Abstract
The health profile of Bangladesh has improved remarkably, yet gaps in delivering quality health care remain. In response to the need for evidence to quantify resources for providing health services in Bangladesh, this study estimates unit costs of providing the essential services package (ESP) in the not-for-profit sector. This study used a stratified sampling approach to select 18 static clinics, which had fixed facilities, from 330 non-profit clinics under Smiling Sun network in Bangladesh. Costs were estimated from the providers' perspective, using both top-down and bottom-up methods, from July 2014 to June 2015. In total, there were 1115 observations (clients) for the 13 primary care services analysed. The estimated 2015 average costs per visit were: antenatal care ($7.03), postnatal care ($4.57), control of diarrheal diseases ($1.32), acute respiratory infection ($1.53), integrated management of child illness ($2.02), sexually transmitted infections ($4.70), reproductive tract infections ($3.56), tuberculosis ($41.65), limited curative care ($4.30), immunization ($2.23), family planning ($0.72), births by normal delivery ($29.45) and C-section ($114.83). Unit costs varied widely for each service, both between individual patients and among clinic level means. The coefficient of variation for the 13 services averaged 66%, implying potential inefficiencies. In addition, 32.9% of clients were not offered any lab test during the first antenatal visit. The unit cost of essential services differed by the type and location of clinics. Ultra clinics, on average, incurred 37% higher costs than vital (outpatient type) clinics, and urban clinics spent 40% more than rural clinics to deliver a unit of service. The study suggests that inefficiency and quality concerns exist in health service delivery in some facilities. Increasing the volume of clients through demand-side mechanisms and standardization of services would help address those concerns. Unit costs of services provide essential information for estimating resource needs for scaling up the ESPs.