One size does not fit all: an application of stochastic modeling to estimating primary healthcare needs in Ethiopia at the sub-national level

Abstract

Background: Primary healthcare systems require adequate staffing to meet the needs of their local population. Guidelines typically use population ratio targets for healthcare workers, such as Ethiopia's goal of two health extension workers for every five thousand people. However, fixed ratios do not reflect local demographics, fertility rates, disease burden (e.g., malaria endemicity), or trends in these values. Recognizing this, we set out to estimate the clinical workload to meet the primary healthcare needs in Ethiopia by region.

Methods: We utilize the open-source R package PACE-HRH for our analysis, which is a stochastic Monte Carlo simulation model that estimates workload for a specified service package and population. Assumptions and data inputs for region-specific fertility, mortality, disease burden were drawn from literature, DHS, and WorldPop. We project workload until 2035 for seven regions and two charted cities of Ethiopia.

Results: All regions and charted cities are expected to experience increased workload between 2021 and 2035 for a starting catchment of five thousand people. The expected (mean) annual clinical workload varied from 2,930 h (Addis) to 3,752 h (Gambela) and increased by 19-28% over fifteen years. This results from a decline in per capita workload (due to declines in fertility and infectious diseases), overpowered by total population growth. Pregnancy, non-communicable diseases, sick child care, and nutrition remain the largest service categories, but their priority shifts substantially in some regions by 2035. Sensitivity analysis shows that fertility assumptions have major implications for workload. We incorporate seasonality and estimate monthly variation of up to 8.9% (Somali), though most services with high variability are declining.

Conclusions: Regional variation in demographics, fertility, seasonality, and disease trends all affect the workload estimates. This results in differences in expected clinical workload, the level of uncertainty in those estimates, and relative priorities between service categories. By showing these differences, we demonstrate the inadequacy of a fixed population ratio for staffing allocation. Policy-makers and regulators need to consider these factors in designing their healthcare systems, or they risk sub-optimally allocating workforce and creating inequitable access to care.

Keywords: Ethiopia; Human resources for health; Monte Carlo; PACE-HRH; Primary healthcare; Stochastic; Sub-national; Workload.

Fig. 1

Predicted clinical workload, by region. The solid line in the center of each plot represents the median value of 100 simulation trials, for the annual hours of direct clinical care time required to meet the needs of a region’s population, from 2021 to 2035. Does not include any non-clinical time such as training, outreach and education, or supply management. Shadings represent the 25th and 75th (dark shading), and 5th and 95th (light shading) percentiles of simulated results. Panel A Estimates are for a catchment area population of 5,000 in 2020 that increases with population growth over the 15 years modeled. Panel B Estimates are for a fixed population of 5,000 individuals, with an evolving population pyramid (based on local demographics, fertility, and mortality), but without total population growth

Fig. 2

Breakdown of service categories as a percentage of the total clinical hours. Values are the median of the percentage calculations from 100 simulation trials. Bars represent values for 2021. Triangles represent values for 2035. The breakdown in 2021 reflects the current state of contributions to total clinical needs by each service category based on existing evidence of disease incidence rates and demographic characteristics of each region. The breakdown in 2035 reflects both the predicted change in demographics and the expected trend in local disease burden based on development observed in the past decades. Pregnancy related services include antenatal care, delivery, childbirth complications, postnatal care, and newborn care. Sexual health includes treatment of menstrual problems, response to and medical treatment for gender-based violence, and sexually transmitted infections management. Sick child care includes case management of fever, diarrhea, parasites, scabies, and pediatric palliative care in children under five. NCDs include screening and care for non-communicable diseases, including hypertension, diabetes, cancer, asthma, and ophthalmic diseases. The service categories not ranked in the top six in any region are not displayed, including first aid, HIV, mental health, neglected tropical diseases, and tuberculosis

Fig. 3

Annual fertility per 100 women, DHS 2019. Bars show the fertility rates by 5-year age band reported in the DHS 2019, by region. Shading indicates the rate at which those baseline values are changing. Annualized rates of year-over-year change are calculated from DHS 2019 and DHS 2011 for all regions except for Addis Ababa. DHS 2019 and DHS 2000 are used to calculate year-over-year change for Addis Ababa

Fig. 4

Total annual hours of direct clinical care for Addis Ababa (Panel A) and Oromia (Panel B), under different fertility assumptions. The first two panels show estimates based on annual growth rates in fertility calculated across different time points in DHS. The third panel shows estimates based on an assumed low-level 0.05% decline per year for comparison purposes. The solid lines in the center of each panel represent median of predicted workload, and the shaded areas represent a 50% simulation interval. Estimates are for a catchment area population of 5,000 in 2020 that increases with population growth over the 15 years modeled. Does not include any non-clinical time such as training, outreach and education, or supply management

Fig. 5

Ratio of clinical workload for the peak month to the average monthly clinical workload, by region. The peak month is the month predicted with the highest clinical workload of each year. The solid lines indicate the median value from 100 simulation trials, and the shaded areas represent a 90% simulation interval

Fig. 6

Predicted clinical workload based on age group classifications, for Amhara, Oromia, and SNNPR. Does not include any non-clinical time such as training, outreach and education, or supply management. Values are the mean of calculations from 50 simulation trials. Population between the age of 0 and 100 are assigned into following groups: infants (i.e., age under 1), under 5 (i.e., age of 1 to 4), 5–14, 15–24, 25–34, 35–44, 45–54, 55–64, 65–100. Panel A Proportion of total clinical workload contributed by each age group, female, and male. Panel B Average annual clinical workload per capita in minutes, by age groups, female, and male