A Cross Sectional Study on Out of Pocket Expenditures Among Cardiovascular Disease Affected Household in Khartoum State 2019.

Background: The burden of non-communicable diseases is escalating rapidly, cardiovascular diseases (CVDs) are the leading cause of mortality and morbidity worldwide. In Sudan, the incidence of cardiovascular diseases is 2.5. The nancial hardship of CVDs is devastating on both patients and households. The aim of this study is to estimate the out of pocket expenditures (OOP), quantify the catastrophic health expenditures (CHE), and identify the coping methods and costs among cardiovascular diseases affected households in Khartoum state. Methods: This is a cross-sectional study that took place in three main CVDs hospitals (n=122, response rate=88%). Descriptive statistics were used to calculate the OOP expenditures, the incidence and intensity of CHE and coping methods. Inferential statistics were used to assess factors associated with CHE incidence. Results: The annual out of pocket expenditures were found to be 2026INT$, distributed mainly between outpatients’ visits cost and inpatients care cost. The incidence of CHE was found to be 70% at the 10% threshold, the lower-income quartiles suffered the greatest incidence and intensity. The income, number of inpatients visits and certain diagnoses were found to increase the odds of incurring CHE. In order to cope with the disease expenses households resorted to borrowing, selling assets, and nearly half of the participants sacriced basic needs. Conclusion: study forces This paper protects


Background
Universal Health Coverage (UHC) means that all people should receive health services without exposing them to nancial hardship (1). One of the elements that imposes patients to nancial hardship is the increasing treatment costs paid by the patients themselves through out of pocket (OOP) payments at the time of health delivery (2). Accordingly, countries that over-rely on OOP payment to nance their health system pose a huge nancial burden on households (3), and consume large amount of their income (4). This leads to nancial catastrophe to both patients and households. It forces individuals to allocate their budget from basic human needs to pay for health expenditures. This comes in the expense of other necessities such as food, clothe, and education, thus pushing them more towards the edge of poverty (5). The impact of these payments goes beyond nancial catastrophe and push families to postpone or desert health services (2,5,6). Families cope with different mechanisms to deal with these expenses like saving, borrowing from relatives or friends, selling assets, and taking loans (7).
Sudan follows the OOP model with minimum national health insurance like many other African countries which only bene ts the wealthy (8). The OOP health expenditures in Sudan (% of total health expenditure) expanded from 61% in 2004 to 70% in 2011 and to 75.51% in 2014 (6,8) which is the highest among the region and exceeds the WHO uppermost level(40%) (8).
Two-third of world death is assigned to non-communicable diseases; it also comprises 40% of all death below 70. Most of which reside in low and middle income countries (9). Cardiovascular diseases (CVD) is the chief cause of mortality and among the leading causes of morbidity worldwide (10). Heart diseases are an important cause of morbidity and mortality in Sudan, with prevalence of 2.5 (10,11).
Furthermore, chronically ill patients like CVDs patients confront higher nancial catastrophe due to their long-term treatment, loss of productivity and disability in the long run (2,5,12). They also face higher costs in the acute phase of the disease (13). The economic effects of CVDs are devastating; leading to billions of dollars lost as a result of healthcare cost, decreased productivity, and disabling or fetal outcomes (14).
In spite of the high prevalence of CVDs and OOP healthcare nancing in low-middle income countries(LMIC), there are limited information evaluating the association between acute CVDs events and their economic impact (15), as few economic analyses related to CVDs took place in Africa (14). Similarly, the numbers of studies that report the cost of diseases in Sudan are quite small. The knowledge provided by this study will be the ground and the reference for future studies evaluating CVDs economic burden in the country.
Sudan witnessed drastic economic changes after its 2019 revolution. The country has been dealing with an economic crisis and shifted from being a LMIC to a low income country (LIC) according to the last world bank countries classi cation by income level (16). More recently, subsides on fuel has been lifted and the in ation rate reached 81.28% (17), which consequently affected the health care services and medication prices (18), that was already high due to drug prices liberation and privatization policies (19).
All these expenses rely directly at the patient on both the public and private health facilities. The numbers provided in this paper are anticipated to rise tremendously in 2020 compared to 2019.

Aims:
The aim of this study is to estimate the out of pocket expenditure, quantify the catastrophic expenditure (CHE), and identify the coping methods and costs among cardiovascular disease affected households at Khartoum state.

Study setting and design:
This is a quantitative, descriptive, cross-sectional, facility-based study. It took place in the three main public cardiac centers in Khartoum state which are; Al-Shaab Hospital, Ahmed Qasim Hospital, and Sudan Heart Center. A convenience sample of 122 male and female adults diagnosed with IHD, stroke, HTN, and dyslipidemia for more than one year. Patients less than 18 years old and newly diagnosed patients of less than a year were excluded. Co-patients were included when the patients were severely ill that prevented them from contributing to the data collection process. Individuals were recruited from both the outpatient clinic and the inpatient ward.

Study population and sampling:
The sample size was calculated using this formula: n = Z²PQ/d² where n = sample size, Z = statistical certainty Z = 1.8 (at 93% of con dence), P = Fraction of phenomena of interest (OOP expenditure prevalence in Sudan) p = 0.7(13),Q = 1 -p and d = desired margin of error d = 0.07.

Data collection method and tool:
The data collection period started in February and continued in August. It stopped in between due to political situations in the country. A face-to-face interview was administered by four medical students. The interview lasted approximately 15 minutes.
A detailed well-structured questionnaire was employed in the interview that was developed from the previous literature. The questionnaire consisted of 5 parts; the rst part for socio-demographic and economic indicators, the second part for insurance status, the third part for cardiovascular disease status, the fourth part to assess OOP spending on cardiovascular disease and the last part was to identify different types of coping strategies and costs. Before starting the data collection period, a pilot survey was conducted on one hospital to test for questionnaire practicality and acceptability, correct tool errors, and check for sampling procedure applicability.
The household income was calculated as the net income of all actively employed family members and other sources of nancing, then households were classi ed into four income quartiles (Q1, Q2, Q3, Q4) with Q1 being the poorest. OOP expenditure for the last 12 months was calculated as the sum of OOP payments for outpatients visit and OOP payments for hospital admission and OOP payments of special diet for CVDs patients in the last 12month. These payments included direct medical cost like fees for laboratory investigations, medications, consultations, procedures and hospital bed days and direct nonmedical cost for food and transportation and co-patients expenses. Informal payments were excluded because were found to be minimal. Outpatients' expenses were calculated from the last visit, the inpatient' expenses were calculated from the current hospitalization period and some patients had both inpatients and outpatients' expenses. OOP payments were estimated in Sudanese SDGs and then transferred to US dollars, 1 US$ = 46 SDGs at the time of data collection (20). 138 individuals were recruited to this study with a response rate of 88%.

Data analysis
The data was analysed using SPSS 25. The annual OOP expenditures were calculated per household and it was the sum of inpatients and outpatients' expenses for the past 12 months from the date of data collection. The annual OOP payments for outpatient visits were the result of the mean OOP of the last visit multiplied by the number of the annual outpatient visits. The annual OOP for inpatient care was the mean of the last hospitalization care expenses multiplied by the number of hospital admission during the last year. Expenditures were viewed as catastrophic when it exceeded 10% of the household's income (5), although other thresholds of 5%, 15%, 20% were calculated for comparative purpose. The magnitude of CHE was the proportion of the household that experienced CHE. The intensity of the CHE was calculated as the exceedance percentage of the income threshold among household experiencing CHE which is known as mean positive overshoot (21). The distribution and the intensity of the CHE were identi ed among the income Quartiles. To study the impact of means other than income household used to nance out of pocket payments on CHE, the amount of coping cost of borrowed money and sold asset was deducted from the total OOP expenditures (5, 7).
Inferential statistics using a logistic regression model were used to assess the covariates affecting CHE. The covariates were chosen according to the previous studies which are; hospital type, income, education, occupation, residence, household size, number of household members paid for working, insurance acquisition, disease presentation and period/number of inpatient visits (5-7, 15, 22). The association was considered to be signi cant when P-value is less than or equal to0.05.

Major characteristics of the sample populations:
A total of 122 households participated in this study, the majority of the respondents were females (59.8%), almost half of them (50.8%) were housewives, while only (9%) were formally employed. 40 patients (32.8%) were educated up to primary level. One-third of the patients were the primary income provider (32.8%). The median income of CVDs affected households was 6000SDG = 130US$ per month.
Regarding the insurance status, most of the patients held insurance coverage (76%), nearly all participants had a public one (91.3%). The insurance did not cover all the disease expenses (84.8%) of the participants. Despite the insurance coverage, more than half of the participants (60%) were not able to afford all the CVDs related expenses (lab tests, drugs, procedures, …etc.). Since the vast majority (91%) were covered by Social Health Insurance (SHI), this gives a clue about the failure of SHI to achieve Universal Health Coverage. The majority of the study participants were diagnosed with hypertension (43.4%) followed by stroke (23%). (Table 1)  3.2 Out of pocket expenditure in CVDs patients: The estimated annual mean for OOP expenditures on outpatients' visits was 16,008 SDGs = 348 US$. The annual out of pocket expenditures on cardiac admission for the last 12 months was estimated to be 76,738 SDGs = 1,668 US$. The estimated monthly out of pocket paid for special CVDs diet was found to be 450 SDGs = 9.7 US$, accordingly the collective annual out of pocket expenditure for CVDs patients for the last 12 months was found to be 2026IN$ = 93,197 SDGs. Table (2). As shown in the gure the greatest share of outpatients visit cost was spent on drugs which is a direct medical cost and the second-largest share was spent on the direct non-medical cost which is co-patient's expenses, while the cost of inpatients care was driven mainly by food and drugs expenses. (Fig. 1). As seen in the graph operations cost share was less than anticipated.
As depicted in the gure, the mean of annual OOP expenditures varied across income groups with no steady increase, it was higher among the two highest income quartiles compared to the lowest income quartiles. (Fig. 2) 3.3 Incidence and intensity of catastrophic health expenditure: From the below table, 70% of households experienced CHE above the 10% income threshold. The incidence is highest among the low-income groups at any given threshold, meaning, the incidence of CHE is decreased with an increase in income. The P-value for this was found to be (P value = 0.001) indicating a statistically signi cant difference. The magnitude of CHE reduced by 12% when the amount of OOP payments nanced through different coping mechanisms was subtracted from it. This formed the Adjusted CHE which was experienced by 58%. (Table 3)

Factors playing role in catastrophic health expenditures:
The results of logistic regression which is the relationship between the incidence of CHE and other covariates, indicated that income groups are one of the determinants of CHE incidence. The poorest households had increased odds of incurring CHE by 38 folds compared to the rest economic quartiles (OR = 37.8, 95%CI (3.609-396.149), P value = .002) with odds decreasing as the income increases. Also, households had more odds of incurring nancial health catastrophe if the patient's diagnosis was HTN or valvular heart disease (OR = .067, 95%CI (.008-.556), P value = .012), (OR = .020, 95%CI (.001-.329), P value = .006) respectively.
The number of inpatients visits is signi cantly associated with CHE incidence, it increases the odds of facing nancial catastrophe directly (OR = 5.559, 95%CI (2.010-15.370), P value = .001).
Other covariates like (educational level, occupation, residence, insurance acquisition, duration of illness and number of households paid for working) weren't signi cantly associated with CHE incidence. The fact that insurance acquisition didn't play a role in protecting households from nancial catastrophe support the claim of its failure to achieve Universal Health Coverage. (Table 5). The number of patients who received nancial assistance from charity organizations (e.g. Dewan Al-Zaka) was 22.1%, while 12% of households mentioned that they received governmental assistance. On the other hand, more than half of the patients 61% preferred to receive nancial assistance on drug costs to ease the disease burden. (Table 6).

Discussion
This is the rst study in Sudan to estimate the amount of out of pocket expenditures paid for cardiovascular diseases health care. The aim of this study was to estimate the out of pocket expenditures, to quantify the incidence and intensity of catastrophic health expenditure for CVDs and to identify the coping methods used by households to nance their health care expenses by means other than their main income. The cross-sectional descriptive design was suitable to meet these targets with a response rate of 88%.
In this study, we calculated the direct cost of cardiovascular disease by measuring the OOP expenditures; it was found to be 93,197 SDGs = 1,553 US$ a mean value and 16,154 SDGs = 269 US$ a median value.
By comparison, the mean value of OOP expenditure is less than what was documented in a study for atherosclerotic cardiovascular disease patients in the United States which was INT$ 4425 (12), and more than in-hospital out-of-pocket expenditures for CVDs in the United States which was INT$1,229 in 2006 (23). And the median value is much lower than what was found in Tanzania INT$374 and India INT$2,917 (15).
This high number of OOP payments re ects the status of the healthcare nancial system in Sudan which relies directly on the user's fees at the point of health delivery and lacks pre-payments and risk pooling mechanisms.
Another point worth mentioning; medications costs were found to be the major factor for outpatient costs, a nding that was reported elsewhere (5,22,24), due to the high medication's prices in Sudan. This may resulted from the liberalization and privatization policies of drug prices (19) and the high in ation rate in 2019 (17).
This study showed that households affected by CVDs in Khartoum suffer from great nancial hardship with 70% of Sudanese households seeking CVDs care experiencing nancial catastrophe. The incidence of catastrophic health expenditure was found to be 70% when the threshold was set at 10% of total expenditure and 54.5% when the threshold was set at 20% threshold. This high number likely occurred due to the fact that Sudan has the highest OOP expenditure in the region, higher than WHO uppermost level (8). This incidence is much higher than what was reported in other countries (5,15,22) but still lower than the incidence of CHE of acute coronary syndrome in India (25). When we calculated CHE incidence we used household income rather than consumption expenditure. Which was found to be 6000 SDGs per month compared to the national household income of 852 SDGs per month, a nding from the national baseline survey (NBLS2009) (4).
The incidence of CHE was higher among poor households; 93% of the lowest income groups experienced CHE compared to 53% of the richest group. The low-income households also suffered from the greatest intensity of CHE, on average they spent 22% of their income on health care expenses compared to the richest group who spent 10.7%. The ndings from the logistic regression support this fact; the increase in income protects households from incurring CHE while low income increases its chances. This nding is consistent with the previous literature (5,22,25,26).
One of the factors that played a role in increasing the likelihood of incurring CHE is the number of inpatient visits. An increase in the number of visits increases the odds of incurring CHE by 5.5% (OR=. 5.559 ,95%CI (2.010-15.370), P-value = 0.00), this nding was also identi ed by others (5,15,26). Moreover, the diagnose of HTN and valvular heart disease increased the risk of nancial catastrophe.
Not as anticipated, results from the logistic regression revealed no signi cant association between health insurance acquisition and incurring CHE. This might be explained by different ways; First, half of the study participants are from rural areas and 15% had an informal job and these groups the SHI has failed to recruit (6,27). Second, the narrow package of services provided by the SHI that overlook many important and expensive medications, as a result the patients resort to the private sector to obtain it. The last reason, it might be related to a recent national program by the government that provides free treatments, removing some nancial restrains on CVDs patients, regardless of their insurance status (28).
However, SHI in Sudan might have decreased the disease burden on some households. Results of another literature from Kenya and Hong Kong were similar in their ndings; showing that health insurance doesn't appear to be a signi cant determinant of catastrophic health expenditure (26,29).
Moreover, in order to cope with these catastrophic expenditures, many households had to fund these expenses through different coping mechanisms. The main method was borrowing money followed by selling assets. It's also worth mentioning that nearly half of the participants sacri ced some essential needs by reducing their budget for electricity, food, and clothes in order to cope with these expenses. Less than anticipated, only 21% received money from charity and 12% mentioned that they received nancial assistance from the government. These methods are consistent with the ones mentioned in the Sudan households utilization survey in 2009 (4) and consistent with the previous literature (5,30).
Almost 20% stated that they lost an income source by selling their assets. Coping methods might provide a temporary relief, but this comes at the expense of future earnings. Draining the patients from these important assets and sources of income can expose the household to a future nancial crises (30).
The coping-adjusted catastrophic health expenditure incidence is lower than the unadjusted catastrophic health expenditure incidence by 12%. This reduction in number came mainly from borrowing money and selling assets.
The results of this study should be interpreted with caution in view of these limitations; the method used in this study overlooked the under-utilization of cardiovascular health services at the point of health delivery by low income families who are more prone to experience CHE. Also, by choosing public health facilities this study ignored a huge amount of out of pocket payments in private sectors. Moreover, we chose household income rather than consumption expenditure in order to measure the CHE. All these factors led to the underestimation of both the incidence and intensity of CHE among CVDs affected households. This study might fall into recall errors because we required detailed OOP payments reporting for both outpatient and inpatient care for 12 months recall period. This might either overestimated or underestimated CVDs health cost, but these details are one of the strengths of this study.

Conclusions
Seeking treatment for CVDs in Khartoum state puts families at great nancial hardship. The high amount of OOP payments for CVDs health services resulted in 70% households experiencing CHE, with the lowest income groups suffering the greatest magnitude and intensity. Families had to undergo different coping mechanisms like borrowing money, selling assets and more than half of the household had to cut their budget on basic needs to cope with the high disease costs. Public health insurance failed to provide health coverage for cardiovascular disease costs and to protect the patients from the impact of these payments.
According to this study results, a set of recommendations appeared to be important and should be passed to policymakers; the National Health System strategies should be designed in a way that protects households from nancial catastrophe associated with cardiovascular diseases, by reducing the OOP payments at the point of service delivery. This can be done through: a. Increasing prepayment mechanisms like social health insurance, tax-based nancing of health care, and risk pooling mechanisms (31).
b. Social health insurance scheme should be expanded to include vulnerable groups and the informal sector.
c. The bene ts package provided by the social health insurance and tax-based prepayment system should be expanded to include all the medications required for CVDs treatments as they were the main factor for the high OOP payments.
d. Further research should tackle the eld of CVDs costs to address the re ection of the economic crises that happened in 2019-2020 on CVDs burden in Sudan. separately. Verbal informed consent from the participants was more appropriate to this study for the following reasons; First, the low literacy rate in the Sudanese population. Second, the low response rate that was noticed with written consent. This was approved by all above mentioned committees.

List Of Abbreviation
Verbal informed consent was attained from the patients before the start of data collection process.
Con dentiality of patients' information was ensured throughout the different stages of this study.
2. Consent for publication: Not applicable.
3. Availability of data and materials: The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests:
The authors declare that they have no competing interests.

5.
Funding: This research didn't receive any speci c grant from funding agencies in public, commercial or non-pro t sectors.
6. Authors' contributions: M. was a major contributor to the conception and designing of the research, collecting the data and writing the manuscript. While O.E. made a substantial contribution in designing the study and editing the work steps. All the authors read and approved the nal manuscript.
7. Acknowledgements: Our appreciation goes to the patients and their families who suffer from both the disease and its costs, to our colleague who helped during the data collection process. And to anyone who provided guidance and support during the study.