OBJECTIVES: The aim of this study was to understand the costs, outcomes and cost-drivers associated with COVID-19 in-patient care within the South African public-sector during the first-wave of the pandemic.

METHODS: We analysed patient-level resource usage and outcomes through retrospective medical-record reviews and bottom-up micro-costing (provider perspective) in adult patients (>18years) hospitalised for COVID-19 at two tertiary-level hospitals in Gauteng (South Africa) during the first-wave (March-2020 to October-2020). A combination of random and purposive sampling (targeting patients requiring oxygen-support or ICU-care) was used. Admission data was restricted to the first COVID-19 hospitalization per patient. The primary outcome of interest was inpatient costs per patient day (PPD).

RESULTS: The median-cost PPD for 574 admissions was USD 405.21 (interquartile range(IQR): 384.38-460.67) with median length-of-stay(LOS) of 9-days(IQR: 5-14) and in-hospital mortality of 30% (175/574). Median-cost PPD was higher for non-survivors, USD 442.38 (IQR: 395.53-582.03) with median LOS 6-days (IQR: 3-12) in comparison to USD 399.91 (IQR: 380.89 – 433.59) and median LOS 10-days (IQR: 6-15) in patients discharged-alive. The mean-cost PPD was USD 483.58 (standard-deviation(SD): 208.01) in comorbid patients and USD 478.04(SD: 287,80) in patients without comorbidities. A cost-analysis by level-of-care found the median-cost PPD for the general, high-care and ICU wards to be: USD 397.39 (IQR: 385.70- 410.13, n/N=498/574), USD 490.20 (IQR: 467.86-536.24, n/N=166/574) and USD 1 530.59 (IQR: 1 486.34-1 585,55, n/N=72/574) respectively. Patients requiring mechanical-ventilation in ICU had the highest median-cost PPD of USD 1 587.36 (IQR: 1 554.42-1 607.18). Staff costs were the largest driver of cost PPD, followed by infection-prevention and control costs.

CONCLUSIONS: This is the first South African based study to estimate resource requirements for COVID-19 inpatient care using real-world data. Understanding these cost characteristics during the pandemic’s first-wave, which saw an exponential rise in cases and hospital admissions, can help support future risk-management planning in low- and middle-income countries.