Factors Associated to the Enrollment in Health Insurance : An Experience from Selected Districts of Nepal

The enrollment in Health Insurance (HI) is considered as a sustainable way of financing for health and preparedness for catastrophic health care cost during receiving health services. Various socio-demographic factors are still unanswered regarding their influence. A study aiming to assess the factors associated with the enrollment of HI was conducted in 2018 in two districts of Nepal namely Kailali and Baglung. The study was cross-sectional covering 810 (405 enrolled and 405 not-enrolled) randomly selected households (HH). Socio-demographic variables were considered as independent variables and enrollment in HI as dependent variable. An interview schedule was used as a tool for data collection. Univariate, bivariate and multivariate analyses were performed to analyze the data. The data show that various socio-demographic characteristics are associated with the enrollment of HI. A significant statistical difference is seen between enrollment to HI and HH headship, age group of respondents, ability to feed the family, presence of chronic diseases in family, knowledge on HI, willingness to pay (WTP) for HI, having HI guidelines or books, participation in HI related training, interactions with neighbours, access to communication media: the radio/FM and TV, hoarding boards (HB), newspapers, posters/pamphlets/brochures; and access to health facilities. The results further show that female heads appear more likely to enroll (aOR = 1.47) in HI than the male. HH headship of the respondents also seem more likely to enroll. Higher age respondents are less likely to enroll. Interestingly, literate respondents and joint families are less likely to enroll than illiterate and nuclear families respectively. However, respondents having knowledge in HI seem more likely to enroll (aOR = 28.97, p<0.001) than those who are unaware about HI. Those with higher WTP for HI are more likely to enroll (aOR = >1.673, p<0.05) than low WTP. Respondents having guidelines or books, interactions with neighbours or relatives, exposure to the radio/FM, TV and HB seem significantly more likely to enroll. Respondents who feel susceptible to diseases are more likely to enroll (aOR = 1.484, p<0.05) compared to those who do not. Knowledge on HI, WTP, having HI related books or guidelines, exposure to the media (the radio/FM, TV, HB), interactions with neighbours appear to be the positive predictors for enrollment. Appropriate interventions should be implemented considering the factors for increased participation in HI.


Introduction
Though the Constitution of Nepal, 2015 (Article 35) has offered the basic health service as a fundamental right of the citizen (Nepal Law Commission, n. d.), the Government of Nepal (GoN) had allocated less than four (3.86) percent of the total budget (Department of Health Services, 2018) in the fiscal year 2016/17 for the health sector which accounts for less than two percent of gross domestic products.The investment seems insufficient for public health service delivery as per the constitutional provision and international and national commitments.Therefore, an alternative approach was searched for health financing.Besides these, the GoN declared commitment to meet the targets of Sustainable Development Goals (SDG) including Universal Health Coverage (UHC), health insurance (HI) is one and the strategies to meet the health related targets of SDG and the sentiment of UHC.Following these provisions, the GoN has implemented health insurance programme (HIP) as Social Health Security at the initial phase in three districts namely Kailali, Baglung and Illam since 2016.Now the programme has been expanded in other districts too (Health Insurance Board, 2017).
Many low and middle income countries' health care expenditure is mostly covered by out-of-pocket (OOP) while receiving health services.The high OOP leads to increasing financial risks and inequity in access to quality health care (Adebayo, 2014).It is claimed that a high level of OOP is one of the causes of poverty.Therefore, strengthening the financial security is one of the main strategies to reduce poverty.Available data shows that 60.4 percent of current health expenditure was paid by OOP (World Health Organization, 2016) which is a barrier of UHC.It means that if the people or patients have money during illness, they can afford or receive treatment.An appropriate mechanism is needed for pre-financing approaches to reduce the uncertainty of health illness or catastrophic cost for health (Adebayo et al., 2015;Panda et al., 2016).Since the health status of the citizen is one of the major indicators of the human development index, community based health is also considered as a means of poverty reduction (Tesfay, 2014).The GoN started HIP to ensure the access to quality health services without financial hardship by means of the HI (Health Insurance Board, n.d.) and to reduce the gap of health service utilization by the poor and the rich.HI is considered as social protection that aims to reduce poverty and vulnerability as well (Koehler & Hoffmann, 2014).World Health Organization (WHO) claims that two percent poverty increases every year due to high out-of-pocket expenditure while receiving the health services especially from the private sectors.It is also assumed that HI can make the health services accessible to all (Koehler & Hoffmann, 2014).HI is regarded as 'a ticket' for good health care (Abdel-Ghany & Wang, 2001).It is claimed that patients who do not have HI, receive fewer services and less care have lower chance of experiences of positive clinical outcomes (Fowler et al., 2010).It is anticipated that healthy citizens can contribute to economic and social development of the country significantly.However, the HIP may be suffered to run smoothly due to inadequate homework to address the factors that are associated with the enrollment as experienced from United Mission to Nepal in 1976 and B P Koirala Institute of Health Sciences in 2000 respectively (KOICA-Nepal Health Insurance Support Project, 2014).It is believed that HI can enhance the access to health services for those who live in low and middle income countries (Adebayo et al., 2015).
In Nepal, the enrollment rate in HI seems very low in many districts though the causes of poor enrollment is still unanswered (Health Insurance Board, 2017).Even in the United States of America, one third of the Americans aged under 65 years did not have HI in 2010 (Fowler et al., 2010).Only the lunching of the programme is not enough but participation of the people is equally important.Therefore it is crucial to identify the factors that influence in the enrollment.Many studies show that socio-demographic characters are the major predictors for enrollment in HI (Adebayo et al., 2015;Panda et al., 2016).The study aimed to assess the factors that are linked to the enrollment of health insurance in selected districts of Nepal.

Methods
A descriptive study was conducted in Baglung and Kailali Districts of Nepal.The Government of Nepal (GoN) initiated the Health Insurance (HI) programme as Social Health Security (SHS) in Kailali, Baglung and Illam Districts in the initial phase in 2016.The study chose Baglung from the Hill and Kailali from the Terai.All the enrolled households (HH) were the population of the study for enrolled HH sample.The data were collected by using an interview schedule at respondents' home or where they were available and convenient to them.Generally, responses were collected from household heads.In his/her absence or him/her refusing to respond, another senior member of the HH was requested to respond.Only the HHs who had enrolled before January 15 th , 2018 were included in the sample.Enrolled or participated by insurance companies other than Health Insurance Board (HIB) were excluded in the study.Various socio-demographic characteristics, access to information sources, and knowledge about health and HI were considered as independent variable and enrollment in health insurance (EHI) as a dependent variable.The sample size was calculated by using Daniel's formula (Naing, Winn, & Rusli, 2006): where n = sample size, z = level of confidence,  = expected prevalence (assuming 50/50 probability or 50%) (50%,  = 0.5), and d = accepted margin of error (5%, d = 0.05).
So the sample size = 384.16.
By adjusting the non-response rate of five percent experienced in latest survey (Ministry of Health; New ERA; and ICF, 2017) which accounts for 405 for enrolled HH and same size of sample was allocated for non-enrolled HH.Altogether a total of 810 samples were selected for the study.The sample was distributed as per the population proportion to size following the latest national census (Central Bureau of Statistics, 2014).As a result, 566 households (283 enrolled and 283 non-enrolled HH) for Kailali and 244 (122 enrolled and 122 non-enrolled HH) for Baglung.The list/number of the enrolled HHs was obtained from HIB district offices respectively from Kailali and Baglung and the sample was selected by using random number generator; a software available at Google for enrolled HH.For the non-enrolled sample, the nearest HH of the selected enrolled HH was selected.The interview schedule was used to collect the data.Data were collected for three months since March, 2018 in Kailali and Baglung.An ethical approval was taken from Nepal Health Research Council and permission was obtained from the concerned districts and local authorities as required.Consent was taken from respondents before interviewing.Data was entered into the SPSS version 20 software and the sample was cross checked for consistency.Some attributes of variables were merged due to the poor responses.Univariate, bivariate then multivariate analyses were performed for statistical output.

Socio-demographic Background of Respondents/Households
Of the total respondents, more than half (51%) were females and two third (66%) were household heads.Nearly 60 percent respondents were from 21 to 40 age group while the mean age of the respondents was 37 years.More than 92 percent of the respondents were literate.Forty-one percent of the respondents were from the nuclear family.The average size of the family was 5.6 whereas 56 percent of the HHs had upto five members.Just over the half (51.2%) of the HHs had enough food throughout the year.More than one third (34.6%) of the respondents expressed that presence of some kinds of diseases in their families, of them one third had heart related problems.Of the respondents, 28 percent were unnoticed by the HI.More than 42 percent of respondents said that they had felt financial trouble due to health problems.An average of willingness to pay for HI was 1429 Nepalese Rupees whereas 74 percent of the participants expressed that willingness to pay for HI was almost three times of the current contribution amount NRs.500 mentioned by HIB.Only 17 percent of the respondents had read HI related guidelines or books or information flyers and five percent of the participants participated in HI related training or meetings.Nearly one third (32%) of the respondents had discussed about HI with their neighbours or relatives whereas just 19 percent of respondents had gained knowledge about HI from the social media.Nearly 48 percent of the respondents had listened to HI related information from the radio/FM and 38 percent from television.Nearly 27% of the respondents had seen HI related messages from the hoarding boards but just 13% of the respondents read HI related information from newspapers.Similarly, 18% of the respondents had seen HI related brochures, leaflets, pamphlets or posters.Less than half (48%) of the respondents felt that they were susceptible to health problems.More than 72% of the HHs had access to health institutions within half an hour and the mean time to visit health facilities was nearly the same -30 (SD 22) minutes.Nearly one third of the respondents informed that their family members were aboard during data collection.

Socio-demographic Characteristics of the Respondents/Households and Enrollment in Health Insurance
The data shows some interesting results that more than 53 percent of the male respondents were enrolled in HI compared to 47 percent of the females.The male household head had higher enrollment (53%) compared to female HHH (45%) (p<0.05).The enrollment was higher in age group, 41 to 60 years which accounts for 59 percent (p<0.001).Remarkably, 55 percent of the respondents who could not read and write were enrolled in HI.
More than half (51%) of the HHs from joint families were enrolled compared to 49 percent of nuclear families.Data shows that the higher the family size the lower the enrollment rate.Only 39 percent of the HHs having more than 10 members were enrolled in HI.In the case of wealth, the percent of enrollment in HI were 58, 52, 51, 46 and 43 percent respectively from the richest, rich, poorest, poor and middle income.Nearly half (49.9%) of the HHs having enough food were enrolled and 64 percent who had enough food for just three to six months were enrolled in HI (p<0.01).Nearly 60 percent of the HHs having ailments were enrolled in HI (p<0.001).More than two third (68%) of the respondents having knowledge about HI were enrolled and less than half (42%) of the respondents faced financial problems during health care and 52 percent of them were enrolled.Nearly six out of ten (59%) of the respondents who were enrolled were found willing to pay for HI that was more than 1500 Nepalese Rupees (p<0.001).More than 83 percent of the respondents who had HI related books or guidelines were enrolled (p<0.001) to the HI.Nearly three fourths (73%) of the respondents who participated in training or workshops were enrolled (p<0.01).Similarly, 73 percent of the respondents enrolled in HI expressed that they discussed with neighbour and relatives about HI related matters (p<0.001).More than half (56%) of the respondents who knew from social media were enrolled.Consequently, 62 and 63 percent of the respondents were enrolled in HI who listened HI related information from radio/FM and watched Television respectively (p<0.001).More than two thirds (68%) of the respondents who noticed the message from the hoarding board were enrolled in HI (p<0.001).Likewise, 64 percent of the respondents were enrolled who read HI related information from newspapers (p<0.01).Another similar result showed that 68 percent of the respondents were enrolled who read or seen HI related leaflet, brochure, posters and pamphlet (p<0.001).Of the respondents, about 59 percent were enrolled in HI who felt that they were susceptible to health problems (p<0.001).More than half (51%) of the enrolled respondents expressed that they had access to health facilities within 30 minutes.The data show that HHs having family members aboard had equal chance to enroll in HI or not.

Multivariate Logistic Regression on the Factors Associated with the Enrollment of Health Insurance
Findings form logistic regression indicate that female respondents were more likely to enroll (aOR = 1.047) in HI than males.Similarly, the respondents who were household head were less likely to enroll in HI (aOR = 0.934).Nearly the same result was seen in age groups, the respondents age 21 to 40 years were less likely (aOR = 0.177, p<0.001) to enroll while those age 41 to 60 years (aOR = 0.282, p<0.05) and more than 60 years (aOR = 0.324) compared to age less than 20 years.Interestingly, literate respondents were also less likely to enroll in HI compared to the illiterate.The respondents, who could simply read and write, basic education, secondary education and higher education were less likely to enroll in HI (aOR = 0.224, 0.233, 0.227, and 0.179 p<0.001) respectively.Joint families also seemed less likely (aOR = 0. 921) to enroll in HI than nuclear families.HHs having more than 10 members were more likely (aOR = 1.194) to enroll compared to small family having less than five members.Moreover, poor, middle income and the richest HHs were less likely to enroll than the poorest HHs.HHs having food enough for 9 to 12 months and three to 6 months were more likely to enroll compared to HHs that had enough food throughout the year.However, the respondents having knowledge about HI were more likely to enroll in HI (aOR = 28.970,p<0.001) compared to those who were unaware about HI.
Families with financial crisis due to health problems were less likely (aOR = 0.812) to enroll in HI compared to those who had not.Willingness to pay seemed a positive predictor for enrollment in HI.Respondents who wanted to pay NRs. 501 to 1500 and more than 1500 were more likely to enroll in HI (aOR = 1.673 & 1.793, p<0.05) than those who wanted to pay equal or less than 500 per year per person.Respondents having HI related books or guidelines at home were more likely (aOR = 4.379, p<0.001) to enroll compared to those who had not whereas, respondents' who had participated in training and workshop were less (aOR = 0.503) likely to enroll in HI than those who did not participate.The respondents who had some discussion with neighbours were more likely to enroll (aOR = 1.851, p<0.01) compared to those who did not discuss with neighbours about HI but exposure to social media about HI was not a strong driver to enroll (aOR = 0.539, p<0.05) in HI.HHs that received HI related information from radio/FM and television seemed more likely to enroll (aOR = 1.115 & 1,346).Similarly, respondents who got HI related messages from hoarding boards were more likely to enroll (aOR = 1.492) than those who did not have exposure.
Likewise, respondents who got information form brochure, leaflet, poster and pamphlet seemed to have lower chances to enroll in HI.Respondents who felt susceptible to diseases were more likely to enroll (aOR = 1.484, p<0.05) compared to those who did not.HHs that had access to health facilities within 30 to 60 minutes were more likely to enroll (aOR = 1.076) than those who had access within half an hour.HHs having family member aboard were less likely to enroll in HI (aOR = 0.751) compared to HHs that did not have.

Discussion
The data show that different factors were associated with the enrollment in HI.Some factors that were significantly associated include household headship, age group of respondents, ability to feed the family, presence of chronic diseases, knowledge on HI, experiences of financial trouble due to diseases, willingness to pay, exposure to HI related books or guidelines, participation in HI related training and workshops, interaction with neighbours, information received from the radio/FM, TV, hoarding boards, newspapers, poster or pamphlet, and feeling susceptible to health problems.Results from different studies show that educational level, age as well as self-employment are positively associated with the enrollment of HI (Abdel-Ghany & Wang, 2001).
Some studies have found that the enrollment in HI seemed high in some ethnic groups, others claim that age group and education level of respondents are the predictors for enrollment.Household size and enrollment in HI have positive association but other studies did not support it.A systematic review shows that educational level (high), sex (male), age (younger), and HH size (larger) have positive association with willingness to pay and enrollment in HI (Adebayo et al., 2015).The review also claims that lack of funds, lack of trust and poor quality of health care are major causes of low enrollment in HI.Another study from Ethiopia shows that presence of diseases in HH, income, educational status and first point of treatment were some influencing factors for enrollment in HI and utilization of services as well (Tilahun, Atnafu, Asrade, Minyihun & Alemu, 2018).However, another study indicates that presence of HI is associated with the prompt utilization of health services (Skinner, Foster, Mitchell, & Haynes, 2014).Illiteracy or low level of education, poor social support and homelessness are considered as determinants for non-enrollment in HI (Fowler et al., 2010).Interestingly, nearly the same results are shown in a systematic review that income of HH, educational level of HHH, female headed HH, age of HHH, size of HH and presence of chronic diseases in a family member were positively associated with the enrollment of HI (Panda et al., 2016).In this study, HHs having high wealth status seemed less likely to enroll in HI.However, a study in Nigeria shows a contrasting result that HHs from lowest wealth quintiles had higher risk of catastrophic health expenditure (Ilesanmi, Adebiyi, & Fatiregun, 2014).

Conclusion
Different socio-demographic characteristics appear as the influencing factors for the enrollment in HI.Age group of the respondents, knowledge regarding HI, willingness to pay for HI, interaction with neighbours, feeling of susceptibility to diseases or health problems, and HI related information from the radio/FM, television and hoarding boards were the major predictors for enrollment in HI.Since the interactions with neighbours had a positive association with the enrollment in HI, appropriate interaction should be made with neighbours and the public.Besides these, the influencing factors should be considered at the time of planning of the intervention.

Table 1 .
Background Characteristics of Households and Respondents

Table 2 .
Background Characteristics of Households/Respondents and the Enrollment in Health Insurance

Table 3 .
Adjusted Odds Ratio from Logistic Regression Model at 95 % Confidence Interval by Background Characteristics of Households/Respondents and Enrollment in Health Insurance