Background
According to the World Health Organization (WHO), out-of-pocket (OOP) payments for health care at the point of service are an inequitable means of financing a health system [
1]
. Nevertheless, globally, OOP payments contribute a significant and increasing share of current health spending [
2]. On average, OOP payments for people in low- and middle-income countries (LMICs) represent around 40% of health spending and present a significant barrier to access and utilization of health services [
2].
Although health care financing reform in Ethiopia has been in process since 1998, OOP payments remain a significant financial burden on households [
3]. User fees are waived for poor people as a means of financial protection; however, the implementation of this programme is weak [
4]. Furthermore, few households are eligible for a waiver as there is a quota for each village (
kebele) leading to limited benefit for most poor households. As a result, households often resort to borrowing money from family, friends, money-lenders, or to selling their assets and reducing consumption [
5,
6].
There is compelling evidence that OOP payments act as sources of catastrophic expenditures and impoverishment [
7]. Catastrophic OOP payments occur when spending on health care is above a certain threshold that results in financial distress [
8]. However, the threshold levels for estimating catastrophic payments differ in the literature, ranging from 10% [
9] to 25% of total consumption expenditure [
10]. In a study on catastrophic health payments from 2010,the global incidence of catastrophic spending at the 10% threshold was estimated to be 11·7% [
10].
Nonetheless, even small amounts of health expenditure may result in catastrophe and impoverishment for vulnerable households [
11,
12]. Impoverishment occurs when a household that is above the poverty line pre-payment crosses the poverty line after paying (post-payment) for health care, shifting from non-poor to poor [
13]. In addition, some households who were poor become even poorer after paying for health care [
14]. A recent global analysis of impoverishment in 122 countries reported that 1.4% of the world’s population were impoverished by OOP payments on health care at the $1·90 per day poverty line [
15].
Households that include a member with depression face disproportionately high catastrophic costs because of OOP payments [
16]. In India, the incidence of catastrophic OOP payments for depression in women was 14.6% compared to 4.9% in the control group [
16]. In Pakistan, OOP payments for the treatment of depressive disorder also led to significant costs for households [
17]. In the limited studies from Ethiopia, depression has been found to be associated with increased use of healthcare, which may result in increased costs [
18,
19]. While these studies contribute to the knowledge base of the impact of depression, they have not considered the magnitude of household catastrophic health expenditures and impoverishment associated with depression because of health care use. Moreover, many studies rely on self-report symptom scales rather than clinical diagnoses to define depression and do not consider the associated level of disability [
20‐
23]. International diagnostic criteria include disability as a key criterion for diagnosis of depression [
24,
25], and the World Health Organization clinical guidelines for depression require the presence of disability to define ‘moderate-severe depression’ which would require a clinical intervention [
26]. Therefore, disability could be a key factor affecting health care costs in a person with depression, but the impact of disability has seldom been investigated to date.
The study reported in this paper is part of the multi-country Emerald programme (Emerging mental health systems in low- and middle-income countries) which sought to provide rigorous, population-based evidence about the adequacy and fairness of mental healthcare financing [
27]. The objective of this sub-study was to estimate catastrophic OOP healthcare payments, level of impoverishment and financial coping strategies adopted by households of persons with depression of differing levels of disability compared to control households without an index person with depression.
Results
A total of 129 households of a person with depression and 129 comparison households were recruited. Of these, one household of a person with depression was excluded due to missing data on household consumption expenditure, income and OOP payments for health care. In the analysis, 128 households with depression and 129 households without depression were included.
Household characteristics
Household socio-demographic, economic and selected health outcome variables of the 257 participants are presented in Table
1. The percentage of household heads with no formal education in the three groups did not differ significantly (
p = 0.164). The mean age of the head of the household was higher in households of persons with depression and high disability (
p = 0.046).
Table 1
Characteristics of study households by mental health condition and severity
Socio-demographic and economic |
Age of HH head (years), mean (SD) |
48.7 (11.8)
| 44.1 (12.9) | 44.2 (13.8) |
HH size, mean (SD) | 5.2 (2.1) | 4.9 (1.9) | 5.0 (2.0) |
Adult Equivalent Size, mean (SD) | 2.7 (0.9) | 2.6 (0.8) | 2.6 (0.8) |
HH with at least one older person ≥60 years old, n (%) | 15 (23.0) | 14 (22.9) | 28 (21.8) |
HH with at least one child younger than 15 years, n (%) | 56 (86.1) | 49 (80.3) | 112 (87.5) |
Residence, n (%) |
Rural | 46 (70.8) | 57 (90.5) | 103 (79.8) |
Urban | 19 (29.2) | 6 (9.5) | 26 (20.2) |
Gender, n (%) |
Male | 52 (80.0) | 49 (80.3) | 96 (75.0) |
Female | 13 (20.0) | 12 (19.7) | 32 (25.0) |
HH Head marital status, n (%) |
Never married | 1 (1.5) | 3 (5.0) | 6 (4.7) |
Married | 52 (80.0) | 50 (83.3) | 98 (76.6) |
Separated/divorced/widowed | 12 (18.5) | 7 (11.7) | 24 (18.7) |
HH Head education, n (%) |
No formal education, n (%) | 44 (67.67) | 38 (60.3) | 67 (52.4) |
Primary education | 13 (20.0) | 19 (30.2) | 36 (28.1) |
More than primary | 8 (12.3) | 6 (9.5) | 25 (19.5) |
HH with health insurance, n (%) | 2 (3.1) | 1 (1.6) | 9 (6.9) |
Annual total consumption, median (IQR) ††$ |
369.7 (253.8, 519.0)
| 485.9 (320.4, 795.5) | 495.6 (339.6, 778.3) |
Annual health payments, mean (SD) ††$ |
45.3 (111.4)
| 37.6 (50.7) | 28.9 (39.2) |
Clinical Characteristics |
Functioning |
Index patient WHODAS, median (IQR) |
47.2 (38.9, 61.1)
| 16.6 (11.1, 25.0) | – |
Symptom scores |
Index patient PHQ-9, median (IQR) |
12.0 (9.0, 15.0)
| 9.0 (6.0, 11.0) | – |
Median annual consumption was significantly lower in households with depression and high disability (US$ 369.7) compared with households with depression and low disability (US$ 485.9) and households without depression (US$ 495.6), p = 0.007.There was no significant difference between the percentage of households enrolled in any form of health insurance: 1.6% in households of a person with depression and 6.9% in comparison households.
Mean household OOP payments on healthcare were highest in households with depression and high disability (US$45.3) compared with households with depression and low disability (US$37.6) and households without depression (US$28.9) (p < 0.001).
In Additional file
1: Table S1, the household budget share of food, non-food and health care payments is presented. An average of 74.9% (95%CI:70.7,79.2) of total household consumption was spent on food by households of persons with depression and high disability compared with 76.5% (95%CI: 72.7, 80.4) in households with depression and low disability and 78.7% (95%CI: 75.9, 81.4) in households without depression, which was non-significant (
p = 0.201). A comparison of the share of annual OOP health payments relative to annual consumption was 7.5% (95%CI: 4.3,10.7) for households of persons with depression and high disability, 6.7% (95%CI: 4.4, 9.0) for households of persons with depression and low disability and 4.9% (95%CI:3.7,6.2) for households without a person with depression (
p = 0.230).
Incidence and intensity of catastrophic health payments
The sensitivity analysis for the incidence and intensity of catastrophic health payments is reported in Table
2. There was an inverse relationship between catastrophic incidence (percentage of households experiencing catastrophic OOP health expenditure) and the various thresholds. For example, 24.0% of households of persons with depression and high disability, 15.3% of households containing depression and low disability, and 12.1% of households without persons with depression reported total OOP payments exceeding 10% of total consumption expenditure (
p = 0.041). Increasing the threshold to 25% reduced the catastrophic incidence to 5.7% in households of person with depression and high disability,5.5% in households of persons with depression and low disability and 1.2% in households without depression (
p = 0.284).
Table 2
Sensitivity analysis of catastrophic out-of-pocket healthcare payments at various threshold levels
Depression and high disability | Headcount (%) | 46.1 |
24.0
| 11.5 | 5.7 |
Overshoot (%) | 4.0 | 2.6 | 2.0 | 1.2 |
Mean positive overshoot (%) | 8.8 |
17.3
| 17.5 | 21.5 |
Depression and low disability | Headcount (%) | 38.8 | 15.3 | 9.2 | 5.5 |
Overshoot (%) | 3.5 | 1.9 | 1.2 | 0.5 |
Mean positive overshoot (%) | 8.1 | 9.1 | 10.4 | 13.9 |
Comparison without depression | Headcount (%) | 32.9 | 12.1 | 6.0 | 2.4 |
Overshoot (%) | 1.9 | 0.9 | 0.4 | 0.02 |
Mean positive overshoot (%) | 5.9 | 7.2 | 7.6 | 8.7 |
The intensity of catastrophic out-of-pocket payments at all threshold levels was also larger for households of persons with depression and high disability compared to households with depression and low disability and households without a person with depression. For instance, the mean positive overshoot shows that, on average, OOP health payments for households of persons with depression and high disability were 17.3% higher than the 10% of total consumption. At the same threshold, in households with depression and low disability, the percentage of households spending more than the threshold for health care was 8.1, and 7.6% in comparison households without depression. The corresponding values for the 25% threshold of total consumption were 21.5, 10.4 and 1.7% higher than the threshold for households containing a member with depression and high disability, with depression and low disability and without depression, respectively.
Impact of out-of-pocket payments on poverty measures
Table
3 presents information on impoverishment. OOP health expenditure (post-payment head count) increased the percentage of poor households irrespective of the presence of a person with depression and disability in the household. Poverty increased by 5.8% in households with depression and high disability, by 3.5% in households with depression and low disability, and 2.3% in households without depression(
p = 0.039).
Table 3
Impoverishing effect of out-of-pocket payments based on pre and post payments on health care
Poverty head count |
Pre-payment head count A | 12.6% | 12.3% | 10.8% |
Post payment head count B | 18.4% | 15.8% | 13.1% |
Absolute percentage point change (impact) C(=B-A) |
5.8%
†
| 3.5% | 2.3% |
Relative percentage change (=C/A*100) |
46.0%
†
| 28.4% | 21.2% |
Poverty gaps |
Prepayment poverty gap A$ | 8.7 | 5.3 | 6.8 |
Post payment poverty gap B$ | 9.9 | 5.6 | 6.9 |
Absolute point change (impact)C(=B-A) $ |
1.2
†
| 0.3 | 0.1 |
Relative percentage change(=C/A*100) |
13.7%
†
| 5.6% | 1.5% |
Normalized poverty gaps |
Pre-payment normalized gap A | 4.7% | 2.9% | 3.7% |
Post-payment normalized gap B | 5.4% | 3.0% | 3.8% |
Absolute percentage point change (impact) C(=B-A) | 0.7% | 0.1% | 0.1% |
Relative percentage change(=C/A*100) |
14.8%
†
| 3.4% | 2.7% |
The poverty gap increased from US$ 8.7 to US$ 9.9 for depression and high disability, US$ 5.3 to US$ 5.6 for depression and low disability, and US$ 6.8 to US$ 6.9 for control households without depression. This change in poverty gap following out-of-pocket payments was significantly higher(p = 0.047) for households having a member with depression and high disability.
Factors associated with catastrophic out-of-pocket payments
The incidence of catastrophic payment accounting for differences in socio-demographic and economic status is shown in Table
4. The risk of catastrophic OOP payments was significantly related to the level of disability. Households of persons with depression and high disability were three times more likely to experience catastrophic OOP payments than households without persons with depression (RR:2.1, 95%CI: 1.1, 4.6). Likewise, urban households (RR:1.6; 95%CI: 1.0, 3.3) were significantly more likely to report catastrophic OOP payments. In contrast, households with no children were less likely to experience catastrophic payments (RR: 0.2; 95% CI: 0.06, 0.8).
Table 4
Predictors of catastrophic health expenditure among households of a person with depression and comparison households
Mental health condition |
Depression and high disability | 13 (24.0) |
1.9 (1.0–4.1)
|
2.1 (1.1–4.6)
|
Depression and low disability | 8 (15.3) | 1.2 (0.4–2.9) | 1.3 (0.5–3.1) |
Comparison without depression | 10 (12.2) | 1.00† | 1.00† |
Area of residence |
Urban | 8 (21.6) |
1.4 (1.0–2.9)
|
1.6 (1.0–3.3)
|
Rural | 23 (15.2) | 1.00† | 1.00† |
Gender of the household head |
Male | 27 (17.2) | 1.2 (0.5–2.6) | 0.9 (0.3–2.1) |
Female | 6 (14.6) | 1.00† | 1.00† |
Consumption quintile |
Quintile 1 (lowest) | 6 (21.4) | 1.5 (0.5–4.1) | 1.1 (0.4–3.1) |
Quintile 2 | 2 (6.6) | 0.4 (0.1–2.1) | 0.3 (0.08–1.5) |
Quintile 3 | 7 (18.4) | 1.3 (0.5–3.5) | 0.8 (0.3–2.2) |
Quintile 4 | 9 (21.9) | 1.5 (0.6–3.9) | 1.1 (0.4–2.6) |
Quintile 5 (highest) | 7 (13.7) | 1.00† | 1.00† |
Children in the household |
0 | 2 (7.4) |
0.3 (0.05–0.9)
|
0.2 (0.06–0.8)
|
1 or 2 | 13 (15.4) | 0.7 (0.3–1.4) | 0.5 (0.2–1.2) |
3 or more | 16 (20.7) | 1.00† | 1.00† |
Household head education |
No formal education | 19 (17.4) | 0.9 (0.3–2.4) | 0.8 (0.3–2.2) |
Primary education | 7 (14.0) | 0.7 (0.2–2.2) | 0.5 (0.2–1.6) |
More than primary education | 5 (17.8) | 1.00† | 1.00† |
Household having a member above 60 years old |
Yes | 26 (17.4) | 1.2 (0.5–3.1) | 1.0 (0.4–2.5) |
No | 5 (13.5) | 1.00† | 1.00† |
Financial coping strategies
Higher proportions of households with depression and high disability implemented various coping strategies compared with households without depression. For example, drawing up accounts at shops (34.6% vs. 28.0%), taking a loan from a bank or financial institution (28.8% vs.18.2%),reducing food consumption (36.5%vs.23.1%), reducing medical visits (36.5%vs.8.5%) and withdrawing children from school (15.3% vs. 6.1%).See Table
5.
Table 5
Financial coping strategies adopted by households with and without a member living with depression
Drew up accounts at shops | 18 | 34.6 (21.2–47.9) | 12 | 22.2 (10.7–33.6) | 23 | 28.0 (18.1–37.9) |
Loan from Bank or financial institution | 15 | 28.8 (16.1–41.5) | 8 | 14.8 (5.0–24.6) | 15 | 18.2 (9.7–26.8) |
Reduced food consumption | 19 | 36.5 (23.0–50.0) | 15 | 27.7 (15.4–40.1) | 19 | 23.1 (13.8–32.4) |
Reduced medical visits | 19 | 36.5 (23.0–50.0) | 12 | 22.2 (10.7–33.6) | 7 | 8.5 (2.3–14.7) |
Received support from relatives | 18 | 34.6 (21.2–47.9) | 14 | 25.9 (13.8–37.9) | 19 | 23.1 (13.8–32.4) |
Withdrew children from school | 8 | 15.3 (5.2–.25.5) | 6 | 11.1 (2.4–19.7) | 5 | 6.1 (0.8–11.3) |
Took on paid extra work | 17 | 32.6 (19.5–45.8) | 14 | 25.9 (13.8–37.9) | 26 | 31.7 (21.4–41.9) |
Used savings | 4 | 7.6 (0.2–15.1) | 9 | 16.6 (6.3–26.9) | 10 | 12.2 (4.9–19.4) |
Sold assets | 37 | 74.0 (61.4–86.5) | 48 | 88.8 (80.2–97.5) | 60 | 78.9 (69.5–88.3) |
Table
6 shows the results from multivariate analysis of the relative risk of financial coping strategies among households with and without depression. Households of persons with depression and low disability were at significantly higher risk (RR: 1.1; 95% CI: 1.1, 1.3) of selling assets to cope with financial difficulties compared with households without a member with depression. In households of persons with depression and high disability (RR:4.4; 95% CI:2.1, 9.3) and households of persons with depression and low disability (RR: 2.3; 95%CI: 1.0,5.7),the risk of reducing medical visits was significantly higher than among households without a person with depression. Households having a member with depression and high disability were at significantly higher risk of cutting food consumption for financial coping (RR:2.0; 95%CI: 1.1, 4.5) compared with households without a member with depression. Withdrawing children from school was about three times higher for households with depression and high disability compared with households without depression (RR: 3.0; 95%CI: 1.0, 8.5).
Table 6
Un- adjusted and adjusted risk ratios (RR) for coping strategies for financial constraints by households with and without a member living with depression
A) Unadjusted model |
| RR (95% CI) | RR (95%CI) | RR (95%CI) | RR (95%CI) | RR (95% CI) | RR (95% CI) | RR (95% CI) | RR (95% CI) |
Depression and high disability | 0. 9 (0.7–1.1) | 1.2 (0.7–2.0) |
1.6 (1.0–2.7)
|
2.5 (1.1–7.3)
| 1.5 (0.8–2.6) |
4.2 (1.9–9.4)
| 0.6 (0.2–1.9) | 1.0 (0.6–1.7) |
Depression and low disability | 1.1 (0.9–1.3) | 0.7 (0.4–1.4) | 1.2 (0.6–2.1) | 1.8 (0.5–5.6) | 1.1 (0.6–2.0) |
2.6 (1.0–6.1)
| 1.3 (0.6–3.1) | 0.8 (0.4–1.4) |
Comparison group without depression | 1.00† | 1.00† | 1.00† | 1.00† | 1.00† | 1.00† | 1.00† | 1.00† |
b) Adjusted model # |
| RR (95% CI) | RR (95%CI) | RR (95%CI) | RR (95%CI) | RR (95% CI) | RR (95% CI) | RR (95% CI) | RR (95% CI) |
Depression and high disability | 0.9 (0.7–1.1) | 1.1 (0.4–2.5) |
2.0 (1.1–4.5)
|
3.0 (1.0–8.5)
| 1.5 (0.9–2.6) |
4.4 (2.1–9.3)
| 0.4 (0.1–1.5) | 0.8 (0.5–1.3) |
Depression and low disability |
1.1 (1.0–1.3)
| 0.7 (0.4–1.4) | 1.1 (0.5–2.7) | 1.5 (0.4–5.1) | 1.0 (0.5–2.0) |
2.3 (1.0–5.7)
| 1.3 (0.6–3.1) | 1.1 (0.8–1.5) |
Comparison group without depression | 1.00† | 1.00† | 1.00† | 1.00† | 1.00† | 1.00† | 1.00† | 1.00† |
Discussion
Ethiopia has taken steps towards health care financing reform in addressing the universal health coverage components of financial protection [
42]. However, progress has been slow, and many health facilities continue to rely on user fees. There has been limited focus on the financial burden borne by households due to out-of-pocket healthcare costs, including for people with mental health conditions.
According to the results of this study, households of persons with depression and high disability had a higher incidence of catastrophic expenditure irrespective of threshold levels. At the 10% threshold, about one in every four households with high disability depression reported catastrophic OOP payments in the preceding one month.
The higher incidence of catastrophic payments in households of persons with depression and high disability may result from the tendency for people with depression to present repeatedly to health services with somatic complaints and receive non-specific treatments which do not address the underlying problem. This leads to further help-seeking and expenditure on investigations and medications, which may account for higher OOP payments. Previous studies from Ethiopia reported similar findings [
18,
19]. A higher prevalence of disabilities is known to be associated with increased OOP health payments in LMICs [
43]. In a study in Tanzania, functional disability led to increased OOP health expenditures [
44]. Nonetheless, our finding of 24.0% catastrophic OOP payments for households of persons with depression and high disability at the 10% threshold is higher than the 14.6% reported in an earlier study from India [
16]. The higher rate in our study may be attributable to the predominantly rural, low-income setting of the current study while the Indian study was conducted in a more affluent area, with high literacy. Furthermore, India has different health care financing and financial protection mechanisms.
At a 25% threshold of household total consumption expenditure, our findings of 5.7% of catastrophic costs with depression and high disability and 5.5% with depression and low disability are relatively higher than the multi-country report of 2.6% [
10]. A plausible explanation is that depression is very costly/burdensome. Moussavi et al. (2007) compares disability from depression to other chronic conditions and finds that depression is always more burdensome [
45]. Thus, persons with depression and disability face higher financial risk due to the need for treatment and care.
Our findings highlight that the burden of OOP payments on households with a person with depression is heavy. The mean positive overshoot of catastrophic OOP payments at the 10% threshold shows that, on average, catastrophic expenditure was higher by 17.3% for households with depression and high disability compared with 7.6% for households of persons without depression. This translates into households with depression and high disability spending 27.3% of their total expenditure on health compared to 17.6% for households without depression. Our results relating to mean positive overshoot for households with persons with depression and high disability appear to be higher than the 24%reported by Tolla et al. (2016) for Ethiopia on OOP costs for cardiovascular diseases in cardiac hospitals [
46]. This indicates that households with depression and high disability are, in most cases, exposed to a greater risk of catastrophic health spending.
Our analysis demonstrated that catastrophic OOP payments were more prevalent in urban residents and in the poorest households. A previous study in Ethiopia reported a similar finding [
47]. This might be because urban residents visit health clinics more often. Urban residents may also be more likely to visit private health facilities that charge significantly higher fees. The higher proportion of catastrophic OOP payments among the poorest shows that OOP payments are regressive and that there is a lack of financial risk protection for poor households against illness. This is explained by the low number (3%) of households enrolled in the government’s financial protection scheme. This study has demonstrated that households that have no children had a lower risk of incurring catastrophic OOP payments compared with households having children. The reason may be that children are more likely to get ill and incur medical costs. This finding is in line with earlier studies conducted in low -and middle-income countries [
48,
49]. Our study findings also reinforce the well documented relationship between poverty and mental illness [
50]. We found prepayment poverty headcount to be more prevalent, about 2% points higher in households with depression and high disability compared with households without depression.
Both poverty headcount and poverty gap become higher after OOP payments for health care. We found that 5.8% of households of persons with depression and high disability, 3.5% with depression and low disability, and 2.3% of households without depression, fell into poverty due to OOP payments for healthcare. Although it is not always possible to directly compare findings due to methodological differences, this finding is higher than that reported for the African region (1.4% in 2010) [
15]. The difference might be due to the poverty line measurement used: Wagstaff and colleagues used the 2011 $1.90-a-day poverty line while in our study we used the national poverty line (US$0.497-a-day). We argue that estimating poverty impact using a locally defined poverty line is more appropriate as the local price levels of household consumables are thereby considered.
The average shortfall from the poverty line (poverty gap) following OOP payments was substantial. On average, OOP increased the poverty gap by 13.7% in households with depression and high disability, by 5.6% in households with depression and low disability and by 1.5% in household without depression. These findings highlight that the pre-payment poor became even poorer and some who were not poor before payment became so after paying for health care.
The risk of adopting financial coping strategies varied by the presence of a person with depression and with the level of disability. We found that depression and higher disability increased the risk of interrupting medical visits. Other studies have also reported that non-adherence to prescribed medications and loss to follow-up is common in the treatment of depression [
21,
51]. In a recent study from Sodo, Ethiopia, drop-out from care was reported to be mostly due to poverty in people with severe mental disorder [
52]. We argue that non-adherence to treatment minimizes health care payments (i.e. OOP costs were relatively low because of low use of hospital care) and underestimates the incidence of catastrophic expenditure. Furthermore, as reported elsewhere, such practices will have a negative consequence on outcomes of the condition [
51]. Our finding on selling assets by households of persons with depression and low disability for financial constraint is similar to what was reported in other study [
53].
This study also found that depression and high disability significantly increased the risk of withdrawing children from school. This is consistent with a previous study that reported school dropout and absenteeism as being associated with maternal common mental disorders, mostly comprising depression and anxiety [
54]. Withdrawing children from school might be for intra-household labour substitution, whereby in resource-poor settings children are obliged to take on the work activities of a sick parent. Similarly, children are also involved in providing care for the sick household member. However, school withdrawal may have broader livelihood impacts on future generations, leading to inter-generational transmission of poverty, a key target of the Sustainable Development Goals. In line with this, in our previous study in the same population, we found similar findings in households of persons with severe mental disorders such as psychosis and bipolar disorder [
55].
Our study findings support the robust evidence base regarding the association between illness and reduction in food consumption through pathways of medical expenditure [
56]. Indeed, in the situation of illness, financial constraints may trigger reduction of food consumption which will result in household food insecurity. In South Africa, maternal depression was associated with household food insecurity [
57]. In Ethiopia, Hadley et al. (2008) reported depression was independently associated with food insecurity [
58]. In India, a common response by households to financial difficulties during illness was to change consumption patterns in order to cover health care costs [
59].
The strength of our study is that, to our knowledge, this is the first study to examine catastrophic OOP payments and impoverishment in households of a person with and without depression and associated financial coping strategies from any sub-Saharan African country. We estimated expenditures in a comprehensive and systematic way and analyzed poverty using a local poverty line. The inclusion of a comparison group without depression in our study enabled us to estimate the net effect of depression and associated disability on out-of-pocket health expenditures.
Nevertheless, our study is not without limitations. The study samples were drawn from health facilities which may not be generalizable to the general population. The cross-sectional design of the study precludes any interpretation of causality. Our comparison households for depression may have included a person with depression, as we did not screen all members of the household. Information on transportation costs was not included in the estimates although such costs represent a substantial economic burden. We did not assess for co-morbid physical and mental health conditions in households in our samples. Our sample size was relatively small, meaning that we may have been under-powered to detect moderate effects. In view of the small sample size and methodological challenges, our findings can at most be taken as indicative of the burden of OOP payments on households. Moreover, the study should be seen in light of being one of the very first attempts at assessing catastrophic health expenditure and impoverishment in households of persons with depression in LMIC. Multiple hypothesis testing could have led to spurious findings; however, the pattern of associations across outcomes strengthens confidence in the findings.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.