Factors associated with self-reported mental health symptoms among adult community members in Siaya and Kisumu Counties, Kenya

¹School of Health Sciences, Jaramogi Oginga Odinga University of Science and Technology, Bondo, Kenya, ²Department of Family Medicine and Population Health (FAMPOP), University of Antwerp, Wilrijk, Belgium

^&Corresponding author
Jane Adhiambo Owenga, School of Health Sciences, Jaramogi Oginga Odinga University of Science and Technology, Bondo, Kenya

Introduction    

Mental health is a critical aspect of overall health, but it receives less attention compared to physical health. Depression, anxiety, and bipolar disorder are the most common mental disorders and can occur in anyone. Globally, mental health disorders represent a significant burden, especially in low- and middle-income countries (LMICs) such as Kenya. Approximately one in eight people globally live with a mental disorder, highlighting a significant public health concern [1]. The prevalence of various mental disorders varies by gender and age. Anxiety disorders and depressive disorders are the most common in both males and females [1].

The World Health Organization (WHO) estimates that approximately 1 billion people worldwide suffer from mental disorders [2]. These disorders often coexist with noncommunicable diseases (NCDs) like cardiovascular diseases (hypertension), diabetes, and cancer [3]. Studies indicate that one out of every five patients diagnosed with coronary artery disease or heart failure experiences depression, while one in three stroke survivors also suffers from depression following a stroke [4]. Mental health issues are also prevalent among those with diabetes and cancer [5].

Mental disorders are the leading cause of years lived with disability (YLDs), accounting for one in every six YLDs globally [1]. Most people with mental health conditions do not receive effective care due to unavailable, inaccessible, or unaffordable services, as well as widespread stigma. Cultural differences in belief systems, language, and expressions around mental health affect how and where people seek assistance and whether they recognize mental health issues in themselves and others. Studies have shown a high prevalence of mental health disorders in various countries and settings. A systematic review of 174 surveys across 63 countries found that 17.6% of respondents had a common mental disorder in the past year, and 29.2% had experienced one at some point in their lives. Women showed higher rates of mood (7.3%) and anxiety (8.7%) disorders, while men had higher rates of substance use disorders (7.5%) within the past year. Regional variations were noted, with North and Southeast Asia having the lowest prevalence rates, and English-speaking countries having the highest lifetime prevalence [6].

A South African study reported a lifetime prevalence of common mental disorders at 30.3%, underscoring the significant mental health burden in sub-Saharan Africa [7]. Another study in Australia, analyzing data from the Crossroads II survey, examined the prevalence and factors associated with mental health problems, specifically psychological distress and depression, among rural Victorians. The results indicated that 16.2% of adults experienced psychological distress and 13.6% had depression. Key factors linked to higher mental health issues included being unmarried, smoking, and obesity, while physical activity and community participation were protective. Despite similar mental illness rates in rural and urban areas, rural regions face workforce shortages and higher chronic disease rates. The study suggested that targeted lifestyle interventions may help reduce mental illness risk and distress [8].

Key factors influencing mental health include age, sex, socio-economic status, and health-related factors. Women are more likely to report mental health issues than men, and mental health problems tend to increase with age [9]. Higher education levels are generally associated with better mental health outcomes [10], while low socio-economic status is a significant risk factor for mental health disorders [11]. Chronic diseases such as diabetes and hypertension also increase the prevalence of mental health issues [12]. Social support systems, including treatment supporters and visits by community health promoters, are essential for managing mental health [13]. The economic impact of mental health conditions is enormous, with productivity losses and other indirect costs far exceeding healthcare costs. Schizophrenia is the most costly mental disorder per person, while depressive and anxiety disorders, though less costly per person, are more prevalent and significantly contribute to overall national costs [1].

The World Health Organization (WHO) reports that Kenya ranks fifth among African countries for depression prevalence, with around two million individuals affected [14]. In Kenya, mental health issues are widespread yet underreported and undertreated due to stigma, lack of awareness, and inadequate healthcare infrastructure [15]. One in four individuals seeking healthcare in Kenya has a mental health condition [15]. Estimates indicate that up to 25% of outpatients and 40% of inpatients in Kenyan health facilities suffer from mental conditions, which are influenced by the presence of non-communicable diseases (NCDs) [16].

Studies done within Kenya have also indicated a high prevalence of mental health disorders. For example, a study in Nandi County found that 45% of participants had a lifetime diagnosis of a mental disorder, but only 1.7% were receiving treatment. The findings highlight the need for improved strategies to diagnose and treat mental disorders at the community level and suggest further research on screening methods similar to home-based counseling and testing for HIV [17].

Another study conducted in western Kenya, focusing on four counties (Uasin-Gishu, Trans Nzoia, Bungoma, and Busia), found that mental healthcare services were primarily provided at higher-level facilities, with minimal structures at primary care facilities. None of the counties had a stand-alone policy or dedicated budget for mental healthcare. The national hospital offered a variety of mental health medications, while the other counties had limited options, mainly antipsychotics. All four counties reported submitting mental health data to the Kenya Health Information System (KHIS). There were no clearly defined mental healthcare structures at the primary care level, except for funded projects under the national referral hospital, and the referral mechanisms were poorly defined. Additionally, there was no established mental health research in the counties, except for research affiliated with the national referral hospital [15].

Kisumu faces rising rates of hypertension, diabetes, and cardiovascular morbidity [18-20], while Siaya ranks fourth in HIV prevalence and second in HIV incidence nationally, with increasing NCD-related morbidity [12,21]. Chronic diseases like diabetes and hypertension also raise the prevalence of mental health issues. National survey data further confirm that Western Kenya counties, including Kisumu and Siaya, carry elevated NCD risk factors compared to other regions [12,22].

The Kenyan government has initiated efforts to improve mental health services, such as the Kenya Mental Health Policy 2015-2030 and the Mental Health Action Plan 2021-2025, which emphasize integrating mental health into primary care [15]. Despite these efforts, there is still a significant gap in mental health service provision at the county level through to the community levels. This study, therefore, proposes to characterize the burden of mental health symptoms in Kisumu and Siaya counties and determine the associated factors.

Methods     

Study setting: this study was conducted in Kisumu and Siaya counties, located in the western region of Kenya along the shores of Lake Victoria. Kisumu County covers an area of approximately 2,085.9 square kilometers and has a population of over 1.1 million people, according to the 2019 Kenya Population and Housing Census. Kisumu is characterized by both urban and rural settings, with high levels of urbanization in Kisumu city and rural agricultural communities on the outskirts. The county experiences a tropical climate with two rainy seasons and supports farming, fishing, and trade as major economic activities.

In contrast, Siaya County is predominantly rural and lies to the west of Kisumu. It covers an area of about 2,530.4 square kilometers and has a population of approximately 993,000 people, according to the 2019 Kenya Population and Housing Census. Agriculture, particularly small-scale farming and fishing, forms the backbone of the local economy. Both counties face a high burden of communicable diseases such as HIV/AIDS, malaria, and tuberculosis, alongside a rising trend of non-communicable diseases (NCDs) such as hypertension and diabetes.

Access to healthcare services varies across the two counties. Kisumu has better healthcare infrastructure, including referral hospitals like Jaramogi Oginga Odinga Teaching and Referral Hospital (JOOTRH), sub-county hospitals, and private health facilities. Siaya County, while making strides in improving healthcare delivery, still grapples with infrastructural and human resource limitations, especially in remote rural areas.

Study design: the study employed a quantitative cross-sectional survey design.

Study participants: the population comprised all adults aged 18 years and above in the selected households in Kisumu and Siaya Counties.

Sample size determination: the sample size was calculated using the Kish Leslie formula (1965), where n refers to the estimated minimum sample size required, p is the prevalence of a characteristic in a sample, e is the acceptable margin of error (5%), and z is the confidence interval (CI) set at 1.96 for a 95% CI.

Given the population-based cross-sectional study by Jenkins et al. [23]; carried out in 2015 among households in Nyanza Province, the prevalence of common mental health issues among households was 10.3% [23] therefore, setting the n to 141. Accounting for a 10% non-response rate (14 participants), the final sample size required for this study was calculated to be 155.

Sampling procedure: Siaya and Kisumu Counties, which are two of Kenya´s six counties located within the Lake Victoria Basin, were purposively selected for this study due to their notably high disease burden. Specifically, Siaya ranks fourth in HIV prevalence and second in HIV incidence nationwide. To ensure a representative and systematic approach, a multistage sampling strategy was implemented, starting at the sub-county level and proceeding to individual households. At each administrative level, a random selection method was applied.

Of the six sub-counties in Siaya, Alego-Usonga, Bondo, Rarieda, Gem, Ugunja, and Ugenya, Alego-Usonga was randomly selected. Within Alego-Usonga, Central Alego Ward, Kakum-Kombewa sublocation was randomly selected for the study. All 11 villages within Kakum-Kombewa in Central Alego Ward were included in the sampling frame. Of the seven sub-counties in Kisumu, Kisumu Central, Kisumu East, Kisumu West, Seme, Muhoroni, Nyakach, and Nyando, Seme was randomly selected. Within Seme, Central Seme Ward, East Othany Sub-location was randomly selected for the study. All the 9 villages within East Othany in Central Seme ward were included in the sampling frame.

Households in each village were selected through systematic random sampling. The households are listed by village. Village and household numbers were used to randomly select the first household, then every household with an even number was selected until the selected sample size for the village was reached. The number of households selected per village was proportionate to its size (probability proportional to size-PPS) to ensure geographic representativeness across the sub-location. In Alego-Usonga Siaya County, 112 participants were selected, and in Seme, Kisumu County, 128, totalling 240 participants. All eligible participants who provided informed consent were enrolled in the study. After the calculated sample size of 155 participants was reached, recruitment continued through systematic random sampling until a total of 240 participants.

Data collection tool and procedure: data were collected using a structured questionnaire. The items employed were adapted from the Patient Health Questionnaire-9 (PHQ-9), which is a widely used and validated instrument for screening depression. The PHQ-9 was developed in the late 1990s as part of the ‘Primary Care Evaluation of Mental Disorders (PRIME-MD)´ and has been extensively validated in both clinical and community settings. It is endorsed by the U.S. National Institute of Mental Health (NIMH) and the World Health Organization (WHO). The tool was adapted from PHQ-9 (Kroenke et al. 2001) [24], validated in African settings (Monahan et al. 2009 [25]; Akena et al. 2012 [26]), and expanded to 11 items across physical, affective, and cognitive domains. Scoring used a 4-point Likert scale (range 11-44), with symptoms defined as ≥23; reliability was acceptable (Cronbach´s α=0.77-0.80). The study was conducted from September 2023 to November 2023. The questionnaire was administered face-to-face by trained research assistants to 240 community household members in Kisumu and Siaya Counties. The researcher randomly picked the first respondent in every village studied, then calculated the sampling interval based on the number of households in that village; the subsequent respondents were then picked based on the sampling interval.

Data analysis: descriptive statistics were computed for all variables, with categorical variables presented as frequencies and percentages, and continuous variables as means with standard deviations. Chi-square tests (or Fisher's exact test where appropriate) were used to examine associations between socio-demographic characteristics and mental health status. Exploratory factor analysis (EFA) using polychoric correlations was conducted to identify the underlying factor structure of mental health symptoms and establish the contribution of various symptoms to overall variance. Factors with eigenvalues >1 were retained. Factor loadings >0.40 were considered substantial. Internal consistency reliability was assessed using Cronbach's alpha, with values ≥0.70 indicating acceptable reliability. Logistic regression models were fitted to identify factors associated with mental health status. Unadjusted odds ratios (OR) were computed for each predictor separately. An adjusted multivariable model was then fitted, including all predictors simultaneously, to obtain adjusted odds ratios (aOR). Results are presented with 95% confidence intervals (CI) and p-values, with statistical significance set at p<0.05. All analyses were performed using Stata version 16.

Ethical approval: the ethical approval to carry out the study was also sought from the Ethical Review Board of Jaramogi Oginga Odinga Teaching and Referral Hospital (Ref No: ISERC/JOOTRH/026/24). A permit to carry out research was sought from the National Commission for Science, Technology and Innovation (NACOSTI/P/25/414763). A letter of acceptance to carry out the study was sought from the Kisumu and Siaya County health departments. Informed consent from the respondents was sought by adequately explaining to the study participants the objectives of the study and completing a consent form. The respondents were assured of the confidentiality of all the information that they provided. There were no potential risks involved since there was no invasive procedure done.

Results     

In total, 240 respondents completed the survey. Kisumu had 115 (47.9%) respondents, with 90 (50.6%) reporting no mental health symptoms and 25 (40.3%) reporting having mental health symptoms. Siaya had 125 (52.1%) of respondents, with 88 (49.4%) reporting no mental health issues and 37 (59.7%) reporting having mental health symptoms (p=0.165). Significant differences were observed across age categories (p=0.001). The highest prevalence of mental health issues was 42 (67.7%) among those aged 60+ years. Males comprised 76 (31.7%) of the sample, with 61 (34.3%) having no mental health symptoms and 15 (24.2%) having mental health issues. Females comprised 164 (68.3%), with 117 (65.7%) having no issues and 47 (75.8%) having mental issues (p=0.142). Married individuals represented 191 (79.6%) of the sample, with 146 (82.0%) having no mental health issues and 45 (72.6%) having mental symptoms (p=0.002), while widowed individuals, 10 (16.1%), had a notably higher prevalence of mental health issues. There were significant differences in education levels (p<0.005), with no formal education, 17 (27.4%), reporting more cases with mental health symptoms. No significant difference was observed in socioeconomic status (SES) (p=0.594), while a significant association was found with mental health symptoms (p<0.005), with those diagnosed with NCDs (diabetes mellitus (DM) and hypertension (HTN)) 38 (61.3%) reporting having mental health symptoms. Having a caregiver was significantly associated with mental health issues (p<0.005), with 22 (35.5%) of those with symptoms having a caregiver. Similarly, being visited by a community health promoter had significant differences (p=0.004), with those not visited reporting more mental health issues 17 (27.4%). Table 1 illustrates characteristics of community household members in Siaya and Kisumu counties, stratified by mental health symptoms.

Factor analyses for the symptoms construct of mental health: in order to establish the contribution of various factors to the variance in mental symptoms, factor analysis was performed. Loss of energy (factor 1) had an eigenvalue of 3.97433, indicating that it explains most of the variance in mental health symptoms. The rest of the factors explained less and less variance in mental health symptoms. Table 2 provides the results of factor analysis for mental health symptoms.

The exploratory factor analysis results for the mental health symptoms and scale reliability: regarding symptom prevalence in the physical domain, 90 (37.5%) reported never having a loss of energy, while 112 (46.7%) reported sometimes experiencing a loss of energy, and 14 (5.8%) always had a loss of energy (factor loading 0.7795, alpha 0.7687). In the affective domain, 135 (56.3%) never lost interest in doing things they used to enjoy, 50 (20.8%) sometimes lost interest, and 16 (6.7%) always lost interest (factor loading 0.5032, alpha 0.7897). In the cognitive domain, the study also reported that 213 (88.8%) had never had suicidal thoughts, while two (0.8%) always had suicidal thoughts (factor loading 0.4964, alpha 0.8044). One hundred thirty-five (56.3%) never felt nervous, with 62 (25.8%) sometimes feeling nervous while three (1.3%) always felt nervous (factor loading 0.7603, alpha 0.7715), similarly, 137 (57.1%) never had trouble concentrating, 47 (19.6%) sometimes had trouble concentrating while four (1.7%) always had trouble concentrating (factor loading 0.6771, alpha 0.7793). Physical domain reported 117 (48.8%) never had a change in sleeping patterns, while 78 (32.5%) sometimes had a change in sleeping patterns, with 17 (7.1%) always having a change in sleeping patterns (factor loading 0.6195, alpha 0.7774). Table 3 illustrates the frequencies of mental health symptoms and their factor loadings, along with scale reliability (alpha).

Factors associated with mental health status: to establish factors associated with mental health status, logistic regression analysis was performed for each variable separately (unadjusted model) and collectively (adjusted model), with estimated odds ratios (OR) and 95% confidence intervals (CI) computed.

In the unadjusted analyses, several factors showed significant associations with mental health status. Age category 30-39 years was significantly associated with lower odds of mental health symptoms compared to the youngest age group (10-19 years) (OR=0.03, 95% CI: 0.001-0.87, p=0.041). Being widowed was significantly associated with increased odds of mental health symptoms compared to being married (OR=6.49, 95% CI: 2.11-19.97, p=0.001).

Higher educational attainment showed a strong protective association against mental health symptoms. Compared to those with no formal education, individuals with primary education had 82% lower odds (OR=0.18, 95% CI: 0.07-0.44, p<0.001), those with secondary education had 88% lower odds (OR=0.12, 95% CI: 0.05-0.34, p<0.001), and those with university/college education had 87% lower odds (OR=0.13, 95% CI: 0.03-0.50, p=0.003) of experiencing mental health symptoms. Having been diagnosed with a non-communicable disease (diabetes mellitus or hypertension) was strongly associated with increased odds of mental health symptoms (OR=5.29, 95% CI: 2.85-9.82, p<0.001). Conversely, having been visited by a community health promoter showed a protective association, with 65% lower odds of mental health symptoms (OR=0.35, 95% CI: 0.17-0.73, p=0.005).

In the adjusted model, which controlled for all variables simultaneously, none of the associations remained statistically significant at the conventional p<0.05 level. However, some trends were notable: being widowed showed elevated but non-significant odds (aOR=14.48, 95% CI: 0.57-366.8, p=0.105), and having been visited by a community health promoter continued to show a protective trend (aOR=0.33, 95% CI: 0.05-2.09, p=0.238). The loss of statistical significance in the adjusted model suggests potential multicollinearity among predictors and/or insufficient statistical power due to the relatively small sample size (N=240), particularly when multiple variables are included simultaneously. Table 4 presents the complete odds ratios (OR) and adjusted odds ratios (aOR) with 95% confidence intervals (CI) for factors associated with mental health status in both unadjusted and adjusted models.

Discussion     

This study among adults in Kisumu and Siaya Counties (Western Kenya) found a high burden of self-reported mental health symptoms (25.8%). In unadjusted analyses, symptoms were significantly associated with being widowed, lower education, and having an NCD (e.g., hypertension/diabetes), while recent engagement with community health promoters (CHPs) was protective. Individuals with primary, secondary, or tertiary education had 82-88% lower odds of symptoms than those with no formal schooling, suggesting education may confer protection via health literacy, coping skills, and socioeconomic opportunities. The strong unadjusted association between NCD diagnosis and symptoms (OR=5.29, p<0.001) is consistent with the bidirectional links between chronic illness and mental health [12], and with evidence that mental disorders co-occur with NCDs such as hypertension, diabetes, and cancers [5,12]. Patterns by age and marital status align with broader literature reporting variation by age and socioeconomic status (WHO 2022) [1] and Patel et al. [9], although gender was not statistically significant in this sample. Our results differ from Dashputre et al. [8], who reported higher symptoms among the unmarried, likely reflecting contextual differences and an older sample here.

In fully adjusted models, associations lost statistical significance, though directions were consistent; wide confidence intervals (e.g., widowed aOR=14.48; 95% CI:0.57-366.8) suggest limited power and collinearity among sociodemographic, health, and service engagement factors. Symptom profiles spanned physical, affective, and cognitive domains, underscoring the need for integrated responses. Programmatically, findings support embedding routine mental health screening and brief psychosocial care within NCD clinics and primary care and strengthening community health promoters´ (CHPs) platforms for household-level awareness, stigma reduction, basic screening, and referral, particularly for widowed individuals, older adults, and people living with chronic illness. Leveraging caregivers, faith-based and community groups, and NCD support groups may enhance acceptability and sustainability, while periodic quality of life assessments can help surface unmet needs and guide targeted community mental health interventions.

Study, strengths, and limitations: to our knowledge, this is the first post-COVID study exploring mental health symptoms and associated factors in these Kenyan communities, providing valuable epidemiological insights. However, several limitations should be noted. First, the small sample size limits statistical power and precision of effect estimates, particularly in multivariable models. Second, data were self-reported, which may introduce response bias, though such measures remain useful for capturing perceptions and experiences. Third, potential selection bias exists, as participants who consented may differ from those who declined, affecting representativeness. Fourth, the screening tool may not fully capture cultural nuances or clinical diagnoses. Finally, the study focused on older adults, excluding adolescents and youth, a population increasingly affected by mental health challenges. Future research should include larger, more diverse samples and validated, culturally adapted tools to provide a comprehensive picture of mental health status in the region.

Conclusion     

This study reveals a high burden of mental health symptoms (25.8%) among adults in Western Kenya, which is strongly linked to widowhood, low education, and NCDs in unadjusted analyses, while engagement with community health promoters showed protective effects. Symptoms spanned physical, affective, and cognitive domains, with loss of energy as the most prominent. These findings highlight the need for integrated mental health responses within existing health systems. Immediate actions include embedding mental health screening in NCD and primary care services, equipping community health promoters with simple symptom checklists, and creating targeted support for high-risk groups such as widowed individuals and those with chronic illnesses. Strategic priorities should focus on stigma reduction campaigns, capacity building for primary care providers in psychosocial interventions, and strengthening referral systems. Future research should involve larger samples, include adolescents and youth, use culturally adapted screening tools, and explore implementation pathways for integrated mental health and NCD care. Embedding mental health into routine NCD care and leveraging community health platforms can help address the dual burden of chronic disease and mental health, improving outcomes and quality of life for vulnerable populations.

Competing interests     

The authors declare no competing interests.

Authors' contributions     

Jane Adhiambo Owenga led the conceptualization and study design, coordinated the survey and data collection, and drafted the manuscript; Ivy Akinyi contributed to the literature review and manuscript writing; Sylvester Okumu Ogutu supported study design and tool development and participated in the literature review; Japheth Ogol Ouma programmed the data collection tools, conducted data analysis, contributed to manuscript drafting, and provided technical input to improve the paper. All the authors read and approved the final version of this manuscript.

Acknowledgments     

We are deeply grateful to the households and study participants in Siaya and Kisumu Counties for their time and insights, the research assistants for their dedication during data collection, and the community health promoters for their invaluable support throughout the study.

Tables     

Table 1: socio-demographic characteristics of community household members, recruited from Kisumu and Siaya Counties, Kenya, September to November 2023 (N=240)

Table 2: factor analysis of the symptom constructs of mental health among community household members in Kisumu and Siaya Counties, Kenya, September to November 2023 (N=240)

Table 3: exploratory factor analysis results for mental health symptoms and scale reliability among community household members in Kisumu and Siaya Counties, Kenya, September to November 2023 (N=240)

Table 4: factors associated with mental health status among community household members in Kisumu and Siaya Counties, Kenya, September to November 2023 (N=240)

References