Examining national and district-level trends in neonatal health in Peru through an equity lens: a success story driven by political will and societal advocacy

Peru met the fourth Millennium Development Goal (MDG) of reduction of under-five mortality rate by two thirds between 1990 and 2015 [1, 2], and is on track to meet the fifth goal of reduction of maternal mortality by three quarters by 2015 [1, 2]. It has also reduced dramatically under-five stunting and underweight prevalence, and has been ranked as the first country among all low and middle-income countries (LMIC) in reducing its neonatal mortality rate from 1990 onwards [3].

Initial global emphasis on child survival has recently shifted to focus on maternal and neonatal health, as an imperative to reach sustainable progress at country and local level [4‐6]. Systematic analyses on factors driving progress in LMIC are critical to guide future policies and programmes. A recent multi-country review of health-systems bottlenecks and success factors related to country progress of neonatal mortality has identified health workforce planning, financial protection measures and dynamic leadership as key facilitating strategies for reduction of neonatal mortality in eight of the 13 countries accounting for the highest burden of neonatal deaths [7].

In-depth analysis disaggregated by wealth quintiles, urban and rural residences and by districts (equivalent to departments in Peru) may highlight a differing pace of progress and may help to identify key areas for further policy and programmatic refinement to ensure continued progress in newborn health.

We aimed to analyse national and district level neonatal mortality rate (NMR) from 2000 to 2013. We specifically aimed to: (a) describe national and district NMR variations over time; (b) assess NMR trends by wealth quintile and place of residence; (c) describe evolution of mortality causes; (d) assess completeness of registered mortality; (e) assess coverage and equity of NMR-related interventions; and (f) explore underlying driving factors.

To describe national time trends, we compared neonatal deaths and rates from different sources including the United Nations Inter-agency Group for Child Mortality Estimation (IGME) [8], the Institute of Health Metrics and Evaluation (IHME) [9], the Child Health Epidemiology Reference Group [10, 11], those estimated through pooled data on births and deaths by calendar year for newborns born to women interviewed in multiple Demographic and Health Surveys (DHS) [12], and official registered & corrected neonatal mortality data [13, 14]. The study period was 2000–2013 for data from IGME, IHME and for individual DHS and, 2001–2011 for official registered and corrected rates. For national level birth cohorts-based NMR, baseline and last period were 1999–2001 and 2011–2013, respectively.

For NMR equity analyses by wealth quintiles and by urban and rural domains [15], NMRs based on birth cohorts were calculated from multiple DHS for five three-year periods, from 1999–2001 to 2011–2013 [16].

For district-level analyses, NMRs were also calculated through the pooled birth cohorts’ method, but in this case for four three-year periods, from 2001–2003 to 2010–2012, to accumulate sufficient data available at that level.

Average annual rates for coverage and equity of interventions related to NMR (e.g., family planning needs satisfied, at least 4 antenatal care visits, and skilled birth attendance) and for NMR by wealth quintiles and by urban and rural residences were determined through variance-weighted regression of log-rates on year, using the whole time series from the Peruvian DHS [15]. We compared average annual rates between geographic areas and between poorest and richest quintiles by testing for interaction between year and region.

To assess the completeness of registered and corrected neonatal deaths from the official registration system, we used the Child Health Epidemiology Reference Group (CHERG) information as the reference standard at national level [11], and the most recent registered and corrected neonatal deaths available for the 2011–2012 period as the reference at district-level [17]. We correlated the district under-registration degree with the proportion of urban population estimated by the Peruvian DHS [16], and with the human development index (HDI) [9] reported by the United Nations [18].

Data related to expenditure on reproductive, maternal and neonatal interventions were derived from the Ministry of Finance official website [19]. We used currency exchange rates from the World Development Indicators database and the population estimated by the Peruvian National Institute of Statistics to calculate per capita expenditure in constant 2012 US$ [20, 21].

To account for possible driving factors that may explain the neonatal mortality reduction over the study period, we estimated crude and adjusted coefficients through simple and multiple regressions, where the dependent variable was NMR and the covariates were individual factors from the different dimensions of an explanatory conceptual framework that we designed in advance. The conceptual framework encompassed four dimensions of factors that could explain the progress achieved in neonatal mortality, namely: contextual factors (GDP per capita in USD 2012 US$, Gini for income index, percentage of families living under the poverty line, percentage of families with at least one unmet basic need, percentage or urban population, median years of women’s schooling) obtained from the annual national household surveys and from the Peruvian DHS [12, 16], non-health sector changes (total fertility rate, percentage of households with piped water inside the house, rural coverage of JUNTOS conditional cash transfer programme) obtained from the DHS and from government reports [16, 22], health-sector changes (utilisation of the Comprehensive Health Insurance System SIS-number of annual under-five outpatient preventative and clinical attendances per total under-five population, density of doctors, nurses and midwives per 10,000 population) also obtained from government reports (http://​www.​sis.​gob.​pe/​portal/​index.​html, [23]), and coverage of NMR-related interventions (percentage of women aged 15–49, either married or in union, who have their needs for family planning satisfied; percentage of pregnant women with at least four antenatal care visits; percentage of live births attended by skilled birth attendant) obtained from the Peruvian DHS [16].

We additionally included a qualitative analysis that encompassed several rounds of individual interviews and group discussions with people familiar with maternal and newborn health, complemented by a desk review. We targeted members of the academia, central and regional governments, non-governmental organisations, civil society representatives, and other stakeholders. The discussions focused on the potential underlying factors that could explain the progress in NMR reduction achieved by Peru, widely categorised in social determinants of health, non-health sector changes, health sector changes including RMNCH financing, and coverage of NMR-related interventions.

We sought to build a diagrammatic summary of the health policy and system changes that occurred in Peru, based on the information gathered, to support our analysis of contextual, policy, programmes and financing factors influencing the achieved NMR reduction.

Remarkable reductions of NMR have been achieved in Peru, with the greatest reductions in the rural areas and poorest quintiles. Reductions for NMR have been achieved within a framework of sustained economic growth and reduction of poverty, sustained political commitment and leadership, strong civil society advocacy, and of crosscutting integrated implementation of effective RMNCH interventions with a strong component of culturally appropriate antenatal and delivery practices.

In agreement with recent evidence suggesting that NMR reduction is more related to improvement in social determinants than to increase of specific interventions coverage [33], and to health system strengthening-related changes [7], our study suggests that a combination of improvement of contextual factors, non-health sector changes such as scaling-up of an antipoverty conditional cash transfer programme that incentivises health service utilisation [34], and an equitable increase in coverage of interventions may be a powerful and sustained way to reduce NMR. Improvement in social determinants and out-of-health sector factors is necessarily mediated by enabling health policy and system innovation, instrumental for implementation of comprehensive integrated maternal and neonatal interventions, including strengthening of obstetric and neonatal solving capacity of health facilities [7], to cope with complex problems such as pregnancy and delivery complications, and high risk newborns. Peru has achieved substantial progress in such efforts [7, 35], and has also been placing progressively more emphasis to attraction and retention of qualified health professionals, and to putting in place general and neonatal intensive care units across the country, although this is still largely an ongoing effort [36].

During the early 2000s, as a result of civil society demand, poverty roundtables and other institutionalised spaces of dialogue were established, which facilitated muti-sectoral dialogue and citizen participation in the design of social policies, and increased transparency, governance and accountability in the implementation and evaluation of policies and programmes. Within this context of strengthened civil society participation, an increased awareness on the importance of avoiding preventable maternal and neonatal deaths led to implementation of cross-cutting programmes such as the Strategic Maternal-Neonatal Programme, which was introduced in 2007 and scaled-up shortly afterwards (http://​www.​mef.​gob.​pe/​index.​php?​option=​com_​content&​view=​article&​id=​2144:​salud-materno-neonatal&​catid=​211&​Itemid=​101528). The sustained emphasis on a rights-based approach to pregnancy and delivery that took into account respect for cultural diversity has also surely played an important role in the progress achieved, by promoting culturally appropriate practices. Such practices include maternity waiting homes for pregnant women and vertical deliveries with the husband’s or partner’s presence in the delivery room, and are particularly important in rural areas of the Amazon and the Andes of Peru to increase utilisation of health facilities for maternal and neonatal care [28, 37‐39]. Financing also played a key role in the progress achieved in neonatal health in Peru, as is attested by an impressive increase of domestic expenditure on reproductive health from 2002 to 2013, and on maternal and neonatal health from 2000 to 2013, by about 9 and 25 times respectively.

Although Peru has decreased the coverage and NMR gap between rural and urban areas and between the richest and the poorest segments of the population, there is still a remaining gap, particularly for facility-dependent interventions such as antenatal care and skilled birth attendance. Similarly, neonates born into wealthier and urban families still show a substantial survival advantage, which should be addressed to achieve further and sustainable reductions in NMR, as a recent study performed in countries of low and middle income suggested [40].

More specifically, in Peru maternal and neonatal health interventions have been implemented as part of an integrated and synergistic approach, confirming an increasing trend within countries committed to push forward an effective convergence to reach real progress in RMNCH [41].

Even if all causes of neonatal deaths were reduced by about 50 % from 2000 to 2013, all infectious and related causes account for a quarter of all neonatal deaths, while pre-term birth complications account for about 44 % and intra-partum related events and congenital abnormalities for a third of the total neonatal mortality. It must be emphasised that deaths occurring within the first week of birth account for about three quarters of all neonatal deaths, highlighting the need to address the main causes of early neonatal mortality. Hence the Peruvian health system is facing the double challenge of successfully tackling traditionally prevalent neonatal causes of death such as infections, as well as causes needing further and substantial investment in appropriate infrastructure, equipment and qualified resources such as prematurity and intra-partum related events, for quality provision of both preventative and curative interventions.

Almost all districts reduced their NMR during the study period. Among the few that increased their rates, Lima, Arequipa and La Libertad had already reached low figures, and therefore we cannot exclude statistical fluctuations rather than real increases, given the low number of neonatal deaths in those districts. Of note, Lima is the capital city of Peru, with the highest urbanisation progress and higher access to tertiary level health facilities and highly specialised health professionals, followed by Arequipa and La Libertad, which are also highly urbanised departments. In such settings, NMR may artificially increase, due to improved mortality registration. By contrast, San Martin and Ayacucho, the other departments where NMR increased, have a substantial rural population, and the increases found are very likely showing a remaining urban-rural equity gap that needs to be urgently addressed.

The limitations of this study are that the regressions are exploratory, and therefore cannot prove causal associations to explain the factors behind the progress achieved by Peru. This is due to two main reasons, firstly neonatal mortality reduction is multi-causal where each potential factor is interrelated, and therefore quantitative approaches may not be able to fully explain this complex causality network. Secondly, the number of years for which we had data available is limited, and may account at least partially for the lack of consistent associations in the multiple regression analysis.

Complete and reliable registered data are part of the health system changes needed to effectively monitor the progress of NMR. Peru has recently achieved a substantial reduction in the under-registration of neonatal deaths, due to the implementation of a national surveillance system tracing neonatal deaths at district-level, as well as complete and correct registration [17] - although further improvement is needed, particularly in the less developed departments, which also have higher proportions of rural population. Thus we acknowledge that relying on DHS data for analysing trends and specific patterns of NMR has well known limitations [42]. We tried to overcome this drawback by assembling birth cohorts through pooling of several DHS. We foresee that future analyses using better quality of registered data will soon be possible in Peru, thanks to the described improvements in the vital registration system.

The example of Peru offers useful lessons to low and middle-income countries committed to accomplish the MDGs and post-2015 goals, specifically those related to improvement of maternal and neonatal health, which will need further and substantial reductions of NMR and also securing a healthy start for every newborn to guarantee a future free of disabilities [6].

In conclusion, Peru has achieved a substantial and equitable reduction of NMR due to a unique combination of improvement of social determinants, political will and commitment, and civil society advocacy, which translated into equitable and sustained implementation of effective maternal and neonatal health interventions. Additional efforts will be needed to close the remaining gaps posed by a double burden of neonatal deaths that encompass traditionally prevalent and more complex causes, which will need a combined implementation of primary care level and high technology interventions such as neonatal intensive care units at national and sub-national levels.

CHERG, Child health epidemiology reference group; DHS, Demographic and health surveys; HDI, Human development index; IGME, Inter-agency group for child mortality estimation; IHME, Institute of health metrics and evaluation; LMIC, Low and middle-income countries; MDG, Millennium development goal; NMR, Neonatal mortality rate