Financial Implications of Tariffs for Medical Oxygen on Rwandan Public Hospitals’ Finance Management During the Coronavirus Epidemic

BACKGROUND

Medical oxygen therapy is essential in the treatment of respiratory illnesses including severe coronavirus disease (COVID-19), yet supply remains insufficient in many resource-constrained hospitals.¹ Despite the 2017 addition of medical oxygen to the World Health Organization Model List of Essential Medicines for the management of hypoxemia,² the global COVID-19 pandemic increased the global visibility of its therapeutic value and subsequently the need to prioritize its supply. This visibility has in turn exposed the already existing systemic bottlenecks hospitals faced, particularly in resource-limited countries, in providing sustainable access to medical oxygen therapy.³ The scale of oxygen scarcity has been especially significant in resource-limited countries including India⁴ and many countries in sub-Saharan Africa.^5,6 The bottlenecks include limited supply-chain capacity,¹ unreliable electrical power needed to deliver oxygen at the point of care,⁵ inadequate health care provider training at facilities,⁷ and limited capacity and government resources to maintain local oxygen production.^8–10

In Rwanda, oxygen production capacity has grown over time due to the increased need associated with the COVID-19 pandemic. Before the pandemic, there were 9 pressure swing adsorption (PSA) oxygen plants in Rwanda—2 privately owned and 7 owned by public hospitals. The plants at public hospitals produce oxygen for internal consumption and sell the surplus to peripheral hospitals. The revenue from these sales contributes to the hospitals’ internally generated revenue, for which hospitals have full autonomy and fiscal responsibility.

By February 2020, national consumption was 6,760 cylinders (50 L, 150 bar) per month, and public health care facilities accounted for 68% of the supply. Public plants collectively operated only at 64% capacity, producing 156 cylinders (50 L, 150 bar) of oxygen per day versus 242 cylinders (50 L, 150 bar) per day capacity. Only 5 of the 7 public plants operated at full capacity. Both supply and demand factors contributed to this suboptimal performance. Meanwhile, delayed payments from surplus sales, irregular plant maintenance, and reportedly low reimbursement tariffs—which providers argued did not significantly cover costs—disincentivized production, ultimately affecting supply. For example, oxygen consumption data from the University Teaching Hospital of Kigali and Ruhengeri Hospital show that consumption represents only 48% of the expected clinical need for oxygen (based on the per bed by ward and overall bed occupancy rates). This ratio varies by province—the Eastern and Western provinces consume the least oxygen and have the least accessibility to production hubs. In addition, clinicians were underdiagnosing hypoxemia due to a lack of adequate oximeters resulting in low and irregular orders.

During the initial response to the pandemic, production could not be increased immediately, leading to supply gaps, and public hospitals with plants prioritized providing oxygen to COVID-19 designated treatment centers, following demand patterns. Since February 2020, additional PSA plants have been installed in response to the pandemic, primarily at COVID-19 treatment centers to manage and control severe COVID-19 cases. To date, 22 new PSA plants have been added, increasing the number of PSA plants located in public health care facilities to 29.

Health facilities in Rwanda purchase oxygen by volume. The average procurement cost price of oxygen is Rwandan Francs [RWF]15000 (US$16) per cylinder (300 L). Insurers, in contrast, have a single hourly medical oxygen therapy tariff and reimburse hospitals based on the number of hours that beneficiaries spend on oxygen therapy regardless of volume consumed. This reimbursement structure poses problems because the Rwanda Clinical Guidelines for Hypoxemia Screening and Oxygen Therapy Administration in Neonates, Children and Adults recommend different oxygen flow rates (i.e., volumes) to treat various medical conditions.

However, it was uncertain whether the current tariff structure is sufficient to cover the costs incurred by facilities in purchasing and producing oxygen in the long term, especially during periods of increased demand for high-volume therapy. Health care facilities' ability to meet their costs of production influence their ability to continue to produce oxygen sufficiently. In Rwanda, human resources, medical accessories, and fixed costs such as oxygen distribution piping installation across essential wards, maintenance, and repairs are covered by the government’s global budgets from the Ministry of Finance and Economic Development. Variable costs including oxygen are reimbursed by health insurers; the dominant insurer is the Community Based Health Insurance (CBHI), which in 2021 covered 85.6% of the population.¹¹ The CBHI scheme also reimburses 85% of facility oxygen therapy utilization (in hours) and has the lowest tariff (Table 1).

To the best of our knowledge, there has not been a formal assessment of whether the current arrangements for the production and consumption of medical oxygen meet the demand and supply of medical oxygen for health facilities in the long term. In this study, we assessed whether the current tariff rates allow hospitals to break even by comparing the variable oxygen procurement costs against CBHI reimbursement for hospital wards that used medical oxygen therapy. This is an important consideration, as it impacts how much medical oxygen can be produced, procured, and ultimately whether all patients in need of oxygen therapy can have equitable and sustainable access to the therapy they need. We also describe the current financing structures for oxygen in Rwandan health facilities and explore different tariff models that may be fairer for both the health provider and the payer.

METHODS

Rwanda currently operates a well-functioning, decentralized health care public service system comprising 1,060 health posts, 500 health centers, 42 district and provincial hospitals, and 5 national referral hospitals. Rwanda also has a private health services sector that comprises 2 general hospitals, 2 eye hospitals, 50 clinics and polyclinics, 8 dental clinics, 4 eye clinics, and 134 dispensaries.⁸ Public hospitals make up the larger proportion (64%) of facilities, and utilization is 1.8 visits per patient annually. We compared the oxygen procurement cost and insurance claims for 25 hospitals—20 district hospitals, 4 referral hospitals, and 1 provincial hospital—between June 2019 and June 2020. Importantly, most of the study period occurred before the onset of the COVID-19 pandemic in Rwanda.

In assessing whether the current fixed tariff adequately covers procurement costs of medical oxygen by the facilities, and under what circumstances would a variable tariff help hospitals break even, we conducted 2 assessments. First, we estimate the financial implications of the current fixed tariff that only factors the duration (hours) of oxygen consumed to hospitals at the individual ward level. Second, we estimated a variable tariff that factors both the duration (hours) and the volume (liters) that closely links the amount of oxygen consumed and the variable cost of providing the therapy.

To estimate the clinical oxygen therapy utilization of patients, we collected administrative data on admissions, diagnosis, treatments, therapy flow rates, and treatment duration from electronic hospital management records¹² for 37,809 patients in the 25 hospitals between June 2019 and June 2020, across an average of 9 wards per hospital including neonatology, internal medicine, pediatric, emergency, surgery, anesthesia, maternity, and outpatient department. Consumption per patient was estimated using the average recommended oxygen consumption per ward, as stipulated in the national clinical guidelines (Table 2). The Clinton Health Access Initiative’s good data handling practice procedure was adopted for all patient-identifiable information in receiving, transiting, and analyzing data, as well as upon completion of data analysis.

We estimated the financial costs incurred by hospitals based on their revenue from insurance reimbursement claims and the total oxygen procurement costs. The recurring procurement costs (oxygen cylinder refills and transport costs only, excluding the gas cylinder cost) and reimbursement revenues of oxygen were obtained from the finance departments of the same hospitals based on the unit sales price per cylinder and the reimbursements received from insurers.

The hospital revenue and margins from oxygen therapy were estimated based on the reported hours of treatment, average flow rates from 2012 clinical guidelines, the tariffs from different insurance schemes and noninsured patients, and the average price of oxygen procurement. The sufficiency and financial implications of the tariffs for hospitals were estimated based on the margins between revenue data and total oxygen procurement costs.

In Rwanda, oxygen is reimbursed at a fixed rate determined by time per hour at RWF635 (US$0.68) per hour for a patient on the CBHI scheme) and procured in volumes per liter from the supplier at an average of RWF300 (US$0.32) per liter of compressed gas oxygen. Since the pandemic resulted in a surge in volume of high flow oxygen therapy, we created an alternative tariff model to compare against the existing tariff model. An alternative model would consider both volume and time to address the variances in usage across wards (Box).

Studies show that approximately 10%–15% of the patients with COVID-19 will require oxygen therapy.^13–15 As of September 24, 2021, the Rwanda Biomedical Center reported the cumulative total of positive COVID-19 cases in Rwanda at 96,337. Extrapolating from the experience in India, for inference, we implied that between 9,634 and 14,451 cases in Rwanda would have required oxygen therapy in this period.

Survival analysis findings established that the median time to complete supplemental oxygen therapy among the studied population for COVID-19 patients was 6 days (144 hours).¹⁶ While oxygen therapy is recommended for the management of COVID-19 patients, the methods of administration vary according to severity. The oxygen flow rates typically vary from 2 to 15 LPM.¹⁷ Therefore, we assessed the potential revenue based on use of the variable tariff rate in comparison to the fixed tariff rate within the oxygen flow rate range of 2 to 15 LPM.