Article Text

Improving the demand for birth registration: a discrete choice experiment in Ethiopia
1. Mahari Yihdego1,
2. Ayanaw Amogne1,
3. Selamawit Desta2,
4. Yoonjoung Choi3,
5. Solomon Shiferaw4,
6. Assefa Seme4,
7. Li Liu2,
8. Stéphane Helleringer2
2. 2School of Public Health, Johns Hopkins University, Baltimore, Maryland, USA
3. 3iSquared, Information x Insight, Severna Park, MD, USA
4. 4Department of Reproductive Health and Health Service Management, Addis Ababa University, Addis Ababa, Ethiopia
1. Correspondence to Dr Stéphane Helleringer; sheller7{at}jhu.edu

## Abstract

Introduction Birth registration remains limited in most low and middle-income countries. We investigated which characteristics of birth registration facilities might determine caregivers’ decisions to register children in Ethiopia.

Methods We conducted a discrete choice experiment in randomly selected households in Addis Ababa and the Southern Nations, Nationalities, and People’s Region. We interviewed caregivers of children 0–5 years old. We asked participants to make eight choices between pairs of hypothetical registration facilities. These facilities were characterised by six attributes selected through a literature review and consultations with local stakeholders. Levels of these attributes were assigned at random using a fractional design. We analysed the choice data using mixed logit models that account for heterogeneity in preferences across respondents. We calculated respondents’ willingness to pay to access registration facilities with specific attributes. We analysed all data separately by place of residence (urban vs rural).

Results Seven hundred and five respondents made 5614 choices. They exhibited preferences for registration facilities that charged lower fees for birth certificates, that required shorter waiting time to complete procedures and that were located closer to their residence. Respondents preferred registration facilities that were open on weekends, and where they could complete procedures in a single visit. In urban areas, respondents also favoured registration facilities that remained open for extended hours on weekdays, and where the presence of only one of the parents was required for registration. There was significant heterogeneity between respondents in the utility derived from several attributes of registration facilities. Willingness to pay for access to registration facilities with particular attributes was larger in urban than rural areas.

Conclusion In these regions of Ethiopia, changes to the operating schedule of registration facilities and to application procedures might help improve registration rates. Discrete choice experiments can help orient initiatives aimed at improving birth registration.

• child health
• cross-sectional survey
• other study design

This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/.

## Statistics from Altmetric.com

### Key questions

• Birth registration is incomplete in most low and lower-middle income countries.

• It also prevents the establishment of reliable vital statistics about fertility and mortality.

#### What are the new findings?

• In two regions of Ethiopia, there were important barriers to birth registration related to costs, distances and wait times.

• Caregivers of young children also expressed strong preferences for registration facilities that had convenient opening schedules (eg, evenings and weekends), and delivered birth certificates in a single visit.

• In urban areas, only requiring one of the parents to be present at the time of registration might also help improve registration rates.

#### What do the new findings imply?

• Interventions that modify the opening schedule of registration facilities, as well as registration procedures, might complement current initiatives to improve birth registration.

• Discrete choice experiments have the potential to help inform the development of birth registration systems in low and middle-income countries.

## Introduction

Birth registration is the process of recording a child’s birth in governmental registers or databases. It is necessary to establish a birth certificate, which gives each child a number of rights and protections.1 2 For example, it helps establish filiation and inheritance rights. Ownership of a birth certificate is associated with fewer school dropouts, reduces exposure to child trafficking, labour or early marriage and often improves access to healthcare services.3–5 Birth registration is also a key component of the production of annual estimates of fertility and mortality rates. These vital statistics are essential in planning and evaluating social services such as healthcare or education.3 6 7

The coverage of birth registration varies greatly throughout the world.6 In high-income countries, birth registration is timely and (nearly) universal. In poorer countries, many births are never registered.3 Others are registered only several years after the birth, for example, when a birth certificate is needed to enrol in school. Within countries, the most disadvantaged social groups have lower registration rates than more affluent groups.8–10

Reaching universal birth registration in low income and lower-middle income countries (LLMIC) has recently become a key objective of LMICgovernments and various global actors.11 12 The birth registration rate is one of the indicators used to track progress towards the 16th and 17th Sustainable Development Goals, that is, the promotion of more inclusive societies and the strengthening of systems contributing to sustainable development. Major global initiatives have been launched to strengthen civil registration and vital statistics (CRVS) systems.13 14 They focus on promoting legislative changes required to expand birth registration, developing new tools to facilitate the production of vital statistics and/or strengthening the administrative systems that implement birth registration.

Improving the coverage of birth registration in LMICs also requires stimulating the demand for, and removing barriers to, birth registration among local populations, particularly in settings where significant numbers of births occur at home. This is so because CRVS systems are predominantly ‘passive’: the caregiver(s) of a child must contact a CRVS agent to report the occurrence of the birth and complete the required paperwork.

We investigated the preferences of caregivers for the registration of births in two regions of Ethiopia, a country with some of the lowest birth registration rates worldwide.15 We used a discrete choice experiment (DCE), that is, a survey methodology in which respondents repeatedly choose between hypothetical versions of a service characterised by a small number of randomly selected attributes.16 Statistical analysis of DCE data allows assessing the relative importance of each of these attributes in influencing decisions to obtain a particular service. DCEs are widely used in marketing and management,17 18 and have recently helped guide health systems strengthening in LMICs.19–23 This methodology has however not been used to inform the development of CRVS systems.

## Results

We selected 840 caregivers for the birth registration study and 715 consent to participate (response rate=85.1%). Among those, 705 completed the DCE section of the interview. Four hundred and fifty-nine DCE respondents resided in urban areas (65.1%, table 2) versus 246 in rural areas (34.9%).

Table 2

Characteristics of participants in the discrete choice experiment, Addis Ababa and SNNPR of Ethiopia 2018/2019

In urban areas, approximately 4 out of 10 respondents resided in Addis Ababa (194/459, 42.3%). All rural respondents resided in the SNNPR. There were large differences in background characteristics by place of residence. Urban respondents were younger, more educated and often members of wealthier households than rural residents. Only 10.9% of urban respondents (50/459) had never been to school versus 49.6% of rural respondents (122/246). Similarly, more than two-thirds of urban respondents resided in a household that belonged to the wealthiest quintile of the Ethiopian population (314/465, 67.5%), whereas this was the case for <1% of rural respondents (2/250, 0.8%). A larger proportion of urban respondents had ever heard messages about birth registration (49.9% vs 16.7%). DCE participants were predominantly women (96.3% in urban areas and 99.2% in rural areas), who were currently married (85.7% in urban areas and 90.3% in rural areas).

Participants failed to complete 26 of the 5640 total DCE choices they were asked to make (0.45%). Among valid DCE answers, respondents opted out of the choice between the two hypothetical facilities 726 out of 5614 times (12.9%).

This proportion was slightly higher in rural areas (295/1960, 15.1%) than in urban areas (431/3654, 11.8%). Two-thirds of respondents never opted out of the choice they were asked to make (470/705, 66.7%). This proportion was higher among urban respondents than among rural respondents (69.5% vs 61.4%). Only six respondents (0.85%) opted out of every choice. Among those, five resided in urban areas and one resided in rural areas. In the choice set with a dominant option, 656 respondents selected the objectively most desirable registration facility (out of 703 respondents having provided valid answers to this choice set, 93.3%). There were no differences in the likelihood of selecting the dominant option between urban and rural respondents (93.2% in urban areas vs 93.5% in rural areas).

The parameter estimates for the effects of registration facility attributes on utility are shown in table 3. Positive coefficient estimates indicate that respondents favour a particular attribute or level of that attribute. Conversely, negative estimates indicate that an attribute and/or its level create disutility for the respondents.

Table 3

Results from random parameter logit models of DCE data, Addis Ababa and SNNPR of Ethiopia 2018/2019

In both urban and rural areas, respondents were less likely to select facilities that had longer waiting times (β=−0.130 in urban areas, and β=−0.127 in rural areas), or were located further away from their residence (β=−0.581 in urban areas, and β=−0.569 in rural areas). They preferred facilities that completed all registration procedures and delivered birth certificates in a single visit (β=0.731 in urban areas, and β=0.631 in rural areas) and facilities that were open on weekends (β=0.555 in urban areas, and β=0.543 in rural areas).

Higher costs of birth certificates negatively affected the utility of caregivers in both urban and rural areas (β=−0.010 and β=−0.020, respectively). However, the disutility resulting from higher costs was larger in rural areas. Respondents in urban areas expressed preferences for facilities that were open for extended hours on weekdays (β=0.205), and that only required one of the parents to be present for registration (β=0.427). Respondents in rural areas did not display similar preferences in their choices.

There was heterogeneity between respondents in preferences relating to several attributes (table 3). We rejected the null hypothesis that there was no variation across participants in the effects of costs (SD=0.007 in urban areas, and SD=0.012 in rural areas), waiting time (SD=0.224 in urban areas, and SD=0.217 in rural areas) and distance (SD=0.590 in urban areas, and SD=0.449 in rural areas) on choice patterns. In urban areas, there was heterogeneity in preferences related to the number of visits required to register a birth (SD=0.527), and with application procedures (SD=0.931). There was no evidence of heterogeneity in preferences relating to the opening hours of the registration facility, both in urban and rural areas. In rural areas, there was no evidence of heterogeneity in preferences relating to the number of visits required to obtain the birth certificate and to application procedures.

Results from the WTP analysis are shown in table 4. Negative WTP estimates indicate that respondents would require compensation to select registration facilities with such attribute levels, whereas positive WTP estimates represent the implicit price that respondents are willing to incur to access registration facilities with an attractive attribute. DCE participants would thus require compensation to use registration facilities that are further away from their residence (WTP=−57.79 birr or −US$2.02 per additional hour in urban areas, and WTP=−28.37 birr or −US$1.00 in rural areas), or that require longer waiting times (WTP=−12.93 birr or −US$0.45 per additional hour in urban areas, and WTP=−6.31 birr or −US$0.22 in rural areas). Participants would be willing to incur a cost of 72.61 birr (US$2.54) in urban areas, and 31.44 birr (US$1.10) in rural areas, to access a facility that delivered the birth certificate in a single visit. Relative to a facility open at regular weekday hours, participants were willing to incur additional costs in order to attend a registration facility that opens on weekends (55.15 birr or US$1.93 in urban areas, vs 27.02 birr or US$0.95 in rural areas).

Table 4

Estimates of willingness to pay for facility attributes from mixed logit models, Addis Ababa and SNNPR of Ethiopia 2018/2019

## Discussion

We documented the preferences of caregivers in registering births in two regions of Ethiopia. We used an established experimental survey method (ie, a DCE) in a population-based sample. The DCE confirmed the existence of significant barriers to birth registration in these two regions. DCE participants were less likely to opt for registration facilities that were further away from their homes, that had longer wait times to obtain services and/or that charged a fee for the acquisition of the birth certificate. This is consistent with findings from other studies that have investigated barriers to birth registration in other settings, using an array of other methodologies.49 50

We found strong preferences for registration facilities that deliver a birth certificate in a single registration visit. The current CRVS policy in Ethiopia requires that the administrative facilities implementing birth registration deliver the birth certificate immediately to the parents/caregivers. However, in our survey, more than a third of caregivers who had registered the birth of their child reported having to return to the registration facility several times to complete that process. More consistently implementing the current policy might help improve registration rates.

DCE respondents in urban areas also preferred registration facilities that only required one of the two parents to be present at the time of registration. The main procedure outlined by the current legal framework in Ethiopia however requires both parents to be present at the registration facility in order to register a birth. This might constitute a barrier to birth registration, as also indicated by prior studies that have investigated reasons reported by caregivers for not registering a birth in other settings.50 Indeed, this might make birth registration more complex for children who have at least one parent with rigid work schedules or who is engaged in migration, or for children whose parents might no longer be in a relationship/union.

Our study highlighted other characteristics of registration facilities that might play a key role in the registration-related behaviours of caregivers. In particular, choices were influenced by the opening schedule of the facility: DCE participants expressed consistent preferences for registration facilities that remained opened on weekends. In urban areas, they also expressed preferences for facilities that remained opened for extended hours on weekdays. This might be because the current opening schedule of registration facilities conflicts with work schedules or with times during which economic activities of caregivers are ongoing.

There were differences in the preferences revealed by the DCE between residents of urban and rural areas. In particular, the effects of costs on patterns of choices were larger in rural areas than in urban areas. This is likely due to the fact that rural residents were much poorer than urban residents. As a result, WTP estimates were lower in rural areas. For example, caregivers in urban areas were willing to pay more than 55 birr (ie, approximately US$2.0 on 1 January 2019) to access a registration facility that was opened on weekends, whereas caregivers in rural areas were willing to incur only half of that implicit price (27.02 birr or US$0.99).

Our DCE has several limitations. First, we only investigated the main effects of each attribute of registration facilities, without considering potential interactions between attributes. This is problematic because the effects of an attribute might depend on the levels of another: for example, facilities that remain open on weekends might be particularly attractive in settings where both parents are required to be present at the time of registration, because it is more likely that both parents will be available on weekends. Investigating interactions between attributes would however require respondents to make a larger number of choices during the DCE than we deemed feasible in this setting.51

Second, due to limited sample sizes, we only investigated whether preferences for birth registration varied between urban and rural areas. We did not investigate whether preferences varied across other subgroups, for example, by poverty or educational level. In urban areas, we also did not investigate whether preferences varied between Addis Ababa and the smaller cities of the SNNPR. Third, we only presented DCE respondents with choices that were characterised by a limited set of attributes. Other aspects of the registration facilities/process might affect registration choices, for example, whether the registration office is located in an administrative setting or in a healthcare setting.

Fourth, some of the choice patterns in our DCE indicated that some respondents might not have fully understood the choices they were asked to make, or experienced fatigue. A small fraction of the respondents consistently opted out of the choices they were presented (<1%); whereas others (6.7%) failed to select the objectively most appealing alternative in a dominant choice set. However, these proportions were consistent with the experience of other high-quality DCEs,36 including with more educated populations.20 52 We also replicated our analyses of the DCE after excluding respondents with inconsistent choices, and we found similar patterns of preferences (online supplementary file 2).

### Supplemental material

Fifth, the caregivers we interviewed were in large majority women (>96%). However, as for other services (eg, family planning), men likely play important roles in the decision process about birth registration, in particular in settings where their presence is required for registration. Future studies should thus ensure that men are included in DCEs designed to elicit preferences towards birth registration. Sixth, our statistical analyses of DCE data made several assumptions that might have impacted our results. For example, we used mixed logit models to represent heterogeneity in preferences within the population. Other recent work has used latent class models to represent such heterogeneity.23 43 Similarly, we accounted for opt-out effects by including an alternative specific constant in our models. Other approaches (eg, nested logit models) might yield slightly different estimates of the WTP for various attributes of a service.34

Finally, our work was limited to two regions of Ethiopia, and thus does not represent the preferences of residents of other regions of the country where birth registration is also low. Furthermore, Ethiopia is a country where birth registration has only recently been reorganised and implemented nationwide. Preferences for birth registration might differ in countries with higher background rates of event registration (eg, Kenya).

Despite these limitations, our work indicates several strategies that might help further accelerate the scale-up of birth registration in Addis Ababa and in the SNNPR. It appears warranted to explore whether altering the opening schedule of registration facilities to allow evening and weekend openings might help improve birth registration rates. This is feasible within the current legislative framework for birth registration in Ethiopia and could thus be tested during a cluster-randomised trial in those two regions. Such a change might stimulate the demand for birth registration and complement initiatives that aim to strengthen and streamline the administrative systems that implement civil registration.14 Other strategies highlighted by our DCE (eg, reducing the legal requirements for parental presence at the time of registration) might also have an impact on birth registration rates, but would not be possible without amendments to the legislative framework that regulates birth registration in Ethiopia. Finally, our work shows that DCEs might also be a useful methodology to help orient initiatives to increase birth registration rates, similar to their role in health systems strengthening.

## Footnotes

• Handling editor Seye Abimbola

• Contributors MY, YC, SS, AS, LL and SH designed the study. MY, AA and SD collected the data and oversaw the quality control. AA and SH prepared and analysed the data. SH wrote the first draft of the paper. All authors reviewed and contributed to the draft paper, contributed to the interpretation of the analysis and approved the final submission.

• Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

• Competing interests None declared.

• Patient and public involvement We consulted with staff from the Vital Events Registration Agency (VERA) in designing the survey experiment. Dissemination events are being planned at VERA in the coming months.

• Patient consent for publication Not required.

• Ethics approval The Institutional Review Boards of the School of Public Health at Johns Hopkins University and of Addis Ababa University approved the protocol. All respondents provided oral informed consent before participating.

• Provenance and peer review Not commissioned; externally peer reviewed.

• Data availability statement Data are available upon request. They will soon be made publicly available via the PMA website.

## Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.