The Health Extension Program and its association with change in utilization of selected maternal health services in tigray region, Ethiopia: A segmented linear regression analysis

Background: In 2003, the Ethiopian Ministry of Health established the Health Extension Program (HEP), with the goal of improving access to health care and health promotion activities in rural areas of the country. This paper aims to assess the association of the HEP with improved utilization of maternal health services in Northern Ethiopia using institution-based retrospective data. Methods: Average quarterly total attendances for antenatal care (ANC), delivery care (DC) and postnatal care (PNC) at health posts and health care centres were studied from 2002 to 2012. Regression analysis was applied to two models to assess whether trends were statistically significant. One model was used to estimate the level and trend changes associated with the immediate period of intervention, while changes related to the post-intervention period were estimated by the other. Results: The total number of consultations for ANC, DC and PNC increased constantly, particularly after the late-intervention period. Increases were higher for ANC and PNC at health post level and for DC at health centres. A positive statistically significant upward trend was found for DC and PNC in all facilities (p<0.01). The positive trend was also present in ANC at health centres (p = 0.04), but not at health posts. Conclusion: Our findings revealed an increase in the use of antenatal, delivery and post-natal care after the introduction of the HEP. We are aware that other factors, that we could not control for, might be explaining that increase. The figures for DC and PNC are however low and more needs to be done in order to increase the access to the health care system as well as the demand for these services by the population. Strengthening of the health information system in the region needs also to be prioritized. Copyright: Tigray regional state is located in the northern part of the country and has an estimated total population of 4.3 million, of whom 50.8% are female. Around 80% of the population live in rural areas and the majority of the inhabitants are orthodox Christians [23]. The region is divided into seven zones and 47 districts (weredas), of which 35 are rural. There is one specialized referral hospital in the region, as well as five zonal hospitals, seven district hospitals, 208 health centres (HCs) and more than 600 health posts (HPs). Coverage estimates from the Tigray Health Bureau (THB) indicate 75% for the first visit of ANC, 32% for skilled delivery (those attended by a nurse, midwife, health officer [non-physician clinician] or a physician), 13% for clean and safe delivery (those attended by HEWs), 51% for PNC (women who, after delivery, are assisted at health facilities by nurses/midwives or visited by HEWs at households) and 90% for family planning use [24]. The HEP was first implemented in Tigray in the middle of 2005. New health posts were built and at least one HEW was assigned to each health post. HEWs’ tasks included delivering basic curative services at the health post and promotive and preventive services in the community. HEWs may also attend deliveries, although if a complication emerges they have to refer to the health centre, which might be two to three hours’ walking distance. However, nowadays, it is mainly recommended that childbirth in Tigray should occur at least at health centre level, despite the fact that this is inaccessible to most rural women. Based on the HEP guidelines, essential supplies have been delivered to health posts (Table 1). Since maternal health was one of the major components of the HEP packages, capacity-building activities have been performed periodically. After the program starts to deliver the promotive and preventive health services, major efforts have been carried out to refill supplies, equipment and materials. Occasionally, refresher training on clean and safe delivery has also been provided to HEWs. Prior to data collection, a situational analysis was carried out to assess the availability of data through reviewing health profiles and reports and conducting discussions with concerned officials. Of the 34 rural district health offices, six were selected based on geographical distribution. When visiting the districts, documents were reviewed for completeness of the data based on a checklist prepared to answer the research question. Too many missing data was found in three of the districts and these were excluded. Due to the limited resources it was not possible to replace the excluded districts. The study districts Kilte-Awlaelo, Ganta-Afeshum and Hintalowajirat, which are located in the eastern, north-eastern and south-eastern parts of the state of Tigray, respectively (45, 120 and 35 kilometres from the capital Mekelle, respectively) were finally selected. The variables in the checklist included the name of the district and health facility, date, facility code, total population of the catchment area and type of services provided, namely: 1) antenatal care, defined as the number of women who attended a HP or HC at least once during their pregnancy; 2) delivery care, meaning the number of women assisted by skilled health personnel at health centres and clean and safe deliveries by HEWs at health posts; and 3) post-natal care, included the number of women checked by a health worker at a HC or by an HEW at a HP or during a household visit at least once during the 45-day period after delivery. The format of the registration book and the reporting template was checked for its compliance to each other, where both of the records were found to accommodate the checkups of ANC and PNC from first visit till four which possibly avoids double counting of women. Two pairs of people at each district carried out the data collection process. From the district health office, the person who was in charge of the archive office and the maternal health expert were fully engaged in sorting the relevant documents and handing over to the PI. Then, the first author and a MPH student were involved in labelling and recording the data after a proper check. The data from the period between July 2002 and June 2012 was extracted from the district offices and checked for accuracy, cleaned in an Excel spreadsheet and then exported to the Stata software for analysis (clean and raw data can be found S1 Table, S2 Table, S3 Table and S4 Table). In total, data from 16 health centres and 45 health posts was collected. The 10-year data of the three outcome variables (antenatal care, delivery and post-natal care) was aggregated on a quarterly basis. The data was analyzed separately by HC, by HP and as total. The period was subdivided into three phases, named as pre-intervention (July 2002–June 2005), immediate intervention (July 2005–June 2008) and late-intervention (after July 2008). The pre intervention phase shows the period when the program had not yet started. The proximity of health posts and the presence of HEWs were not ensured. The promotive and preventive services were neither provided at the health posts nor at household level. The data source for the period of the pre-intervention was from poorly designed health posts (built by the community) that were staffed by primary health care workers and primary midwifes who trained for six months after they accomplished high school and from few health centres and moderate number of poorly renovated clinics with less number of human resources. Immediate intervention was defined as when the HEP was initiated and late-intervention as when the HEP was consolidated in the region. The year 2005 was selected as the starting point of the intervention because of two reasons: 1) the minimum essential supplies were ensured at health posts, and 2) the health extension workers started to provide promotive and preventive services at health posts and households that year. Data were analyzed using segmented linear regression. This technique was applied to control for secular trends and to adjust for potential serial correlation of the data [25–27]. The linear regression can be represented as an equation: Where Yt is the outcome variable at time t and time is a continuous variable indicating the period from the beginning of the study up to the end of the observation. The intervention variable was coded 0 for the pre-intervention time (from July 2002-June 2005) and 1 for the post-intervention time (July 2005-June 2012). The post-slope variable was coded 0 up to the last point of the pre-intervention period (till June 2005) and coded from 1 onwards sequentially after the intervention started. In this model, B0 captures the baseline level of the outcome at time 0 (beginning of the period); B1 estimates the structural trend in utilization, independently from the intervention; B2 estimates the immediate impact of the intervention or the change in level in the outcome of interest after the intervention; B3 reflects the change in trend after the intervention; and et reflects the residual error of the calculated value. During the analysis two models were developed. Model 1 was used to estimate the level and trend changes associated with the immediate period of intervention (after July 2005). Changes related to the late-intervention period (after July 2008) were estimated by Model 2. Since autocorrelation is commonly detected in longitudinal data, the Prais–Winsten estimator was applied [25]. The study received ethical approval from the University of Mekelle, Health, Research Ethical Review Committee of the College of Health Sciences, Northern Ethiopia (approval letter ERC 0122/2012). A letter of permission was issued from the Tigray Regional Health Bureau and the district health authorities. Written informed consent from the archive officers and maternal health experts was obtained. The records was anonymized and deidentified prior to analysis.