The impact of post-election violence on HIV and other clinical services and on mental health-Kenya, 2008.

INTRODUCTION: In December 2007, civil disruption and violence erupted in Kenya following national elections, displacing 350,000 people and affecting supply chains and services. The Kenyan government and partners were interested in assessing the extent of disruption in essential health services, especially HIV treatment.
METHODS: A two-stage cluster sampling for patients taking antiretroviral therapy (ART) was implemented ten weeks after elections, March 10-21, 2008, at twelve health facilities providing ART randomly selected in each of the three provinces most affected by post-election disruption-Rift Valley, Nyanza, and Central Provinces. Convenience samples of patients with tuberculosis, hypertension, or diabetes were also interviewed from the same facilities. Finally, a convenience sampling of internally displaced persons (IDPs) in the three provinces was conducted.
RESULTS: Three hundred thirty-six IDPs in nine camps and 1,294 patients in 35 health facilities were interviewed. Overall, nine percent of patients reported having not returned to their routine health care facility; 9%-25% (overall 16%) reported a temporary inability for themselves or their children to access care at some point during January-February 2008. Less than 15% of patients on long-term therapies for HIV, tuberculosis, diabetes, or hypertension had treatment interruptions compared with 2007. The proportion of tuberculosis patients receiving a ≥45-day supply of medication increased from five percent in November 2007 to 69% in December 2007. HIV testing decreased in January 2008 compared with November 2007 among women in labor wards and among persons tested through voluntary counseling and testing services in Nyanza and Rift Valley Provinces. Patients and their family members witnessed violence, especially in Nyanza and Rift Valley Provinces (54%-59%), but few patients (2.5%-14%, 10% overall) personally experienced violence. More IDPs reported witnessing (80%) or personally experiencing (38%) violence than did patients. About half of patients and three-quarters of IDPs interviewed had anxiety or depression symptoms during the four weeks before the assessment. There was no association among patients between the presence of HIV, tuberculosis, diabetes, and hypertension and the prevalence of anxiety or depression symptoms.
CONCLUSION: More than 85% of patients in highly affected provinces avoided treatment interruptions; this may be in part related to practitioners anticipating potential disruption and providing patients with medications for an extended period. During periods of similar crisis, anticipating potential limitations on medication access and increased mental health needs could potentially prevent negative health impacts.