Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Predictors of catastrophic out-of-pocket health expenditure in rural Egypt: application of the heteroskedastic probit model

Predictors of catastrophic out-of-pocket health expenditure in rural Egypt: application of the... Background: Out-of-pocket (OOP) health expenditure is a pressing issue in Egypt and far exceeds half of Egypt’s total health spending, threatening the economic viability, and long-term sustainability of Egyptian households. Targeting households at risk of catastrophic health payments based on their characteristics is an obvious pathway to mitigate the impoverishing impacts of OOP health payments on livelihoods. This study was conducted to identify the risk factors of incurring catastrophic health payments hoping to formulate appropriate policies to protect households against financial catastrophes. Methods: Using data derived from the Egyptian Household Income, Expenditure, and Consumption Survey (HIECS), a multiplicative heteroskedastic probit model is applied to account for heteroskedasticity and avoid biased and inconsistent estimates. Results: Accounting for heteroskedasticity induces notable differences in marginal effects and demonstrates that the impact of some core variables is underestimated and insignificant and in the opposite direction in the homoscedastic probit model. Moreover, our results demonstrate the principal factors besides health status and socioeconomic characteristics responsible for incurring catastrophic health expenditure, such as the use of health services provided by the private sector, which has a dramatic effect on encountering catastrophic health payments. Conclusions: The marked differences between estimates of probit and heteroskedastic probit models emphasize the importance of investigating homoscedasticity assumption to avoid policies based on incorrect evidence. Many policies can be built upon our findings, such as enhancing the role of social health insurances in rural areas, expanding health coverage for poor households and chronically ill household heads, and providing adequate financial coverage for households with a high proportion of elderly, sick members, and females. Also, there is an urgent need to limit OOP health payments absorbed by private sector to achieve an acceptable level of fair financing. Keywords: Out-of-pocket health payments, Catastrophic health expenditure, Multiplicative heteroskedastic probit model * Correspondence: suzan.mohmed75@cu.edu.eg Department of Biostatistics and Demography, Faculty of Graduate Studies for Statistical Research, Cairo University, Giza, Egypt Full list of author information is available at the end of the article © The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 2 of 15 1 Introduction the latest round of the Egyptian Household Income, Ex- Diseases could cause catastrophic health payments, penditure, and Consumption Survey (HIECS), which en- undermine the process of income generation, and ables us to incorporate many fundamental covariates jeopardize future economic welfare [1]. Catastrophic neglected by other studies, including the type of health health expenditure (CHE) occurs when out-of-pocket insurance, health service providers, and household com- (OOP) health payments absorb a large portion of the re- position [12]. sources of a household to the extent that they severely affect the household’s living standard [2]. In the short 2 Methods term, catastrophic payments disrupt household con- 2.1 Data sumption of other essential needs, whereas, in the long Data were obtained from the Egyptian Household In- term, they may force households to deplete assets, use come, Expenditure and Consumption Survey) HIECS), a savings, sell livestock, and incur debts [3–5]. nationally representative survey conducted by Central Substantial health expenditures push households into Agency for Public Mobilization and Statistics) CAPM poverty or deeper [6, 7]. Approximately 30% of house- AS) in 2015 [14]. The HIECS uses three different types holds in 15 African countries had to borrow and sell as- of questionnaires: Expenditure and Consumption Ques- sets to finance health payments; this percentage tionnaire, Assisting Questionnaire, and Income Ques- increased to 50% when facing hospitalization expenses tionnaire. The survey collects data on demographic, and to 40% when using other health services [8]. Even socioeconomic, and health characteristics of rural and modest health payments cause indebtedness with exorbi- urban households, selected using a two-stage stratified tant interest rates besides the sale of consumables and random sampling design. HIECS measures consumption assets, which ultimately lead to poverty [9]. patterns and provides detailed information on house- Two approaches are commonly applied to measure the hold’s income sources and health status. Individuals incidence of catastrophic health payments: in the first were asked about suffering any illness during 6 months approach, catastrophic health payments are defined as preceding the survey, having a chronic disease or a dis- health expenditures exceeding a specific fraction of the ability, healthcare providers, and health insurance cover- total household expenditure or nonfood expenditures age and its type. [10]. In the second approach, they are defined as the Out-of-pocket healthcare expenditures (OOP) com- ability to pay, which is measured after deducting subsist- prise inpatient and outpatient health services and other ence spending [6]. Both approaches asserted that OOP reported expenditures (medicine, laboratory tests, X- health payments accelerate and exacerbate poverty; cop- rays, medical equipment) and do not include insurance ing strategies with health payments create a vicious cir- premiums. We used household as the unit of analysis in cle of indebtedness and impoverishment, and must be order to account for the intra-household distribution of considered to understand the trend and impetus of pov- resources and coping strategies with health payments, erty over time [11]. assuming that the economic impact of diseases affects all Although public health investments dwindle annually household members. Our analysis focuses on rural areas in Egypt, urban areas have captured the largest percent- including 6670 households. Rural households are distrib- age of public health spending, whereas rural areas have uted mainly in Rural Lower Egypt and Rural Upper invariably received less attention and are poorly funded. Egypt. All estimates are adjusted to represent national The gap between health spending in urban and that in figures using appropriate sampling weights. Both de- rural areas has reached roughly 67%, which is reflected scriptive and analytical statistics were performed using R in a substantial disparity in the availability of healthcare software. infrastructure. Moreover, rural areas are characterized by low levels of income, education, and economic devel- opment. All these factors may contribute to the high in- 2.2 Measuring incidence of catastrophic OOP health cidence rates of catastrophic payments [12]. payment This study was conducted to estimate the incidence of Catastrophic health expenditure was defined as “Any CHE, identify its core determinants, and reveal the ex- health expenditure that threatens a household’s financial tent to which health insurance schemes properly target capacity to maintain its subsistence needs and does not vulnerable economic households and whether they suc- necessarily equate to high healthcare costs. Even rela- ceed in protecting them financially. This study contrib- tively small expenditures on health can be financially utes to the literature by exploring the predictors of CHE disastrous for poor households” [2]. According to the using a multiplicative heteroskedastic probit model, that methodology proposed by Xu et al. [6], the most com- accounts for heteroskedasticity and provides unbiased mon approach applied in recent studies, OOP health ex- and consistent estimates [13]. Moreover, this study used penditure is catastrophic whenever it is greater than or Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 3 of 15 equal to 40 percent of the household’s capacity to pay, of rural households whose members were engaged in the where: agriculture sector and exhibited varied seasonal incomes. The probit model assumes that the error distribution – Capacity to pay (ctp ) is the remained household of the latent model is homoscedastic and has unit vari- income after satisfying food subsistence spending. ance. However, many studies that examined the determi- – Subsistence spending (se ) is determined based on nants of catastrophic health payments have neglected to average food expenditures of households with food verify the violation of the homoscedasticity assumption, share within the 45th and 55th percentile of the which could result in substantially biased and inconsist- total sample. Actual food expenditure is used to ent estimates besides misspecified standard errors. Spe- measure the capacity to pay for the household cific covariates may influence the probability of whose food expenditure is less than subsistence incurring catastrophic payments through health payment spending. Expenditures are adjusted using the variance. On average, better-off households may spend economy of scale estimated by Xu et al. (2003) (β= more on healthcare services exposing them to CHE, but 0.56) to reflect food consumption sharing among also the large variance of health payments at high ex- household members. Xu’s methodology has been penditure levels due to uncommon and sophisticated clearly described elsewhere [6]. medical treatments will raise the incidence of CHE. This variance is hypothesized to increase with household ex- 2.3 Multiplicative heteroskedastic probit model penditure as poor households have constrained budgets Misspecification in limited dependent models that may limiting their response to health shocks, whereas rich result from heteroskedasticity, omitted variables, hetero- households that face illness episodes spend too much geneity and nonlinearity leads to inconsistent and compared with rich households with good health. A het- biased estimates. The quasi maximum likelihood esti- eroskedastic probit model is applied to consider that mator (QMLE) of the probit model is inconsistent in health expenditure h is heteroskedastic with unfixed case of violation of homoscedasticity, even using variance σ . A heteroskedastic probit model generalizes white corrected estimator for standard errors is use- Φ(.) to a normal cumulative distribution function (CDF), less in the heteroskedastic probit model because the allowing variance to vary systematically as a multiplica- QMLE itself is biased [13]. tive function of set of m explanatory variables x [13, 16]. Let c denote a binary response (0,1) representing the no occurrence of catastrophic health expenditure (CHE) σ ¼ exp x τ ð3Þ i i when household health expenditure (h ) exceed its cap- acity to pay (ctp ). The subscript i denotes the where x is m × 1 vector of covariates (x , x , x ,.. x ) i 1i 2i 3i mi household: that determine the variance of the error term, while τ is the corresponding coefficient vector. 1 h > ctp i Therefore, the probability of incurring catastrophic c ¼ ð1Þ 0 h ≤ctp i health payments is () The probability of incurring catastrophic health y β PrðÞ c ¼ 1¼ Φ ð4Þ i 0 expenditure, exp x τ 0 0 A multiplicative heteroskedastic probit model relaxes PrðÞ c ¼ 1¼ Pr ε < y β ¼ Φ y β ð2Þ i i i i the homoscedasticity by allowing the scale of the inverse link function to vary systematically from one observation y is a k × 1vector of observed independent variables(y , i 1i to another as a function of explanatory variables. It is fit- y , y …. y ), β is vector of the corresponding parame- 2i 3i ki ted via maximum likelihood where the log-likelihood ters, ε is random error term and Φ(.) is the cumulative function takes the form: distribution function of standard normal variable with () mean 0 and variance 1. y β lnL ¼ h ln Φ Total household consumption is a superior proxy for i 0 exp x τ i∈ F household economic status; it is less subject to fluctua- "# () y β tions and measurement errors than income. Also, it in- þ h ln 1−Φ  ð5Þ i 0 cludes the value of home production, which is a major exp x τ i∉ F source of income in developing countries, especially rural areas [15]. Consequently, household consumption where F is the set of all observations i such that c ≠ was used to avoid underestimating the living standards 0and h denote the weights. i Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 4 of 15 All explanatory variables enter y vector while total ex- composition of the households indicated that 28.2% of penditure variable is included in x vector for modelling household members were children and 8.62% are elderly; nearly 18.2% of the households had at least one elderly the heteroskedastic variance of the error term σ . The member. In respect to household head characteristics, marginal effects are more informative and directly inter- 66.2% of household heads were literate, approximately pretable than multiplicative effects (probit model’s coef- half of them had at least an elementary certificate, only ficients). The heteroskedastic probit model’s marginal 0.8% had higher university degrees, 77.7% were effects are estimated as a function of the parameters β employed, 73.5% had chronic diseases, and only 31.7% and τ [13] as follow: were insured. The marginal effect of continuous variable w is The average OOP payments were significantly higher 0 0 among better-off households than those among poor ∂ PrðÞ c ¼ 1 y β β −y β:g i z i z ¼ ϕ ð6Þ 0 0 ones. The richest households spend nearly five times ∂w exp x τ exp x τ i i that of what the poorest households spend on healthcare services (LE 6954.28 vs. LE 1483.5). Meanwhile, the where ϕ(.) is the probability density function (pdf). If the health share of the total budget ranged from 8.17% at variable w is an element of both y and x vectors, its co- the poorest economic level to 11.08% at the richest level, efficient in the mean equation (1)is β and in the vari- indicating that poor households cannot deduct a large ance equation (4)is g , while g equals 0 if w appears z z z percentage of their budget for healthcare services. The only in y vector. health share of the household budget has an upward The marginal effect of dummy variable w is trend with the economic level (Fig. 1). Disaggregated 0 0 health expenditures enable us to distinguish between ∂ PrðÞ c ¼ 1 y β y β 1 0 ¼ Φ −Φ ð7Þ 0 0 budget shares allocated to inpatient and outpatient care ∂w expðÞ x τ expðÞ x τ 1 0 across economic levels. The results showed that the ex- w is set to 1 in y and x and set 0 in y and x . Stand- penditure share on inpatient services in rich households z 1 1 0 0 ard errors of the marginal effects are derived to allow in- was nearly four times that in poor households (2.55% vs. ference and hypothesis testing, using delta method 0.6%), whereas both poor and rich households allocated which is the most commonly used [17]. similar shares of their budgets to drugs (4.4% and 4.7%, To investigate the determinants of CHE, dummy vari- respectively). ables for the characteristics of household heads were in- Egypt has a narrow geographic distribution of health cluded in the analysis, such as gender, illiteracy status, services, which is more concentrated in urban areas. and education level. Besides, the analysis comprised age Households across the socioeconomic strata in urban and squared age of the household heads to control for areas witness extensive healthcare infrastructure in con- the nonlinearity of age. Several household-level covari- trast to rural households, which resulted in heavy reli- ates were also incorporated, such as the presence of ance on the purchasing of non-prescribed drugs in rural chronic diseases and disability; health insurance cover- areas. Drug purchases account for a larger share of OOP age; and proportions of sick members, insured members, health payments of rural households (54.7%), followed insured children, insured elderly, wage earners, and edu- by outpatient health services (38.7%), whereas inpatient cated members, in addition to household total expend- services acquire the smallest share (6.6%). iture and geographic area. To capture the effects of An investigation of the components of health expend- household composition, we included household size and iture across various economic levels indicated the over- the proportions of children, elderly, and females. use of medicines among rural households, particularly poor households; pharmaceutical products had the lar- 3 Results gest share of OOP health expenditure in the poorest The distribution of annual health expenditure was right- level compared with other quintiles (62.25% in the poor- skewed (skewness = 7.27), the mean annual health ex- est level vs. 48.1% in the richest level). Meanwhile, in- penditure (LE 2903.4) was higher than the median (LE patient services account for 11.97% of the total health 1706.3). The positive excess kurtosis (101) indicated that expenditure of rich households compared with only 4% health expenditure distribution is leptokurtic. After tak- of the total health expenditure of poor households. ing the log transformation, the mean value became The amount of OOP health payments depends mainly 7.316, whereas the median value was 7.42, skewness was on where households seek healthcare. The data revealed 0.27, and kurtosis was 1.4. The average annual house- a significant variation in the paid health shares by the hold income was LE 37625. The average household size type of healthcare providers; rural households using pri- was five members, with 25.6% being males. One of five vate hospitals and clinics paid on average 12.92% of their rural households was headed by a female. The age budgets on healthcare services. Also, households seeking Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 5 of 15 Fig. 1 The mean out-of-pocket health share by economic level healthcare in public hospitals and social health insurance counterparts (36% vs. 12%(; the proportion of insured hospitals allocate a high share, accounting for 10.5% and household heads increases sharply as the household eco- 10%, respectively, of their total budget. Households that nomic level increases, rising from 15% in the poorest level depend on medical consultations from pharmacies allo- to 52.2% in the richest level (Fig. 3). Meanwhile, at the cate 9.48% of their budgets. However, civil society health household level, on average, half of the household mem- services were associated with a relatively small share bers are insured, which is due to the compulsory insur- (6.19%). Although private healthcare providers demand ance for children in schools where the average number of higher payments than public healthcare providers and insured children in the household is 3, whereas, on aver- are hypothesized to be less used by households at low age, only one elderly member in the household is insured. economic levels, the data indicated that, contrary to ex- A significant difference in the budget share allocated pectations, private healthcare providers are the most to health was observed between insured and uninsured used across all economic levels; 65.58% of the poorest households. Although uninsured households spend, on households use private hospitals and clinics, and similar average, less than insured households (LE 2827.6 vs. LE proportions were found at other economic levels (Fig. 2). 3426.9), they allocate larger shares to health spending However, risk-pooling mechanisms are significant (8.8% vs. 8.3%). Also, health shares of households with tools for protecting vulnerable households from unex- uninsured heads were significantly higher than those of pected health risks. The HIECS indicates that health in- households with insured heads across all economic levels surance coverage is low in rural areas. Nearly 68.25% of (Fig. 4). A two-way analysis of variance (ANOVA) was household heads are uninsured, even though 71.64% of performed to simultaneously assess the effects of both them have at least one chronic disease. The high propor- health insurance coverage and economic level on health tion of informal workers, which characterizes rural areas, expenditure and revealed that both variables have sig- explains the low health insurance coverage for house- nificant effects on health expenditure at a significance hold heads. Additionally, apparent inequalities exist in level of 0.01. health insurance coverage across genders, education A comparison of health expenditure on inpatient and levels of household heads, and other socioeconomic outpatient services across insured and uninsured house- characteristics of households. The proportion of insured holds revealed the languid role of health insurance in male household heads is three times that of their female protecting rural households against high health costs. Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 6 of 15 Fig. 2 Distribution of rural households by economic level and healthcare provider Outpatient care expenditures among insured households insurance coverage was significant with a Pillai test stat- are significantly higher than those among uninsured istic of 0.03 and a p value of below 0.001. households; insured households spend, on average, ap- Also, the share distribution for outpatient services proximately LE 1422.3 on outpatient healthcare services highlights the relatively larger shares incurred by insured compared with LE 888.97 by uninsured households households (Fig. 5). A different conclusion is drawn for (3.58% vs. 2.79% of household budget). To simultan- the effect of insurance coverage on pharmaceutical eously test whether expenditures on both inpatient and spending; the pharmaceutical share spent by insured outpatient services vary between insured and uninsured households is significantly less than that incurred by un- households, multivariate analysis of variance (MAN- insured households (4.61% vs. 4.86%). OVA) was performed. The global multivariate test was Data revealed that rural areas have a high prevalence significant and indicated that the effect of health rate of chronic diseases; every rural household has at Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 7 of 15 Fig. 3 Distribution of rural households by health insurance coverage of household head across economic levels least one member with chronic disease, and the average catastrophic payment gap between households with and number of household members with chronic diseases is without chronic diseases, the presence of chronic dis- 3. Households suffering from chronic diseases spend eases increases the overall incidence rate of CHE from more than twice that of what households without 7.88 to 9.18%. Our data indicated also that having health chronic diseases spend. The health share of households insurance decreases the incidence of CHE among house- experiencing chronic diseases has a clear increasing pat- holds with chronic diseases to 5.44% compared with tern with the economic level (Fig. 6). 10.90% among households without health insurance The incidence of CHE is small in rural households as coverage. a whole (7.91%), but large among the poorest house- A homoscedastic probit model was used along with a holds (14.59%). The incidence of CHE has regressive dis- multiplicative heteroskedastic probit model in which the tribution across economic quintiles: it declines sharply total household expenditure becomes a part of the scale from 14.59% in the poorest quintile to 5.93% in the poor sub-model. The results of both models are provided in quintile, and it reaches 5.26% and 4.92% among house- Table 1. The likelihood ratio test of heteroskedasticity holds in the middle and rich economic levels, respect- substantiates a violation of homoscedasticity; the total ively; then it rises to 7.81% among the richest household expenditure is more correlated with variance. households. The presence of chronic diseases signifi- Coefficients are significant and have the hypothesized cantly increases the incidence of CHE (p < 0.01) and signs in the mean function of the heteroskedastic probit makes the pattern more serious at some economic levels. model. The likelihood ratio test on the nested models At the lowest economic level, the incidence rate of CHE prefers the heteroskedastic probit model over the ordin- among households with chronic diseases is 16.22% com- ary probit model. Moreover, comparing the two models pared with 9.46% among households without chronic using information metrics also confirmed that the het- diseases. While a better economic status decreases the eroskedastic probit model is the better model as it has Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 8 of 15 Fig. 4 The mean out-of-pocket health share by economic level and health insurance status of household heads lower Akaike’s information criteria. Also, the Bayesian probit model, whereas it was highly significant in both information criteria that strongly penalizes including mean and latent scale models (heteroskedastic probit additional variables to the model demonstrates the same model), indicating that the inference based on the ordin- conclusion. ary probit model is inaccurate in case of violation of the Significant differences in coefficients and marginal ef- homoscedasticity assumption. Moreover, household size fects were observed between the probit and heteroprobit has a reversal sign and becomes significant after ac- models, emphasizing how large the estimates are biased counting for heteroskedasticity. in the ordinary model. First, the impact of most variables While the two models differ in terms of the number of increases dramatically in the heteroskedastic probit significant variables, parameter estimates, and standard model, such as the effects of proportions of wage errors, some variables remain significant with the same earners, educated members, and insured children in the sign in both models, such as the proportions of female, household, whereas the impact of some variables dimin- elderly, and sick members, which are positively associ- ishes, such as the type of healthcare provider (e.g., public ated with CHE. Moreover, based on the estimates of the hospitals and hospitals affiliated with the health insur- two models, households headed by an illiterate individ- ance organization (HIO hospitals)). Second, some essen- ual, those headed by a man, those located in rural Lower tial variables recover their significance, such as Egypt, and those with chronically ill members suffer household size, the proportion of wage earners, health from CHE. Meanwhile, households headed by an edu- insurance by occupational syndicates, and outpatient cated, employed, and insured person are protected healthcare services. The most striking difference be- against CHE in both models. The two models demon- tween the two models was that the total household ex- strated that age and age squared of household heads penditure was not significant in the homoscedastic have no significant effects on the likelihood of incurring Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 9 of 15 Fig. 5 Distribution of outpatient services share across health insurance status CHE. Also, the proportions of children, insured children, increases the probability of CHE by 0.39. The heteroske- and insured elderly were not significant variables in both dasticity probit model showed that the infinitesimal models. change in total household expenditure reduces the likeli- Marginal effects and their standard errors are derived hood of CHE by 0.07. from both models (Table 1). Based on the estimated marginal effects of the heteroskedastic probit model, the 4 Discussion probability of incurring CHE increases by 0.06 and 0.04 Household characteristics have a significant effect on in male- and illiterate-headed households, respectively, health expenditure and the probability of encountering whereas it decreases by 0.05, 0.09, and 0.08 if the house- CHE. Poor households have the highest incidence rate hold head has a secondary degree, a university degree, of CHE; the distribution of CHE is regressive across eco- and a postgraduate degree, respectively. The employ- nomic quintiles. This is attributed to poor households’ ment status of the household head has a positive effect inability to afford healthcare. As commonly observed in on the probability of CHE: it decreases the probability of the literature [18], consumption necessities, such as food CHE by 0.07. An increase in the proportion of educated and shelter, exhaust the main bulk of a poor household’s members and wage earners in the household reduces the budget, leaving a small share to healthcare. Meanwhile, probability of CHE by 0.33 and 0.27, respectively. Mean- better-off households could finance health payments and while, the proportions of female, elderly, and sick mem- are more protected against CHE than poor households. bers increase the likelihood of incurring CHE by 0.08, Our empirical results demonstrated that a large-size 0.03, and 0.06, respectively. The presence of chronic dis- household is more likely to incur CHE. This is consistent eases raises the probability of CHE by 0.1, whereas the with the findings of [19], who have suggested that the probability drops by 0.04 in households with health in- large size of households increases the probability of at surance coverage. Using private sector services greatly least one member getting sick, especially in the presence Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 10 of 15 Fig. 6 The mean out-of-pocket health share according to the presence of chronic diseases of overcrowding and infectious diseases. Besides, the lim- 22, 27, 28]; besides, they are more expected to face sev- ited economies of scale in using medical care compared eral episodes of illness [29]. In addition, female propor- with other consumption items increase health payments. tion in households is also positively associated with However, some studies have found that having a larger CHE. One key finding was that educated household size protects households from CHE as it creates an oppor- members successfully shelter their households from tunity for accumulating earnings [18, 20–22]. CHE. This effect could be explained by the significance The high proportion of wage earners significantly pro- of education in reducing the anxiety associated with tects rural households from CHE. It is an indicator of a sickness and preventing making an irrational decision. household’s ability to pool financial resources and in- Household standard of living determines the amount formative about health insurance coverage, especially if of affordable health payments and the risk of facing the wage earners are officially employed and receive CHE. The results revealed that the total household ex- regular income [23, 24]. This finding is in contrast to penditure is negatively associated with the risk of incur- those of previous studies [25], which have revealed that ring CHE. This association has been established by a high proportion of wage earners is positively associated many studies [2, 21, 27, 30]. However, some studies have with CHE as they are more likely to seek healthcare and stated that the likelihood of incurring CHE decreases in provide financial support to other household members. response to an increase in household expenditure as The household composition reflects its health needs poor households are forced to forgo or delay seeking and required expenditure [26]. The results demonstrated healthcare to avoid fees, whereas better-off house- that households with a high proportion of elderly mem- holds are at higher risk of CHE because they have bers are highly likely to face CHE than those with high high responsiveness to health needs, and simultan- proportions of children. This is mainly due to their well- eously, they prefer the highest quality private health- known health needs and their inability to work [7, 21, care facilities [7, 19, 22, 24, 31]. Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 11 of 15 Table 1 Estimates of homoscedastic probit model and heteroskedastic probit model, HIECS, 2015 Variables Probit model Heteroskedastic probit model Coef. Marginal effect Coef. Marginal effect "Mean model" Intercept − 2.407*** (0.504) − 2.398* (1.239) Male-headed household 0.363*** (0.087) 0.045*** (0.009) 0.816*** (0.204) 0.061** (0.053) Age of household head − 0.012 (0.0127) − 0.001 (0.002) − 0.010 (0.026) − 0.001 (0.001) Squared age of household head 0.000 (0.000) 0.000 (0.000) − 0.010 (0.026) 0.001 (0.001) Illiterate household head 0.126* (0.071) 0.018* (0.010) 0.239 (0.159) 0.049** (0.071) Educational degree of household head (Dummy variables) Having a primary or lower secondary degree − 0.196 (0.091) − 0.015 (0.011) − 0.151 (0.212) − 0.013 (0.023) Having a secondary or post-secondary degree − 0.151* (0.091) − 0.019* (0.010) − 0.352* (0.199) − 0.053* (0.170) Having a university degree − 0.469** (0.154) − 0.048*** (0.011) − 1.273** (0.182) − 0.094*** (0.021) Having a postgraduate degree − 0.028 (0.474) − 0.003 (0.063) − 1.311 (2.176) − 0.083* (0.063) None (omitted category) Employed household head − 0.255*** (0.474) − 0.038** (0.012) − 0.528** (0.162) − 0.072** (0.069) Geographical area: rural lower Egypt 0.722* (0.367) 0.126* (0.065) 1.868* (1.003) 0.321** (0.091) Household size − 0.0206 (0.027) − 0.002 (0.003) 0.158** (0.060) 0.032** (0.021) Proportion of females 0.437*** (0.132) 0.059*** (0.018) 0.592** (0.283) 0.082** (0.031) Proportion of children in household 0.186 (0.169) 0.025 (0.023) 0.297 (0.397) 0.064 (0.317) Proportion of elderly in household 0.271* (0.135) 0.037* (0.018) 0.200** (0.260) 0.029** (0.012) Presence of chronic diseases 0.324*** (0.064) 0.039*** (0.007) 0.769*** (0.152) 0.097*** (0.032) Presence of disability 0.164 (0.012) 0.025 (0.021) 0.118 (0.261) 0.101 (0.073) Proportion of sick members in household 0.218* (0.19) 0.031* (0.017) 0.506* (0.269) 0.063** (0.072) Presence of insurance coverage − 0.213*** (0.064) − 0.027*** (0.007) − 0.447** (0.159) − 0.042*** (0.029) Household total expenditure − 0.003 (0.002) − 0.009 (0.007) − 0.167*** (0.012) − 0.071*** (0.47) Proportion of wage earners in household − 0.219 (0.178) − 0.027 (0.022) − 4.052** (0.495) − 0.271*** (0.149) Proportion of educated members in household − 0.551*** (0.164) − 0.066*** (0.019) − 1.373** (0.473) − 0.332*** (0.017) Proportion of children covered by Health Insurance − 0.355 (0.266) − 0.039 (0.029) − 1.595 (1.014) − 0.071 (0.017) Proportion of elderly covered by Health Insurance 0.238 (0.912) 0.035 (0.721) 0.934 (1.127) 0.011 (0.23) Type of health insurance coverage (Dummy variables) − 0.028 (0.174) − 0.004 (0.023) − 0.229 (0.423) − 0.061 (0.093) Covered by HIO Covered by private insurance 0.391 (0.433) 0.068 0.093) 0.617 (1.02) 0.252 (0.093) Covered by employer-provided private insurance − 0.279 (0.259) − 0.031 (0.024) − 0.594 (0.656) − 0.081 (0.074) Covered by occupational syndicates 0.391 (0.284) 0.068 (0.061) 1.285* (0.684) 0.168* (0.051) Other (omitted category) Health provider (Dummy variables) HIO hospitals 4.793** (0.169) 0.027 (0.028) 0.161* (0.400) 0.012* (0.021) Public hospitals 4.275*** (0.101) 0.085 (0.017) 0.153* (0.224) 0.031* (0.021) Private hospitals 0.531*** (0.064) 0.066*** (0.009) 1.097* (0.146) 0.389** (0.071) Pharmacy − 0.081 (0.124) − 0.026 (0.012) − 0.418 (0.245) − 0.076 (0.032) Other(omitted category) Using outpatient health services 0.131 (0.101) 0.019 (0.017) 0.049* (0.242) 0.091* (0.021) Het-test χ ------ 52.24*** LogL − 1566.968 − 1541 Likelihood ratio test 229.13*** McFadden’s pseudo R 0.069 Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 12 of 15 Table 1 Estimates of homoscedastic probit model and heteroskedastic probit model, HIECS, 2015 (Continued) Variables Probit model Heteroskedastic probit model Coef. Marginal effect Coef. Marginal effect "Mean model" AIC 3175.94 3125.69 BIC 3317.12 3273.603 Number of fisher scoring iterations 6 10 N 6143 6143 ***p<0.001, **p<0.01, *p<0.05, ( ): robust standard errors against misspecification. Coefficient and standard error of household total expenditure in latent scale model is{0.003*** (0.002)}. LogL is the maximum value of the likelihood function. Likelihood ratio test measures the significance of the overall model. Pseudo R shows the amount of variance explained by explanatory variables. Wald χ test the significance of the full model; if all slopes equal 0. Het-test χ is the likelihood ratio test of the null hypothesis of homoscedasticity; it tests the model with heteroskedasticity against the full model with out. AIC and BIC are information metrics that penalize the inclusion of additional variables and used to select the appropriate model. Number of fisher scoring iterations indicates how quickly the iterative weighted least squares (IWLS) algorithm terminates. Illness is the main driver of health expenditures and been higher among insured rural households than unin- pushes households to reallocate considerable shares of sured ones. In addition, 49.2% of Iranian households had their resources to OOP health payments, whereas demo- incurred debts, 21.7% sold their jewelry, and 15.9% used graphic–economic characteristics control the burden of their savings to finance health costs [36]. These findings these expenditures. Our findings corroborate earlier lit- confirm that the provision of health insurance should erature that the presence of chronic diseases significantly coincide with adequate financial coverage for all re- increases the risk of incurring CHE [21, 27, 28, 32]. quired health needs, in addition to providing accessible Meanwhile, the presence of disabilities does not have a health services of acceptable quality to deter recourse to significant effect on the risk of facing CHE, which con- the private sector. tradicts the findings of [18, 33]. To control potential confounders related to health in- In the absence of health insurance coverage, health surance status, we used the available data about the type expenditures can absorb a substantial fraction of the of health insurance. Although the selection of health in- household resources and severely disrupt living standards surance providers may depend on unobservable covariates [33]. Health insurance coverage significantly reduces the [37], our analysis did not have this issue as the choice of chance of incurring CHE. A similar conclusion was health insurance in Egypt is highly restricted. There is no reached by many studies. For example, Reddy (2013) has adverse selection as public sector employees and students found that health insurance is negatively correlated with are covered semi-compulsorily by Egypt’s primary insur- the risk of incurring CHE and indebtedness in three Asia- ance provider) HIO hospitals), and employees of other op- Pacific countries: China, Malaysia, and the Philippines erational bodies are covered by occupational syndicates, [34]. Also, Van Minh and Tran (2012) have indicated that whereas private health insurance is less prevalent in Egypt, health insurance has a modest positive effect on CHE. particularly rural areas. Moreover, our model incorporated Contrary to the widely held hypothesis about the role of the measures of health status as the presence of chronic health insurance in reducing financial hardships [21], diseases or disability and the socioeconomic covariates to some studies have shown that that health insurance has a control for any selection effect. contradictory effect on health payments: health insurance The type of health insurance does not have a signifi- caused larger health payments as it facilitates access to cant effect on protecting insured households from finan- healthcare services creating an induced demand for ser- cial hardships, except that provided by occupational vices [4, 18, 27, 35]. syndicates which, in contrast, is positively associated Our most intriguing finding is that increasing the pro- with CHEs. Some studies have shown that health insur- portion of insured children and elderly members had in- ance coverage has contrasting effects according to the significant effects on the incidence of CHE. These benefits package design [18, 23]. Limited insurance pack- findings support that increasing health insurance cover- ages with low depth of coverage, high co-payment rates, age does not necessarily reduce OOP health spending and relatively low-reimbursement ceilings increase the and achieve effective protection against CHE. For ex- risk of CHE. For example, Shahrawat and Rao [38] have ample, Iran has undertaken promising steps to expand revealed that Indian insurance schemes with limited health insurance coverage for most rural households, coverage (hospital expenses only) did not contribute to and 90% of the population has access to primary health- reducing CHE as outpatient care expenses occupied the care services; however, the incidence rate of CHE has main share of health expenditure. Kimani et al. [39] also Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 13 of 15 have explained that health insurance fails to protect facing CHE in rural areas that can provide the basis for households due to its limited coverage of only bed costs analyzing policy options to alleviate the burden of cata- for inpatient stays. Others have demonstrated that pri- strophic payments. Moreover, it explains the role of vate health insurance offering a broad benefits package existing health insurance schemes in providing financial succeeds in protecting households from CHE compared protection in rural areas. The key strength of this study with social insurance schemes [29]. lies in obtaining unbiased and consistent estimates, and Another factor influencing the likelihood of facing declaring the implications of ignoring heteroskedasticity CHE is the healthcare provider; households who receive by highlighting the dramatic differences in marginal ef- health services from private hospitals are most likely to fects estimates between the ordinary and heteroskedastic afford CHE. It is worth noting that most OOP health probit models. payments go to the private sector. Although more than 90% of insured households are covered by HIO hospi- 4.1 Limitations of the the study tals, only 19% use them, whereas 67.8% use private hos- It is worth mentioning the limitations of the study. First, pitals. This could be because the private health system the households that are unable to finance health costs offers better amenities, lower waiting time, and varied and do not seek healthcare were missed during analysis. and high-quality healthcare services. It is quite a surprise This could result in underestimating the proportion of to find that households using HIO or public hospitals households that incurred CHE [26]. However, our over- have CHE. all estimates are not at risk because the percentage of Inpatient services resulting from unforeseen health households with zero health payments did not exceed shocks could be sizable and exhaust more resources in 3% of the total sample. Second, the lack of longitudinal the short term, whereas, in the long term, outpatient ser- data is a major limitation as in other studies, longitu- vices capture a relatively great amount of health pay- dinal data were most suited to adequately capture the ef- ments in addition to drugs, especially for patients with fects of health payments and make a causal analysis. chronic diseases. Our estimates corroborate the findings Third, we cannot explore the other impoverishing effects of previous studies that households using inpatient of OOP health payments, such as using savings, incur- health services contribute to the increase in the risk of ring debts, and depleting productive assets as the survey incurring CHE, especially those who sought treatment in did not include these effects. private hospitals [29]. However, this study revealed the influential predictors Regarding the characteristics of the household head, the of incurring CHE. CHE is also related to the perform- likelihood of facing CHE is anticipated to increase as the ance of the health system and its health insurance tools. age of the household head increases due to age-related CHE and the resultant impoverishment have been wide- health needs [18, 20, 23], the age of the household head spread in countries with a poorly functioning national and its squared value are not significant variables in our health system, inadequate social welfare schemes, and model. Also, in contrast to earlier findings that female- poor economic performance. The characteristics of headed households are at higher risk of CHE [18, 28, 29], health services can exacerbate the financial burdens of we detected that male-headed households are more likely illness; availability of health services at unlimited costs to face CHE than those headed by females. Our estimates and low coverage can deter the poor from using health emphasized that households with educated heads are less services or cause regressive burdens [10]. Therefore, fur- likely to encounter CHE than those with illiterate heads. ther investigation is necessary to detect the role of the Educated heads consciously invest in health without mak- health system in shaping CHE. ing irrational decisions or incurring futile payments. In line with other studies, in this study, the employment sta- 5 Conclusion tus of the household head serves as a protective tool Heavy reliance on OOP health payments causes financial against financial catastrophes [18, 28, 29]. burdens for households and creates inequitable access to Although reducing poverty rates has a high priority on healthcare services, particularly in light of the growing the development agenda of national governments and presence of profit-seeking healthcare providers. CHE has international agencies, several studies have focused only received considerable attention in many countries and on poverty estimates in Egypt, whereas scarce empirical have been substantiated as a significant additional deter- studies shave investigated the burdensome of OOP minant of poverty; however, to date, the concern on health payments and their impoverishing impact at the CHE has remained insufficient in Egypt. OOP health household level, which provides little evidence on the in- payments remain the principal source of healthcare fi- cidence of CHE and their determinants [12]. There are nancing in Egypt as in most low-income countries. Iden- several notable strengths of this study. This study deter- tifying the characteristics that push households to be mined the underlying factors associated with the risk of more vulnerable to CHE is of great importance for Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 14 of 15 designing efficient health systems. In this context, this interpret the data. All the authors read, approved the manuscript, and consented to publish. study was conducted to produce conclusions to help policymakers in conducting meaningful reforms in rural Funding areas based on the incidence rates of CHE and their in- This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit organization. fluencing factors. Investigating the determinants of CHE in rural areas Availability of data and materials using a heteroskedastic probit model demonstrated that The data that support the findings of this study are available from Central Agency for Public Mobilization and Statistics (CAMPAS), but restrictions apply inference based on an ordinary probit model is inaccur- to the availability of these data, which were used under license for the ate due to the violation of the homoscedasticity assump- current study, and so are not publicly available. tion. In the heteroskedastic probit model, larger households positively are associated with CHE, and hav- Declarations ing high proportions of elderly, sick, and female mem- Ethics approval and consent to participate bers positively contribute to the probability of incurring We have used a secondary dataset (HIECS, 2015). The data has been CHE. Moreover, having high proportions of educated collected by Central Agency for Public Mobilization and Statistics (CAMPAS) maintaining all ethical approval and consent to participate. This study members, wage earners, and insured members helps received exempt status because it was a secondary analysis of an existing rural households escape from CHEs, whereas households data set; it does not deal on any human subject. with less educated, unemployed, and male heads exhibit Consent for publication higher probabilities of incurring CHE. Not applicable CHEs are substantially attributed to the increased de- mand of rural households for private sector services and Competing interests The authors declare that they have no competing interests. little use of health services provided by public hospitals. Even Egypt’s primary insurance provider does not have a Author details significant effect in protecting households against CHE. Department of Biostatistics and Demography, Faculty of Graduate Studies for Statistical Research, Cairo University, Giza, Egypt. Department of Applied Using health services provided by HIO hospitals or pub- Statistics and Econometrics, Faculty of Graduate Studies for Statistical lic hospitals in rural areas increases the likelihood of en- Research, Cairo University, Giza, Egypt. countering CHE. Received: 14 January 2021 Accepted: 21 June 2021 Many policies could be developed from our findings, such as enhancing the role of social health insurance in rural areas and expanding the health coverage for chron- References ically ill household heads to reduce the burden from 1. Gertler P, Gruber J. Insuring consumption against illness. Am Econ Rev. 2002;92(1):51–70. their shoulders and protect them from falling into pov- 2. Su TT, Kouyaté B, Flessa S. Catastrophic household expenditure for health erty. There is also an urgent need to limit and control care in a low-income society: a study from Nouna District, Burkina Faso. Bull OOP health payments absorbed by the private sector to World Health Organ. 2006;84(1):21–7. 3. Islam A, Maitra P. Health shocks and consumption smoothing in rural achieve an acceptable level of fairness in financing. households: does microcredit have a role to play? J Dev Econ [Internet]. Expanding the scope of financial protection has become 2012;97(2):232–43. a priority in rural areas to lessen the incidence of CHE 4. Sparrow R, de Poel EV, Hadiwidjaja G, Yumna A, Warda N, Suryahadi A. Coping with the economic consequences of ill health in Indonesia. Health while ensuring access to healthcare services for disad- Econ. 2014;23(6):719–28. vantaged groups. 5. Alam K, Mahal A. Economic impacts of health shocks on households in low and middle income countries: a review of the literature. Global Health. 2014;10:1. Abbreviations 6. Xu K, Evans DB, Kawabata K, Zeramdini R, Klavus J, Murray CJL. Household OOP health expenditure: Out-of-pocket health expenditure; catastrophic health expenditure: a multicountry analysis. Lancet. 2003; CHE: Catastrophic health expenditure; HIECS: Household Income, Expenditure 362(9378):111–7. and Consumption Survey; CAPMAS: Central Agency for Public Mobilization 7. Xu K, Evans DB, Carrin G, Aguilar-Rivera AM, Musgrove P, Evans T. and Statistics; HH Head: Household head; NHA: National health accounts; Protecting households from catastrophic health spending. Health Aff. QMLE: Quasi maximum likelihood estimator; ctp : Capacity to pay; se : Food h h 2007;26(4):972–83. subsistence spending 8. Leive A, Xu K. Coping with out-of-pocket health payments: empirical evidence from 15 African countries. Bull World Health Organ. 2008;86(11): Acknowledgements 849–56. We are greatly thankful to the reviewers of this paper, critical comments and 9. Van Damme W, Van Leemput L, Por I, Hardeman W, Meessen B. Out-of- suggestions make this paper ready for publication. We are grateful to Central pocket health expenditure and debt in poor households: evidence from Agency for Public Mobilization and Statistics, CAMPAS, Cairo, Egypt for Cambodia. Trop Med Int Heal. 2004;9(2):273–80. access to the data. 10. Wagstaff A, Van DE. Catastrophe and impoverishment in paying for health care: with applications to Vietnam1993--1998. Health Econ. 2003; Authors’ contributions 12(11):921–34. SA contributed to the conception and design of the study; helped to 11 van Doorslaer E, O’Donnell O, Rannan-Eliya RP, Somanathan A, Adhikari SR, acquire, analyze, and interpret the data; drafted the manuscript; and revised Garg CC, et al. Effect of payments for health care on poverty estimates in 11 the manuscript for important intellectual content. FS and MNA revised the countries in Asia: an analysis of household survey data. Lancet. 2006; manuscript for important intellectual content. MRA helped to analyze and 368(9544):1357–64. Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 15 of 15 12. Rashad A, Sharaf M. Catastrophic and impoverishing effects of out-of-pocket 37 Deb P, Munkin MK, Trivedi PK. Bayesian analysis of the two-part model with health expenditure: new evidence from Egypt. Am J Econ. 2015;5(5):526–33. endogeneity: application to health care expenditure. J Appl Econom. 2006; 13. Greene WH. Econometric analysis 7th Edition. International ed. New Jersey: 21(7):1081–99. Prentice Hall; 2012. 38. Shahrawat R, Rao KD. Insured yet vulnerable: out-of-pocket payments and India’s poor. Health Policy Plan. 2012;27(3):213–21. 14. Central Agency for Public Mobilization and Statistics, Egypt. Income, 39. Kimani DN, Mugo MG, Kioko UM. Catastrophic health expenditures and Expanditure & Consumption Survey 2015, Survey methodology. 2016. impoverishment in Kenya. Eur Sci J. 2016;12:434. 15. Deaton A, Zaidi S. Guidelines for constructing consumption aggregates for welfare analysis. Washington (DC): World Bank Publications, LSMS Working Paper No. 135; 2002. Publisher’sNote 16. Harvey AC. Estimating regression models with multiplicative Springer Nature remains neutral with regard to jurisdictional claims in heteroscedasticity. Econom J Econom Soc. 1976;48(3):461–5. published maps and institutional affiliations. 17. Cornelißen T. Standard errors of marginal effects in the heteroskedastic probit model. Discussion paper no. 320, Institute of Quantitative Economic Research, University of Hannover; 2005. [cited 2020 April, 20]. Available from: https://www.econstor.eu/bitstream/10419/22432/1/dp-320.pdf. 18. Rezapour A, Arabloo J, Tofighi S, Alipour V, Sepandy M, Mokhtari P, et al. Determining equity in household’s health care payments in Hamedan Province, Iran. Arch Iran Med. 2016;19(7):480–7. 19. O’Donnell O, Eddy van D, Rannan-Eliya RP, Somanathan A, Charu CG, Piya Hanvoravongchai MNH, et al. Explaining the incidence of catastrophic expenditures on health care: comparative evidence from Asia. EQUITAP [Internet]. 2005;5. Available from: http://www.equitap.org/publications/docs/ EquitapWP5.pdf. 20. Li Y, Wu Q, Xu L, Legge D, Hao Y, Gao L, et al. Factors affecting catastrophic health expenditure and impoverishment from medical expenses in China: policy implications of universal health insurance. Bull World Health Organ. 2012;90(9):664–71. 21. Van Minh H, Xuan Tran B. Assessing the household financial burden associated with the chronic non-communicable diseases in a rural district of Vietnam. Glob Health Action [Internet]. 2012;5(1):18892. 22. Van Minh H, Kim Phuong NT, Saksena P, James CD, Xu K. Financial burden of household out-of-pocket health expenditure in Viet Nam: findings from the National Living Standard Survey 2002-2010. Soc Sci Med. 2013;96:258–63. 23. Buigut S, Ettarh R, Amendah DD. Catastrophic health expenditure and its determinants in Kenya slum communities. Int J Equity Health [Internet]. 2015;14(1):1–12. 24. Narcı HÖ, Şahin İ,Yıldırım HH. Financial catastrophe and poverty impacts of out-of-pocket health payments in Turkey. Eur J Heal Econ. 2015;16(3):255–70. 25. Hjortsberg C. Why do the sick not utilise health care? The case of Zambia. Health Econ. 2003;12(9):755–70. 26. Pradhan M, Prescott N. Social risk management options for medical care in Indonesia. Health Econ. 2002;11(5):431–46. 27. Kien VD, Van Minh H, Giang KB, Dao A, Tuan LT, Ng N. Socioeconomic inequalities in catastrophic health expenditure and impoverishment associated with non-communicable diseases in urban Hanoi, Vietnam. Int J Equity Health [Internet]. 2016;15(1):1–11. 28. Choi J, Choi J, JH K, KB Y, EC P. Association between chronic disease and catastrophic health expenditure in Korea. BMC Health Serv Res. 2015;15(1): 1–8. 29. Cavagnero E, Carrin G, Xu K, Aguilar-Rivera Am A. Health financing in Argentina: an empirical study of health care expenditure and utilization. Innovations in Health Financing. 2006; working paper series No. 8. Available from: https://www.who.int/health_financing/documents/argentina_cava gnero.pdf. 30. Chuma J, Maina T. Catastrophic health care spending and impoverishment in Kenya. BMC Health Serv Res. 2012;12:413. 31. Somkotra T, Lagrada LP. Which households are at risk of catastrophic health spending: experience in Thailand after universal coverage. Health Aff. 2009;28:3. 32. Arsenijevic J, Pavlova M, Groot W. Measuring the catastrophic and impoverishing effect of household health care spending in Serbia. Soc Sci Med [Internet]. 2013;78(1):17–25. 33. Waters HR, Anderson GF, Mays J. Measuring financial protection in health in the United States. Health Policy (New York). 2004;69(3):339–49. 34. Reddy SR, Ross-Degnan D, Zaslavsky AM, Soumerai SB, Wagner AK. Health care payments in the asia pacific: validation of five survey measures of economic burden. Int J Equity Health. 2013;12(1):1–12. 35. Wagstaff A. Measuring financial protection in health. The World Bank. 2008. 36. Daneshkohan A, Karami M, Najafi F, Karami MB. Household catastrophic health expenditure. Iran J Public Health. 2011;40(1):94–9. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of the Egyptian Public Health Association Springer Journals

Predictors of catastrophic out-of-pocket health expenditure in rural Egypt: application of the heteroskedastic probit model

Loading next page...
 
/lp/springer-journals/predictors-of-catastrophic-out-of-pocket-health-expenditure-in-rural-JGU6HVIyuO
Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2021
eISSN
2090-262X
DOI
10.1186/s42506-021-00086-x
Publisher site
See Article on Publisher Site

Abstract

Background: Out-of-pocket (OOP) health expenditure is a pressing issue in Egypt and far exceeds half of Egypt’s total health spending, threatening the economic viability, and long-term sustainability of Egyptian households. Targeting households at risk of catastrophic health payments based on their characteristics is an obvious pathway to mitigate the impoverishing impacts of OOP health payments on livelihoods. This study was conducted to identify the risk factors of incurring catastrophic health payments hoping to formulate appropriate policies to protect households against financial catastrophes. Methods: Using data derived from the Egyptian Household Income, Expenditure, and Consumption Survey (HIECS), a multiplicative heteroskedastic probit model is applied to account for heteroskedasticity and avoid biased and inconsistent estimates. Results: Accounting for heteroskedasticity induces notable differences in marginal effects and demonstrates that the impact of some core variables is underestimated and insignificant and in the opposite direction in the homoscedastic probit model. Moreover, our results demonstrate the principal factors besides health status and socioeconomic characteristics responsible for incurring catastrophic health expenditure, such as the use of health services provided by the private sector, which has a dramatic effect on encountering catastrophic health payments. Conclusions: The marked differences between estimates of probit and heteroskedastic probit models emphasize the importance of investigating homoscedasticity assumption to avoid policies based on incorrect evidence. Many policies can be built upon our findings, such as enhancing the role of social health insurances in rural areas, expanding health coverage for poor households and chronically ill household heads, and providing adequate financial coverage for households with a high proportion of elderly, sick members, and females. Also, there is an urgent need to limit OOP health payments absorbed by private sector to achieve an acceptable level of fair financing. Keywords: Out-of-pocket health payments, Catastrophic health expenditure, Multiplicative heteroskedastic probit model * Correspondence: suzan.mohmed75@cu.edu.eg Department of Biostatistics and Demography, Faculty of Graduate Studies for Statistical Research, Cairo University, Giza, Egypt Full list of author information is available at the end of the article © The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 2 of 15 1 Introduction the latest round of the Egyptian Household Income, Ex- Diseases could cause catastrophic health payments, penditure, and Consumption Survey (HIECS), which en- undermine the process of income generation, and ables us to incorporate many fundamental covariates jeopardize future economic welfare [1]. Catastrophic neglected by other studies, including the type of health health expenditure (CHE) occurs when out-of-pocket insurance, health service providers, and household com- (OOP) health payments absorb a large portion of the re- position [12]. sources of a household to the extent that they severely affect the household’s living standard [2]. In the short 2 Methods term, catastrophic payments disrupt household con- 2.1 Data sumption of other essential needs, whereas, in the long Data were obtained from the Egyptian Household In- term, they may force households to deplete assets, use come, Expenditure and Consumption Survey) HIECS), a savings, sell livestock, and incur debts [3–5]. nationally representative survey conducted by Central Substantial health expenditures push households into Agency for Public Mobilization and Statistics) CAPM poverty or deeper [6, 7]. Approximately 30% of house- AS) in 2015 [14]. The HIECS uses three different types holds in 15 African countries had to borrow and sell as- of questionnaires: Expenditure and Consumption Ques- sets to finance health payments; this percentage tionnaire, Assisting Questionnaire, and Income Ques- increased to 50% when facing hospitalization expenses tionnaire. The survey collects data on demographic, and to 40% when using other health services [8]. Even socioeconomic, and health characteristics of rural and modest health payments cause indebtedness with exorbi- urban households, selected using a two-stage stratified tant interest rates besides the sale of consumables and random sampling design. HIECS measures consumption assets, which ultimately lead to poverty [9]. patterns and provides detailed information on house- Two approaches are commonly applied to measure the hold’s income sources and health status. Individuals incidence of catastrophic health payments: in the first were asked about suffering any illness during 6 months approach, catastrophic health payments are defined as preceding the survey, having a chronic disease or a dis- health expenditures exceeding a specific fraction of the ability, healthcare providers, and health insurance cover- total household expenditure or nonfood expenditures age and its type. [10]. In the second approach, they are defined as the Out-of-pocket healthcare expenditures (OOP) com- ability to pay, which is measured after deducting subsist- prise inpatient and outpatient health services and other ence spending [6]. Both approaches asserted that OOP reported expenditures (medicine, laboratory tests, X- health payments accelerate and exacerbate poverty; cop- rays, medical equipment) and do not include insurance ing strategies with health payments create a vicious cir- premiums. We used household as the unit of analysis in cle of indebtedness and impoverishment, and must be order to account for the intra-household distribution of considered to understand the trend and impetus of pov- resources and coping strategies with health payments, erty over time [11]. assuming that the economic impact of diseases affects all Although public health investments dwindle annually household members. Our analysis focuses on rural areas in Egypt, urban areas have captured the largest percent- including 6670 households. Rural households are distrib- age of public health spending, whereas rural areas have uted mainly in Rural Lower Egypt and Rural Upper invariably received less attention and are poorly funded. Egypt. All estimates are adjusted to represent national The gap between health spending in urban and that in figures using appropriate sampling weights. Both de- rural areas has reached roughly 67%, which is reflected scriptive and analytical statistics were performed using R in a substantial disparity in the availability of healthcare software. infrastructure. Moreover, rural areas are characterized by low levels of income, education, and economic devel- opment. All these factors may contribute to the high in- 2.2 Measuring incidence of catastrophic OOP health cidence rates of catastrophic payments [12]. payment This study was conducted to estimate the incidence of Catastrophic health expenditure was defined as “Any CHE, identify its core determinants, and reveal the ex- health expenditure that threatens a household’s financial tent to which health insurance schemes properly target capacity to maintain its subsistence needs and does not vulnerable economic households and whether they suc- necessarily equate to high healthcare costs. Even rela- ceed in protecting them financially. This study contrib- tively small expenditures on health can be financially utes to the literature by exploring the predictors of CHE disastrous for poor households” [2]. According to the using a multiplicative heteroskedastic probit model, that methodology proposed by Xu et al. [6], the most com- accounts for heteroskedasticity and provides unbiased mon approach applied in recent studies, OOP health ex- and consistent estimates [13]. Moreover, this study used penditure is catastrophic whenever it is greater than or Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 3 of 15 equal to 40 percent of the household’s capacity to pay, of rural households whose members were engaged in the where: agriculture sector and exhibited varied seasonal incomes. The probit model assumes that the error distribution – Capacity to pay (ctp ) is the remained household of the latent model is homoscedastic and has unit vari- income after satisfying food subsistence spending. ance. However, many studies that examined the determi- – Subsistence spending (se ) is determined based on nants of catastrophic health payments have neglected to average food expenditures of households with food verify the violation of the homoscedasticity assumption, share within the 45th and 55th percentile of the which could result in substantially biased and inconsist- total sample. Actual food expenditure is used to ent estimates besides misspecified standard errors. Spe- measure the capacity to pay for the household cific covariates may influence the probability of whose food expenditure is less than subsistence incurring catastrophic payments through health payment spending. Expenditures are adjusted using the variance. On average, better-off households may spend economy of scale estimated by Xu et al. (2003) (β= more on healthcare services exposing them to CHE, but 0.56) to reflect food consumption sharing among also the large variance of health payments at high ex- household members. Xu’s methodology has been penditure levels due to uncommon and sophisticated clearly described elsewhere [6]. medical treatments will raise the incidence of CHE. This variance is hypothesized to increase with household ex- 2.3 Multiplicative heteroskedastic probit model penditure as poor households have constrained budgets Misspecification in limited dependent models that may limiting their response to health shocks, whereas rich result from heteroskedasticity, omitted variables, hetero- households that face illness episodes spend too much geneity and nonlinearity leads to inconsistent and compared with rich households with good health. A het- biased estimates. The quasi maximum likelihood esti- eroskedastic probit model is applied to consider that mator (QMLE) of the probit model is inconsistent in health expenditure h is heteroskedastic with unfixed case of violation of homoscedasticity, even using variance σ . A heteroskedastic probit model generalizes white corrected estimator for standard errors is use- Φ(.) to a normal cumulative distribution function (CDF), less in the heteroskedastic probit model because the allowing variance to vary systematically as a multiplica- QMLE itself is biased [13]. tive function of set of m explanatory variables x [13, 16]. Let c denote a binary response (0,1) representing the no occurrence of catastrophic health expenditure (CHE) σ ¼ exp x τ ð3Þ i i when household health expenditure (h ) exceed its cap- acity to pay (ctp ). The subscript i denotes the where x is m × 1 vector of covariates (x , x , x ,.. x ) i 1i 2i 3i mi household: that determine the variance of the error term, while τ is the corresponding coefficient vector. 1 h > ctp i Therefore, the probability of incurring catastrophic c ¼ ð1Þ 0 h ≤ctp i health payments is () The probability of incurring catastrophic health y β PrðÞ c ¼ 1¼ Φ ð4Þ i 0 expenditure, exp x τ 0 0 A multiplicative heteroskedastic probit model relaxes PrðÞ c ¼ 1¼ Pr ε < y β ¼ Φ y β ð2Þ i i i i the homoscedasticity by allowing the scale of the inverse link function to vary systematically from one observation y is a k × 1vector of observed independent variables(y , i 1i to another as a function of explanatory variables. It is fit- y , y …. y ), β is vector of the corresponding parame- 2i 3i ki ted via maximum likelihood where the log-likelihood ters, ε is random error term and Φ(.) is the cumulative function takes the form: distribution function of standard normal variable with () mean 0 and variance 1. y β lnL ¼ h ln Φ Total household consumption is a superior proxy for i 0 exp x τ i∈ F household economic status; it is less subject to fluctua- "# () y β tions and measurement errors than income. Also, it in- þ h ln 1−Φ  ð5Þ i 0 cludes the value of home production, which is a major exp x τ i∉ F source of income in developing countries, especially rural areas [15]. Consequently, household consumption where F is the set of all observations i such that c ≠ was used to avoid underestimating the living standards 0and h denote the weights. i Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 4 of 15 All explanatory variables enter y vector while total ex- composition of the households indicated that 28.2% of penditure variable is included in x vector for modelling household members were children and 8.62% are elderly; nearly 18.2% of the households had at least one elderly the heteroskedastic variance of the error term σ . The member. In respect to household head characteristics, marginal effects are more informative and directly inter- 66.2% of household heads were literate, approximately pretable than multiplicative effects (probit model’s coef- half of them had at least an elementary certificate, only ficients). The heteroskedastic probit model’s marginal 0.8% had higher university degrees, 77.7% were effects are estimated as a function of the parameters β employed, 73.5% had chronic diseases, and only 31.7% and τ [13] as follow: were insured. The marginal effect of continuous variable w is The average OOP payments were significantly higher 0 0 among better-off households than those among poor ∂ PrðÞ c ¼ 1 y β β −y β:g i z i z ¼ ϕ ð6Þ 0 0 ones. The richest households spend nearly five times ∂w exp x τ exp x τ i i that of what the poorest households spend on healthcare services (LE 6954.28 vs. LE 1483.5). Meanwhile, the where ϕ(.) is the probability density function (pdf). If the health share of the total budget ranged from 8.17% at variable w is an element of both y and x vectors, its co- the poorest economic level to 11.08% at the richest level, efficient in the mean equation (1)is β and in the vari- indicating that poor households cannot deduct a large ance equation (4)is g , while g equals 0 if w appears z z z percentage of their budget for healthcare services. The only in y vector. health share of the household budget has an upward The marginal effect of dummy variable w is trend with the economic level (Fig. 1). Disaggregated 0 0 health expenditures enable us to distinguish between ∂ PrðÞ c ¼ 1 y β y β 1 0 ¼ Φ −Φ ð7Þ 0 0 budget shares allocated to inpatient and outpatient care ∂w expðÞ x τ expðÞ x τ 1 0 across economic levels. The results showed that the ex- w is set to 1 in y and x and set 0 in y and x . Stand- penditure share on inpatient services in rich households z 1 1 0 0 ard errors of the marginal effects are derived to allow in- was nearly four times that in poor households (2.55% vs. ference and hypothesis testing, using delta method 0.6%), whereas both poor and rich households allocated which is the most commonly used [17]. similar shares of their budgets to drugs (4.4% and 4.7%, To investigate the determinants of CHE, dummy vari- respectively). ables for the characteristics of household heads were in- Egypt has a narrow geographic distribution of health cluded in the analysis, such as gender, illiteracy status, services, which is more concentrated in urban areas. and education level. Besides, the analysis comprised age Households across the socioeconomic strata in urban and squared age of the household heads to control for areas witness extensive healthcare infrastructure in con- the nonlinearity of age. Several household-level covari- trast to rural households, which resulted in heavy reli- ates were also incorporated, such as the presence of ance on the purchasing of non-prescribed drugs in rural chronic diseases and disability; health insurance cover- areas. Drug purchases account for a larger share of OOP age; and proportions of sick members, insured members, health payments of rural households (54.7%), followed insured children, insured elderly, wage earners, and edu- by outpatient health services (38.7%), whereas inpatient cated members, in addition to household total expend- services acquire the smallest share (6.6%). iture and geographic area. To capture the effects of An investigation of the components of health expend- household composition, we included household size and iture across various economic levels indicated the over- the proportions of children, elderly, and females. use of medicines among rural households, particularly poor households; pharmaceutical products had the lar- 3 Results gest share of OOP health expenditure in the poorest The distribution of annual health expenditure was right- level compared with other quintiles (62.25% in the poor- skewed (skewness = 7.27), the mean annual health ex- est level vs. 48.1% in the richest level). Meanwhile, in- penditure (LE 2903.4) was higher than the median (LE patient services account for 11.97% of the total health 1706.3). The positive excess kurtosis (101) indicated that expenditure of rich households compared with only 4% health expenditure distribution is leptokurtic. After tak- of the total health expenditure of poor households. ing the log transformation, the mean value became The amount of OOP health payments depends mainly 7.316, whereas the median value was 7.42, skewness was on where households seek healthcare. The data revealed 0.27, and kurtosis was 1.4. The average annual house- a significant variation in the paid health shares by the hold income was LE 37625. The average household size type of healthcare providers; rural households using pri- was five members, with 25.6% being males. One of five vate hospitals and clinics paid on average 12.92% of their rural households was headed by a female. The age budgets on healthcare services. Also, households seeking Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 5 of 15 Fig. 1 The mean out-of-pocket health share by economic level healthcare in public hospitals and social health insurance counterparts (36% vs. 12%(; the proportion of insured hospitals allocate a high share, accounting for 10.5% and household heads increases sharply as the household eco- 10%, respectively, of their total budget. Households that nomic level increases, rising from 15% in the poorest level depend on medical consultations from pharmacies allo- to 52.2% in the richest level (Fig. 3). Meanwhile, at the cate 9.48% of their budgets. However, civil society health household level, on average, half of the household mem- services were associated with a relatively small share bers are insured, which is due to the compulsory insur- (6.19%). Although private healthcare providers demand ance for children in schools where the average number of higher payments than public healthcare providers and insured children in the household is 3, whereas, on aver- are hypothesized to be less used by households at low age, only one elderly member in the household is insured. economic levels, the data indicated that, contrary to ex- A significant difference in the budget share allocated pectations, private healthcare providers are the most to health was observed between insured and uninsured used across all economic levels; 65.58% of the poorest households. Although uninsured households spend, on households use private hospitals and clinics, and similar average, less than insured households (LE 2827.6 vs. LE proportions were found at other economic levels (Fig. 2). 3426.9), they allocate larger shares to health spending However, risk-pooling mechanisms are significant (8.8% vs. 8.3%). Also, health shares of households with tools for protecting vulnerable households from unex- uninsured heads were significantly higher than those of pected health risks. The HIECS indicates that health in- households with insured heads across all economic levels surance coverage is low in rural areas. Nearly 68.25% of (Fig. 4). A two-way analysis of variance (ANOVA) was household heads are uninsured, even though 71.64% of performed to simultaneously assess the effects of both them have at least one chronic disease. The high propor- health insurance coverage and economic level on health tion of informal workers, which characterizes rural areas, expenditure and revealed that both variables have sig- explains the low health insurance coverage for house- nificant effects on health expenditure at a significance hold heads. Additionally, apparent inequalities exist in level of 0.01. health insurance coverage across genders, education A comparison of health expenditure on inpatient and levels of household heads, and other socioeconomic outpatient services across insured and uninsured house- characteristics of households. The proportion of insured holds revealed the languid role of health insurance in male household heads is three times that of their female protecting rural households against high health costs. Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 6 of 15 Fig. 2 Distribution of rural households by economic level and healthcare provider Outpatient care expenditures among insured households insurance coverage was significant with a Pillai test stat- are significantly higher than those among uninsured istic of 0.03 and a p value of below 0.001. households; insured households spend, on average, ap- Also, the share distribution for outpatient services proximately LE 1422.3 on outpatient healthcare services highlights the relatively larger shares incurred by insured compared with LE 888.97 by uninsured households households (Fig. 5). A different conclusion is drawn for (3.58% vs. 2.79% of household budget). To simultan- the effect of insurance coverage on pharmaceutical eously test whether expenditures on both inpatient and spending; the pharmaceutical share spent by insured outpatient services vary between insured and uninsured households is significantly less than that incurred by un- households, multivariate analysis of variance (MAN- insured households (4.61% vs. 4.86%). OVA) was performed. The global multivariate test was Data revealed that rural areas have a high prevalence significant and indicated that the effect of health rate of chronic diseases; every rural household has at Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 7 of 15 Fig. 3 Distribution of rural households by health insurance coverage of household head across economic levels least one member with chronic disease, and the average catastrophic payment gap between households with and number of household members with chronic diseases is without chronic diseases, the presence of chronic dis- 3. Households suffering from chronic diseases spend eases increases the overall incidence rate of CHE from more than twice that of what households without 7.88 to 9.18%. Our data indicated also that having health chronic diseases spend. The health share of households insurance decreases the incidence of CHE among house- experiencing chronic diseases has a clear increasing pat- holds with chronic diseases to 5.44% compared with tern with the economic level (Fig. 6). 10.90% among households without health insurance The incidence of CHE is small in rural households as coverage. a whole (7.91%), but large among the poorest house- A homoscedastic probit model was used along with a holds (14.59%). The incidence of CHE has regressive dis- multiplicative heteroskedastic probit model in which the tribution across economic quintiles: it declines sharply total household expenditure becomes a part of the scale from 14.59% in the poorest quintile to 5.93% in the poor sub-model. The results of both models are provided in quintile, and it reaches 5.26% and 4.92% among house- Table 1. The likelihood ratio test of heteroskedasticity holds in the middle and rich economic levels, respect- substantiates a violation of homoscedasticity; the total ively; then it rises to 7.81% among the richest household expenditure is more correlated with variance. households. The presence of chronic diseases signifi- Coefficients are significant and have the hypothesized cantly increases the incidence of CHE (p < 0.01) and signs in the mean function of the heteroskedastic probit makes the pattern more serious at some economic levels. model. The likelihood ratio test on the nested models At the lowest economic level, the incidence rate of CHE prefers the heteroskedastic probit model over the ordin- among households with chronic diseases is 16.22% com- ary probit model. Moreover, comparing the two models pared with 9.46% among households without chronic using information metrics also confirmed that the het- diseases. While a better economic status decreases the eroskedastic probit model is the better model as it has Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 8 of 15 Fig. 4 The mean out-of-pocket health share by economic level and health insurance status of household heads lower Akaike’s information criteria. Also, the Bayesian probit model, whereas it was highly significant in both information criteria that strongly penalizes including mean and latent scale models (heteroskedastic probit additional variables to the model demonstrates the same model), indicating that the inference based on the ordin- conclusion. ary probit model is inaccurate in case of violation of the Significant differences in coefficients and marginal ef- homoscedasticity assumption. Moreover, household size fects were observed between the probit and heteroprobit has a reversal sign and becomes significant after ac- models, emphasizing how large the estimates are biased counting for heteroskedasticity. in the ordinary model. First, the impact of most variables While the two models differ in terms of the number of increases dramatically in the heteroskedastic probit significant variables, parameter estimates, and standard model, such as the effects of proportions of wage errors, some variables remain significant with the same earners, educated members, and insured children in the sign in both models, such as the proportions of female, household, whereas the impact of some variables dimin- elderly, and sick members, which are positively associ- ishes, such as the type of healthcare provider (e.g., public ated with CHE. Moreover, based on the estimates of the hospitals and hospitals affiliated with the health insur- two models, households headed by an illiterate individ- ance organization (HIO hospitals)). Second, some essen- ual, those headed by a man, those located in rural Lower tial variables recover their significance, such as Egypt, and those with chronically ill members suffer household size, the proportion of wage earners, health from CHE. Meanwhile, households headed by an edu- insurance by occupational syndicates, and outpatient cated, employed, and insured person are protected healthcare services. The most striking difference be- against CHE in both models. The two models demon- tween the two models was that the total household ex- strated that age and age squared of household heads penditure was not significant in the homoscedastic have no significant effects on the likelihood of incurring Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 9 of 15 Fig. 5 Distribution of outpatient services share across health insurance status CHE. Also, the proportions of children, insured children, increases the probability of CHE by 0.39. The heteroske- and insured elderly were not significant variables in both dasticity probit model showed that the infinitesimal models. change in total household expenditure reduces the likeli- Marginal effects and their standard errors are derived hood of CHE by 0.07. from both models (Table 1). Based on the estimated marginal effects of the heteroskedastic probit model, the 4 Discussion probability of incurring CHE increases by 0.06 and 0.04 Household characteristics have a significant effect on in male- and illiterate-headed households, respectively, health expenditure and the probability of encountering whereas it decreases by 0.05, 0.09, and 0.08 if the house- CHE. Poor households have the highest incidence rate hold head has a secondary degree, a university degree, of CHE; the distribution of CHE is regressive across eco- and a postgraduate degree, respectively. The employ- nomic quintiles. This is attributed to poor households’ ment status of the household head has a positive effect inability to afford healthcare. As commonly observed in on the probability of CHE: it decreases the probability of the literature [18], consumption necessities, such as food CHE by 0.07. An increase in the proportion of educated and shelter, exhaust the main bulk of a poor household’s members and wage earners in the household reduces the budget, leaving a small share to healthcare. Meanwhile, probability of CHE by 0.33 and 0.27, respectively. Mean- better-off households could finance health payments and while, the proportions of female, elderly, and sick mem- are more protected against CHE than poor households. bers increase the likelihood of incurring CHE by 0.08, Our empirical results demonstrated that a large-size 0.03, and 0.06, respectively. The presence of chronic dis- household is more likely to incur CHE. This is consistent eases raises the probability of CHE by 0.1, whereas the with the findings of [19], who have suggested that the probability drops by 0.04 in households with health in- large size of households increases the probability of at surance coverage. Using private sector services greatly least one member getting sick, especially in the presence Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 10 of 15 Fig. 6 The mean out-of-pocket health share according to the presence of chronic diseases of overcrowding and infectious diseases. Besides, the lim- 22, 27, 28]; besides, they are more expected to face sev- ited economies of scale in using medical care compared eral episodes of illness [29]. In addition, female propor- with other consumption items increase health payments. tion in households is also positively associated with However, some studies have found that having a larger CHE. One key finding was that educated household size protects households from CHE as it creates an oppor- members successfully shelter their households from tunity for accumulating earnings [18, 20–22]. CHE. This effect could be explained by the significance The high proportion of wage earners significantly pro- of education in reducing the anxiety associated with tects rural households from CHE. It is an indicator of a sickness and preventing making an irrational decision. household’s ability to pool financial resources and in- Household standard of living determines the amount formative about health insurance coverage, especially if of affordable health payments and the risk of facing the wage earners are officially employed and receive CHE. The results revealed that the total household ex- regular income [23, 24]. This finding is in contrast to penditure is negatively associated with the risk of incur- those of previous studies [25], which have revealed that ring CHE. This association has been established by a high proportion of wage earners is positively associated many studies [2, 21, 27, 30]. However, some studies have with CHE as they are more likely to seek healthcare and stated that the likelihood of incurring CHE decreases in provide financial support to other household members. response to an increase in household expenditure as The household composition reflects its health needs poor households are forced to forgo or delay seeking and required expenditure [26]. The results demonstrated healthcare to avoid fees, whereas better-off house- that households with a high proportion of elderly mem- holds are at higher risk of CHE because they have bers are highly likely to face CHE than those with high high responsiveness to health needs, and simultan- proportions of children. This is mainly due to their well- eously, they prefer the highest quality private health- known health needs and their inability to work [7, 21, care facilities [7, 19, 22, 24, 31]. Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 11 of 15 Table 1 Estimates of homoscedastic probit model and heteroskedastic probit model, HIECS, 2015 Variables Probit model Heteroskedastic probit model Coef. Marginal effect Coef. Marginal effect "Mean model" Intercept − 2.407*** (0.504) − 2.398* (1.239) Male-headed household 0.363*** (0.087) 0.045*** (0.009) 0.816*** (0.204) 0.061** (0.053) Age of household head − 0.012 (0.0127) − 0.001 (0.002) − 0.010 (0.026) − 0.001 (0.001) Squared age of household head 0.000 (0.000) 0.000 (0.000) − 0.010 (0.026) 0.001 (0.001) Illiterate household head 0.126* (0.071) 0.018* (0.010) 0.239 (0.159) 0.049** (0.071) Educational degree of household head (Dummy variables) Having a primary or lower secondary degree − 0.196 (0.091) − 0.015 (0.011) − 0.151 (0.212) − 0.013 (0.023) Having a secondary or post-secondary degree − 0.151* (0.091) − 0.019* (0.010) − 0.352* (0.199) − 0.053* (0.170) Having a university degree − 0.469** (0.154) − 0.048*** (0.011) − 1.273** (0.182) − 0.094*** (0.021) Having a postgraduate degree − 0.028 (0.474) − 0.003 (0.063) − 1.311 (2.176) − 0.083* (0.063) None (omitted category) Employed household head − 0.255*** (0.474) − 0.038** (0.012) − 0.528** (0.162) − 0.072** (0.069) Geographical area: rural lower Egypt 0.722* (0.367) 0.126* (0.065) 1.868* (1.003) 0.321** (0.091) Household size − 0.0206 (0.027) − 0.002 (0.003) 0.158** (0.060) 0.032** (0.021) Proportion of females 0.437*** (0.132) 0.059*** (0.018) 0.592** (0.283) 0.082** (0.031) Proportion of children in household 0.186 (0.169) 0.025 (0.023) 0.297 (0.397) 0.064 (0.317) Proportion of elderly in household 0.271* (0.135) 0.037* (0.018) 0.200** (0.260) 0.029** (0.012) Presence of chronic diseases 0.324*** (0.064) 0.039*** (0.007) 0.769*** (0.152) 0.097*** (0.032) Presence of disability 0.164 (0.012) 0.025 (0.021) 0.118 (0.261) 0.101 (0.073) Proportion of sick members in household 0.218* (0.19) 0.031* (0.017) 0.506* (0.269) 0.063** (0.072) Presence of insurance coverage − 0.213*** (0.064) − 0.027*** (0.007) − 0.447** (0.159) − 0.042*** (0.029) Household total expenditure − 0.003 (0.002) − 0.009 (0.007) − 0.167*** (0.012) − 0.071*** (0.47) Proportion of wage earners in household − 0.219 (0.178) − 0.027 (0.022) − 4.052** (0.495) − 0.271*** (0.149) Proportion of educated members in household − 0.551*** (0.164) − 0.066*** (0.019) − 1.373** (0.473) − 0.332*** (0.017) Proportion of children covered by Health Insurance − 0.355 (0.266) − 0.039 (0.029) − 1.595 (1.014) − 0.071 (0.017) Proportion of elderly covered by Health Insurance 0.238 (0.912) 0.035 (0.721) 0.934 (1.127) 0.011 (0.23) Type of health insurance coverage (Dummy variables) − 0.028 (0.174) − 0.004 (0.023) − 0.229 (0.423) − 0.061 (0.093) Covered by HIO Covered by private insurance 0.391 (0.433) 0.068 0.093) 0.617 (1.02) 0.252 (0.093) Covered by employer-provided private insurance − 0.279 (0.259) − 0.031 (0.024) − 0.594 (0.656) − 0.081 (0.074) Covered by occupational syndicates 0.391 (0.284) 0.068 (0.061) 1.285* (0.684) 0.168* (0.051) Other (omitted category) Health provider (Dummy variables) HIO hospitals 4.793** (0.169) 0.027 (0.028) 0.161* (0.400) 0.012* (0.021) Public hospitals 4.275*** (0.101) 0.085 (0.017) 0.153* (0.224) 0.031* (0.021) Private hospitals 0.531*** (0.064) 0.066*** (0.009) 1.097* (0.146) 0.389** (0.071) Pharmacy − 0.081 (0.124) − 0.026 (0.012) − 0.418 (0.245) − 0.076 (0.032) Other(omitted category) Using outpatient health services 0.131 (0.101) 0.019 (0.017) 0.049* (0.242) 0.091* (0.021) Het-test χ ------ 52.24*** LogL − 1566.968 − 1541 Likelihood ratio test 229.13*** McFadden’s pseudo R 0.069 Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 12 of 15 Table 1 Estimates of homoscedastic probit model and heteroskedastic probit model, HIECS, 2015 (Continued) Variables Probit model Heteroskedastic probit model Coef. Marginal effect Coef. Marginal effect "Mean model" AIC 3175.94 3125.69 BIC 3317.12 3273.603 Number of fisher scoring iterations 6 10 N 6143 6143 ***p<0.001, **p<0.01, *p<0.05, ( ): robust standard errors against misspecification. Coefficient and standard error of household total expenditure in latent scale model is{0.003*** (0.002)}. LogL is the maximum value of the likelihood function. Likelihood ratio test measures the significance of the overall model. Pseudo R shows the amount of variance explained by explanatory variables. Wald χ test the significance of the full model; if all slopes equal 0. Het-test χ is the likelihood ratio test of the null hypothesis of homoscedasticity; it tests the model with heteroskedasticity against the full model with out. AIC and BIC are information metrics that penalize the inclusion of additional variables and used to select the appropriate model. Number of fisher scoring iterations indicates how quickly the iterative weighted least squares (IWLS) algorithm terminates. Illness is the main driver of health expenditures and been higher among insured rural households than unin- pushes households to reallocate considerable shares of sured ones. In addition, 49.2% of Iranian households had their resources to OOP health payments, whereas demo- incurred debts, 21.7% sold their jewelry, and 15.9% used graphic–economic characteristics control the burden of their savings to finance health costs [36]. These findings these expenditures. Our findings corroborate earlier lit- confirm that the provision of health insurance should erature that the presence of chronic diseases significantly coincide with adequate financial coverage for all re- increases the risk of incurring CHE [21, 27, 28, 32]. quired health needs, in addition to providing accessible Meanwhile, the presence of disabilities does not have a health services of acceptable quality to deter recourse to significant effect on the risk of facing CHE, which con- the private sector. tradicts the findings of [18, 33]. To control potential confounders related to health in- In the absence of health insurance coverage, health surance status, we used the available data about the type expenditures can absorb a substantial fraction of the of health insurance. Although the selection of health in- household resources and severely disrupt living standards surance providers may depend on unobservable covariates [33]. Health insurance coverage significantly reduces the [37], our analysis did not have this issue as the choice of chance of incurring CHE. A similar conclusion was health insurance in Egypt is highly restricted. There is no reached by many studies. For example, Reddy (2013) has adverse selection as public sector employees and students found that health insurance is negatively correlated with are covered semi-compulsorily by Egypt’s primary insur- the risk of incurring CHE and indebtedness in three Asia- ance provider) HIO hospitals), and employees of other op- Pacific countries: China, Malaysia, and the Philippines erational bodies are covered by occupational syndicates, [34]. Also, Van Minh and Tran (2012) have indicated that whereas private health insurance is less prevalent in Egypt, health insurance has a modest positive effect on CHE. particularly rural areas. Moreover, our model incorporated Contrary to the widely held hypothesis about the role of the measures of health status as the presence of chronic health insurance in reducing financial hardships [21], diseases or disability and the socioeconomic covariates to some studies have shown that that health insurance has a control for any selection effect. contradictory effect on health payments: health insurance The type of health insurance does not have a signifi- caused larger health payments as it facilitates access to cant effect on protecting insured households from finan- healthcare services creating an induced demand for ser- cial hardships, except that provided by occupational vices [4, 18, 27, 35]. syndicates which, in contrast, is positively associated Our most intriguing finding is that increasing the pro- with CHEs. Some studies have shown that health insur- portion of insured children and elderly members had in- ance coverage has contrasting effects according to the significant effects on the incidence of CHE. These benefits package design [18, 23]. Limited insurance pack- findings support that increasing health insurance cover- ages with low depth of coverage, high co-payment rates, age does not necessarily reduce OOP health spending and relatively low-reimbursement ceilings increase the and achieve effective protection against CHE. For ex- risk of CHE. For example, Shahrawat and Rao [38] have ample, Iran has undertaken promising steps to expand revealed that Indian insurance schemes with limited health insurance coverage for most rural households, coverage (hospital expenses only) did not contribute to and 90% of the population has access to primary health- reducing CHE as outpatient care expenses occupied the care services; however, the incidence rate of CHE has main share of health expenditure. Kimani et al. [39] also Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 13 of 15 have explained that health insurance fails to protect facing CHE in rural areas that can provide the basis for households due to its limited coverage of only bed costs analyzing policy options to alleviate the burden of cata- for inpatient stays. Others have demonstrated that pri- strophic payments. Moreover, it explains the role of vate health insurance offering a broad benefits package existing health insurance schemes in providing financial succeeds in protecting households from CHE compared protection in rural areas. The key strength of this study with social insurance schemes [29]. lies in obtaining unbiased and consistent estimates, and Another factor influencing the likelihood of facing declaring the implications of ignoring heteroskedasticity CHE is the healthcare provider; households who receive by highlighting the dramatic differences in marginal ef- health services from private hospitals are most likely to fects estimates between the ordinary and heteroskedastic afford CHE. It is worth noting that most OOP health probit models. payments go to the private sector. Although more than 90% of insured households are covered by HIO hospi- 4.1 Limitations of the the study tals, only 19% use them, whereas 67.8% use private hos- It is worth mentioning the limitations of the study. First, pitals. This could be because the private health system the households that are unable to finance health costs offers better amenities, lower waiting time, and varied and do not seek healthcare were missed during analysis. and high-quality healthcare services. It is quite a surprise This could result in underestimating the proportion of to find that households using HIO or public hospitals households that incurred CHE [26]. However, our over- have CHE. all estimates are not at risk because the percentage of Inpatient services resulting from unforeseen health households with zero health payments did not exceed shocks could be sizable and exhaust more resources in 3% of the total sample. Second, the lack of longitudinal the short term, whereas, in the long term, outpatient ser- data is a major limitation as in other studies, longitu- vices capture a relatively great amount of health pay- dinal data were most suited to adequately capture the ef- ments in addition to drugs, especially for patients with fects of health payments and make a causal analysis. chronic diseases. Our estimates corroborate the findings Third, we cannot explore the other impoverishing effects of previous studies that households using inpatient of OOP health payments, such as using savings, incur- health services contribute to the increase in the risk of ring debts, and depleting productive assets as the survey incurring CHE, especially those who sought treatment in did not include these effects. private hospitals [29]. However, this study revealed the influential predictors Regarding the characteristics of the household head, the of incurring CHE. CHE is also related to the perform- likelihood of facing CHE is anticipated to increase as the ance of the health system and its health insurance tools. age of the household head increases due to age-related CHE and the resultant impoverishment have been wide- health needs [18, 20, 23], the age of the household head spread in countries with a poorly functioning national and its squared value are not significant variables in our health system, inadequate social welfare schemes, and model. Also, in contrast to earlier findings that female- poor economic performance. The characteristics of headed households are at higher risk of CHE [18, 28, 29], health services can exacerbate the financial burdens of we detected that male-headed households are more likely illness; availability of health services at unlimited costs to face CHE than those headed by females. Our estimates and low coverage can deter the poor from using health emphasized that households with educated heads are less services or cause regressive burdens [10]. Therefore, fur- likely to encounter CHE than those with illiterate heads. ther investigation is necessary to detect the role of the Educated heads consciously invest in health without mak- health system in shaping CHE. ing irrational decisions or incurring futile payments. In line with other studies, in this study, the employment sta- 5 Conclusion tus of the household head serves as a protective tool Heavy reliance on OOP health payments causes financial against financial catastrophes [18, 28, 29]. burdens for households and creates inequitable access to Although reducing poverty rates has a high priority on healthcare services, particularly in light of the growing the development agenda of national governments and presence of profit-seeking healthcare providers. CHE has international agencies, several studies have focused only received considerable attention in many countries and on poverty estimates in Egypt, whereas scarce empirical have been substantiated as a significant additional deter- studies shave investigated the burdensome of OOP minant of poverty; however, to date, the concern on health payments and their impoverishing impact at the CHE has remained insufficient in Egypt. OOP health household level, which provides little evidence on the in- payments remain the principal source of healthcare fi- cidence of CHE and their determinants [12]. There are nancing in Egypt as in most low-income countries. Iden- several notable strengths of this study. This study deter- tifying the characteristics that push households to be mined the underlying factors associated with the risk of more vulnerable to CHE is of great importance for Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 14 of 15 designing efficient health systems. In this context, this interpret the data. All the authors read, approved the manuscript, and consented to publish. study was conducted to produce conclusions to help policymakers in conducting meaningful reforms in rural Funding areas based on the incidence rates of CHE and their in- This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit organization. fluencing factors. Investigating the determinants of CHE in rural areas Availability of data and materials using a heteroskedastic probit model demonstrated that The data that support the findings of this study are available from Central Agency for Public Mobilization and Statistics (CAMPAS), but restrictions apply inference based on an ordinary probit model is inaccur- to the availability of these data, which were used under license for the ate due to the violation of the homoscedasticity assump- current study, and so are not publicly available. tion. In the heteroskedastic probit model, larger households positively are associated with CHE, and hav- Declarations ing high proportions of elderly, sick, and female mem- Ethics approval and consent to participate bers positively contribute to the probability of incurring We have used a secondary dataset (HIECS, 2015). The data has been CHE. Moreover, having high proportions of educated collected by Central Agency for Public Mobilization and Statistics (CAMPAS) maintaining all ethical approval and consent to participate. This study members, wage earners, and insured members helps received exempt status because it was a secondary analysis of an existing rural households escape from CHEs, whereas households data set; it does not deal on any human subject. with less educated, unemployed, and male heads exhibit Consent for publication higher probabilities of incurring CHE. Not applicable CHEs are substantially attributed to the increased de- mand of rural households for private sector services and Competing interests The authors declare that they have no competing interests. little use of health services provided by public hospitals. Even Egypt’s primary insurance provider does not have a Author details significant effect in protecting households against CHE. Department of Biostatistics and Demography, Faculty of Graduate Studies for Statistical Research, Cairo University, Giza, Egypt. Department of Applied Using health services provided by HIO hospitals or pub- Statistics and Econometrics, Faculty of Graduate Studies for Statistical lic hospitals in rural areas increases the likelihood of en- Research, Cairo University, Giza, Egypt. countering CHE. Received: 14 January 2021 Accepted: 21 June 2021 Many policies could be developed from our findings, such as enhancing the role of social health insurance in rural areas and expanding the health coverage for chron- References ically ill household heads to reduce the burden from 1. Gertler P, Gruber J. Insuring consumption against illness. Am Econ Rev. 2002;92(1):51–70. their shoulders and protect them from falling into pov- 2. Su TT, Kouyaté B, Flessa S. Catastrophic household expenditure for health erty. There is also an urgent need to limit and control care in a low-income society: a study from Nouna District, Burkina Faso. Bull OOP health payments absorbed by the private sector to World Health Organ. 2006;84(1):21–7. 3. Islam A, Maitra P. Health shocks and consumption smoothing in rural achieve an acceptable level of fairness in financing. households: does microcredit have a role to play? J Dev Econ [Internet]. Expanding the scope of financial protection has become 2012;97(2):232–43. a priority in rural areas to lessen the incidence of CHE 4. Sparrow R, de Poel EV, Hadiwidjaja G, Yumna A, Warda N, Suryahadi A. Coping with the economic consequences of ill health in Indonesia. Health while ensuring access to healthcare services for disad- Econ. 2014;23(6):719–28. vantaged groups. 5. Alam K, Mahal A. Economic impacts of health shocks on households in low and middle income countries: a review of the literature. Global Health. 2014;10:1. Abbreviations 6. Xu K, Evans DB, Kawabata K, Zeramdini R, Klavus J, Murray CJL. Household OOP health expenditure: Out-of-pocket health expenditure; catastrophic health expenditure: a multicountry analysis. Lancet. 2003; CHE: Catastrophic health expenditure; HIECS: Household Income, Expenditure 362(9378):111–7. and Consumption Survey; CAPMAS: Central Agency for Public Mobilization 7. Xu K, Evans DB, Carrin G, Aguilar-Rivera AM, Musgrove P, Evans T. and Statistics; HH Head: Household head; NHA: National health accounts; Protecting households from catastrophic health spending. Health Aff. QMLE: Quasi maximum likelihood estimator; ctp : Capacity to pay; se : Food h h 2007;26(4):972–83. subsistence spending 8. Leive A, Xu K. Coping with out-of-pocket health payments: empirical evidence from 15 African countries. Bull World Health Organ. 2008;86(11): Acknowledgements 849–56. We are greatly thankful to the reviewers of this paper, critical comments and 9. Van Damme W, Van Leemput L, Por I, Hardeman W, Meessen B. Out-of- suggestions make this paper ready for publication. We are grateful to Central pocket health expenditure and debt in poor households: evidence from Agency for Public Mobilization and Statistics, CAMPAS, Cairo, Egypt for Cambodia. Trop Med Int Heal. 2004;9(2):273–80. access to the data. 10. Wagstaff A, Van DE. Catastrophe and impoverishment in paying for health care: with applications to Vietnam1993--1998. Health Econ. 2003; Authors’ contributions 12(11):921–34. SA contributed to the conception and design of the study; helped to 11 van Doorslaer E, O’Donnell O, Rannan-Eliya RP, Somanathan A, Adhikari SR, acquire, analyze, and interpret the data; drafted the manuscript; and revised Garg CC, et al. Effect of payments for health care on poverty estimates in 11 the manuscript for important intellectual content. FS and MNA revised the countries in Asia: an analysis of household survey data. Lancet. 2006; manuscript for important intellectual content. MRA helped to analyze and 368(9544):1357–64. Abdel-Rahman et al. Journal of the Egyptian Public Health Association (2021) 96:23 Page 15 of 15 12. Rashad A, Sharaf M. Catastrophic and impoverishing effects of out-of-pocket 37 Deb P, Munkin MK, Trivedi PK. Bayesian analysis of the two-part model with health expenditure: new evidence from Egypt. Am J Econ. 2015;5(5):526–33. endogeneity: application to health care expenditure. J Appl Econom. 2006; 13. Greene WH. Econometric analysis 7th Edition. International ed. New Jersey: 21(7):1081–99. Prentice Hall; 2012. 38. Shahrawat R, Rao KD. Insured yet vulnerable: out-of-pocket payments and India’s poor. Health Policy Plan. 2012;27(3):213–21. 14. Central Agency for Public Mobilization and Statistics, Egypt. Income, 39. Kimani DN, Mugo MG, Kioko UM. Catastrophic health expenditures and Expanditure & Consumption Survey 2015, Survey methodology. 2016. impoverishment in Kenya. Eur Sci J. 2016;12:434. 15. Deaton A, Zaidi S. Guidelines for constructing consumption aggregates for welfare analysis. Washington (DC): World Bank Publications, LSMS Working Paper No. 135; 2002. Publisher’sNote 16. Harvey AC. Estimating regression models with multiplicative Springer Nature remains neutral with regard to jurisdictional claims in heteroscedasticity. Econom J Econom Soc. 1976;48(3):461–5. published maps and institutional affiliations. 17. Cornelißen T. Standard errors of marginal effects in the heteroskedastic probit model. Discussion paper no. 320, Institute of Quantitative Economic Research, University of Hannover; 2005. [cited 2020 April, 20]. Available from: https://www.econstor.eu/bitstream/10419/22432/1/dp-320.pdf. 18. Rezapour A, Arabloo J, Tofighi S, Alipour V, Sepandy M, Mokhtari P, et al. Determining equity in household’s health care payments in Hamedan Province, Iran. Arch Iran Med. 2016;19(7):480–7. 19. O’Donnell O, Eddy van D, Rannan-Eliya RP, Somanathan A, Charu CG, Piya Hanvoravongchai MNH, et al. Explaining the incidence of catastrophic expenditures on health care: comparative evidence from Asia. EQUITAP [Internet]. 2005;5. Available from: http://www.equitap.org/publications/docs/ EquitapWP5.pdf. 20. Li Y, Wu Q, Xu L, Legge D, Hao Y, Gao L, et al. Factors affecting catastrophic health expenditure and impoverishment from medical expenses in China: policy implications of universal health insurance. Bull World Health Organ. 2012;90(9):664–71. 21. Van Minh H, Xuan Tran B. Assessing the household financial burden associated with the chronic non-communicable diseases in a rural district of Vietnam. Glob Health Action [Internet]. 2012;5(1):18892. 22. Van Minh H, Kim Phuong NT, Saksena P, James CD, Xu K. Financial burden of household out-of-pocket health expenditure in Viet Nam: findings from the National Living Standard Survey 2002-2010. Soc Sci Med. 2013;96:258–63. 23. Buigut S, Ettarh R, Amendah DD. Catastrophic health expenditure and its determinants in Kenya slum communities. Int J Equity Health [Internet]. 2015;14(1):1–12. 24. Narcı HÖ, Şahin İ,Yıldırım HH. Financial catastrophe and poverty impacts of out-of-pocket health payments in Turkey. Eur J Heal Econ. 2015;16(3):255–70. 25. Hjortsberg C. Why do the sick not utilise health care? The case of Zambia. Health Econ. 2003;12(9):755–70. 26. Pradhan M, Prescott N. Social risk management options for medical care in Indonesia. Health Econ. 2002;11(5):431–46. 27. Kien VD, Van Minh H, Giang KB, Dao A, Tuan LT, Ng N. Socioeconomic inequalities in catastrophic health expenditure and impoverishment associated with non-communicable diseases in urban Hanoi, Vietnam. Int J Equity Health [Internet]. 2016;15(1):1–11. 28. Choi J, Choi J, JH K, KB Y, EC P. Association between chronic disease and catastrophic health expenditure in Korea. BMC Health Serv Res. 2015;15(1): 1–8. 29. Cavagnero E, Carrin G, Xu K, Aguilar-Rivera Am A. Health financing in Argentina: an empirical study of health care expenditure and utilization. Innovations in Health Financing. 2006; working paper series No. 8. Available from: https://www.who.int/health_financing/documents/argentina_cava gnero.pdf. 30. Chuma J, Maina T. Catastrophic health care spending and impoverishment in Kenya. BMC Health Serv Res. 2012;12:413. 31. Somkotra T, Lagrada LP. Which households are at risk of catastrophic health spending: experience in Thailand after universal coverage. Health Aff. 2009;28:3. 32. Arsenijevic J, Pavlova M, Groot W. Measuring the catastrophic and impoverishing effect of household health care spending in Serbia. Soc Sci Med [Internet]. 2013;78(1):17–25. 33. Waters HR, Anderson GF, Mays J. Measuring financial protection in health in the United States. Health Policy (New York). 2004;69(3):339–49. 34. Reddy SR, Ross-Degnan D, Zaslavsky AM, Soumerai SB, Wagner AK. Health care payments in the asia pacific: validation of five survey measures of economic burden. Int J Equity Health. 2013;12(1):1–12. 35. Wagstaff A. Measuring financial protection in health. The World Bank. 2008. 36. Daneshkohan A, Karami M, Najafi F, Karami MB. Household catastrophic health expenditure. Iran J Public Health. 2011;40(1):94–9.

Journal

Journal of the Egyptian Public Health AssociationSpringer Journals

Published: Aug 3, 2021

Keywords: Out-of-pocket health payments; Catastrophic health expenditure; Multiplicative heteroskedastic probit model

References