Depressive Symptoms and Associated Factors Among Kenyan Health Care Workers

The mental health of health care workers (HCWs) is an important global concern, particularly in low- and middle-income countries (LMICs) such as Kenya, where unique stressors may place HCWs at disproportionate risk for depression.¹ However, depression prevalence and risk factors in this population are not well understood. To address this gap, we evaluated the prevalence of depressive symptoms among Kenyan HCWs and identified associated demographic, psychological, and workplace factors.

Methods |

The Aga Khan University Institutional Scientific and Ethics Review Committee and recruitment sites approved this survey study. Participants provided electronic informed consent and were compensated 500 Kenyan shillings (US $3.86). The study followed the AAPOR reporting guideline.

We used longitudinal data collected from HCWs from 5 hospitals in Nairobi, Kenya, based on the Intern Health Study protocol.² HCWs who enrolled in the longitudinal UZIMA-DS study beginning in April 2023 completed surveys at enrollment and quarterly for 12 months. The surveys assessed depressive symptoms using the 9-item Patient Health Questionnaire (PHQ-9) in English³ and other demographic, psychosocial, and workplace factors (eMethods in Supplement 1).⁴

We used a weighting strategy to reduce participant attrition biases (eMethods in Supplement 1). To capture potential variation, we assessed mean PHQ-9 scores and calculated the proportion of participants meeting PHQ-9 criteria for depression at each survey time point. We used Pearson correlations and χ² tests to examine associations between demographic variables and PHQ-9 scores. Risk factors were entered into a stepwise linear regression model. To explore workplace factors associated with depressive symptoms, we included repeated measures of work hours, workplace discrimination, medical errors, and stressful life events in a generalized estimating equation (GEE) model. Significant demographic and personality variables were added as covariates. Analyses were conducted in R, version 4.4.1 (R Project for Statistical Computing). Two-sided P < .05 was considered significant.

Results |

Among 527 enrolled participants, 514 (97.5%) completed the PHQ-9 questionnaire at enrollment and were included in the analysis (mean [SD] age, 34.1 [8.1] years; 340 women [66.1%] and 174 men [33.9%]) (Table 1 and eFigure in Supplement 1). At 3, 6, 9, and 12 months, 197 (38.3%), 211 (41.1%), 210 (40.9%), and 177 (34.4%) participants completed the depressive symptom surveys, respectively.

Table 1.

Participant Characteristics^a

Characteristic	Unweighted value (N = 514)	Weighted value (N = 514)b
Demographic factors
Age, mean (SD), y	34.1 (8.1)	34.0 (9.7)
Sex
Female	340 (66.1)	339 (66.0)
Male	174 (33.9)	175 (34.0)
Profession
Physician	39 (7.6)	41 (8.0)
Nurse	303 (58.9)	317 (61.7)
Other	172 (33.5)	156 (30.4)
Relationship status
In a committed relationship	388 (75.5)	389 (75.7)
Not in a committed relationship	126 (24.5)	125 (24.3)
No. of children, mean (SD)	1.5 (1.1)	15 (1.3)
Neuroticism, mean (SD)	30.8 (7.7)	30.8 (8.5)
Difficult early family environment, mean (SD)	19.0 (13.1)	19.0 (16.2)
Years of work experience, mean (SD), y	9.7 (6.9)	9.7 (8.2)
Psychosocial and workplace factors
Work hours, mean (SD)	54.2 (24.8)	55.0 (33.9)
Medical errors
Yes	31 (6.0)	31 (6.0)
No	483 (94.0)	483 (94.0)
Workplace discrimination
Yes	220 (42.8)	242 (47.1)
No	294 (57.4)	272 (52.9)
Stressful life events
Yes	321 (62.5)	330 (64.2)
No	193 (37.5)	184 (35.8)

^aUnless noted otherwise, values are reported as No. (%) of participants.

^bA weighting strategy was used to reduce participant attrition biases.

After attrition weighting, 175 (34.0%), 51 (26.0%), 44 (21.0%), 47 (22.5%), and 32 (17.9%) participants met PHQ-9 criteria (≥10) for major depression at enrollment and at 3, 6, 9, and 12 months, respectively. A total of 222 participants (43.1%) met PHQ-9 criteria at 1 or more quarterly assessments, with 87 (16.9%) and 28 (5.5%) meeting criteria for moderately severe (≥15) and severe (≥20) depression.

Among demographic and personality factors established as risk factors for PHQ-9 score increase among medical residents in previous studies² and 1 context-specific factor, age (r = −0.09; P = .04), neuroticism (r = 0.54; P < .001), difficult early family environment (r = 0.42; P < .001), relationship status (r = −0.16; P < .001), and years of work experience (r = −0.08; P = .04) were associated with depressive symptoms. In GEE analysis, neuroticism, difficult early family environment, relationship status, stressful life events, and workplace discrimination were associated with increased depressive symptoms (Table 2).

Table 2.

Factors Associated With Increased Depressive Symptoms

Factor	β (95% CI)	p value
Demographic factors
Age	−0.02 (−0.08 to 0.04)	.48
Female sex	0.23 (−0.46 to 0.91)	.52
Profession
Nurse vs physician	−0.61 (−1.60 to 0.39)	.23
Other vs physician	−0.14 (−1.22 to 0.95)	.80
In a committed relationship	−0.96 (−1.82 to −0.11)	.03
Neuroticism	0.22 (0.18–0.27)	<.001
Difficult early family environment	0.07 (0.04–0.10)	<.001
Years of work experience	0.06 (−0.41 to 0.54)	.79
Psychosocial and workplace factors
Work hours	0.16 (−0.05 to 0.38)	.14
Medical errors	1.13 (−0.17 to 2.43)	.09
Workplace discrimination	1.95 (1.36–2.54)	<.001
Stressful life events	2.51 (1.98–3.05)	<.001

Discussion |

In this 12-month survey study of Kenyan HCWs using validated measures and an established protocol, 43.1% of participants met PHQ-9 criteria for major depression—a rate notably higher than that reported for HCWs in Europe and North America and in the general Kenyan population.^5,6 Consistent with previous studies,² risk factors for depression were higher neuroticism, not being in a committed relationship, having a difficult early family environment, experiencing stressful life events, and experiencing workplace discrimination. Unlike findings from US studies,² work hours and medical errors were not associated with depression.

Potential limitations include underreporting and lack of generalizability beyond nurses and physicians in Nairobi. These findings underscore the need for culturally tailored mental health interventions that address key individual and contextual risk factors for HCWs in LMICs.

Data Sharing Statement:

See Supplement 2.