A prospective cohort study was conducted in a Tertiary care hospital, in North India from October 2015 to December 2016, to determine the causes of MNM.

The cases which met the criteria of MNM as per the Maternal Near Miss Operational Guidelines, [4] were included for audit by the Maternal mortality review committee along with maternal death cases. The Institutional Ethical Committee approved the study (IEC/VMMC/ SH/Thesis/October/2015). All the participants signed the consent form (Hindi or English).

During the study period, all the women who met the criteria for MNM were identified at the time of discharge and enrolled in the study. Maternal mortality cases were identified as per WHO definition of maternal death [8] from amongst all the female deaths in the hospital.

Facility based MNM Review form and Facility based maternal death form were filled for MNM cases and maternal deaths, respectively by the treating doctors. Patient characteristics including age, parity, gestational age at the time of admission, booked (at least 3 antenatal visits at the institute), mode of delivery, ICU admission, lifesaving intervention and foetal outcome were analysed for both the groups. Caregivers were interviewed and medical records reviewed to identify the gaps in both the groups by the members of the departmental committee members. In some near miss cases, women were also interviewed. Audit was conducted for both by the departmental committee, during the weekly and monthly meet.

Patients were categorised by final diagnosis with respect to haemorrhage, hypertension, sepsis, dystocia (direct causes). Anemia, thrombocytopenia, and other medical disorders were considered as indirect causes contributing to MNM and deaths.

The above data was analysed to calculate the following near miss indices [9]:

Data entry was done on Microsoft Excel spreadsheet and it was analysed using Statistical Package for Social Sciences (SPSS) version 21.0. For qualitative data, proportions were calculated. Mean score was calculated for quantitative data. Test of significance of differences between proportions and mean were calculated. Qualitative data was analysed by Chi-square test and t-test was applied for quantitative data. p-value <0.05 was considered significant.

During the study period, there were 31,925 deliveries at the Institute. Total live births, maternal mortality and near miss cases are shown in [Table/Fig-1].

[Table/Fig-2] shows the characteristics of women in both the groups. Mean age of the women in both the groups were comparable, majority of the women were in their twenties. More number of women with teenage pregnancies died than survived (12.2% versus 5.6%). Most of the women were unbooked. Many women had more than one level of delay. Delay in women seeking care (level 1) was most common in both the groups. A high proportion of women also experienced a delay in referrals to the current Health facility (Level 3 delay), 70% and 74.8%, respectively. More number of women in the maternal death group (34.3% versus 27.7%) presented to the hospital with postpartum complications, however, the difference was not statistically significant. Many had undergone LSCS at another health facility and referred with multiple complications to the present facility (27.5% and 37.8%, in MNM and MM group, respectively).

A high proportion of women in MNM group had live births (61.8%) compared to MM group (38.9%, p-value <0.001). However, the incidence of Intra-uterine fetal deaths were comparable in both groups (23.4% vs 29.0%).

Most common causes of MNM in the study were haemorrhage (53.8%) [Table/Fig-3]. Majority of Maternal deaths were due to direct causes (71.8%). The most common causes were hypertensive disorders of pregnancy (31.2%). In the MNM group, admission to ICU was the most common intervention (63.9%), [Table/Fig-4,5 and 6].

Over the last decade; there is a gaining momentum to use MNM as an indicator of obstetric care. Several studies have been conducted to audit the MNM cases [10-27] both in India and across the globe. However, unlike maternal death, identifying MNM is complex. Different criteria have been used in the past to define MNM, like disease specific criteria, management specific criteria or organ system dysfunction criteria. WHO have defined near miss using organ system dysfunction criterion. While, many authors have used the WHO Criteria for near miss audit [Table/Fig-7], some had to make minor changes to this approach [14,16,23, 24] as the need was felt to modify according to the lack of some facilities at their centre or an underestimation of near miss cases was perceived using this approach.

Moreover, many previous studies are retrospective [11-15] or have a small sample size [10,12,17,23,24], making it difficult to arrive at a statistically powered conclusion. This study aimed to determine the incidence of MNM cases as defined in Maternal Near Miss Review Operational Guidelines, released by the MOHFW, Government of India in 2014 [4]. For identification of an MNM case according to this criterion (minimum three, one from each category) must be present:

Clinical findings (either symptoms or signs), Investigations, Interventions or any single criteria which signifies cardio respiratory collapse from different categories of various adverse events and disorders associated with them. Very few Indian studies have used the MOHFW criteria for MNM [10] to audit these cases.

The strength of the current study is that it has a large sample size, is prospective and is well-designed, also comparing the socio-demographic profile of MNM cases with maternal deaths. Roopa PS et al., [10], was a multi-centric, pilot study using the Operational guidelines by MOHFW [4], however, no comparison was made with maternal mortality cases.

The MNM Incidence Ratio in this study (0.8%) was lowest except for Gupta S et al., [Table/Fig-7] [17]. This difference can be explained due to different settings in which the studies were conducted. The latter caters to a selected population of medically insured patients who are often booked and the hospital is not open to general public. While the hospital in the current study is a tertiary care centre, receiving unbooked and complicated cases from across the state and the neighbouring states. Disease specific criterion as used by some studies [11,13], is known to report a higher no. of cases as near miss. Many studies have used the WHO criterion [9] or their modifications [Table/Fig-7] which may explain the difference in the incidence ratio. WHO criterion is organ system dysfunction based, while, the Indian guidelines incorporates all the 3 criterion; disease based, management based and organ system dysfunction based (1 from each category is required to label as near miss, unless there is evidence of cardiovascular collapse).

The low MNM/MD ratio in this study [Table/Fig-7], may be wrongly interpreted as a poor quality of care, but it actually is due to strict adherence to Indian guidelines and also, most of these patients were unbooked (91.2-93.9%) and referred (71.4-74.8%) to the centre after multiple complications.

In spite of the heterogeneity in the study design and the setting, most of the studies like the current study have reported haemorrhage followed by hypertensive disorders as the most common cause of MNM, while hypertensive disorders was the leading cause of maternal deaths [Table/Fig-7]. Even studies from developed countries like Australia and Netherlands have reported obstetric haemorrhage as the most frequent cause of severe acute maternal morbidity [20,21].

A detailed analysis of women in haemorrhage group reveals that most of the critically bleeding women were in the postpartum phase with PPH (39.3%). This observation is similar to other Indian studies by Roopa PS et al., and an Australian study by Jayaratnam S et al., which have also observed PPH as the most common cause of MNM [10,20].

In the hypertensive group, there were higher proportion of eclampsia versus pre-eclampsia (13.2 and 8.4%, respectively), hence reaffirming that timely initiation of magnesium sulphate therapy and termination of pregnancy can salvage these women.

Obstetrical sepsis (54.2%) and medical disorders like cardiac disease (53.6%), continue to pose a major threat to the health of the women. They have the highest mortality index [Table/Fig-5] also reported in other studies [Table/Fig-7]. Puerperal sepsis was most common (6.8%) cause of sepsis. This is in spite, of launch of Janani Suraksha Yojana (JSY), a cash incentive scheme launched by the Government of India to promote institutional deliveries [7,28]. Though there has been an increase in the institutional deliveries to 84.44% as per the recent National Family Health Survey-4 (NFHS 4 2015-16) in Delhi state, the increased load to health facility may compromise the quality of care due to limited infrastructure or trained health personnel [29].

Indirect causes (medical disorders) too pose a major challenge to the health of the women. A multi-disciplinary approach and timely detection and referral of the high risk women to a tertiary care centre, through good quality antenatal coverage and referral linkage, may help optimise the health of these women.

In this study, a significant association was demonstrated with lower socioeconomic status and illiteracy and occurrence of maternal mortality, (p<0.0001, p<0.01, respectively). Also, the most common delay identified were type 1 delay (delay in the decision to seek care; 67.9-77.9%) and type 2 delay (delay arrival at the health facility;70-74.8%) [26]. Thus, highlighting the role of social factors and cost in a women’s health and how education may play a role in changing attitudes and promoting a health seeking behaviour.

The interview of all the maternal near cases and verbal autopsy for maternal death cases could not be done.

The MNM IR in this study was 8/1000 live births and the MMR was 421/100,000 live births. The MNM: MM ratio was 1.9:1. Overall Mortality Index (MI) was 34%. Severe Maternal Outcome Ratio (SMOR) was 12.2/1000 live births. The most common cause of MNM was haemorrhage, mostly, post-partum haemorrhage. Lower socio-economic status and illiteracy were associated with occurrence of maternal mortality. The most common cause of maternal deaths were hypertensive disorders of pregnancy. Hence, from the present study it is concluded, that mothers may benefit by upgradation of the infrastructure of the peripheral health centres (like ensuring availability of blood banks, round the clock operation theatre facility, magnesium sulphate for seizure prophylaxis etc.) along with a network of referral linkage to ensure speedy and appropriate referrals. Inspite of a high percentage of institutional delivery, obstetrical sepsis and medical disorders continue to challenge the health system as they have the highest mortality index.