Background: Postnatal care helps prevent neonatal deaths. This study aims to examine factors associated with non-utilisation of postnatal care in Indonesia.
Methods: A cross-sectional analytic study was conducted using data from the 2002–3 Indonesia Demographic and Health Survey (IDHS), which used multistage cluster random sampling. Contingency table and logistic regression analyses were used to determine the factors associated with non-utilisation of postnatal care services. The population attribution risk percentage (PAR%) for non-utilisation of postnatal care services was also calculated.
Results: Data were available for 15 553 singleton live-born infants. The prevalence of non-attendance at postnatal care services was consistently higher in rural areas than in urban areas. Maternal factors associated with lack of postnatal care included low household wealth index, low education levels, lack of knowledge of pregnancy-related complications or where distance from health services was a problem. Infants of high birth rank and those reported to be smaller than average were less likely to receive postnatal care. Other indicators of access to healthcare services which were associated with non-utilisation of postnatal care services included few antenatal care checks, use of untrained birth attendants and births outside healthcare facilities.
Conclusion: Public health interventions to increase the utilisation of postnatal care services should target women who are poor, less educated, from rural areas and who use untrained birth attendants. Strategies to improve the availability and accessibility of antenatal care services and skilled birth attendance including focused financial support and health promotion programmes, particularly in the rural areas, should increase utilisation of postnatal care services in Indonesia.
Statistics from Altmetric.com
Of the approximately 130 million infants born annually, four million infants die in the neonatal period, representing almost 40% of deaths of children under 5 years of age.1 More than two-thirds of neonatal deaths occur in the first 7 days of life, and of these, more than half die within the first 24 hours.2
Strategies to prevent and reduce neonatal deaths have been implemented worldwide, including the provision of postnatal care to newborns within the first 42 days of life. Postnatal care delivered through facility-based clinical care, outreach or family community care enhance both maternal and infant survival.3 Owing to the dramatically increased risk of newborn death in the first hours and the first days of life, newborns are recommended to receive postnatal healthcare immediately after delivery.4 Early care enables health professionals to identify potential complications in newborns, and to provide treatments promptly. Previous analysis of the 2002–3 Indonesia Demographic and Health Survey (IDHS) confirmed the protective effect of postnatal care on neonatal deaths.5 The risk of dying for infants who received postnatal care within the first 40 days of life was significantly lower (OR 0.63, p = 0.03) than infants who never received postnatal care services.
Despite a beneficial impact, postnatal care services are underutilised particularly among those who are in the greatest need.6 7 Although in Indonesia neonates are recommended to receive a health check between 1 and 7 days, and again between 8 and 30 days after birth, in 2005 the national coverage of postnatal care visits was only 65% and ranged from 2% to 88%.8 This indicates the need for public health interventions directed at increasing use of postnatal care services. A clear understanding of the factors associated with non-utilisation of postnatal care services is therefore important to help in the development and the implementation of evidence-based approaches to increase the use of postnatal care services. This study aims to examine the factors associated with non-utilisation of postnatal care services in Indonesia, using the 2002–3 IDHS.
Data from 29 483 ever-married women aged 15–49 years were obtained from the 2002–3 IDHS, a population-based multistage stratified sampling survey.9 This survey collected information on demographic characteristics, such as fertility and mortality rates, knowledge and practices of family-planning methods, the main child health indicators, coverage of maternity care services, men’s involvement in reproductive health and other determinants of maternal and child health in Indonesia. It was conducted in 26 of the 30 provinces in Indonesia from October 2002 to April 2003, but four provinces, Nanggroe Aceh Darussalam, Maluku, North Maluku and Papua, were excluded because of poor security. In the 26 selected provinces, the samples were stratified by urban and rural areas, and within each stratum census blocks (CBs) were the primary sampling units and were selected using systematic random sampling. Twenty-five households were then randomly selected within each CB. In total, 1592 CBs were included in this survey, with 33 419 available households. The sampling methods for this survey have been reported in detail elsewhere.9 10
In this study, the data included 15 553 singleton infants born alive between 1997 and 2002, and who survived through at least the first week of life. Figure 1 shows the conceptual framework used in this analysis, which is an adaptation of the behavioural model for use of health services, proposed by Andersen.11 The potential risk factors were categorised into five main groups: (1) external environment factors including the socioeconomic and health services characteristics of the cluster; (2) household and individual characteristics that existed before the onset of the need for postnatal care services (predisposing factors); (3) factors which would enable the infants to receive the services; (4) factors associated with adverse postnatal outcomes (need factors); and (5) previous utilisation of maternity care services.
The primary outcome of this study was the non-utilisation of postnatal care services (never- or ever-received postnatal care services). Postnatal care was any healthcare service which the mother reported was received by the infant during the first 40 days of life. A combined variable was used to describe maternal residence by region and type of residence. Three provinces, Lampung, DI Yogyakarta and North Sulawesi, were selected as the reference group because of their high rates of postnatal care attendance.
The household wealth index was obtained from IDHS and incorporated the household assets, such as the ownership of durable goods, household facilities, housing conditions and ownership of transportation devices.12 The quintiles of the sum of the weighted asset scores were used as the indicator of household wealth in this study. Birth weight was not used in this analysis because of the high proportion of cases where birth weight was not available and recorded (more than 20%).
Using contingency table analyses and multivariable logistic regression, the association between all potential risk factors and non-utilisation of postnatal care services was examined.
In the multivariate analysis, evidence of collinearity was initially assessed by examining the changes in the direction of effect between the univariate and multivariate analysis, or the identification of implausible standard errors when the variables were entered simultaneously into the model. Maternal age at child birth was omitted in this study as it was highly correlated with the infant’s birth rank.
An hierarchical modelling approach13 was applied in the multivariate logistic regression, so that the effects of the more distal factors could be assessed properly without interference from the more proximate factors. In this first model, the external environmental factors were entered, and only the significant risk factors (p<0.05) were retained. The effects of these factors were assessed prior to the inclusion of the next group of factors in the model. In the second step, the predisposing factors were entered into the first model, and their effects were assessed in the presence of the external environment factors obtained from the first model, prior to the inclusion of the enabling factors. This process was repeated for the enabling, need and previous maternal use of healthcare services factors.
ORs and 95% CIs were determined and all estimates were weighted by the sampling probabilities. The statistical analysis in this study was performed using STATA/MP version 10.0 (2007) (Stata Corporation, College Station, Texas), and survey commands were used for the logistic regression, which took into account the sampling weight and the cluster sampling design.
The study population for this analysis was infants who had survived the first week of life, because of the potential of illness leading to death to have prevented the infant from receiving postnatal care. This assumption was tested with two sensitivity analyses: the first in which only the infants who survived the neonatal period were examined, and the second in which all survivors beyond the first day of life were examined.
Selected significant risk factors were also used to calculate the population attributable risk percentage (PAR%), which estimated the contribution of the risk factor to the total risk for non-utilisation of postnatal care services.14 Using the proportion of the population exposed to the risk factor (Pe) and the OR of the risk factor, the PAR% was calculated according to the following formula:
More than half the infants born between 1997 and 2002 resided in the Java region, and approximately 56% were born to mothers aged 20–29 years (see online table). Almost 60% of the deliveries occurred outside the healthcare facility, although around 66% of all the deliveries were assisted by trained birth attendants. Overall, 67% of the infants born within this period received postnatal care, of which 94% (95% CI 93.2 to 95.4) received it in the first week of life.
At the community level, the prevalence of non-utilisation of postnatal care services was consistently higher in the rural than in the urban areas for each region (see online table). At the household and individual level, low attendance at postnatal care services was associated with high birth rank, low household wealth index, poor maternal knowledge of complications during pregnancy, childbirth or postnatal period and no maternal exposure to mass media such as newspapers, radio or television.
The multivariate logistic regression analysis showed that four factors were strongly associated with non-utilisation of postnatal care services: region and type of residence, percentage of infants receiving four or more antenatal care checks in the cluster, household wealth index and the use of untrained delivery attendants. This analysis demonstrated that significant differences in the utilisation of postnatal care services across regions persisted after adjusting for covariates. Compared with infants from Lampung/DI Yogyakarta/North Sulawesi, infants from the rural areas of Java and Bali/Nusa Tenggara Barat (NTB)/Nusa Tenggara Timur (NTT) were less likely to receive postnatal care. Among mothers who lived in rural areas of Java, 47% of deliveries were not assisted by trained birth attendants, compared with only 25% in urban Java areas. A higher proportion of deliveries assisted by untrained attendants were also found in rural areas (55%) of Bali/NTB/NTT than in urban areas (21%). A summary variable examining type of residence (urban/rural) found that compared with infants from the urban areas, rural infants had an increased risk for not utilising postnatal care services (OR 2.00, 95% CI 1.54 to 2.60, p = 0.00). Similar findings were observed concerning maternal educational attainment, and household wealth index which favoured those who lived in the urban areas.
Poor attendance at postnatal care services was related to poor utilisation of antenatal care services. A progressive decrease of postnatal care non-utilisation was observed as the percentage of infants receiving more than four antenatal care services in the cluster increased. The odds of not attending postnatal care services increased as the household wealth index decreased. Infants whose mother had less than secondary level education were also less likely to attend postnatal care services. Maternal exposure to mass media also emerged as a significant predictor. There was a significantly increased odds of not using postnatal care services among mothers without any knowledge of pregnancy-related complications. With increasing birth rank, the likelihood for not attending the postnatal care services also increased.
Among the enabling factors, the odds of not using postnatal care increased with mothers who admitted that the distance to the health facility was a major problem. Moreover, from those who considered the distance as a major problem, approximately 82% (95% CI 77.2 to 86.3) lived in the rural areas. Surprisingly, infants whose mothers intended to become pregnant at the time of pregnancy were more likely not to utilise postnatal care services.
The odds of not using postnatal care were significantly higher for smaller than average-sized infants by around 31% than for larger than average-sized infants. The finding was similar when the birth size variable was replaced by birth weight. Although not significant, low birthweight infants (<2500 g) were also found to be more likely not to receive any postnatal care (OR 1.41, 95% CI 0.93 to 2.14, p = 0.10) than infants who weighed more than 3500 g. As expected, infants who were not weighed had highly significant odds for not receiving healthcare services (OR 8.28, 95% CI 6.14 to 11.16, p = 0.00).
Among the indicators of previous maternal exposure to healthcare services, a highly significant effect was observed in infants delivered by untrained birth attendants. The odds of not utilising postnatal care services for infants delivered by untrained births attendants increased remarkably by almost 31 times the odds for infants delivered by doctors. Additionally, the likelihood for not receiving postnatal care also increased among infants born outside of healthcare facilities.
In the sensitivity analyses, the three models found similar factors and strength of association with all neonatal deaths, neonatal deaths excluding those on day 1, and neonatal deaths excluding those on days 0–6 (data not shown). In this study, we omitted 175 deaths that occurred in the first week (0–6 days) of life, of which 43% (95% CI 31.6 to 55.5) did not receive postnatal care.
Table 1 shows the crude and adjusted PAR% for a selection of risk factors. Out of the total risk in this population for not attending postnatal care services, 92% was attributable to infants who were delivered by untrained birth attendants. It was also revealed that the adjusted PAR% progressively increased with the reduction of household wealth index, to around 47% in the lowest quintile, and compared with the wealthiest households the combined PAR% for household wealth index was 72%.
This study has demonstrated a significant influence of environmental, predisposing, enabling, need and previous exposure to other healthcare services factors on non-utilisation of postnatal care services in Indonesia. We found four factors to be strongly associated with non-utilisation of postnatal care services: region and type of residence, percentage of infants receiving four or more antenatal care checks in the cluster, household wealth index and the use of untrained delivery attendants. These findings demonstrate a close linkage between access to antenatal care, skilled birth attendance and subsequent use of postnatal services. The identification of these factors will be important in developing public health policies and interventions aimed at increasing the utilisation of postnatal care services.
An important strength of this study is that the analyses used a nationally representative survey with high individual (98%) and household (99%) response rates.9 To reduce recall bias by the interviewed mothers, the data analysed were restricted to infants born in the 5-year period preceding the survey.15 16 17 A wide range of different factors were examined with a hierarchical approach to provide better estimates of the effect of the factors associated with not utilising postnatal care.
The study does have several limitations that should be considered when interpreting the findings. The selection of factors examined was driven by their availability in the IDHS dataset. As a consequence, this analysis did not include several possible determinants such as the content of postnatal care services and the users’ satisfaction with services. It was also not possible to examine the effects of the utilisation of antenatal care services on the utilisation of postnatal care services at the individual level, since this information was only available for the most recent delivery of a mother. Some factors in the analysis were not infant specific, such as parental occupation, which covered the parent’s employment over the 12 months preceding the survey.
Our analysis showed highly significant associations between the pregnancy and delivery healthcare services and the utilisation of postnatal care services. Infants delivered outside a healthcare facility and by untrained birth attendants were significantly less likely to utilise postnatal care services. This finding might reflect lack of access and availability of local delivery healthcare services or maternal choice. The high PAR% of these factors indicates a strong need to improve the access, the coverage and acceptability of all maternity care services and increase community awareness about the importance of delivery at healthcare services. Moreover, strategies to educate untrained birth attendants, particularly the traditional birth attendants, about the importance of postnatal care services may help to improve the uptake of these services. The involvement of local community health workers and traditional birth attendants within the healthcare system in India18 and Bangladesh19 has been shown to have a significant impact on neonatal health.
Another indicator of health service access and availability was mothers’ attendance for antenatal care. Antenatal care counselling could improve mothers’ awareness and knowledge of the importance and the availability of postnatal care services, and motivate them to utilise postnatal care services.20
Household economic status also emerged as a strong predictor of utilisation of postnatal care services. A progressive increase in the odds of not using postnatal care services was found as the household wealth index decreased. Previous literature has identified a significant role of household economic status on the utilisation of health services, and our findings also reflected a similar association with low household economic status and poor uptake of postnatal care in Indonesia.21 22 23 The high PAR% among the poorer households underlines the importance of developing a healthcare-financing mechanism that could enable women from these households to better access perinatal healthcare services. Conditional cash transfer schemes, which have a positive effect on the utilisation of health services in low-income and middle-income countries, might be an approach to targeting women from poorer households to increase their use of postnatal care services and other pregnancy and delivery healthcare services.24 25 However, this intervention should be complemented by strategies directed at raising community awareness of the importance of different pregnancy and delivery care services.
As found in other countries, infants born in rural areas consistently had higher odds for not using these services, which might be due to the lack of these services or reduced access to healthcare services from the inadequate transportation infrastructure.26 Additionally, lower maternal educational attainment and the household wealth index, which have been shown in our study to be associated with lower postnatal care uptake, may have contributed to the underutilisation of postnatal care services in rural areas. As found in our study, infants born to mothers who considered that distance to the healthcare facility was a “major problem” were less likely to receive postnatal care. A combination of outreach, family community and facility-based clinical care needs to be applied to ensure better access to postnatal care services, particularly in the rural areas of Indonesia.3
An unexpected finding was that smaller than average-sized infants were less likely to receive postnatal care. This might be attributable to the perceived vulnerability of small babies, which therefore led their parents to delay visiting the clinic-based postnatal care services. Home-based postnatal care services might help provide essential newborn care to these high-risk infants. Another unexpected finding was that women who wanted the baby at that time were less likely to use postnatal care than women who wanted to have the baby later. The reasons for this finding were not clear; however, it was not confounded by household wealth index or maternal educational level as these potential confounders were adjusted in the multivariate analysis model.
In addition to social disadvantage reflected by household wealth index, we also found that limited maternal education, lack of exposure to information and lack of health knowledge about pregnancy were significantly associated with non-utilisation of postnatal care services. This finding is supported by other studies reporting women’s knowledge and educational attainment as important factors for utilisation of healthcare services.27 28 Education increases women’s knowledge of the importance of health, empowers them, facilitates their healthcare seeking behaviour, increases their confidence and improves their ability to seek the most appropriate healthcare services.29
We found that infants of high birth rank were less likely to utilise postnatal care services. The prior experience of these women with birth might have influenced their perceptions about the need for this type of healthcare service. The decline in health services utilisation among higher birth rank infants has also shown in Peru27 and the Philippines.30 Household economic constraints and lack of time as a result of having many other children might be other reasons for the low utilisation among these groups.6 30 31 These findings underlined the importance of raising community awareness through health promotion strategies targeting vulnerable groups to increase their utilisation of postnatal care services.
This study shows the significant role of external environment, household and individual level factors in non-utilisation of postnatal care services in Indonesia. The estimates of PAR% revealed the potential for comprehensive public health interventions, particularly strategies that target mothers from poorer households. Focused financial support is required for mothers from economically disadvantaged households to reduce the inequitable access to pregnancy and delivery healthcare services with trained healthcare providers in Indonesia. This intervention should be complemented with health promotion programmes to increase community awareness of the benefit of pregnancy and delivery healthcare services and efforts to improve access to primary care services, especially for the rural and remote communities.
What is already know on this subject
Infants receiving postnatal care are significantly protected against neonatal death.
Postnatal care services are under-utilised particularly among those who are in the greatest need.
The national coverage of postnatal care visits in Indonesia in 2005 was only 65% and ranged from 2% to 88%
What this study adds
Evidence of a strong association between not attending postnatal care services and decreasing household wealth.
Mothers from the poorest households were five times more likely not to utilise postnatal care services than those from the wealthiest households.
Evidence of a highly significant influence of type of delivery attendance on non-utilisation of postnatal care services.
Being delivered by untrained births attendants increased the odds of not receiving postnatal care by 31 times compared with infants delivered by doctors.
▸ An additional table is published online only at http://jech.bmj.com/content/vol63/issue10
Funding We are indebted to the Australian Agency for International Development (AusAID) for funding CRT PhD scholarship in International Public Health at the University of Sydney, Australia, and this analysis is part of CRT thesis to fulfil the PhD requirement. CLR is supported by a NHMRC Research Fellowship.
Competing interests None declared.
Provenance and Peer review Not commissioned; externally peer reviewed.
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.