04, p = 0.852
Education of the caregiver	402			χ² (4, N = 402) = 13.46, p = 0.012
Uneducated		48/186	30/216
Primary		28/186	30/216
Secondary		52/186	85/216
Higher		12/186	8/216
Diploma/Bachelor’s and Above		46/186	63/216
Current employment status of the caregiver	402			χ² (3, N = 402) = 4.55, p = 0.212
Unemployed		68/186	82/216
Employed Full-Time		62/186	77/216
Employed Part-Time		54/186	49/216
Retired		2/186	8/216
Monthly income of the caregiver	402	Q1:13916.7 Q2:25000.0 Q3:32250.0	Q1:12416.7 Q2:30000.0 Q3:50000.0	F (1, 400) = 8.67, p < 0.013
Monthly income of the caregiver	402			χ² (5, N = 402) = 31.19, p < 0.012
Up to 30,000 PKR		138/186	121/216
30,000–50,000 PKR		38/186	44/216
50,000–75,000 PKR		6/186	16/216
75,000–125,000		0.0 3/186	6/216
125,000–200,000 PKR		0.0 0/186	4/216
Greater than 2 lacs		1/186	25/216
Residential area of the caregiver: urban	402	90/186	125/216	χ² (1, N = 402) = 3.61, p = 0.062
Marital status of the caregiver: married	402	144/186	163/216	χ² (1, N = 402) = 0.21, p = 0.652
BMI of the caregiver	402	Q1 = 22.0 Q2 = 25.0 Q3 = 28.0	Q1 = 21.7 Q2 = 23.0Q3 = 26.0	F (1, 400) = 6.21, p = 0.013
BMI of caregiver - categories	402			χ2 (1, N = 402) = 8.70, p = 0.032
Underweight		10/186	0.1 17/216
Normal		45/186	72/216
Overweight		32/186	43/216
Obesity		99/186	84/216
Current tobacco smoking: Smoker	402	33/186	29/216	χ2 (1, N = 402) = 1.43, p = 0.232
Average sleep duration in caregivers	402	Q1:7.0 Q2: 8.0 Q3: 9.0	Q1:7.0 Q2: 8.0 Q3: 9.0	F (1, 400) = 0.01, p = 0.933
Average sleep duration in caregivers - categories	402			χ2 (3, N = 402) = 1.68, p = 0.642
Insufficient Sleep		26/186	25/216
Suboptimal Sleep		42/186	60/216
Optimal Sleep		93/186	102/216
Oversleeping		25/186	29/216
Type of family structure caregiver lives in: Joint	402	96/186	100/216	χ2 (1, N = 402) = 1.13, p = 0.292
Number of siblings of the caregiver	402	Q1:4.0 Q2: 5.0 Q3: 7.0	Q1:4.0 Q2: 5.0 Q3: 7.0	F (1, 400) = 0.31, p = 0.583.
Number of siblings of the caregiver	402			χ2 (4, N402) = 0.86, p = 0.932
Only child		4/186	3/216
Small family		16/186	15/216
Medium-sized family		89/186	107/216
Large family		62/186	75/216
Very large family		15/186	16/216
Overall health status of the caregiver	402			Χ2(3, N = 402) 11.55, P = 0.012
Poor		4/186	4/216
Fair		36/186	21/216
Good		92/186	100/216
Excellent		54/186	91/216
Chronic health conditions diagnosed in the caregiver	402			χ2(5, N = 402) = 3 0.98,p = 0.552
Hypertension		146/186	170/216
Diabetes		15/186	17/216
Asthma		5/186	3/216
Cardiac Related Disease		3/186	7/216
Hepatitis		5/186	2/216
Other Chronic Conditions		12/186	17/216
Whether the caregiver takes sleeping pills or not: No	402	15/186	15/216	χ2 (1, N = 402) = 0.18, p = 0.672
Whether the caregiver lives with the patient: No	402	180/186	196/216	χ2(1, N = 402) = 6.01, p = 0.012
Relationship between caregiver and patient	402			χ2(4, N = 402) = 6.42, p = 0.172
Spouse		88/186	84/216
Offspring		53/186	62/216
Parent		13/186	22/216
Sibling		19/186	20/216
Others		13/186	28/216
Duration of physical activity in caregivers - categories	402			χ2(3, N = 402) = 42.72, p < 0.012
Sedentary		126/186	77/216
Light activity		32/186	81/216
Moderate activity		19/186	46/216
High activity		9/186	12/216
Past depression history of the caregiver: No	402	168/186	197/216	χ2(1, N = 402) = 0.09, p = 0.762
Any relative of the caregiver diagnosed with depression: No	388	163/181	180/207	χ2(1, N = 402) = 0.90, p = 0.342
Assistance provided by the caregiver in household chores	402			χ2(2, N = 402) = 1.71, p = 0.432
None		36/186	36/216
Some		70/186	95/216
Significant		80/186	85/216
Financial support provided by the caregiver for the patient	402			χ2(2, N = 402) = 3.46, p = 0.182
None		47/186	72/216
Some		34/186	31/216
Significant		105/186	113/216
Transport assistance provided by a caregiver for the patient	402			χ2(2, N = 402) = 0.34, p = 0.842
None		29/186	36/216
Some		28/186	36/216
Significant		129/186	144/216
The hours the caregiver spends per day	401			χ2(3, N = 401) = 16.33, p<0.012
Low		62/185	89/216
Moderate		70/185	45/216
High		20/185	21/216
Very High		33/185	61/216
Age of the patient in years	402	Q1:38.0 Q2: 45.0Q3: 54.0	Q1:38.0 Q2:45.0Q3:56.6	F (1, 400) = 0.07, p = 0.793
Distance of caregiver residence from the hospital	402	Q1:11.9 Q2:50.0Q3: 100.0	Q1:10.0 Q2:30.0Q3:75.6	F (1, 400) = 2.41, p = 0.123
Distance of caregiver residence from the hospital	402			χ2(4, N = 402) = 8.15, p = 0.092
Very close		45/186	65/216
Close		29/186	45/216
Moderate Distance		34/186	41/216
Far		45/186	31/216
Very far		33/186	34/216
How often does the patient visit the hospital in a week	401	Q1: 1.0 Q2: 1.0 Q3: 2.0	Q1: 1.0 Q2: 1.0 Q3: 2.0	F (1, 399) = 1.04, p = 0.313
How often does the patient visit the hospital in a week	401			χ2 (3, N = 402) = 6.46, p = 0.092
No hospital visits		0/186	2/215
Occasional visits		160/186	172/215
Frequent visits		23/186	29/215
Very frequent visits		3/186	12/215
Stage of the patient’s illness	402			χ2(4, N = 402) = 5.80, p = 0.212
Stage 1		11/186	20/216
Stage 2		66/186	68/216
Unknown		5/186	13/216
Stage 3		70/186	69/216
stage 4		34/186	46/216

N is the number of non-missing values. ¹: Kruskal-Wallis. ²: Pearson. ³Wilcoxon.

The further details of the baseline socio-demographic, health, and caregiving-related characteristics of the caregiver population (N = 402) are summarized in the supplementary material. Multicollinearity diagnostics were also conducted using the Variance Inflation Factor (VIF) and tolerance statistics (VIF < 5 and tolerance > 0.2 deemed acceptable), ensuring that predictors such as income, education, and residence were not collinear. Confounding was addressed by including all significant bivariate predictors in the multivariate model.

In multivariable logistic regression analysis, several factors were independently associated with moderate-to-severe depressive symptoms among caregivers. Caregivers with secondary education had nearly twofold higher odds of experiencing moderate-to-severe depressive symptoms compared with those with no formal education (adjusted OR = 1.96; 95% CI: 1.03–3.72; p = 0.040). Greater duration of physical activity was significantly associated with increased odds of depressive symptoms (adjusted OR = 1.56; 95% CI: 1.21–2.03; p < 0.001). Similarly, higher monthly income, expressed per 10,000 PKR increment, was associated with higher odds of moderate-to-severe depression (adjusted OR = 1.12; 95% CI: 1.05–1.19; p < 0.001).

Other variables, including overall health status, body mass index, frequency of hospital visits, caregiving hours per day, sleep duration, and co-residence with the patient, were not significantly associated with depressive symptoms after adjustment, as their confidence intervals included unity. Although caregivers not living with the patient demonstrated lower odds of depression, this association did not reach statistical significance (adjusted OR = 0.40; 95% CI: 0.15–1.09; p = 0.074).

All adjusted odds ratios and corresponding 95% confidence intervals are presented in Table 5.

Table 5.

Multivariable logistic regression analysis of factors associated with moderate-to-severe depressive symptoms (PHQ-9 ≥ 10) among caregivers. Outcome: Depression (Yes vs. No) (Adjusted Odds Ratios from multivariable logistic regression model)

Predictor	Adjusted OR	95% CI	p-value
Education of caregivers (Ref: Uneducated)
Primary	1.29	0.60–2.76	0.518
Secondary	1.96	1.03–3.72	0.040
Higher	1.04	0.35–3.07	0.939
Diploma/bachelor’s and above	1.18	0.59–2.36	0.648
Overall health status (Ref: Poor)
Fair	0.54	0.11–2.77	0.463
Good	0.85	0.18–3.99	0.834
Excellent	1.31	0.27–6.26	0.733
BMI (per 1 kg/m² increase)	0.97	0.92–1.02	0.175
Physical activity (hours/week)	1.56	1.21–2.03	< 0.001
Hospital visits (per week)	1.08	0.90–1.29	0.425
Caregiving hours (per day)	1.01	0.98–1.05	0.440
Sleep duration (hours/day)	1.04	0.95–1.14	0.386
Living with patient (No vs. Yes)	0.40	0.15–1.09	0.074
Monthly income (per 10,000 PKR increase)	1.12	1.05–1.19	< 0.001

Data are presented as adjusted odds ratios (ORs) with 95% confidence intervals (CIs), derived from a multivariable binary logistic regression model. The outcome variable was moderate-to-severe depressive symptoms (PHQ-9 ≥ 10 vs. < 10). ORs > 1 indicate higher odds of moderate-to-severe depressive symptoms, while ORs < 1 indicate lower odds. Physical activity is measured in hours per week, and monthly income is scaled per 10,000 PKR increment. PHQ-9 scores reflect depressive symptom severity and do not constitute a clinical diagnosis

Detailed results of the logistic regression analysis, including omnibus tests for individual predictors and overall model fit indices, are presented in the Supplementary files Tables.

Discussion

This study identified a culturally informed three-factor structure of the PHQ-9 (comprising somatic, cognitive–affective, and suicidality–psychomotor symptom domains) among caregivers of patients with breast cancer in Pakistan. Although the chi-square test was statistically significant, multiple complementary fit indices (CFI = 0.96, TLI = 0.94, RMSEA = 0.064, SRMR < 0.08) indicated good model fit, supporting the robustness of the three-factor solution. This structure diverges from the commonly reported unidimensional PHQ-9 model but aligns with findings from other non-Western and medically burdened populations, where depressive symptoms frequently manifest across distinct domains [14–16].

The three-factor solution was retained based on multiple complementary criteria. Parallel analysis indicated retention of three factors with observed eigenvalues exceeding those generated from random data, and scree plot inspection demonstrated a clear inflection after the third factor. This solution accounted for a substantial proportion of the total variance and showed a stable pattern of salient loadings with minimal cross-loading. Moreover, the three-factor model demonstrated conceptual coherence, with each factor representing a distinct domain of depressive symptomatology, supporting its interpretability beyond purely statistical considerations.

The weaker loading of the suicidality item and the strong correlation between somatic and cognitive–affective factors (r = 0.68) suggest that cultural norms, stigma, and caregiving-related stress may influence how depressive symptoms are expressed and reported. In particular, the emergence of a distinct somatic symptom factor is consistent with evidence that psychological distress in caregiving populations is often somatized, especially in settings where mental health stigma limits direct reporting of emotional symptoms. These findings highlight the importance of considering symptom domains beyond global depression scores when screening caregivers.

Item-level patterns further support the clinical relevance of the multidimensional structure. The isolation of sleep and appetite symptoms suggests that caregivers may initially express psychological distress through bodily complaints, warranting heightened clinical vigilance. The clustering of suicidal ideation (Item 9) with self-criticism (Item 6) highlights the potential role of guilt and perceived failure in driving suicidal thoughts among caregivers. Additionally, the loading of psychomotor symptoms (Item 8) alongside emotional symptoms indicates that observable agitation or retardation may represent a salient feature of depression in this population.

Furthermore, the identification of a three-factor structure supports a multidimensional conceptualization of depression among caregivers of patients with breast cancer, reflecting both the statistical structure of the data and the clinical realities of prolonged caregiving stress. Beyond psychometric validation, the high prevalence of moderate-to-severe depressive symptoms observed in this study highlights a substantial unmet mental health need. While the PHQ-9 captures core depressive symptoms consistent with DSM-5-TR criteria, it remains a screening tool rather than a diagnostic instrument, necessitating the need for appropriate clinical follow-up and caregiver-focused mental health interventions.

Previous studies conducted in similar low and middle-income countries have reported comparable findings, where caregivers exhibit high levels of depressive symptoms, anxiety, and stress. A study conducted in India found that 52.5% of caregivers of breast cancer patients experience depression symptoms, which is quite similar to our finding [9]. Due to the similarity in cultural and social contexts between India and Pakistan, the variation in prevalence rate may be influenced by factors such as socioeconomic factors, political instability, and access to mental health support services [17]. A study conducted by Yang et al. evaluated the prevalence and found 8.90% of the population suffered from anxiety and depression among family caregivers [18]. Globally, the pooled depressive symptoms prevalence rate was 25.14% among caregivers of cancer patients [7]. The higher prevalence in our study compared to the above studies, likely reflects systemic gaps in Pakistan, such as limited mental health infrastructure and availability of mental health support services. Consistent with international literature, our study reinforces that caregiving, particularly in the context of chronic and emotionally distressing illnesses like breast cancer, places caregivers at substantial risk for psychological morbidity [19].

The exacerbation of psychological burden, like depressive symptoms, may be influenced by multiple factors such as relationship to the patient, family structure, financial constraints, and socioeconomic status, which may influence the severity of depressive symptoms. In our study, spouses were more likely to develop moderate to severe depressive symptoms. The findings are similar to a study conducted in India, where spousal caregivers and those with financial responsibilities were more vulnerable to depressive symptoms [9].

Financial burden plays a pivotal role in the mental well-being of caregivers. Many breast cancer patients in Pakistan rely on out-of-pocket expenses for treatment, with minimal access to insurance or government aid. This can lead to financial toxicity, adding another layer of psychological strain for caregivers, especially if they are also the primary earners of the household [20, 21]. Like previous findings, caregivers who earn less money are more likely to experience depressive symptoms. The reason could be that people with low monthly income often experience financial toxicity, which refers to the stress and burden of managing expenses with limited resources, ultimately contributing to the development of depressive symptoms [22].

In Pakistani culture, caregiving is primarily a family responsibility, placing significant emotional and physical strain on caregivers. In Pakistan, joint family systems provide shared support, while nuclear family caregivers often manage responsibilities alone, increasing stress and depressive symptoms [23]. In contrast to previous studies, our study reported slight increased prevalence of depressive symptoms in joint families. This can be due to increased domestic and caregiving burden and complex family dynamics [24].

Culturally, caregiving in Pakistan is often considered a familial obligation, predominantly undertaken by female family members, who frequently balance caregiving with household responsibilities. This dual burden may contribute to emotional exhaustion and increased risk of depressive symptoms [18].

In the present study, no statistically significant association was observed between caregiver gender and depressive symptoms. Although a higher proportion of male caregivers in the sample screened positive for depressive symptoms, this likely reflects the gender distribution of the study population rather than a true gender-based difference.

Other factors that might influence the higher prevalence rate of depressive symptoms include lack of awareness regarding mental health care, scarcity of structured psychosocial interventions tailored to caregivers, stigma surrounding mental health, which may discourage caregivers from seeking professional help, and this may explain why depressive symptoms often go unrecognized and untreated [25]. Furthermore, breast cancer itself is heavily stigmatized in many parts of the country, adding to the caregiver’s psychological distress due to fear, social isolation, and uncertainty about the patient’s prognosis [26].

It is crucial to integrate routine psychological assessments, provide access to counselling services, and establish caregiver support groups in patient care plans. Training healthcare providers to recognize signs of caregiver distress and collaborate with mental health professionals to deliver timely interventions. Additionally, public awareness about caregiver mental health and the importance of seeking support is essential. Along with this, the development of caregivers’ support services and mental health resources tailored to Pakistani cultural contexts.

Limitations

In interpreting the prevalence estimates, it is important to recognize the potential influence of reporting and selection biases. Cultural reluctance to disclose mental health symptoms may have contributed to underreporting, suggesting that the true burden of depressive symptoms among caregivers may be higher than observed. Despite this, the prevalence identified in this study remains substantial and highlights the need for systematic mental health screening and support for caregivers within oncological settings.

Several limitations should be considered. The cross-sectional design precludes causal inference, and some caregivers may have experienced depressive symptoms prior to assuming caregiving roles. The use of convenience sampling and recruitment from tertiary care hospitals may limit generalizability, as caregivers from rural areas or those providing home-based care may be underrepresented. Reliance on self-reported measures introduces the possibility of reporting bias, and non-response bias cannot be excluded, as caregivers experiencing severe psychological distress may have declined participation. The high prevalence of hypertension observed in this study may reflect self-reported comorbidities among caregivers in a hospital-based sample, which may overestimate true population prevalence.

An additional limitation of this study is the absence of confirmation of depressive symptoms using structured clinical diagnostic interviews. Although the PHQ-9 is a well-validated screening instrument aligned with DSM-5 criteria, it does not replace a formal psychiatric diagnosis. Consequently, the prevalence estimates reported in this study reflect probable depressive symptoms rather than clinically confirmed major depressive disorder. This limitation should be considered when interpreting the findings, and future studies incorporating diagnostic interviews are warranted to validate and extend these results.

Conclusion

Caregivers of breast cancer patients represent a vulnerable population with a high prevalence of depressive symptoms. The PHQ-9 demonstrated a multidimensional structure comprising affective–cognitive, somatic, and suicidality/self-worth domains. These findings highlight the need for routine psychological screening and targeted mental health support within oncological care settings. Given the cross-sectional design, the use of convenience sampling, and hospital-based recruitment, these findings should be interpreted with caution.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Abbreviations

MDD

Major Depressive Disorder

PHQ-9

Patient Health Questionnaire-9

EFA

Exploratory Factor Analysis

CFA

Confirmatory Factor Analysis

KMO

Kaiser-Meyer-Olkin

RMSEA

Root Mean Square Error of Approximation

CFI

Comparative Fit Index

TLI

Tucker-Lewis Index

χ²

Chi-Square

OR

Odds Ratio

CI

Confidence Interval

SD

Standard Deviation

BMI

Body Mass Index

Author contributions

Conceptualization: Muhammad Usman Hashmi; Fatima Azhar. Methodology: Muhammad Usman Hashmi, Fatima Azhar, Pakeezah Tabasum, Imran Saeed, Zainab Azhar, Muhammad Huzefa Mehmood, Shahzeena Bibi, Hussain Ramzan, Momina Ahmed, Wajeeha Sarwar. Formal analysis and investigation: Muhammad Usman Hashmi, Writing – original draft preparation: Muhammad Usman Hashmi, Fatima, Azhar, Pakeezah Tabasum, Imran Saeed, Zainab Azhar, Muhammad Huzefa Mehmood, Shahzeena Bibi, Hussain Ramzan, Momina Ahmed, Wajeeha Sarwar. Writing – review and editing: Muhammad Usman Hashmi, Fatima, Azhar, Pakeezah Tabasum, Imran Saeed, Zainab Azhar, Muhammad Huzefa Mehmood, Shahzeena Bibi.

Funding

This research received no external funding.

Data availability

The data supporting this study’s findings are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki. Written informed consent was obtained from all participants prior to inclusion in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.