Cost estimation alongside a multi-regional, multi-country randomized trial of antenatal ultrasound in five low-and-middle-income countries

Background: Improving maternal health has been a primary goal of international health agencies for many years, with the aim of reducing maternal and child deaths and improving access to antenatal care (ANC) services, particularly in low-and-middle-income countries (LMICs). Health interventions with these aims have received more attention from a clinical effectiveness perspective than for cost impact and economic efficiency. Methods: We collected data on resource use and costs as part of a large, multi-country study assessing the use of routine antenatal screening ultrasound (US) with the aim of considering the implications for economic efficiency. We assessed typical antenatal outpatient and hospital-based (facility) care for pregnant women, in general, with selective complication-related data collection in women participating in a large maternal health registry and clinical trial in five LMICs. We estimated average costs from a facility/health system perspective for outpatient and inpatient services. We converted all country-level currency cost estimates to 2015 United States dollars (USD). We compared average costs across countries for ANC visits, deliveries, higher-risk pregnancies, and complications, and conducted sensitivity analyses. Results: Our study included sites in five countries representing different regions. Overall, the relative cost of individual ANC and delivery-related healthcare use was consistent among countries, generally corresponding to country-specific income levels. ANC outpatient visit cost estimates per patient among countries ranged from 15 to 30 USD, based on average counts for visits with and without US. Estimates for antenatal screening US visits were more costly than non-US visits. Costs associated with higher-risk pregnancies were influenced by rates of hospital delivery by cesarean section (mean per person delivery cost estimate range: 25–65 USD). Conclusions: Despite substantial differences among countries in infrastructures and health system capacity, there were similarities in resource allocation, delivery location, and country-level challenges. Overall, there was no clear suggestion that adding antenatal screening US would result in either major cost savings or major cost increases. However, antenatal screening US would have higher training and maintenance costs. Given the lack of clinical effectiveness evidence and greater resource constraints of LMICs, it is unlikely that introducing antenatal screening US would be economically efficient in these settings–on the demand side (i.e., patients) or supply side (i.e., healthcare providers). Trial registration: Trial number: NCT01990625 (First posted: November 21, 2013 on https://clinicaltrials.gov). The primary objective of this economic study performed alongside the First Look CRCT was to estimate average patient-level costs in participating countries comparing the US group to standard of care. We collected resource use for ANC visits and delivery-related events based on health facility location and type of delivery (cesarean section or vaginal birth). In addition, we wanted to estimate facility-based costs associated with complications experienced during pregnancy, at the time of delivery, and during a six-week post-partum period. Our approach was to establish unit costs for particular resource use, and to construct key event average costs from these unit cost estimates. We developed these cost estimates through site visits in collaboration with local health researchers, thus providing a clinic or hospital facility perspective for our use and cost assessment. We examined relative prices within countries for different resources and events, and converted them to 2015 United States dollars (USD) for general comparisons across countries, despite differences in health systems in participating countries. Our study population included pregnant women enrolled in the CRTC intervention clusters and women in control (standard of care) clusters participating in the Global Network Maternal Newborn Health Registry (GN MNHR) in five countries: the Democratic Republic of Congo (DRC), Guatemala, Kenya, Pakistan, and Zambia [13]. Each cluster was a defined geographic area generally served by a single health center and its catchment area with roughly 500 births per year. Further details about the CRCT study population and study design are detailed elsewhere [12]. The CRCT adhered to CONSORT guidelines and the CONSORT diagram is presented in Supplemental Figure 1. The overall sample included normal (low-risk) pregnancies and high-risk pregnancies identified as multiple gestation, placenta previa, and breech presentation. Each participant provided written informed consent prior to study participation, including for the use of their study data. We used three primary data sources to inform our economic assessment. First, for intervention clusters, we used the CRCT data on the number of ANC visits, delivery location, type of delivery (vaginal or cesarean section), and select pregnancy-related complications collected during a portion of the trial. Second, similar data for the control-group clusters were obtained from the GN MNHR [13]. Third, we collected average time and cost estimates for typical ANC, delivery, and complication-related interventions or events using consistent, standard data collection tools in each country, with local health experts interviewing hospital administrators and using reimbursement schedules as proxies for costs when available. Due to similar outcomes observed in the CRCT [12], we report country-level averages for ANC and delivery-related resource use for all intervention and control patients combined within each country for the average probability parameters of visit-related or facility-related events. Our cost estimates focused on public health facilities in participating countries, at the local health clinic level and the hospital (facility) level. We also included private facility data collection in Pakistan, DRC, and Kenya to provide a range of estimates. Data collection in Zambia included several public facilities, most often with similar fee structures. We obtained data from one representative rural public hospital in Guatemala. The perspective for the cost assessment was a health facility perspective and this can be viewed as a health delivery-system perspective, accounting for visits, deliveries, and complication events. We focused on facility-related reimbursement for services, despite the fact that some countries do not have public or private health insurance provision for our study participants for ANC or delivery services, implying mothers or families incur the financial burden rather than via payer reimbursement. The period of our data collection included visits or complications during the ANC pre-delivery period, the delivery event, and a 6-week period post-delivery during which complications were managed in a hospital setting. We collected data on average gross-level costs for typical ANC resource use and delivery-related events, as well as low and high ranges for each type of service. We used per patient average cost estimates to calculate ANC and in-hospital costs for patient care. Our data tools included the collection of inpatient and outpatient ANC costs, as well as costs for delivery-related events and pregnancy-related complication event care. For the cost estimation, we did not have data for specific CRCT patient events in the hospital or assess patient-related clinic or hospital records such as for hospital stays related to pregnancy complications or delivery events. All costs were reported in local currencies and converted to 2015 USD for comparisons [14]. We converted average cost estimates to 2015 USD with exchange rates listed in Table 1 [15–19]. We did not calculate costs for US equipment, training sonographers, or maintaining equipment/training. Select population-level statistics for countries participating in the trial (2015) Notes: aConverted to per 100 live births from per 100,000, bConverted to per 100 live births from per 1000 cLatest available mortality statistics for Guatemala were from 2010 We calculated the average amount of resources used associated with ANC visits (with and without antenatal screening US), procedures, or interventions including outpatient and inpatient services, hospital stays and in-facility delivery procedures, and resource use based on reported complications in a portion of study participants (trial and registry). Cost and probability estimates for standard ANC services and complication-related services performed in facility settings for maternal and neonatal complications were used to estimate resource use and event costs. Although antenatal screening US use was primarily based on use in one randomized cohort in the CRCT (i.e., intervention group), we provide an overall estimate of ANC costs based on the average number of antenatal screening US visits recorded for patients in the trial. We calculated a combined ANC visit cost estimate that included non-US visits and antenatal screening US visits recorded for each country-level sample from the CRCT, subtracting the estimated number of US visits from the overall number of ANC visits to avoid double counting. For ANC visits, the average number of visits reported in countries was multiplied by the outpatient facility cost of an ANC visit, for visits without US and for visits including antenatal screening US. The average number of ANC visits calculation used data from the GN MNHR and the CRCT. The average number of overall ANC visits in countries was calculated net of the proportion of patients with one, two, three, or four or more US visits using reported visits from the CRCT. A small percentage of persons (all countries < 1%) with more than four visits with antenatal screening US were assumed to have four US visits for ANC cost calculations. Our facility-level “gross costing” reimbursement-based approach used reported average cost data for each type of visit or other service provided in a health facility, or outpatient setting for ANC visits. Although country-level data collection included estimates from several facilities, available services differed at specific facilities within countries, resulting in some zero dollar cost estimates for services when they were not available in a particular facility. For country-level average cost estimates, only cost estimates greater than zero were used for base-case estimates when a resource item was available at facilities in countries. Some services in countries were bundled as part of overall delivery-related episode-based reimbursement and thus did not have item-level cost estimates. Probabilities for delivery location were obtained from the CRCT and GN MNHR data, indicating whether deliveries occurred at one of three settings: local/village, non-hospital clinic, or a hospital/health center. Average costs for deliveries based on delivery location were estimated using the unit cost estimate for services and the proportion of mothers reported to have delivered using cesarean section. We used CRCT and GN MNHR data for cesarean delivery rates and assumed cesarean deliveries occurred only in the hospital facility for costing purposes. For these deliveries we also used country-level unit costs for hospital stays associated with cesarean deliveries. We did not assume hospital stay costs for non-cesarean deliveries. Delivery costs for a subset of mothers identified with higher-risk pregnancies are also reported. Countries with public reimbursement mechanisms, e.g., Kenya and Zambia, reported “lump sum” estimates for deliveries in public facilities which included payment (cost) for services associated with addressing birth-related complications in mothers and children. Since we did not have a facility unit cost estimate for local-level or clinic-level deliveries, base case estimates assumed a proportion of the hospital-based delivery cost for a non-cesarean delivery, at a 25% rate for local/village deliveries, and 50% for non-hospital clinic deliveries. Base-case estimates used reported unit cost averages and a range for cesarean delivery costs for the hospital services. The use of select hospital/referral services for treatment of complications was reported in a sub-sample of the CRCT for maternal complications in the GN MNHR for neonatal complications, along with major services used. Since we expected hospital-based management for complications to be a higher-cost item, we incorporated pregnancy-related complication collection at sites, to complement similar GN MNHR data on complications. We used these reported proportions of maternal and neonatal complications and applied unit cost estimates to each type of service to estimate average facility-based management costs for these types of typical complications. Maternal health complication management included treatment with antibiotics, hysterectomy or other surgery, use of anesthesia or blood transfusion, and whether inpatient US or chest radiograph was used. Select neonatal services use included hospitalizations, neonatal bath, antibiotics, oxygen, mechanical ventilator, medical eye or cord care, and vitamin K treatment. Unit prices for services were multiplied times the proportion of patients in the CRCT or GN MNHR reporting use of these services specifically for pregnancy-related complications. Countries with public reimbursement mechanisms, e.g., Kenya and Zambia, reported “lump sum” estimates for deliveries in public facilities which represented full reimbursement (cost) for services associated with managing complications in mothers and children. Thus select resource use items in countries are indicated as zero USD cost estimates. In some cases zero USD unit costs indicate the lowest-resourced rural hospitals sampled did not have all services/equipment available, such as in DRC. Sensitivity analysis was conducted for select ANC visit costs and delivery costs for countries. ANC visits with and without antenatal screening US were primary variables of interest and the low and high ranges of these outpatient items were used in testing the impact on ANC visit-related average costs, using the average number of each type of visit in our country-level cohorts. For each country, we first used the low range cost estimates of ANC visits with and without US as well as the high ranges of each type of visit (two-way sensitivity). We then used the low and high range of either ANC with or without US and the base-case cost estimate of the other type of visit (one-way sensitivity analysis). Likewise, low and high ranges of delivery cost variables were tested in sensitivity analysis, for both cesarean and non-cesarean modes of delivery. In addition, since our data collection focused on hospital facility-level cost estimation, we did not have cost estimates for home- or village-level deliveries or clinic-level deliveries. Therefore, we adjusted our proportional cost assumptions for home/village deliveries and clinic-level deliveries from 25 and 50% (base case) of hospital-based non-cesarean deliveries, respectively, and tested 10 and 35% of hospital costs for home−/village-level and 25 and 60% for clinic-level. To provide a high-end cost range for cesarean section deliveries with complications, sensitivity analyses were performed by incorporating additional hospital stay cost estimates associated with 100% of cesarean deliveries, in addition to the standard average costs for cesarean section deliveries.