IndOSS-Assam: investigating the feasibility of introducing a simple maternal morbidity surveillance and research system in Assam, India

Statistical analyses

We compiled the information from the case notification cards and calculated the proportion of cases as a percentage of the total deliveries with 95% CIs and the case fatality rate with 95% CI for each condition in each hospital.

Setting up a steering committee

A multidisciplinary steering committee was successfully constituted in the state, including senior representatives from obstetrics, medical education, public health, medical research and university organisations, as well as individual representatives from professional specialties. A total of six steering committee meetings were held during a period of 18 months, either in person or by teleconference, including in consultation with UKOSS researchers, in order to adapt the UKOSS methodology to the context of Assam.

Establishing priorities for the region

There were 3753 maternal deaths reported through the e-MDR between April 2011 and August 2014, which is approximately half of the estimated 6333 deaths based on the MMR of 301/100 000 live births. The main causes of death were reported to be anaemia, eclampsia, sepsis, haemorrhage, obstructed labour and septic abortion (figure 3). For almost half the cases, the cause was not known, or reported as ‘others’. Almost 64% of the deaths occurred in hospitals (figure 4), which further confirmed that a hospital-based surveillance system would be able to capture a majority of the cases of severe maternal morbidity and mortality.

Request permission

Request permission

Based on the analyses of the causes of death and stakeholder consultation, six conditions were included in the pilot phase (phase-1) of IndOSS-Assam: five causes of direct maternal mortality—eclampsia, postpartum haemorrhage (PPH), puerperal sepsis, septic abortion and uterine rupture, and one suspected cause of indirect maternal death in the state—anaemic heart failure (heart failure during pregnancy and delivery as a consequence of severe anaemia19).

Mapping of reporting hospitals

The major tertiary and community government and private hospitals were identified and manually mapped to assess their geographical distribution. All 14 major tertiary government hospitals were identified (figure 5).

Request permission

Piloting the surveillance system in two selected centres

The pilot system was successfully established, with 100% of monthly reports returned to the steering committee at the Srimanta Sankaradeva University of Health Sciences for the two medical colleges during the pilot phase between March and August 2015. Paper-based case notification cards were received in a timely manner, nil-reporting ensured completeness and there were no identified problems in reporting. The reporting was overseen by either a senior obstetrician or the Head of the Department of Obstetrics and Gynaecology in each medical college.

There were a total of 10 475 deliveries in the two medical colleges participating in IndOSS-Assam in the 6 months of the pilot study; 402 cases and 66 deaths as a result of one of the six conditions under surveillance were reported (table 1). The incidence of eclampsia was 17 per 1000 deliveries (95% CI 14 to 19), PPH was 11 per 1000 deliveries (95% CI 10 to 13) and anaemic heart failure was 3 per 1000 deliveries (95% CI 2 to 5). For each of the other three conditions—puerperal sepsis, septic abortion and uterine rupture—the incidence rate was 2 per 1000 deliveries. Case fatality due to these conditions was high, 16% (95% CI 13% to 20%) of women who were diagnosed with these conditions died, the highest case fatality rate being for puerperal sepsis and anaemic heart failure (table 1).

Strengths and limitations of this study

Despite the fact that Assam has a high MMR in India, little work has been undertaken, to date, to generate good quality evidence about the incidence of severe life-threatening complications during pregnancy and childbirth, with the aim of monitoring the patterns and trends of these conditions in the state. The surveillance platform of the IndOSS-Assam can be used to generate evidence about trends in the incidence and case-fatality of important life-threatening conditions on an ongoing basis. This will help direct research and policies to improve the management of conditions with high incidence and case-fatality, in order to reduce the burden of maternal mortality and morbidity in the state. This study demonstrated the feasibility of establishing IndOSS-Assam as a hospital-based surveillance system in the state. IndOSS-Assam, being an anonymised data collection system, avoids the attribution of blame while generating robust incidence estimates to improve maternal health in the state.

Although we were able to estimate hospital-based incidence rates for six major causes of maternal mortality and morbidity in Assam, these do not necessarily reflect the population-level incidence of these conditions. The estimated incidence and fatality could be lower than the true population incidence, as not all pregnant women deliver in a health facility. It is also possible that our estimates are higher than the true population rate, since pregnant women with serious complications are more likely to be referred to the tertiary hospitals. High fatality could be due to late referral of cases, but it could also be associated with inadequate management and care during pregnancy and delivery. Nevertheless, this pilot study showed that 1 in 25 women who delivered in two tertiary hospitals (FAAMCH and GMCH) suffered from one of the six major causes of maternal deaths in the state and 1 in 130 women died as a result of these conditions. This further justifies the need for establishing IndOSS-Assam as a platform for ongoing surveillance and research in the state. Expanding the network of reporting hospitals for IndOSS-Assam to all tertiary government hospitals and major private maternity hospitals in subsequent phases of the project would ensure that information on the large majority of cases with severe complications are captured.

Findings in relation to other studies

An earlier study in the FAAMCH tertiary hospital identified 66 cases of severe maternal morbidity in 1729 deliveries between June and October 2014, of which 39% were eclampsia and 24% PPH,20 leading to an estimated incidence rate of 1.5% for eclampsia and about 1% for PPH among the total deliveries. These rates are a half to two-thirds lower than that estimated by the IndOSS-Assam active surveillance in the same hospital between March and August 2015. The maternal death reviews in the state show eclampsia, PPH and anaemia to be the major causes of maternal mortality (figure 3), however, maternal mortality rates are not calculable as denominator information about the number of women delivering is not available. The added value of IndOSS is not in estimating the number of deaths due to these known causes of maternal death in Assam, but to monitor the incidence and case-fatality of these important conditions. Allied studies aim to improve their in-hospital management, as the number of facility-based deliveries increase in the state, in order to reduce maternal morbidity and mortality. IndOSS-Assam showed a high incidence of these conditions among pregnant women who delivered in the two pilot centres as well as high fatality rates associated with anaemic heart failure and sepsis.

The million death study estimated that a quarter of the maternal deaths in India were due to PPH.21 ‘Dead women talking’, a civil society report on maternal deaths in India, identified anaemia to be the second largest cause of maternal death in India after PPH.22 Of the 22 women (of 124) who were reported to die from anaemia, 11 had cardiac failure during labour or delivery and in at least 8 of the 29 deaths from PPH, anaemia was a contributing factor.22 Another study in India reported a prevalence of 9.4% cardiac failure among women with severe anaemia.23 This is the first system in Assam to initiate data collection to estimate the incidence of maternal morbidity on an ongoing basis, including cardiac failure due to anaemia during pregnancy. Prior to this study, the incidence and case-fatality rates due to anaemic heart failure among pregnant women in Assam were not known. Through the research platform, we were also able to conduct a retrospective cohort study on the adverse maternal and fetal outcomes of anaemia during pregnancy. As mentioned above, the findings are presented separately.18

The incidence (2%) and case-fatality (12%) rates of eclampsia estimated by the IndOSS-Assam pilot study in the two medical colleges are higher than the rates reported by hospital-based studies in other states in India. A prospective study conducted in a hospital in Nagpur between January 2008 and December 2010 estimated the incidence of eclampsia to be 0.9% and case-fatality 5.5%.24 Similar rates were reported by a teaching hospital in Karnataka (incidence 1% and case-fatality 4%).25 A study in a tertiary hospital in Kolkata identified PPH to be a major cause of maternal morbidity, with an estimated case-fatality rate of 7.5%.26 While puerperal sepsis is a known cause of maternal mortality in India, we found only one study that provided estimated incidence of puerperal sepsis in six hospitals in the state of Gujarat.27 The study reported an incidence of 4%, which is higher than the 0.2% estimated through IndOSS-Assam.27 A study that reviewed the records of all deliveries in a tertiary hospital in Chandigarh over a period of 15 years (January 1988 to December 2002) found 315 cases of septic abortion among 46 417 deliveries, giving an incidence of 0.7%, which is comparable to that estimated through IndOSS-Assam in the two pilot hospitals, although the case-fatality rate in the Assam hospitals (31%) was higher than that estimated in the Chandigarh hospital (25%).28 A hospital in central India estimated an incidence of 0.06% for uterine rupture over a period of 12 years (1989–2000), with a case-fatality rate of 6%.29 It therefore appears that the incidence rates of the six conditions estimated through the IndOSS-Assam are, in general, comparable with the incidence of these conditions estimated by hospital-based studies in others parts of India, but the reported case-fatality rates in the two hospitals in Assam are much higher, further justifying the need for establishing IndOSS-Assam.

Implications and future research

IndOSS-Assam has been shown to be a feasible and simple system to initiate surveillance of the major causes of maternal death in the state, to estimate the incidence and case-fatality rates on an ongoing basis. An examination of the research platform alongside the feasibility study for setting up the surveillance system showed that, through IndOSS-Assam, detailed studies on these conditions can be conducted to identify the specific preventive and management factors that need to be improved to reduce the deaths and severe complications associated with the conditions.18 For example, the study showed that iron deficiency anaemia during pregnancy is associated with increased odds of PPH, and the risk increased by 17-fold among women with moderate-severe anaemia who underwent induction of labour and by 19-fold among women who had infection and moderate-severe anaemia.18 These findings raise important questions about how pregnant women with moderate-severe anaemia should be managed during labour and delivery. At present, there are no clear guidelines, and lessons learnt from further studies can be used to inform local and global guidance about management of labour and delivery among pregnant women with moderate-severe anaemia. The implementation and outcomes of the resultant guidelines/actions can be monitored through the research and surveillance platform of IndOSS-Assam. In addition to generating new hypotheses, the research platform will be used to investigate the gaps in management of women with specific life-threatening conditions on a rolling basis, as well as to identify good practices. The data collected through both platforms, being anonymous, avoid attribution of blame to any particular hospital or healthcare provider.

Involvement of local clinicians, policymakers, researchers, public health specialists and academics from the inception of the project to testing the system shows that there is willingness and commitment to establish and scale up IndOSS-Assam to other public and private maternity hospitals in the state. We will explore the possibility of integrating the surveillance component of IndOSS-Assam with the existing health management information system of the National Health Mission in Assam. Integration would minimise duplication of effort, enable sustainable scaling up of the surveillance platform to all government health centres in the state, streamline data flow and improve government ownership of the data, facilitating timely action. In the longer term, the IndOSS-Assam model could be used in other states in India to encourage the establishment of a national system.