The Effect of Healthcare Worker Density on Maternal Health Service Utilization in Sub-Saharan Africa

View More View Less
  • 1 Stanford University School of Medicine, Stanford, California

Facility births and antenatal care (ANC) are key to improving maternal health. This study evaluates the relationship between physician and nurse/midwife densities and the use of key maternal health services across sub-Saharan Africa (SSA). The authors matched individual-level maternal health service indicators from Demographic and Health Surveys between 2008 and 2017, to country-level physician and nurse/midwife per-capita densities, across 35 SSA countries. The authors performed univariate and multivariate probit regression analyses to evaluate the association between healthcare worker (HCW) densities and facility births as our primary outcome and with additional ANC services as secondary outcomes. We controlled for established maternal health predictors, including literacy, child marriage, reported problems accessing healthcare, GDP per capita, political instability, and government effectiveness scores. HCW density across SSA was low at 0.13 physicians and 0.91 nurses/midwives per 1,000 people, compared with 2010 worldwide mean densities of 1.33 and 3.07, respectively. The probability of facility birth increased by 9.8% (95% CI: 2.1–17.5%) for every additional physician per 1,000 people and 8.9% (95% CI: 7.1–9.7%) for every additional nurse/midwife per 1,000 people. HCW densities were also associated with increased likelihood of ANC by the respective provider type, and with antenatal testing for preeclampsia (urine and blood pressure checks). Other ANC services demonstrated variable relationships with HCW densities based on provider type. In 35 SSA countries, HCW density was positively associated with many key measures of maternal health service utilization including facility birth and ANC testing for preeclampsia.

Author Notes

Address correspondence to Kelly Z. Aluri, Stanford University School of Medicine, 291 Campus Drive, Stanford, CA 94305. E-mail:

These authors contributed equally and are co-first authors.

Disclosure: The data underlying this article are available through the Demographic and Health Surveys at or Software code for the analysis is available upon email request to the corresponding author. Ethics committee approval was not required for this study on publicly available, deidentified data.

Financial support: J. I. R. is supported by NIH Training Grant 5T32AI052073-14 and by a Stanford University Maternal Child Health Research Institute grant. None of the authors received financial or material support for the research and work in this manuscript.

Authors’ addresses: Joelle I. Rosser, Kelly Z. Aluri, Arielle Kempinsky, Shannon Richardson, and Eran Bendavid, Division of Infectious Diseases and Geographic Medicine, Stanford University School of Medicine, Stanford, CA, E-mails:,,,, and