Nsiah‑Boateng et al. BMC Health Services Research BMC Health Services Research (2023) 23:239 https://doi.org/10.1186/s12913‑023‑09236‑7 RESEARCH Open Access Effect of mobile phone‑based health insurance contribution payment system on retention of coverage in the National Health Insurance Scheme in Ghana: an evaluation study Eric Nsiah‑Boateng1,2,3*, Mariam Musah1, Collins Danso Akuamoah1, Francis Asenso‑Boadi1, Francis‑Xavier Andoh‑Adjei1 and Bernard Okoe Boye1 Abstract Background Ghana introduced a mobile phone‑based contribution payment system in its national health insurance scheme (NHIS) in December 2018 to improve the process of enrolment. We evaluated the effect of this digital health intervention on retention of coverage in the Scheme, one year after its implementation. Methods We used NHIS enrolment data for the period, 1 December 2018–31 December 2019. Descriptive statistics and propensity‑score matching method were performed to examine a sample of 57,993 members’ data. Results Proportion of members who renewed their membership in the NHIS via the mobile phone‑based contribu‑ tion payment system increased from 0% to 8.5% whilst those who did so through the office‑based system only grew from 4.7% to 6.4% over the study period. The chance of renewing membership was higher by 17.4 percentage points for users of the mobile phone‑based contribution payment system, compared to those who used the office‑based contribution payment system. The effect was greater for the informal sector workers, males and the unmarried. Conclusions The mobile phone‑based health insurance renewal system is improving coverage in the NHIS particu‑ larly for members who hitherto were less likely to renew their membership. Policy makers need to devise an innova‑ tive way for new members and all member categories to enrol using this payment system to accelerate progress towards attainment of universal health coverage. Further study needs to be conducted using mixed‑method design with inclusion of more variables. Keywords Impact evaluation, Digital health, Health insurance, Retention of coverage, Ghana *Correspondence: Eric Nsiah‑Boateng ensiah‑boateng@st.ug.edu.gh Full list of author information is available at the end of the article © The Author(s) 2023. 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:// creat ivecom mons. org/l icens es/b y/4.0 /. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons.o rg/ public domai n/z ero/1.0 /) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Nsiah‑Boateng et al. BMC Health Services Research (2023) 23:239 Page 2 of 9 Background investigations and medications on the NHIS medicines Many Low- and Middle-Income Countries (LMICs) are list. The benefit package reportedly covers about 95% of implementing digital health innovations in their health disease conditions in Ghana [10]. There are over 4000 systems towards achieving universal health coverage public and private healthcare facilities credentialed to (UHC) [1–3], a fundamental principle of Sustainable provide services for members of the Scheme. Development Goal (SDG) 3.8. Ghana started its UHC The NHIS has improved population coverage and journey in 2003 with the establishment of a National access to healthcare services utilization since its intro- Health Insurance Scheme (NHIS). Prior to this, health- duction almost two decades ago [7, 8]. However, over care was financed strictly from out-of-pocket (OOP) the 2016–2018 period, the scheme recorded consistent payment at the point of service use following a transi- decline in population coverage after reaching a peak of tion from free healthcare at independence in 1957 to 11.3 million members (40%) in 2015. For instance, 10.8 the introduction of user fees in the 1980s [4]. The OOP million people, representing 36% of the population were payment impoverished an already poor population and active members of the scheme in 2018 [11]. This was due consequently excluded a large majority of the people to the inconvenience associated with the registration from accessing care [5, 6]. It also reduced utilization of process at the district offices of the Scheme. Many peo- healthcare services and led to deteriorated health out- ple experienced long queues and hours of waiting time comes. There was no protection for the poor and health- to enrol in scheme. Distance to a district office, and time care financing was highly inequitable. The introduction and travel cost were also found as barriers to enrolling of the NHIS; therefore, removed this financial barrier and and staying enrolled in the NHIS [12–14]. In response, allowed people to access care without paying at the point Management of the scheme implemented a mobile of use [7, 8]. phone-based contribution payment system in December The NHIS provides annual healthcare cover for 2018 to provide an easy, convenient, and affordable way its members at a defined contribution of GHS22.00 for members to renew their membership annually with- (US$3.78)1 and/or processing fee [9] of GHS8.00 out being physically present at an NHIS branch office. (US$1.38) for new members and GHS5.00 (US$0.86) Our review of the literature showed paucity of stud- for existing members. Membership is mandatory by law ies on evaluation of digital health intervention particu- but voluntary in practice due to lack of enforcement. larly in the health insurance industry, indicating a gap in Members are broadly categorized into exempt and non- knowledge in this area of study. A prospective study on exempt group. The exempt group pays no contribution to the use of the NHIS mobile phone renewal system con- the scheme, and they include: 1) persons below the age ducted in one geographic region of Ghana focused on of 18 years: 2) indigent; 3) Social Security and National determinants of renewing membership in the scheme Insurance Trust (SSNIT) pensioners; 4) persons aged through the mobile phone payment system [15]. The 70 years or older; 5) pregnant women; and 6) beneficiar- study found that factors such as living in urban centres; ies of the Livelihood Empowerment Programme against higher level of education; informal sector employees; and Poverty (LEAP). The non-exempt group, however, pays paying premium using the mobile phone were associated contribution directly to the scheme and comprises work- with membership renewal in the Scheme [15]. A cross- ers in the informal sector, including the self-employed. sectional survey in Kenya also found that mobile phone Membership categories are dynamic, in that a member use increased the probability of enrolling in the National may change from one category to another over a period. Health Insurance Fund [2]. However, an intervention The NHIS relies heavily on government funding study on the effect of paying health insurance subscrip- through an earmarked fund, the National Health Insur- tion using mobile money (M-Pesa) in Kenya found no ance Fund (NHIF). This constitutes about 74% of the significant effect on enrolment [3]. Given that Ghana’s total funding of the NHIS. The NHIF is funded through mobile phone-based contribution payment system is a a 2.5% value-added tax (VAT) on selected goods and ser- new digital health solution, and there is little knowledge vices. The remaining funding comes from informal sec- on its effect on enrolment, we sought to evaluate effect tor workers’ contributions; two- and one-half percentage of this intervention on retention of coverage for informed points of each person’s contribution to the SSNIT; return decision-making on its modification and scale-up. on investments; and sector budgetary support from gov- ernment [9]. The NHIS offers a comprehensive health- The NHIS mobile phone‑based contribution payment care cover for outpatient and inpatient services, including system The NHIS mobile phone-based contribution payment system, popularly referred to as “mobile renewal” is a 1 Exchange rate for October 21, 2020: US$1.00 = GHS5.81. two-part system operating two distinct processes: 1) a Nsiah‑Boateng et al. BMC Health Services Research (2023) 23:239 Page 3 of 9 mobile-based renewal process; and 2) a non-biomet- NHIS if they were not at the facility. This serves as a feed- ric authentication process. While the former is used by back mechanism which enables the NHIS to gather data members to renew their membership in the scheme, on attendances which may not be valid. the latter is used at healthcare facilities by providers to authenticate attendance of NHIS members and provide Methods feedback in the form of text message to the purchaser, the Study design and sampling National Health Insurance Authority (NHIA). The mobile The study is a cross-sectional evaluation study of NHIS phone-based contribution payment renewal process is mobile phone-based contribution payment system, based on a simple Unstructured Supplementary Service which was implemented in December 2018 to provide Data (USSD) application integrated onto the existing convenience for members of the scheme to renew their mobile money platform. It begins with an SMS reminder membership. We sourced enrolment data from NHIA, to members, notifying them of their membership expira- covering the period 1 December 2018–31 December tion dates. Two reminders are sent: 1) one month prior to 2019. We randomly sampled a total of 57, 993 out of 21 expiration; and 2) two weeks prior to expiration. million enrolment data of members who had renewed The renewal process starts by dialing a dedicated USSD their membership, either through the mobile phone- short code (*929#), available across the three major tel- based contribution payment system or the in-person ecommunication networks in the country: MTN, Voda- BMS system at the NHIS branch offices nationwide. fone and Airtel-Tigo. The first step requires members Characteristics of the sampled enrolment data were to input their membership number on their member- NHIS member number, type of registration (new or ship cards. Once this is done, the renewal system inter- renewal), and sex. Other characteristics included age, acts with the NHIS membership database to determine defined NHIS member categories such as informal sec- membership category and the corresponding contribu- tor workers, indigents, and SSNIT contributors, etc.; and tion to be paid. The contribution due depends on the platform for registration, which is either through the last renewal location (district). This is due to variations office-based biometric management system or the mobile in contributions across regions. A member then follows phone-based system. the prompts thereafter to complete the renewal process. Once renewal is successful, a confirmation message is Data analysis sent to the member. We performed descriptive analysis to explore socio- In addition to the renewal process, the interface also demographic characteristics and trends in membership allows a member to access information on the benefits renewal of the sampled population by type of enrolment package and the NHIS medicines list. The NHIA charges payment system (in-person biometric enrolment system a GHS1.00 (US$0.17) convenience fee for the use of the versus mobile phone-based renewal payment system). mobile phone-based contribution payment system. This We also employed multivariate logistic regression analy- is added to the contribution before payment is com- sis to estimate the odds of a member retaining coverage pleted. The mobile phone-based contribution payment in the scheme. system is available to all member categories except preg- Moreover, we applied treatment-effects estimation nant women and indigents (core poor). These two cate- method using the logic specification of the propensity gories still need to go to the district offices to renew their score matching to (PMS) evaluate effect of the inter- membership in the scheme. vention on membership renewal. The use of PSM tech- The second part of the mobile phone-based contri- nique for this study is based on the observational data bution payment system begins when a member visits employed and the fact that there was no baseline. The a health facility to access care. The provider also dials a PMS is useful for the selection of comparison/control dedicated short code (*842#) using a pre-registered tel- group ex post from reasonably large administrative data ephone number. Using the unique member number on through a matching procedure based on observed char- the membership card, the provider confirms validity of acteristics [16–18]. the card. Once validity is established, a numerical code, We estimated the probabilities of renewing member- dubbed the Claims Check Code (CCC) is sent to the pro- ship using the mobile phone-based payment system vider. This code is captured on the claims form to validate (propensity scores) for each member of the treatment a member’s attendance at the facility. Thus, claims sub- and control groups based on similarities in observed mitted without CCC are rejected. Again, at the time of characteristics, including sex, age, marital status, CCC generation, a Short Message Service (SMS) is sent and membership category. This technique assumes to the member as confirmation of attendance. The SMS no selection bias based on unobserved characteris- provides an opportunity for the member to contact the tics of the two groups [16, 17, 19]. We then matched Nsiah‑Boateng et al. BMC Health Services Research (2023) 23:239 Page 4 of 9 Fig. 1 Balance plot of propensity scores for control and treatment groups the control group to the treatment group based on Results the propensity scores, that is, the estimated probabili- Characteristics of the sampled population ties of being treated (or renewing membership in the Approximately, 80.2% of the members renewed their NHIS using the mobile phone-based payment system). membership in the NHIS overall (Table  1). Average age This analytical technique is recognized for its ability to of the members was 24.9 years (SD = 21.46); 59.1% were remove biases generated through differences in obser- females; 62.5% had never married (single); and 34% were vational characteristics and increase balance between informal sector workers (non-exempt group). Except treatment and control groups [19, 20]. It has been used for marital status, there were significant differences in in other studies in Ghana [21, 22] and elsewhere [23, socio-demographic characteristics between members 24] to evaluate effect of social protection programmes. who used the office-based payment system and those Members who renewed their membership in the who used the mobile phone-based payment system to scheme through the mobile phone-based payment sys- renew their membership in the scheme (not shown in tem were the treatment group, assigned a value of “1” Table  1). These differences, however, were addressed by and those who used the in-person biometric payment using the propensity score matching method to estimate system were the control group, assigned a value of “0”. and match probabilities of renewing membership in the The explanatory variables were sex, age, marital status, NHIS through the mobile phone-based payment system and membership category, broadly categorised into for each member of the two groups based on similarities premium contributors (non-exempt group) and non- in their observed characteristics. premium contributors (exempt group). The choice of these variables was driven by the NHIS enrolment data. Trends in coverage retention A balance test for the intervention and control groups The proportion of members who renewed their mem- was performed using these explanatory variables and bership over the study period through the mobile the results showed that the two groups were balanced phone-based contribution payment system increased (Fig. 1). consistently from 0% to 8.5% (Fig. 2). Those who renewed We also performed sensitivity analysis using other at the NHIS office, however, went up sharply between the treatment effects estimation methods such as inverse- first two months from 4.7% to 15.2% and declined there- probability weights (IPW), IPW regression adjust- after to 6.4%. ments, and nearest-neighbour matching (NNM) to test robustness of our findings (Supplementary Table  1). Factors associated with membership renewal in the NHIS Microsoft excel and STATA version 14 were used to The multivariate logistic regression model showed that analyse the data. the unmarried (OR = 1.75, 95% CI: 1.54–1.98) and older N siah‑Boateng et al. BMC Health Services Research (2023) 23:239 Page 5 of 9 Table 1 Characteristics of the control and treatment group before matching (n = 57,993) Variable Type of contribution payment system Total Office‑based payment system Mobile phone‑based payment system (Control) (Treatment) Registration type New 11,467 (39.7) 0b 11,467 (19.8) Renewal 17,427 (60.3) 29,099 (100.0) 46,526 (80.2) Sex Female 16,857 (58.3) 17,394 (59.8) 34,251 (59.1) Male 12,037 (41.7) 11,705 (40.2) 23,742 (40.9) Age (Mean, Std.) 23.5(20.3) 26.3 (22.5) 24.9 (21.5) Marital statusa Divorced 64 (0.5) 96 (0.6) 160 (0.6) Married 4,923 (37.9) 5,870 (37.2) 10,793 (37.5) Single 7,826 (60.3) 9,507 (60.2) 17,333 (60.2) Widowed 177 (1.4) 322 (2.0) 499 (1.7) Membership category Exempt group 20,542 (71.1) 17,710 (60.9) 38,252 (66.0) Non‑exempt group 8,352 (28.9) 11,389 (39.1) 19,741 (34.0) a Valid response is 28,785 b The mobile phone‑based contribution payment system was being used for renewal of membership only at time of the study Fig. 2 Trends in membership renewal by type of enrolment payment system, December 2018‑December 2019 adults (OR = 1.04, 95% CI: 1.03–1.04) were significantly Effects of mobile phone‑based contribution payment more likely to retain coverage in the scheme than the system on coverage retention married and young adults, respectively (Table 2). How- The propensity-score matching estimates showed sig- ever, the males (OR = 0.69, 95% CI: 0.64–0.75) were nificant effect of the mobile phone-based contribution significantly less likely to retain coverage in the scheme payment system on membership renewal in the NHIS compared to the females. Likewise, the non-exempt (Table  3). Overall, the chance of renewing membership groups (informal sector workers) (OR = 0.70, 95% CI: in the scheme was higher by 17.4 percentage points for 0.63–0.77) were less likely to retain coverage relative members who used the mobile phone-based application to those exempted from paying contribution to the system than those who used the conventional in-per- scheme. son biometric membership system (BMS) at the district Nsiah‑Boateng et al. BMC Health Services Research (2023) 23:239 Page 6 of 9 Table 2 Multivariate logistic regression estimates for retention of weights regression adjustments), however, had a higher coverage ATET than the nearest-neighbour matching method Variable OR Std. Err [95% C. I] (21.5 percentage points vs 17.4 percentage points) (Sup- plementary Table  1). Results of the subgroup analysis Female Ref also showed that the effect was more pronounced for the Male 0.69*** 0.028 0.64 0.75 informal sector workers, males and the unmarried. Age 1.04*** 0.002 1.03 1.04 Married Ref Discussion Unmarried 1.75*** 0.110 1.54 1.98 This study evaluated effect of the digital health interven- Exempt group (non‑contributors) Ref tion implemented in the NHIS to speed up progress of Non‑exempt group (contributors) 0.70*** 0.036 0.63 0.77 enrolment towards attainment of UHC. It also exam- _cons 3.05*** 0.236 2.63 3.55 = ined factors influencing members’ decision to renew Number of obs 28,785 = their memberships in the scheme. The findings reveal LR chi2(4) 845.23 that the mobile phone-based contribution payment sys- Prob > chi2 = 0.0000 tem increases retention of coverage in the NHIS overall. Log likelihood = ‑8995.1764 As expected, the informal sector workers, the males and Pseudo R2 = 0.0449 the unmarried who hitherto were less likely to enrol in OR Odds Ratio, C.I Confidence Interval the NHIS, are renewing their coverage in the scheme *** p < 0.001 through the newly implemented mobile phone-based contribution payment system more than those using the in-person biometric enrolment payment system at the Table 3 Average treatment effect (ATET) estimates for renewal of membership (n = 28,785) district offices. The possible explanation for the positive impact of the Variable Coef AI Robust [95% C.I] mobile phone-based contribution payment system is that Std. Err it removes barriers to membership renewal, including Overall 0.174*** 0.004 0.167 0.181 time and travel cost, as found in other studies [12, 13]. Male 0.209*** 0.006 0.197 0.221 Prior to the introduction of this digital health interven- Female 0.151*** 0.004 0.143 0.159 tion, individuals had to queue for long hours at the dis- Married 0.147*** 0.005 0.137 0.158 trict offices to either enrol as new members or renew Unmarried 0.189*** 0.005 0.180 0.198 their membership in the NHIS [12, 13], and others expe- Exempt group 0.162*** 0.005 0.153 0.171 rienced delays in the processing of their cards [14, 25]. Non‑exempt group 0.213*** 0.007 0.200 0.226 Consequently, these barriers discouraged them to partic- *** p < 0.001 ipate in the scheme [14]. Our findings corroborate stud- ATET Average treatment effect on the treated, C.I Confidence Interval ies in other development areas where the use of mobile phone and mobile phone-based payments showed posi- tive impact on farm productivity [26], poverty reduction offices. In addition, the probability of renewing coverage and economic growth [27]. Findings of this present study, in the scheme was higher by 21.3 percentage points for however, contradicts one in Kenya [3], where payment of the informal sector workers (premium payers) who used premium through mobile phone yielded no significant the mobile phone-based application system than those effect on enrolment in the National Health Insurance who used the BMS. Similarly, the probability of renewing Fund. The contradiction might be due to the randomized membership in the scheme was higher by 20.9 percent- experiment design employed in the Kenyan study. age points for the males who used the mobile phone- Although the mobile phone-based contribution pay- based contribution payment system than those who used ment system is improving enrolment of males and the BMS. informal sector workers in the NHIS, these groups are less likely to renew membership in the scheme overall, Sensitivity and robustness test compared to their respective counterparts. The feeling Results of the other treatment effects estimation meth- of perceived good health (rare illness) and poor qual- ods demonstrated consistency in the positive impact of ity of services at healthcare provider sites, as found in the mobile phone-based payment system on member- the other studies [14, 25, 28, 29], could be the underly- ship renewal in the NHIS. Two estimation methods ing reasons for this revelation. Our findings are consist- (inverse-probability weights and inverse-probability ent with an earlier study on the NHIS [29] but contradict Nsiah‑Boateng et al. BMC Health Services Research (2023) 23:239 Page 7 of 9 another study that was conducted in one district of the logistic model specification used, provides insight into Greater Accra region [30], where the males and informal associated factors of retention of coverage in the NHIS. sector workers were more likely to retain coverage in the Moreover, the NHIS membership database had fewer scheme. Difference in the findings might be due to scope explanatory variables, which accounted for less than of the two studies with respect to the study area and 5% of the variations in the renewal of membership. The population. limited variables also made the estimation of the ATET Findings of this study, however, show that the older prone to unobservable bias. Characteristics including adults of the scheme are more likely to renew their mem- education and household size, which are found to be bership, probably due to their higher healthcare needs associated with enrolment in health insurance were una- compared to the younger adults. This finding confirms vailable in the enrolment dataset. Nonetheless, the large an earlier study [30] but contradicts a study by Van der sample size used for the study, addresses this limitation. et al. [31], which examined determinants of enrolment in Lastly, the PSM technique helps to address selec- the NHIS using datasets from Global Ageing and Adult tion bias on the observables but not omission of unob- Health (2007–2008) and Ghana Living Standards Survey served differences between the treatment and the control round six (2012–2013). Surprisingly, the unmarried are groups (endogeneity). Therefore, our findings rely on the also more likely to retain coverage in the scheme com- assumption that there was no selection bias based on pared to the married. This finding contradicts an earlier unobserved characteristics of the two groups within the study [29], where the married individuals were rather one-year period of the study. more likely to retain coverage in the scheme compared to the single individuals. The difference in the findings is Conclusions due to the operational definition adopted. In this present The mobile phone-based contribution payment system is study, the unmarried comprised the single, divorced and improving retention of coverage in the NHIS particularly widowed whilst the other study operationalized single as groups who hitherto were less likely to renew their mem- the “never married” individuals. bership in the scheme (males, informal sector workers, Our findings suggest that leveraging electronic-based and the unmarried). Management of the scheme needs payment systems such as mobile phone-based payments to speedy up the process of making the system possible (mobile money), can remove barriers (time and travel for the population to enrol as new members to improve cost) to enrolment and speedy up progress towards reali- enrolment in the NHIS towards UHC. A mixed-method zation of UHC by 2030. Besides, the decreased likelihood study with inclusion of more variables would also be nec- of the informal sector workers and the males renewing essary to understand why these groups have lower chance their memberships indicates that system-wide factors of renewing their membership in the NHIS despite intro- other than time and travel costs, for example, perceived duction of digital health intervention, which addresses poor quality of service and OOP, reported in earlier stud- time and travel costs associated with participation in the ies [14, 25, 28, 29] are important determinants of enrol- scheme. ment, which policy makers need to address to improve enrolment in the scheme. Abbreviations ATET Average Treatment Effect on the Treated Limitations BMS Biometric Membership System We encountered few limitations in the study. First, the CCC C laims Check Code GHS Ghanaian Cedis mobile phone-based contribution payment system has ILO International Labour Organisation only one level (membership renewal). It could not be IPW I nverse‑probability weights used by persons who wished to enrol in the Scheme for LEAP Livelihood Empowerment Programme against Poverty LMICs Low and Middle‑Income Countries the first time. This situation might over-estimate the NHIA N ational Health Insurance Authority treatment effect. However, the matched sample was bal- NHIF National Health Insurance Fund anced, and results of the sensitivity analysis were also NHIS National Health Insurance Scheme NNM Nearest‑neighbour matching consistent, indicating that the findings are robust for pol- OOP O ut‑of‑Pocket Payment icy decision making. Secondly, the one-level design limi- PSM Propensity Score Matching tation made it impossible to separately examine factors SDG Sustainable Development Goals SMS Short Message Service influencing members’ decision to renew their member- SSNIT Social Security and National Insurance Trust ships using this system and the conventional biometric UHC Universal Health Coverage system at the district offices. Nonetheless, the overall USSD U nstructured Supplementary Service Data VAT Value‑Added Tax Nsiah‑Boateng et al. BMC Health Services Research (2023) 23:239 Page 8 of 9 Supplementary Information 5. Waddington CJ, Enyimayew KA. A price to pay: The impact of user charges in ashanti‑akim district, Ghana. Int J Health Plan Manag. The online version contains supplementary material available at https://d oi. 1989;4:17–47. org/ 10. 1186/s 12913‑ 023‑0 9236‑7. 6. Waddington C, Enyimayew KA. A price to pay, part 2: The impact of user charges in the Volta region of Ghana. Int J Health Plan Manag. Additional file 1: Supplementary Table 1. ATET robustness test for 1990;5:287–312. coverage retention. 7. Dalinjong PA, Laar AS. The national health insurance scheme: percep‑ tions and experiences of health care providers and clients in two districts of Ghana. Health Econ Rev. 2012;2:1–13. Acknowledgements 8. National Development Planning Commission. 2008 Citizens’ assess‑ We thank the French Development Agency and International Labour Organi‑ ment of the national health insurance scheme: towards a sustainable sation (ILO) for their financial and technical support during implementation health care financing arrangement that protects the poor. Accra: of the NHIS mobile phone‑based renewal system. We also thank members of National Development Planning Commission; 2009. the project team for their diverse contributions. Finally, we are grateful to the 9. Republic of Ghana. National Health Insurance Act, 2012 (Act 852). reviewers for their insightful comments and time. Accra: Parliament of Ghana; 2012. 10. Ministry of Health. National Health Insurance Policy framework for Authors’ contributions Ghana: Revised Version. Accra: Ministry of Health; 2004. ENB designed the study, analyzed the data, and drafted the manuscript. 11. National Health Insurance Authority. NHIS statistical bulletin 2018. MM participated in the drafting of the manuscript. CDA, FAB, FXAD and BOB Accra: National Health Insurance Authority; 2019. reviewed the manuscript and provided important suggestions to shape it. All 12. Agyepong IA, Nana D, Abankwah Y, Abroso A, Chun C, Nii J, et al. The the authors approved the final version of the manuscript for submission to the “Universal” in UHC and Ghana’s National Health Insurance Scheme: Journal. policy and implementation challenges and dilemmas of a lower mid‑ dle income country. BMC Health Serv Res. 2016;16:1–14. Funding 13. Palermo TM, Valli E, Ángeles‑ G, Milliano M de, Adamba C, Spadafora Not applicable. TR, et al. Impact evaluation of a social protection programme paired with fee waivers on enrolment in Ghana‘ s National Health Insurance Availability of data and materials Scheme. BMJ Open. 2019;9:1–10. The data for this study is publicly available in Mendeley Data Repository with 14. Kotoh AM, Aryeetey GC, Van Der Geest S. Factors That Influence Enrol‑ the Reserved https:// doi. org/ 10. 17632/m yfbx9 g4tg.1, https:// data. mende ley. ment and Retention in Ghana’ National Health Insurance Scheme. Int J com/ datas ets/m yfbx9 g4tg Health Policy Manag Int J Health Policy Manag. 2018;7:443–54. 15. Boaheng JM, Amporfu E, Ansong D, Osei‑Fosu AK. Determinants of pay‑ Declarations ing national health insurance premium with mobile phone in Ghana: a cross‑sectional prospective study. Int J Equity Health. 2019;18:2–9. Ethics approval and consent to participate 16. Leeuw F, Vaessen J. Impact evaluations and development: NONIE All methods were carried out in accordance with relevant guidelines and guidance on impact evaluation. Washington: Network of Networks for regulations. Official approval was obtained from the NHIA for the use of the Impact Evaluation; 2009. enrolment data for this evaluation study. We removed personal identifiers 17. Gertler PJ, Martinez S, Premand P, Rawlings LB, Vermeersch CMJ. Impact such as name, telephone number, and residential address from the data prior Evaluation in Practice. Washington: The World Bank; 2011. to analysis to ensure anonymity of the NHIS members whose data were used 18. Handley MA, Lyles CR, Mcculloch C, Cattamanchi A. Selecting and for the study. Improving Quasi‑Experimental Designs in Effectiveness and Implemen‑ tation Research. Public Health. 2018;39:5–25. Consent for publication 19. Rosenbaum PR, Rubin DB. Reducing Bias in Observational Studies Not applicable. Using Subclassification on the Propensity Score. J Am Stat Assoc. 1984;79:516–24. Competing interests 20. Austin PC. Propensity‑score matching in the cardiovascular surgery The authors are all employees of the National Health Insurance Authority, literature from 2004 to 2006: A systematic review and suggestions for however, their affiliation did not in anyway, influence findings of this study. improvement. Eur J Cardio‑thorac Surg J Cardiovasc Surg J Thorac Car‑ diovasc Surg Between. 2004. https:// doi. org/ 10. 1016/j.j tcvs.2 007.0 7. 021. Author details 21. Blanchet NJ, Fink G, Osei‑Akoto I. The effect of Ghana’s National 1 National Health Insurance Authority, Accra, Ghana. 2 Ministry of Health, Accra, Health Insurance Scheme on health care utilisation. Ghana Med J. Ghana. 3 School of Public Health, University of Ghana, Accra, Ghana. 2012;46:76–84. 22. Bonfrer I, Breebaart L, Van De PE. The Effects of Ghana ‘ s National Received: 10 September 2022 Accepted: 2 March 2023 Health Insurance Scheme on Maternal and Infant Health Care Utiliza‑ tion. PLoS ONE. 2016;11:1–13. 23. Ahmed S, Sarker AR, Sultana M, Chakrovorty S, Ahmed MW, Dorin F, et al. The impact of community‑based health insurance on the utiliza‑ tion of medically trained healthcare providers among informal workers in Bangladesh. PLoS ONE. 2018;13:1–16. References 24. Anindya K, Lee JT, McPake B, Wilopo SA, Millett C, Carvalho N. Impact of 1. Lagomarsino G, Garabrant A, Adyas A, Muga R, Otoo N. Moving Indonesia’s national health insurance scheme on inequality in access towards universal health coverage: health insurance reforms in nine to maternal health services: A propensity score matched analysis. J developing countries in Africa and Asia. Lancet. 2012;380:933–43 Glob Health. 2020;10:1–12. (www. thelan cet. com). 25. Fenny AP, Kusi A, Arhinful DK, Asante FA. Factors contributing to low 2. Obadha M, Colbourn T, Seal A. Mobile money use and social health uptake and renewal of health insurance: a qualitative study in Ghana. insurance enrolment among rural dwellers outside the formal Glob Health Res Policy. 2016;1:1–10. employment sector: Evidence from Kenya. Int J Health Plan Manag. 26. Issahaku H, Abu BM, Nkegbe PK. Does the Use of Mobile Phones by 2020;35:e66‑80. Smallholder Maize Farmers Affect Productivity in Ghana? J Afr Bus. 3. Chemin M. Informal Groups and Health Insurance Take‑up Evidence 2018;19:302–22. from a Field Experiment. World Dev. 2018;101:54–72. 27. Beuermann DW, McKelvey C, Vakis R. Mobile Phones and Economic 4. Nyonator F, Kutzin J. Health for some? The effects of user fees in the Development in Rural Peru. J Dev Stud. 2012;48:1617–28. Volta Region of Ghana. Health Policy Plan. 1999;14:329–41. Nsiah‑Boateng et al. BMC Health Services Research (2023) 23:239 Page 9 of 9 28. Atinga RA, Abiiro GA, Kuganab‑Lem RB. Factors influencing the deci‑ sion to drop out of health insurance enrolment among urban slum dwellers in Ghana. Tropical Med Int Health. 2015;20:312–21. 29. Boateng D, Awunyor‑Vitor D, Scheme N, Atim C, Grey S, Apoya P, et al. Health insurance in Ghana: evaluation of policy holders’ perceptions and factors influencing policy renewal in the Volta region. Int J Equity Health. 2013;12:1–10. 30. Nsiah‑Boateng E, Nonvignon J, Aryeetey GC, Salari P, Tediosi F, Akweongo P, et al. Sociodemographic determinants of health insur‑ ance enrolment and dropout in urban district of Ghana : a cross‑sec‑ tional study. Health Econ Rev. 2019;9:1–9. 31. van der Wielen N, Falkingham J, Channon AA. Determinants of National Health Insurance enrolment in Ghana across the life course: Are the results consistent between surveys? Int J Equity Health. 2018;17:1–14. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. Ready to submit your research ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced rese archers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions