The Impact of Health Worker Absence on Health Outcomes: Evidence from Western Kenya

Absenteeism of health workers in developing countries is widespread with some estimates indicating rates of provider absence of nearly 40% (Chaudhry et. al. 2006). This is the first paper to present evidence of the impact of health provider absence on health outcomes. Using longitudinal data from nearly 600 ante-natal care seekers at a rural ante-natal clinic in Western Kenya, we start by showing that women whose first clinic visit coincides with the nurse’s attendance are nearly 60 percentage points more likely to test for HIV and 13 percentage points more likely to deliver in a hospital or health center. Since the benefits of PMTCT services depend on HIV status, we estimate the heterogeneous impact of absence based on women’s self-reported expectations of being HIV-positive. We find that women with a high pre-test expectation of testing HIVpositive and whose first ANC visit coincides with nurse attendance are 25 percentage points more likely to deliver in a hospital or health center, 7.4 percentage points more likely to receive PMTCT medication, 9 percentage points less likely to breastfeed and 10 percentage points more likely to enroll in the free AIDS treatment program at the clinic than similar women whose first visit coincides with nurse absence. These results suggest that nurse attendance has large effects on the behavior of pregnant women that translate into large gains in child and maternal health. ∗The World Bank, e-mail: [email protected] †University of California, San Diego, International Relations / Pacific Studies, e-mail: [email protected] ‡Georgetown University, Public Policy Institute, e-mail: [email protected] §Columbia University, Department of Economics and SIPA, e-mail: [email protected] ¶UNC-Chapel Hill, School of Public Health, [email protected]