Linking civil registration to the health system. Impact evaluation - - PDF document

1

Linking civil registration to the health system. Impact evaluation study in Burkina Faso Evelina Martelli (evelinamartelli@gmail.com),1 Maria Castiglioni (casti@stat.unipd.it),2 Angela Silvestrini (ansilves@istat.it),3 Francesco Di Domenicantonio (francodido@gmail.com),4 Palmira Gianturco (miragianturco@gmail.com)5

1 Community of Sant’Egidio, Rome, Italy 2 Department of Statistics, University of Padua, Italy 3 Istituto Nazionale di Statistica (ISTAT); Rome, Italy 4 Municipality of Rome, Italy 5 Community of Sant’Egidio, Rome, Italy

Introduction Civil registration is the administrative tool to record occurrence and characteristic of major vital events (mainly births, marriages, and deaths). The primary function of civil registration is to provide individuals with those legal documents needed to officially establish identity, family relationships, and entitlement for rights. Particularly for children such documentation helps to protect them from exploitation and hardship. Furthermore, it produces benefits for the state through the link with other services, such as healthcare, election, tax, and others. Civil registration and Vital statistics (CRVS) systems are increasingly considered crucial for strengthening democracy, governance, and development. CRVS entail fundamental benefits for monitoring progress in basic human rights, including the development goals. The record of vital events is the major source for tracing fertility and mortality, moreover it helps to monitor population size and changes.1 Compared to other evaluation systems, like censuses and surveys, civil registration in the long term is more accurate and cost effective thanks to its greater reliability and its characteristic of tracking changes over time.2 National statistical systems of high income countries are progressively based on exploitation of administrative data of population

1 Setel P W, Macfarlane S B, Szreter S, Mikkelsen L, Jha P, Stout S, AbouZahr C, A scandal of invisibility: making

everyone count by countig everyone, Lancet 2007: published on line October 29. DOI:10.1016/S0140- 6736(07)61307- 5.

2 AbouZahr C, de Savigny D, Mikkelsen L, Setel P W, Lozano R, Lopez AD, Towards universal civil registration and

vital statistics system: the time is now. Lancet 2015; published on line May 11. http://dx.doi.org/10.1016/S0140- 6736(15)60170-2.

2

registers, gradually replacing traditional censuses. Studies present evidence that well functioning CRVS are associated with better population health outcomes3 and development.4 The World Bank, the WHO, UNICEF, the UN, and its Regional Commissions, as long as other UN agencies and Pan African Institutions are very committed in ensuring that countries receive practical guidance on CRVS. Though several studies, reports, and articles stress the ultimate need for innovative approaches on CRVS strengthening in low income countries and for evidence of working strategies and approaches. Effective and sustainable methodologies are very urgently needed. Although evidence of the usefulness of well functioning CRVS systems has been demonstrated in many different fields, it has been noticed that “the lack of major progress in CRVS is probably the most critical development failure during the era of the Millennium Development Goals (MDGs)”. 5 The Community of Sant’Egidio started the Birth Registration for All Versus Oblivion (BRAVO!) Program in order to promote civil registration to safeguarding legal protection and fundamental rights especially for the most vulnerable children. BRAVO! is currently cooperating with the governments of Burkina Faso, Mozambique, and Malawi with the aim to realize universal, continuous, permanent, and compulsory civil registration systems.6 One of the key areas of interventions is the effort to guarantee the sustainability of the system through universality of infant birth registration. To this end, BRAVO! supports the opening of new civil registration offices in health centres providing neonatal services. Up to now 10 registration centres have been opened in Mozambique, 12 in Malawi, and 13 in Burkina Faso. In 2015 the Community of Sant’Egidio started a prospective study with comparison with historical cohort in Burkina Faso to look over the

• utcomes of the linkage between civil registration and health care system at a decentralized level.

Methods Study area Both the action of BRAVO! and the prospective study verifying its effectiveness are fully integrated within the framework of the Burkina Faso laws, regulations, and programmes. Burkina Faso’s legislation provides that every municipality has a principal civil registration centre (CPEC) right at the city hall. The mayor, in its duty of civil registrar, may create secondary civil registration centres (CSEC) in villages and health centres. The acts must be drawn up only in the municipality where the event occurred. In the CSEC only births and deaths can be registered for free in the prescribed period by law - i.e. 61 days. Delayed birth and death registration can occur upon payment of the advised fine only in the CPEC, following a judicial procedure by the local administrative court.

3 Philips D E, AbouZahr C, Lopez A D, Mikkelsen L, de Savigny D, Lozano R, Wilmoth J, Setel P W, Are well

funcioning civil registration and vital statistics systems associated with better health outcomes? Lancet 2015; published

• n line May 11. http://dx.doi.org/10.1016/S0140-6736(15)60172-6.

4 Harbitz, Mia, and Bettina Boekle-Giufrida, ‘Democratic Governance, Citizenship, and Legal Identity’, Working

Paper, Inter-American Development Bank, Washington, D.C., 2009.

5 Shibuya K, Gilmour S, Civil registration as a means to promote human security. Lancet 2015; published on line

• May11. http://dx.doi.org/10.1016/S0140-6736(15)60765-6.

6 United Nations (UN). Principles and recommendations for a vital statistics system, Revision 2. UN, 2001.

3

The prospective study7 with comparison with historical cohort has been conducted in the Centre Ouest Region of Burkina Faso. The country was chosen because it is a low-income ($615 gross per capita income in 2015), landlocked Sub-Saharan country, where CRVS performance is not assessed by the Vital Statistics Performance Index (VSPI) due to the poor production of data.8 Moreover, the area was chosen due to its low birth registration coverage: according to MICS IV, 2010, national birth registration rate among children 0-5 years old is 77%, in the Centre Ouest Region is 62%.9 It is the second region for low child registration rates after the Sahel. The 2006 census detected an overall birth registration rate of 62·3%, with small variances between age classes, and with a difference of 12 points between women (lower) and men.10 The implementation of special free of charge late registration campaigns has dramatically changed the

• situation. Particularly important was the national campaign promoted by the government in

collaboration with the Community of Sant'Egidio in 2009: for 12 months teams of judges and civil registration agents visited every village and urban area in the country to offer free delayed registration of births to those lacking birth certificate. The 2014 multisectoral permanent inquiry11 has shown that the overall birth certification (the possession of the birth certificate by the individual) rate is 79·2%, being that of men 4·7% higher than that of women. Also it pointed out that in the 40-49 age class there is the highest birth certification rate (92·1%), and in the 0-4 years the lowest (59·9%). Even among children under the age of five, birth registration is mostly done with late procedures, as demonstrated in a different context by a study on MICS data processing.12 To improve civil registration system and make it a permanent one, it clearly means to create a system favouring timely registration of new-borns. Health data for the two studied municipalities show a high coverage of neonatal services, with a percentage of assisted deliveries of 73·3% and of immunization for tuberculosis and pentavalent first dose of 100%.13 Moreover, they show that the target population for the health facilities is between 5,000 and 20,000 inhabitants, more than half the cases (52·5%) living less than 5 km from the centre and only one out of five (20·1%) living more than 10 km away from the centre. The distribution of health centres is certainly more widespread than other services provided by the State for early childhood. Study design

7 The study have been approved by the Comité d’ethique pour la recherche en santé of Burkina Faso, decision N° 2016-

5-063 del 4.5.2016 .

8 Mikkelsen L, Phillips D E, AbouZahr C, Setel P W, de Savigny D, Lozano R, lopez A D, A global assessment of civil

registration and vital statistics systems: monitoring data quality and progress. Lancet 2015May 11. http://dx.doi.org/10.1016/S140-6736(15)60171-4.

9 EDSBF- MICS IV, 2010. 10 Institut national de la statistique et de la démographie (INSD), Recensement général de la population et de

l’habitation (RGPH) de 2006 du Burkina Faso.

11 Institut national de la statistique et de la démographie (INSD), Enquête multisectorielle continue (EMC) 2014.

Caractéristiques sociodémographiques de la population, novembre 2015. http://www.insd.bf/n/contenu/enquetes_recensements/Enq_EMC/Caract%e9ristiques_sociodemographiques_de_la_pop ulation.pdf

12 Makinde O A, Olapeju B, Ogbuoji O, Babalola S, Trends in the completeness of birth registration in Nigeria: 2002-

• 2010. Demographic Research 2016 Published August, 17, Vol. 35, Art. 12pp. 315-338. Published Online at:

http://www.demographic-research.org/Volumes/Vol35/12/DOI: 10.4054/DemRes.2016.35.12.

13 District sanitaire de Réo, Plan d’Action 2016, Mai 2015.

4

The prospective study with comparison with historical cohort investigates the increase of the birth civil registrations after the opening of new offices in health centres, the coverage rate compared to the cohort of children attending health centres in the first 61 days of life, and the possible selective effects on the most vulnerable children. It explores the possibilities to strengthen the production of vital statistics and test the procedures for crossing the health data with those of the civil registration

• system. This is to measure and possibly reduce the number of children not registered at birth. It also
• ffered the opportunity to establish and set up a decentralized, low investment, and sustainable

registration system in order to support its possible enlargement on a national scale. In order to study the impact of the action in different socio-economic contexts, it has been chosen to intervene in both urban municipality and in rural one. In every health centre of the two municipalities a secondary registration office has been opened in order to immediately provide with a birth certificate every child born and/or vaccinated there. A municipal civil registration agent keeps the office open 5 days a week. Seven registration centers were opened in the urban municipality of Réo and four in the rural one of Godyr. The two studied municipalities have been matched with two similar and comparable control municipalities – Yako and Gomponsom - with similar characteristics, where no actions have been carried out by the program BRAVO! This to exclude that external variables may have influenced the observed processes. Besides seasonality, adverse climate conditions, political, and administrative changes can affect trends on registrations. For this reason, a comparative long-term approach was necessary to be studied. Data inclusion and exclusion criteria In Godyr the action started on February the 20th, 2015 and in Réo on July the 1st, 2015. Retrospective and current data from civil registers for birth and death have been collected for the four municipalities from January 1st, 2014. For the control municipalities civil registration data have been collected until June 30th, 2016 and for the studied municipalities until April 30th, 2017. Current data from health registers for delivery and vaccination have been collected within the studied municipalities (Réo and Godyr) since the beginning of the pilot and until April 30th, 2017. Health and civil registers identified the eligible children, i.e. those born less than 61 days before the contact in the relevant municipality. All families have been informed of the availability of the registration

• ffice. At the same time, a survey questionnaire was compiled, via direct interview. Data from the

registers, available only in paper form, have been digitalized and integrated through the direct interviews, and appropriately linked, to combine the various sources. The data have been interactive validated, eliminating records with obvious recording errors such as improbable birth date. No probabilistic programs of correction and integration of missing or false data have been used. In the event of a discrepancy of data, civil registers have been considered the predominant source. Not eligible for study are health events (birth and vaccination) regarding children who have been born in a municipality other than the one where the event is registered, those born more than 61 days before contact, those with insufficient child data (absence of date or place of birth). In Réo 10·4% of health events have been excluded from the study, in Godyr 23·6%.

Municipality Inhabitants * Civil registry Health data Eligible events Observation period Birth registrations Observation period Total Deliveries Total Vaccinations Eligible deliveries + vaccinations Réo 62.208 1.01.14- 30.04.17 4.479 1.07.15- 30.04.17 3.622 3.890 6.733 11.212

5

Godyr 19.320 1.01.14- 30.04.17 2.007 20.02.15- 30.04.17 2.043 2.311 3.328 5.335 Yako 80.926 1.01.14- 30.06.16 3.936

• 3.936

Gomponsom 18.268 1.01.14- 30.06.16 495

• 495

Total 180.722 10.917 5.665 6.201 10.061 20.978 * Up to 2006 Census, latest available data

All registration centres are still pursuing their normal civil registration activities. Results Increase in birth civil registrations Figure 1 - Trends in birth registrations in the studied municipalities of Godyr and Réo (January 2014/April 2017)

Note : Monthly declarations of births for 1000 inhabitants (2006 Census, latest available data).

Figure 1 represents the trend of civil registration of newborns 0-61 days from January the 1st, 2014 to April 30th, 2017 in the Godyr and Réo municipalities for 1.000 inhabitants. There is a significant increase in registrations since the opening of secondary registration offices in health centres and then an irregular curve that follows a similar trend in the two municipalities, even though less accentuated for Godyr and more marked for Réo. It is characterized by four negative peaks that coincide with the months of most intense rainfall and with the harmattan wind. While the number of birth certificates has more than doubled every month since the beginning of the pilot. Godyr's increase in the first twelve months compared to the previous twelve months is equal to 196·6%. Indeed, from February the 20th, 2014 to February the 19th, 2015, 267 births were declared; in the next 12 months 792. In Réo there is an even more significant increase in registrations, with a boost in the first year of 314·1% over the previous twelve months: 505 births were registered from July the 1st, 2014 to June the 30th, 2015, while in the following12 months birth registrations have been 2091.

1 2 3 4 5 6 jan-14 feb-14 mar-14 apr-14 may-14 jun-14 jul-14 aug-14 sep-14

• ct-14

nov-14 dec-14 jan-15 feb-15 mar-15 apr-15 may-15 jun-15 jul-15 aug-15 sep-15

• ct-15

nov-15 dec-15 jan-16 feb-16 mar-16 apr-16 may-16 jun-16 jul-16 aug-16 sep-16

• ct-16

nov-16 dec-16 jan-17 feb-17 mar-17 apr-17 Godyr Réo month of starting up

6

In the control municipalities, Gomponsom and Yako, there is a strong month-to-month oscillation for registrations, which are distributed evenly over the months of observation. Recording peaks seems to be linked in these cases to political and administrative hardships. Serious operating problems have been observed in the two control municipalities. Due to disruptions in the provision

• f registers and to lack of adequate human resources, in the two municipalities it has been observed

the practice of pinning on flyers the declarations of birth made by the informant, releasing them a birth certificate, while the official birth registration record in the register has not been formed. This practice, which concerns a smaller number of events in the municipality of Gomponsom (18%), has been unfortunately widely widespread in Yako, where 67% of declarations have not been converted into a valid civil status act. In Figures 2 and 3, valid birth records and invalid birth certificates were added to each other in the red line to indicate the trend in the registrations, while a dotted orange line indicates the birth certificates with no birth registration record in the register. The action has the primary purpose of encouraging users to record births. For this reason, it seems appropriate to enhance, along with true birth certificates, the declarations that show the will of users to register

• newborns. However, it is important to stress also that the secondary purpose of the intervention is to

promote the full functioning of the offices and to equip them with the materials and personnel needed to be fully operational. This practice has not been noticed in Réo and Godyr, where the program BRAVO! facilitated the prompt resolution of any problem that may have occurred. Figure 2 - Trends in birth registrations in the rural municipalities of Godyr (pilot municipality) and Gomponsom (control municipality) (January 2014/June 2016)

Note : Monthly declarations of births for 1000 inhabitants (2006 Census, latest available data).

The increase in the municipality of Godyr at the inauguration of the pilot study does not match in the municipality of Gomponsom, which has an unchanged trend of registrations, with a curve that prior to the commencement of the action is similar to that of Godyr, but which totally differs after it. Figure 3 - Trends in birth registrations in the urban municipalities of Réo (pilot municipality) and Yako (control municipality) (January 2014/June 2016).

0,0 1,0 2,0 3,0 4,0 5,0 jan-14 feb-14 mar-14 apr-14 may-14 jun-14 jul-14 aug-14 sep-14

• ct-14

nov-14 dec-14 jan-15 feb-15 mar-15 apr-15 may-15 jun-15 jul-15 aug-15 sep-15

• ct-15

nov-15 dec-15 jan-16 feb-16 mar-16 apr-16 may-16 jun-16 Godyr Gomponsom (total) Gomponsom (birth certificates without birth records)

7

Note : Monthly declarations of births for 1000 inhabitants (2006 Census, latest available data).

Even in Yako municipality there is no increase in registrations in the observed period, except for a positive peak in April 2014 and a negative one in December 2015. Declarations of birth prior to the commencement of experimentation were in percentage more significant in Yako municipality. In choosing the municipalities of the experimentation, it was opted for intervening in the region with the lowest registration rates for children 0-5 years old, since it is not possible for security reasons to intervene in the Sahel region, which has the lowest registration rates in the country. At the time of the opening of secondary centres, Réo experienced an important increase in registrations, which, despite the monthly oscillations already highlighted above, remain constant. In Yako, however, there are no improvements. The trend of registrations does not seem to have been strongly influenced by the political upheavals that characterized Burkina Faso between 2014 and 2016. There are no general downturns of registrations in the warmer months of the crisis (30.10.14 insurrection that puts an end to Presidency of Blaise Compaore; December '14 replacement of mayors with "special delegation presidents"; 22.09.15 failed coup attempt; 29.11.15 political elections; 22.05.16 administrative elections). Only for Yako the two above mentioned peaks (positive in April 2014 and negative in December 2015) may be the result of political events: the positive peak coincides with a period of gratuity for delayed birth registration in view of the presidential elections, in the course of whom the municipality has taken action to promote registration, including for newborns; and the negative

• ne correspond with the political elections.

Regarding the sex ratio of the registered children, in Godyr and Réo there is an increase in the number of girls compared to males: before the pilot study their number was less than expected, during the action the ratio is balanced. Figure 4 - Sex ratio of newborns registered at birth in Réo and Godyr 12 months prior to action and 12 months during action

0,0 1,0 2,0 3,0 4,0 5,0 jan-14 feb-14 mar-14 apr-14 may-14 jun-14 jul-14 aug-14 sep-14

• ct-14

nov-14 dec-14 jan-15 feb-15 mar-15 apr-15 may-15 jun-15 jul-15 aug-15 sep-15

• ct-15

nov-15 dec-15 jan-16 feb-16 mar-16 apr-16 may-16 jun-16 Réo Yako (total) Yako (birth certificates without birth records)

8

Furthermore, prior to action registered children in Godyr came from 23 villages and in Réo from 18, during action from 48 and 38 respectively. Analyzing the distance between childbirth and registration, prior to experimentation, birth registrations were distributed fairly evenly over the 61 days, while during the experimentation, 48·3% of children in Godyr and 32·1% in Réo were recorded within two days of childbirth, presumably during hospitalization. Within a week, 77·2% of children were registered in Godyr and 60·1% in Réo.

100 200 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 Absolute number of newborns Distance between birth and birth registration (in days)

Godyr

12 months prior to action 12 months during action 150 47% 167 53%

Godyr - 12 months prior to action

Female Male 259 46% 301 54%

Réo - 12 months prior to action

Female Male 401 51% 392 49%

Godyr - 12 months during action

Female Male 1041 50% 1053 50%

Réo - 12 months during action

Female Male

9

Analyzing the propensity of mothers of different ethnic groups to register the child, during the experimentation there is an increase in registrations in all ethnic groups, albeit different in sizes. In Godyr there is a much stronger growth among Mossi moms, in Réo among the Gourousi mothers. Probably this can be explained by the fact that there are no special cultural factors that favor or hamper registration, but the existence of strong community ties has emulative power over women of the same family circle, with a fading effect. This happened in Godyr mainly among Mossi mothers, and in Réo among Gourounsi ones. Comparison of civil registration data with health data The study confirms the high adhesion of newborns to health services. 85·9% of the newborns in Godyr were born and were vaccinated in a health centre, while in Réo, due to historical reasons, 14 this percentage is 65·0%. Among these children, the percentage of civil registration is 97·8% for Godyr and 90·2% for Réo. Among the children who were born in health centers but who did not receive vaccinations, the percentage of birth registration is 62·8% for Godyr and 72·3% for Réo. Among children who received only vaccination in health centers, 90·1% were registered at birth in Godyr and 94·4% in Réo. 30 children in Godyr and 269 in Réo were just registered at birth without them having benefited from any health service. Analyzing the distance between childbirth and registration, among these children there is a reverse trend compared to the totality of newborns: 35·3% were registered at birth at least one month after birth. No death record occurred, though the infant death rate in the Centre West Region according to MICS IV, 2010, is 87‰. From the interviews we know that 16 stillbirths have been recorded in Godyr and 31 in Réo and that at least 28 children died in Godyr and 53 in Réo, only four of whom had been registered at birth in Godyr and one in Réo. Through registers and interviews many variables have been collected for each child. However, none

• f these seems to directly influence the propensity to register children at birth. For some variables,

missing responses are too numerous to be able to evaluate the collected data. The conclusions we

14 Médah R, Analyse du système de santé dans la commune de Réo (province du Sanguié, Burkina Faso), Labo

Citoyennetés, Ètude Recit n. 21, Mai 2008. 200 400 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 Absolute number of newborns Distance between birth and birth registration

Réo

12 months prior to action 12 months during action

10

can draw are indirectly inferred from the analysis of the variables, although they have a high degree

• f likelihood.

Health registers for delivery provide information about birth problems, such as premature birth, low birth weight, resuscitated baby, etc. Among children who have had health problems at birth the civil registration rate is significantly lower than the general one. This too can be an indirect indicator of the death of at least some of these children. Neonatal mortality can partially explain the significantly lower birth registration rate among children just born in the health centres. High vaccination rates in the area suggest that a baby born in a health center and non-vaccinated might have died. Other causes may be the greater mobility of mothers to give birth to health centers of good reputation, or to receive assistance from the family of origin, out of the municipality. This could in part also explain the birth registration of children at health centers without having received any health care. In fact, the law provides that it is only possible to register birth in the municipality where the child was born. But in the Centre Ouest Region there is a high mobility of the population, at local, regional and international levels. Therefore, it is common for children to move in a community other than the birthplace in the first 61 days of life. Some of these children have a birth certificate indicating a place of birth other than the municipality where the registration is made (which should cause the annulment of the act), in others cases families don’t declare the real birthplace of children. A part of the children just registered at birth without them having benefited from any health service could have been born in a health center of another municipality. Moreover, the great majority of the children excluded from the study for eligibility criteria have been left out because at the time of vaccination – within 61 days after birth - families declared that children were born in another municipality. Discussion The longitudinal study before and during the action in the two studied municipalities highlights a noticeable increase in birth recordings, which does not occur in the control municipalities. The bond with the health service has improved the birth registration performance because it has reduced the distance between the village of residence and the registration office. And above all because it has made the procedure available in places that families of newborns attend to take advantage of health services, even travelling long distances. Percentages of approximately 90% of newborn registration were obtained without providing a specific information campaign, but only asking health and civil registration staff to inform families about the possibility of registration. No selective effects were reported for ethnicity, gender of the child, or mother's age. The only factors that seem to adversely affect birth registration are the population's mobility and the child's bad health. With regard to mobility, the problem could be addressed effectively through amendments allowing declaration of birth even in municipalities other than the birthplace, providing for procedures that safeguard the certainty of the data and the uniqueness of the registration. Regarding the registration of births of children with a bad health, it relates to the importance - and absence - of the death records. In this case, regulations to be introduced could envisage a more extensive interpretation of the legislation already in force, entrusting the declaration of birth and death to the health, and administrative staff, and to traditional, and religious authorities, who receive

11

notice of the occurrence. A "clerical" declaration of birth and death could become the task of the above mentioned officials, answerable when they receive notice of an undeclared event. The growth

• f the culture of registration can positively influence, in the medium term, the common perception
• f the uselessness of dead recording, but the transformation of mentality and culture takes longer,

while the need for reliable data is compelling.