Health postservice readiness and use of preventive and curative services for suspected childhood pneumonia in Ethiopia: A cross-sectional study

Objective Pneumonia is the single-leading cause of infectious disease deaths in children under-5. Despite this challenge, the utilisation of preventive and curative child health services remains low in Ethiopia. We investigated the association between health post service readiness and caregivers’ awareness of pneumonia services, care-seeking and utilisation of pneumonia-relevant immunisation in four Ethiopian regions. Design and setting This cross-sectional study was conducted in 52 districts of four regions of Ethiopia from December 2018 to February 2019. The health posts preparedness for sick child care was assessed using the WHO Health Service Availability and Readiness Assessment tool. Multilevel analyses were employed to examine the associations between health post readiness and household-level awareness and utilisation of services. Participants We included 165 health posts, 274 health extension workers (community health workers) and 4729 caregivers with 5787 children 2-59 months. Outcome measures Awareness of pneumonia treatment, care-seeking behaviour and coverage of pentavalent-3 immunisation. Results Only 62.8% of health posts were ready to provide sick child care services. One-quarter of caregivers were aware of pneumonia services, and 56.8% sought an appropriate care provider for suspected pneumonia. Nearly half (49.3%) of children (12-23 months) had received pentavalent-3 immunisation. General health post readiness was not associated with caregivers’ awareness of pneumonia treatment (adjusted OR, AOR 0.9, 95% CI 0.7 to 1.1) and utilisation of pentavalent-3 immunisation (AOR=1.2, 95% CI 0.8 to 1.6), but negatively associated with care-seeking for childhood illnesses (AOR=0.6, 95% CI 0.4 to 0.8). Conclusion We found no association between facility readiness and awareness or utilisation of child health services. There were significant deficiencies in health post preparedness for services. Caregivers had low awareness and utilisation of pneumonia-related services. The results underline the importance of enhancing facility preparedness, providing high-quality care and intensifying demand generation efforts to prevent and treat pneumonia. The Ethiopian primary healthcare system typically consists of a primary hospital, a health centre and five satellite health posts. A health post is the lowest service delivery point staffed by two health extension workers serving around 5000 rural residents. Since 2003, Ethiopia has implemented the health extension programme to achieve universal coverage of primary healthcare for the rural population. This national programme is implemented by health extension workers, and they provide basic promotive, preventive and curative services through outreach and health post-based approaches. In 2010, after a change in policy that allowed the health extension workers to treat child pneumonia, the Ethiopian Ministry of Health and partners initiated the implementation of iCCM of childhood illnesses as part of the health extension programme. Under the iCCM programme, the health extension workers examine, classify and treat pneumonia.34 35 The OHEP intervention had three components, that is, community engagement activities, capacity building of health extension workers and women’s development group leaders, and strengthening of the district health services’ ownership and accountability of the primary newborn and child health services. The intervention was conducted in 26 intervention districts with 26 comparison districts spread in four regions of Ethiopia, namely Tigray, Amhara, Oromia and Southern Nations, Nationalities and Peoples Regions. The intervention started in 2016 and was completed in 2018. For the evaluation, baseline and end line surveys were performed. This paper reports a secondary analysis of end line cross-sectional data.33 This study included all caregivers and children aged 2–59 months, who resided in 52 study districts. It also includes all health posts with one or two health extension workers per health post serving these families. This study was based on secondary analysis of data from the endline survey that was part of the evaluation of the OHEP intervention. Sample size for the end line survey was estimated to measure changes in care-seeking and appropriate treatment for childhood illnesses between intervention and comparison areas at baseline and endline. Assumptions considered for the sample size calculation for the main OHEP evaluation36 were 80% power to detect differences of 15% for the reported level of care-seeking (55%) and 20% for appropriate treatment for childhood illnesses (47%) at the baseline, design effect of 1.001% and 90% completeness. Accordingly, a sample size of 12 000 households was obtained. With this number of households, 6532 children below the age of 5 years were expected to be reached by the survey, of whom 368 were assumed to have any illnesses and 308 to have suspected pneumonia within 2 weeks before the survey. The parent study used a sampling frame generated based on the 2007 Ethiopian Central Statistical Agency housing and population survey. Two hundred enumeration areas, that is, clusters, were selected from 52 study districts with probability proportional to size. A two-staged cluster sampling followed by systematic sampling to select 60 representative households from each study cluster. All caregivers of children aged 2–59 months living in the selected households were interviewed. Moreover, all health posts and one or two health extension workers from each health post serving the population of the study clusters were included.37 Data were collected using structured and pre-tested interviewer-administered questionnaires through an electronic data collection software (CSpro survey software). The questionnaires were translated into local languages (Amharic, Tigrigna and Oromiffa) and back-translated into English. They comprised three main modules; household, healthcare provider and health facility modules (see online supplemental files 1–3). Data collectors and supervisors, who had bachelor’s degree or above, were trained for 2 weeks on data collection techniques, procedures, quality assurance and ethical considerations of the study. Further detailed information about data collection and quality control is available in the published study protocol.33 bmjopen-2021-058055supp001.pdf bmjopen-2021-058055supp002.pdf bmjopen-2021-058055supp003.pdf The outcomes of this study are caregivers’ awareness of pneumonia treatment, care-seeking behaviour and coverage of pentavalent-3 immunisation as defined in the Demographic and Health Surveys.28 The awareness of availability of pneumonia treatment was calculated as the proportion of caregivers who had heard messages regarding pneumonia treatment. Suspected pneumonia was ascertained by asking the caregiver if the child had cough combined with either fast or difficult breathing due to chest problems within 2 weeks before the survey. Care-seeking was defined as children with suspected pneumonia for whom advice or treatment was sought from an appropriate care provider, that is, either government or private providers. The vaccination status of children aged 12–23 months was primarily assessed by reviewing immunisation cards. When cards were not available at home, the caregivers were requested to report the type of vaccines their children had received. Hence, coverage of pentavalent vaccination was estimated as the proportion of children 12–23 months who had received three doses of pentavalent vaccine. The readiness of health posts for sick child care was assessed using the WHO Service Availability and Readiness Assessment tool.38 Using 23 tracer items, the preparedness of facilities was shown in five domains or indices, that is, basic amenities, basic equipment, standard precaution equipment for infection prevention, diagnostic capacity and essential medicines. The mean availability of items across the four domains of readiness was estimated by assigning equal weight to each of the items, and was expressed as a proportion. Health posts’ diagnostic capacity was shown as the proportion of facilities having rapid diagnostic test for malaria. The general service readiness was calculated as the average of percentages depicting mean availability of tracer items in five indices.38 The availability of vaccination card at the health posts was also estimated. The number of health extension workers working at the health post and the percentage of these workers trained in iCCM and who had received supportive supervision during 6 months before the survey were also calculated. The health post demand generation activities were recorded as the proportion of health posts showing opening days or that used community forums to deliver maternal and child health education. The household socioeconomic status was constructed through principal component analysis of household assets, ownership of house, livestock, agricultural land and access to utilities and infrastructures. The factor scores were summed and ranked into quintiles from the poorest to the least poor. The study linked the household, health facility and care provider information. Our analysis was based on linked samples for outcome variables, that is, caregivers’ awareness of pneumonia treatment (n=4934), care-seeking when the child was sick (n=613) and vaccination of 12–23 months old children with a third dose of pentavalent immunisation (n=860). Care-seeking was assessed for all childhood illness episodes, including symptoms of suspected pneumonia as reported by caregivers for the 2 weeks prior to the survey. The effect of clustering on three of the study outcomes was examined by estimating intracluster correlation coefficients (ICC). A multilevel binary logistic regression model was fitted to examine the association between health post readiness and household level awareness, care-seeking and utilisation of three doses of pentavalent vaccinations. We checked for potential household-level confounders. The fitness of the model was checked through Likelihood Ratio Test. Data were analysed using Stata V.14. Patients or the public were not involved in the design or conduct or reporting or dissemination plans of this research.