Does insurance protect individuals from catastrophic payments for surgical care? An analysis of Ghana’s National Health Insurance Scheme at Korle-Bu teaching Hospital

Background: According to the World Health Organization, essential surgery should be recognized as an essential component of universal health coverage. In Ghana, insurance is associated with a reduction in maternal mortality and improved access to essential medications, but whether it eliminates financial barriers to surgery is unknown. This study tested the hypothesis that insurance protects surgical patients against financial catastrophe. Methods: We interviewed patients admitted to the general surgery wards of Korle-Bu Teaching Hospital (KBTH) between February 1, 2017 – October 1, 2017 to obtain demographic data, income, occupation, household expenditures, and insurance status. Surgical diagnoses and procedures, procedural fees, and anesthesia fees incurred were collected through chart review. The data were collected on a Qualtrics platform and analyzed in STATA version 14.1. Fisher exact and Student T-tests were used to compare the insured and uninsured groups. Threshold for financial catastrophe was defined as health costs that exceeded 10% of household expenditures, 40% of non-food expenditures, or 20% of the individual’s income. Results: Among 196 enrolled patients, insured patients were slightly older [mean 49 years vs 40 years P < 0.05] and more of them were female [65% vs 41% p < 0.05]. Laparotomy (22.2%) was the most common surgical procedure for both groups. Depending on the definition, 58-87% of insured patients would face financial catastrophe, versus 83-98% of uninsured patients (all comparisons by definition were significant, p <.05). Conclusion: This study – the first to evaluate the impact of insurance on financial risk protection for surgical patients in Ghana – found that although insured patients were less likely than uninsured to face financial catastrophe as a result of their surgery, more than half of insured surgical patients treated at KBTH were not protected from financial catastrophe under the Ghana's national health insurance scheme due to out-of-pocket payments. Government-specific strategies to increase the proportion of cost covered and to enroll the uninsured is crucial to achieving universal health coverage inclusive of surgical care. Trial registration: Registered at www.clinical trials.gov identifier NCT03604458.

KBTH, founded in 1923, is the largest tertiary teaching hospital and the major referring hospital in Ghana. The hospital is situated in the southwestern part of Accra, the capital of Ghana. The hospital treats patients referred from centers all over Ghana and neighboring West African countries. Inhabitants are mostly urban or suburban with a small proportion of rural and slum dwellers. The Department of Surgery at KBTH has four general surgical units and additional wards encompassing orthopedic surgery, plastic surgery, urology, and neurosurgery. In 2016, a total of 7941 operations were performed; 33% in trauma, 30% in general surgery, 12% pediatric surgery, 8% neurosurgery, 6% urology, and 5% in ophthalmology/oral maxillary facial surgery. The top three general surgery diagnoses in 2016 were symptomatic hernia, appendicitis, and breast cancer [28]. We conducted a cross-sectional survey of a random sample of patients who were admitted and discharged on the general surgery ward between February 1st 2017 and October 1st 2017 in the Department of General Surgery at KBTH. Efforts were made to sample across all four general surgical wards to obtain a representative sample of inpatients. Data collection consisted of two essential components: First, prior to discharge, patients on the general surgery wards were interviewed by two trained research assistants using a questionnaire developed based on the 2014 Ghana Demographic Health Survey and administered in the respondent’s language [29]. It included modules on the respondent’s demographic profile (age, ethnicity, income, employment status, occupation), socio-economic status, household characteristics (number of people in the household, sex of head of household, number of children, household expenditures on food, non-food expenditures), dwelling characteristics, as well as assets ownership. The health utilization module included number of hospitalizations during the last 3 months, number of out-patient visits, and the presence of chronic medical conditions [29]. Participants were also interviewed to collect payment information and receipts for items not billed by the surgical department but incurred as part of the hospitalization. This included payments for laboratory tests, imaging, and medicines billed by other departments or an outside facility. From this, we calculated the total cost to the patient for the hospitalization. Non-medical costs (indirect costs) were calculated as the cost of transportation to the facility and estimated lost wages, both obtained via the respondent’s interview. The lost wages were derived from the average daily wages, obtained from the respondent’s interview, multiplied by the length of stay for the hospitalization, obtained from the hospital record. The questionnaire is available in the appendix. The second component was a chart review and abstraction by the research assistants to obtain all costs of inpatient surgical care. From the hospital record, the primary surgical and any medical diagnoses were obtained as well as all procedures performed. Procedural fees, anesthesia fees, consultation fees, supplies, accommodations fees, and all hospital costs incurred were collected. The surgical cost minus NHIS payments was considered the OOPE for the insured at the surgery department level. The same analysis was repeated for charges outside the surgery department. The uninsured paid 100% of all costs as out-of-pocket payments. The OOPEs are defined from the patient’s perspective, i.e. the cost to the patient or the amount paid by the patient to the provider i.e. the hospital for all related health expenses that is not reimbursed by NHIS. For the primary comparison of financial catastrophe by insurance status, we assumed the insured would be less likely to make catastrophic payments. Previous studies on non-surgical respondents indicated that 6% of the insured compared to 28% of the uninsured made catastrophic payments [27]. Using this difference and a sample ratio of insured to uninsured respondents of 3 to 1, 80% power, and a 95% confidence interval at the level of 0.05 yielded a sample size of 98 participants (25 uninsured, 73 insured) to detect a significant difference in financial catastrophe by insurance status. This sample size was calculated using the OpenEpi, Version 3 (2008) open source calculator-Proper [30].We sampled 203 respondents in order to have power for further subgroup analysis and to account for the possibility of non-response. A total of 196 patients had complete variables of interest (96% response rate) and were included in the final analysis. We used Stata 14.0 to generate frequencies, means and proportion utilizing the Fisher Exact tests. Multiple logistic regression models were used to obtain the relationships between financial catastrophe, and socio-demographic and clinical characteristics. An alpha level of 0.05 was used to determine the statistical significance. We analyzed all available observations and records. Observations and records with missing information on specific variables of interest were excluded from the analysis. Lastly, the principal component analysis was used to generate a household wealth index using 22 household living items including dwelling characteristics, access to utilities, and ownership of household items. Statistical analysis was conducted at the individual and household levels. At the individual level, we examined the impact of NHIS on OOPEs. The OOPEs for the insured were calculated as the sum of: 1) total direct surgery department cost minus calculated payments to be made by NHIS and 2) payments made to other departments and facilities. For the uninsured, NHIS made no contributions and this was the total direct medical cost. Indirect costs were obtained from the transportation cost and calculated lost wages (individual’s reported daily wage multiplied by the length of hospital stay). We compared the mean OOPEs between the insured and uninsured using descriptive statistics. We also compared the means of the socioeconomic and clinical characteristics between the insured and uninsured using Student t tests. Fisher Exact tests were used to identify and compare the frequencies of all variables between the insured and uninsured. At the household level, we explored the impact of NHIS on financial catastrophe using accepted definitions in the literature [31–35]. OOPEs were considered catastrophic if they exceeded 10% of annual total household expenditures, 20% of the individual's income, or 40% of non-food expenditures. Multiple regression models were developed to identify associations between financial catastrophe and socio-demographic and clinical characteristics. The questionnaire was administered and data collected through Qualtrics platform software by research assistants using a secured tablet. Data from the review of the hospital records and payment receipts for all costs incurred were also collected in Qualtrics platform software. The data were managed in Microsoft Excel 2014 and analyzed using Stata 14.1. An alpha level of 0.05 was used to determine the statistical significance. We sought ethical approval from both the University of California San Francisco and Korle-Bu Teaching Hospital. Permission was further obtained from the Department of the General Surgery at KBTH. Informed consent was also obtained from all participants in the study in order the ensure confidentiality by not disclosing any personal identifiers or names in data capturing, analysis and report writing.

N/A

• Dima Health