Factors Influencing Nonrenewal of Health Insurance Membership in Ejisu-Juaben Municipality of Ashanti Region, Ghana

Background. In the quest to prevent households from making catastrophic expenditures at the point of seeking healthcare, the government of Ghana introduced the National Health Insurance in 2003. However, people are reluctant to renew their membership. /is study was, therefore, conducted to identify factors influencing the nonrenewal of National Health Insurance membership in the Ejisu-Juaben Municipality.Methods. A cross-sectional study was conducted among 427 respondents in the Ejisu-Juaben Municipality to ascertain factors influencing the nonrenewal of health insurance membership status. Data were entered and analyzed using Stata version 14. Univariate and multivariate analyses were performed to determine sociodemographic factors, household factors, and systemic factors influencing the nonrenewal of health insurance status. Statistical significance for all testing was set as p≤ 0.05. Results. Sociodemographic factors such as gender (AOR � 0.531; CI � 0.287–0.985) and educational level (AOR � 5.268; CI � 1.130–24.551)) were associated with the nonrenewal of health insurance membership. Income levels in Ghana Cedis were 500–1000 (AOR � 0.216; CI � 0.075–0.617) and 1001–2000 (AOR� 0.085; CI � 0.019–0.383). Systemic decision on factors such as clients’ satisfaction (AOR� 0.149; CI � 0.035–0.640), making copayment (AOR� 0.152; CI � 0,068–0.344), acquiring all prescribed drugs (AOR � 4.191; CI � 2.027–8.668), and awareness of mobile renewal (AOR � 3.139; CI � 1.462–6.739) was associated with nonrenewal of membership. Conclusions. /e nonrenewal of health insurance membership was influenced by sociodemographic, household, and systemic factors. /erefore, the Municipal Health Directorate and the National Health Insurance Authority have to work on these factors to reach the target of 100% active coverage in the municipality.


Background
Health insurance coverage is an imperative aspect of achieving Universal Health Coverage (UHC) [1,2]. In Canada, every citizen and legal residents of the country have basic health insurance coverage which is provided by the government [3]. e public programme, therefore, account for only 37.1% of residents. is shows that public-funded health insurance as practiced by Canada is the way to go to ensure equity and financial protection.
Learning from the developed countries, some low-and middle-income countries have begun introducing health insurance to their populations. A study by Escobar et al (2010) in Costa Rica, Peru, Indonesia, Rural China, Columbia as well as Ghana on the impact of health insurance in low-and middle-income countries ascertained a positive correlation between health insurance status and good health outcomes [4,5].
In 2012, the Ghana Parliament made the membership of National Health Insurance Scheme (NHIS) mandatory for all residents in the country through an Act of Parliament [6]. However, out of the 18.5 million people who are registered into the scheme, only 10 million (40%) are active [7,8]. is, if not addressed, could hamper the country's effort to attain 100% health insurance coverage for the Ghanaian population by the year 2030.
is calls for pragmatic measures in retaining the existing members in addition to enrolling the noninsured in order not to deter Ghana's effort to achieve UHC. Some studies have been conducted in the Eastern, Central, and Volta Regions of Ghana to ascertain factors that influence enrolment and retention into the health insurance scheme [9,10]. However, Ashanti Region, the second largest and the most populated region in Ghana, had not benefited from such studies. In addition, for those regions, which have benefited from such studies, the investigations were facilitybased, scheme-based, and based on individual factors influencing the nonrenewability of NHIS. How sociodemographic factors, household factors, and health system factors influenced the nonrenewability of NHIS membership had not been adequately explored in Ghana. erefore, there was a need for such a study to be conducted in the Ejisu-Juaben Municipality as a lesson from the region could be used in other regions in the country and even in other countries in the subregion. is will help put measures in place to propel the country toward achieving its aim of 100% health insurance coverage.

Design.
A cross-sectional study was conducted from 1 June 2019 to 31 December 2019 among 427 people in Ejisu-Juaben Municipality who have ever registered with NHIS to find out the factors that influence the nonrenewal of health insurance membership.

Sampling Method.
Multistage sampling, involving cluster and simple random sampling methods, was used in this study. Ejisu-Juaben Municipality has five submunicipal health administrative areas. Each health administration was taken as a cluster. One community was selected from each cluster using a simple random sampling method. Here, the names of the communities in each submunicipality were written on pieces of a paper and put into a box and thoroughly shaken. One community was selected in each submunicipality resulting in a total of five communities sampled for the study. In each of these selected communities, 85 household heads were interviewed except in Ejisu, the capital town where 87 were interviewed.
In reaching each of the selected communities, the researcher stood at the center and span a pen. Houses the tip of the pen pointed at were chosen and the household heads were engaged once they were qualified to take part in the study. On leaving each house, the opposite houses were selected and household heads who met the inclusion criteria were interviewed. In a house where there was no opposite house, the researcher span the pen again and repeated the same procedure to engage household heads. ese processes were replicated across all selected communities till the required sample size for each community was attained.

Sample Size.
Since the number of households in the Ejisu-Juaben Municipality from the 2010 census is 33,078, which is a relatively large population, the sample size was calculated using Cochran's formula: where No is the sample size, p is the assumed prevalence of renewal of NHIS membership in the Ejisu-Juaben district, q is (1-p), Z is the significance level at 95% confidence interval read from the z table to be 1.96, and e is the margin of error, which is assumed to be 5%. For this sample calculation, a 50% prevalence rate of renewal was assumed. erefore, our sample size was calculated as Adjusting for a nonresponse rate of 10%, the final sample size was final sample size � effective sample size (1 − nonresponse rate anticipated) , � 427 households heads. (3)

Data Collection Techniques and Tools.
Data were collected using a semistructured questionnaire designed on the Kobo Collect application. Kobo Collect is Kobo Toolbox's data collection application that is used to collect survey data and can be used on any android phone. e questionnaires were administered by four trained research assistants who communicated in English and Twi (the local language of study participants). Data on sociodemographic characteristics, such as age, educational level, ethnicity, marital status, occupation, religion, place of residence, and income, were collected. Data collected on the household level comprised of household income, the health status of a household, household size, the number of dependents in a household, the presence of a person with chronic disease in the household, and the wealth status of a household. Systematic factors assessed include distance to a health facility, road network, quality of care, copayments, and availability of medicines. Data validation was done to ensure data quality. Descriptive data were presented using frequencies and percentages. Univariate and multivariate analyses were performed to establish associations between sociodemographic factors, household factors, and systematic factors influencing the nonrenewal of NHIS membership. Statistical significance for all testing was set as p ≤ 0.05.

Background Characteristics of Respondents.
e mean age of the respondents was 36.3 ± 0.51 years, and a little over half (57.79%) were females. ose who completed Junior High School were 30.48% and traders constituted 26.14%. Vast majority (82.25%) of the respondents were Akan and Christians constituted 89.93% while 70.26% lived in urban areas as shown in Table 1. Table 2, sociodemographic factors such as gender and level of education influenced a person's decisions not to renew their health insurance membership. Females were 0.531 less likely not to renew their health insurance membership compared to males (AOR � 0.531; 95% CI � 0.287-0.985). In addition, people with MSLC, as their highest level of education, were five times more likely not to renew their NHIS membership as compared to those with other levels of education (AOR � 5.268; 95% CI � 1.130-24.551).

Logistics Regression of Household Factors at Influence the Nonrenewal of Health Insurance Membership.
In univariate and multivariate analyses, as shown in Table 3, household factors such as household income and household expenditure influenced a person's decision not to renew their health insurance membership. For instance, those with an average household income of GH¢500-1000 were 0.216 times less likely not to renew compared to those who earn less than GH¢500 (AOR � 0.216; 95% CI � 0.075-0.675) and those earning GH¢1001-2000 were 0.085 times less likely not to renew compared to those who earn less than GH¢500 (AOR � 0.085; 95% CI � 0.019-0.389).

Systematic Factors Influencing the Nonrenewal of Health
Insurance Membership. Table 4 shows that the majority of the respondents (60.67%) lived 1-5 km away from a health facility, 83.45% had good road network to the health facility, 85.85% had a good experience on their last hospital visit, and 65.47% made some form of copayment though they were insured. e majority of the respondents (65.47%) were aware of the use of mobile phones to renew their national health insurance cards, but 75.06% of those aware did not know the code used to renew their membership.
A little over half of the respondents (50.84%) lived 1-5 km from the health insurance office, with 6.24% living less than 1 km, 37.65% lived 6-10 km away from the health insurance office, and 5.28% lived more than 10 km away from the health insurance office.
Only 49.40% of the respondents received all the drugs prescribed while 50.60% not received all the prescribed drugs.

Logistic Regression Analysis of Systemic Factors Influencing Health Insurance Renewal.
e univariate and multivariate logistic regression analyses in Table 5 show that the

Discussion
e results of the study show gender and level of education as sociodemographic factors that influence the renewal of health insurance membership. Females were more likely to renew their insurance compared to their male counterparts.
is could be attributed to the fact that females use healthcare services frequently due to reproductive health issues. Another reason could be assigned to the free maternal health policy under the national health insurance. Under this policy, pregnant women are registered with the health insurance without paying for registration and premium; hence, women appear to be renewing their health insurance membership more than men. is finding is consistent with findings from previous studies which reported that females were more likely to have active health insurance compared to males [11,12].
In addition, in this study, there was no association between age and renewal of health insurance membership. is is inconsistent with previous studies that report that those who were older were more likely to renew their health  [13,14]. On the other hand, educational level showed a significant relationship between renewing health insurance status and educational attainment. is is in consistence with previous studies which suggest that those with some form of formal education were more likely to renew their health insurance membership [11]. Our study, however, showed that those with MSLC were more likely not to renew their insurance status contrary to a finding by a previous study which reports that there was no significant relationship between education and individual's decision to renew insurance membership [9]. e study showed no significant relationship between ethnicity and health insurance status contrary to a previous study [11]. Marital status also did not show any statistically significant association with the health insurance status. is is similar to a finding by a previous study [15] but inconsistent with other studies [9,11] which suggest that females who are not married are less likely to have active health insurance compared to the married. is study also did not find any statistically significant relationship between employment and the renewal of health insurance membership.
is is similar to a finding reported in a previous study [9]. However, findings in our study were not in line with a study which reported that there is an association between employment and the renewal of health insurance membership [12].
Religion did not have an association with the renewal of health insurance membership.
is is inconsistent with other studies that show that Christians are more likely to renew their insurance compared to Muslims and people from other religions [11]. e results of the study did not show statistically significant relationship between the place of residence and the health insurance status.
is is inconsistent with a study conducted in Kenya [16] as well as in Ghana [11].
Household income was found to influence a person's decision to renew or not to renew their health insurance.
ose who earned more were more likely to renew their health insurance compared to those who earned less. is result affirms a previous study conducted in Ethiopia which found that high-income earners are more likely to renew their health insurance than low-income earners [17]. A similar finding is reported in a study carried out in Ghana [11,14]. However, a study carried out in the Dormaa Municipality in Ghana reported findings contrary to those in our study [18]. e study shows a statistically significant relationship between household expenditure and renewal of health insurance status. e multivariate analysis indicates that those who spend more have higher odds of not renewing their health insurance. is could be attributed to the fact that they spend their money on other things which they consider to be more important to them than renewing their health insurance. is is contrary to a previous study which reported that household with higher expenditure have a higher likelihood to renew their health insurance [12].
Health system factors found to influence health insurance nonrenewal were the quality of care at the health facility, making copayments, acquiring all drugs prescribed, closeness to the insurance office, and awareness of mobile renewal. e distance to a health facility did not show any statistical significance. is is contrary to previous studies which showed that those who lived far away from a health facility were less likely to subscribe to the health insurance. If the health facility is far from the individuals, it might cause the person to seek other sources of healthcare, thereby deterring the person from renewing the health insurance [17,19]. e quality of care was found to be statistically significant with the health insurance status. e adjusted odds ratio shows that those who felt they had good and very good quality of care were more likely to renew their health insurance status compared to those who had poor quality of care. e quality of care has been identified as a major factor that affects health insurance use [10,19,20].
In this study, it was found that availability of the prescribed drugs influenced people's decision on the renewal of the health insurance status. ose who did not receive all the Advances in Public Health prescribed drugs were 4 times more likely not to renew their health insurance compared to those who received. is affirms a previous study [9].

Conclusion
e health insurance nonrenewal rate was low and was influenced by sociodemographic, household, and systemic factors. erefore, the Municipal Health Directorate and the National Health Insurance Authority need to work on these factors to reach the target of 100% active coverage in the municipality.

Data Availability
All data generated or analyzed during this study are included within the article.

Ethical Approval
Ethical clearance was sought from the Kwame Nkrumah University of Science and Technology Ethics Review Committee on Human Research Publication and Ethics.

Consent
Permission was sought from the Ejisu-Juaben Health Directorate before data collection began. At the community level, participants were asked to read and sign the consent form prior to the interviews. Participants were assured of the confidentiality of their response and their null association with it at present or in the future. In addition, they were assured that their participation would not affect their relations with the NHIS at present or in the future and that refusal to participate would not attract any penalty.

Conflicts of Interest
e authors declare that they have no conflicts of interest.

Authors' Contributions
KMA and KAM conceived and initiated the study. KMA designed the study, provided training and guidance for data collection and analysis, and further prepared the original draft of the manuscript. PA-B and DO assisted with supervision, editing, and technical support in all areas of the Advances in Public Health 7 study. PT and JKB assisted with the data collection and monitored the entire team for data collection. e draft manuscript was critically reviewed by PT. All authors made significant contributions to the final work and approved the manuscript for submission. All authors have read and approved the manuscript.