Study participants and sampling
This cross-sectional study was carried out on households with children aged 0 to 8 years having physical or mental disabilities in Iran in 2020. The sample size was estimated to be 385 households for each province based on considering the 50% exposure rate of households with disabled children to catastrophic health expenditures (p = 0.5), statistical confidence of 95%, and accuracy rate of 5%. Due to the fact that the study was conducted in 5 provinces, the total sample size was equal to 2000 households (400 households from each province).
Multi-stage clustering method was used for sampling. At first, all provinces of the country were divided into 5 clusters (north, south, west, east and center clusters). Then, one province was randomly selected from each cluster (Kurdistan, Alborz, Kerman, Fars and Gorgan provinces). Among the cities of the selected provinces, the capital city of the province (cities of Sanandaj, Karaj, Kerman, Shiraz and Gorgan) was selected. In the next step, the list of children with disabilities (including visual, hearing, speech, mental and motor disabilities) and with the age of 0 to 8 years was obtained from the welfare organization of selected cities and 400 samples were selected by random sampling.
Data collection
After selecting the households, the families were first contacted by phone, and if they wanted to participate in the study, some trained interviewers, who had at least a bachelor's degree in rehabilitation-related fields, went to the household door to interview and complete the questionnaire. Questionnaires were completed with the help of an informed family member. An informed member of the household was the person who was at least 18 years old and had the most information about income, expenses, and consumption of household health services. Data were collected using the World Health Survey questionnaire, which was developed in 2003 to measure the performance of health systems. This questionnaire has been translated into Persian and its validity and reliability have been confirmed in previous studies [23].
The questionnaire consists of three parts: first part includes demographic characteristics and socio-economic status of the family of a child with disabilities, second part involves the type of disability and the status of benefiting from outpatient, rehabilitation and inpatient health services, and the third part relates to monthly household expenses. In this study we used two recall periods for expenditure questions: The last 30 days for the total household expenditures, and the last 30 days and the last 12 months for outpatient and inpatient expenditures, respectively
Outcome Variable
In this study, the method provided by WHO was used to estimate the incidence of CHE. Accordingly, if the out-of-pocket payment (OPP) for health expenses of a child aged 0 to 8 with a disability was more than 40% of the household's capacity to pay, the household would be in the group of households facing catastrophic health expenditures. The method and details of calculating the incidence of CHE are given in other studies [19].
Independent variables
The socio-economic status of the household was determined using the method proposed by O'Donnell et al. [24]. In this method, the status of household assets including LCD TVs, separate refrigerators, washing machines, cell phones, dishwashers, microwave ovens, Internet access, private cars, private homes, and the number of rooms were assessed. Then, the asset index for each individual was calculated using principal component analysis (PCA) and the study population was classified into 5 quantiles of 1 (the poorest), 2, 3, 4, and 5 (the richest). In some studies, the asset index has been used to determine SES in the Iranian population [6, 25]. The use of inpatient services is a dichotomous variable (yes / no) measured by asking the question "Have you been hospitalized for an illness in the past year?" The use of outpatient services is also a dichotomous variable (yes / no) assessed by asking the question "Have you used outpatient services including visiting a medical doctor, dentist, midwife, nurse, physiotherapist, occupational therapist, speech therapist, audiometrist , optometrist, and traditional therapists, and other rehabilitation services provided by other providers due to illness in the past month.
Other independent variables included gender of head of household (male/female); age group of disabled child (< 5 years/ > 5 years); type of health insurance (Iranian health insurance, armed forces insurance, social security insurance, and other insurances); having supplementary insurance (yes / no); severity of disability (mild, moderate, severe); type of disability (physical, mental, blind, deaf, speech impairment); and number of disabled people in the household (one person, more than one person).
Statistical analysis
Mean (standard deviation) was calculated for the variables of total monthly household expenses, household food expenses, and out-of-pocket payments for health services. Frequency distribution tables were calculated for independent and demographic variables, and frequency percentage was measured for incidence of CHE. The relationship between independent variables and incidence of CHE was evaluated by calculating Odds Ratio (OR) using Multivariate logistic regression. The variables with p< 0.2 in the univariate logistic regression analysis were included in the adjusted logistic regression models. Data analyses were performed using STATA software version 13 (College Station, TX, United States). P<0.05 was considered statistically significant.