Maternal mortality has fallen worldwide, but it is still high. The global maternal mortality ratio fell from 385 per 100,000 live births in 1990 to 216 in 2015, and about 303,000 women died in 2015 [1]. Most deaths occur during or soon after childbirth [2,3]. Haemorrhage, hypertensive disorders and sepsis cause more than half of all maternal deaths, and indirect causes account for more than a quarter [4].

As maternal deaths become fewer in a single hospital, they become less useful for judging the quality of care. Women who survive a life-threatening complication share the same causes and the same failures of care as those who die [5]. Mantel and colleagues showed that severe acute maternal morbidity was nearly five times as common as maternal death and reflected its pattern [5]. Early studies used different definitions, so rates could not be compared [6]. In 2009, the World Health Organization (WHO) defined a maternal near-miss case as a woman who nearly died but survived a complication during pregnancy, childbirth or within 42 days of the end of pregnancy, and it published a single set of identification criteria based on organ dysfunction [7,8].

The WHO criteria include clinical, laboratory and management markers. They identified all maternal deaths and almost all cases of organ failure in a Brazilian validation study [9]. In the WHO Multicountry Survey of 314,623 women in 29 countries, 1.0% had a severe maternal outcome, and postpartum haemorrhage and pre-eclampsia were the commonest underlying conditions [10]. Reported near-miss rates still vary widely between settings [11]. The laboratory-based criteria may miss cases where investigations are limited, as in Malawian district hospitals [12], and may also miss cases in high-income settings [13].

Near-miss review also shows where care fails. The three-delays model describes delay in deciding to seek care, in reaching a facility and in receiving adequate treatment [14]. In a Brazilian multicentre study, any delay was found in 52% of potentially life-threatening conditions, 68.4% of near-miss cases and 84.1% of maternal deaths [15]. WHO and Indian professional bodies recommend near-miss audits for improving the quality of care [16-18].

Indian hospital studies have reported near-miss ratios between 7.56 and 23.85 per 1000 live births [19-21], with haemorrhage, hypertensive disorders, anaemia and sepsis as the main causes [19,22,23]. Most of these studies came from large government medical colleges or university hospitals, and many were retrospective. Data from teaching hospitals that receive referrals from a district are limited. The aim of this study was to describe the causes of maternal near miss and maternal death, the WHO outcome indicators, the WHO criteria met, delays in care and perinatal outcomes at a referral hospital serving Jalgaon district in Maharashtra.

Study design and setting This prospective observational study was conducted in the Department of Obstetrics and Gynaecology, Dr. Ulhas Patil Medical College & Hospital, Jalgaon, Maharashtra. The hospital is a tertiary care teaching hospital that receives obstetric referrals from primary health centres, rural hospitals and private facilities in Jalgaon district and neighbouring areas. It has round-the-clock obstetric, anaesthesia and blood bank services, and an intensive care unit. Written informed consent was obtained from each woman, or from her relative when she was too ill to consent. Participants Women admitted to the department who were pregnant, in labour or within 42 days of the end of pregnancy (delivery, abortion or ectopic pregnancy) were screened daily in the labour room, wards, operation theatre and intensive care unit. All women who met at least one WHO near-miss criterion, and all maternal deaths, were included consecutively. Women who were admitted after 42 days of the end of pregnancy and women with incidental or accidental conditions not related to pregnancy were excluded. Definitions A maternal near-miss case was defined as a woman who nearly died but survived a complication that occurred during pregnancy, childbirth or within 42 days of the end of pregnancy [7]. Women were classed as near-miss cases if they met at least one of the WHO near-miss criteria [8,10]. These were clinical criteria (for example, shock, cardiac arrest, gasping, respiratory rate above 40 or below 6 per minute, oliguria not responding to fluids or diuretics, clotting failure, loss of consciousness for 12 hours or more, stroke, uncontrollable fit or total paralysis, and jaundice in the presence of pre-eclampsia), laboratory criteria (for example, oxygen saturation below 90% for 60 minutes or more, PaO2/FiO2 below 200 mmHg, creatinine 300 µmol/L or 3.5 mg/dL or more, bilirubin above 100 µmol/L or 6.0 mg/dL, pH below 7.1, lactate above 5 mmol/L, acute thrombocytopenia below 50,000/µL, and loss of consciousness with glucose or ketoacids in the urine) and management criteria (continuous use of vasoactive drugs, hysterectomy for infection or haemorrhage, transfusion of 5 or more units of red cells, intubation and ventilation for 60 minutes or more not related to anaesthesia, dialysis for acute renal failure, and cardiopulmonary resuscitation). A maternal death was the death of a woman during pregnancy or within 42 days of its end from any cause related to or aggravated by the pregnancy or its management. A severe maternal outcome was either a near miss or a maternal death. The underlying cause of each severe maternal outcome was classed as haemorrhage, hypertensive disorder, pregnancy with abortive outcome, sepsis, other direct cause or indirect cause. Delays were classed with the three-delays model [14]: delay in deciding to seek care (phase 1), in reaching the hospital (phase 2) and in receiving adequate care after arrival (phase 3). Data collection Each near-miss case and maternal death was recorded on a structured form based on the WHO near-miss tool. The form recorded age, parity, residence, education, socioeconomic status, antenatal care, referral status and distance travelled, gestational age, mode of delivery, the near-miss criteria met, underlying cause, interventions (blood transfusion, magnesium sulphate, uterotonics, antibiotics, surgery and intensive care), delays, duration of hospital stay and perinatal outcome. Information on delays was taken from the woman or her relatives and from referral notes. Each woman was followed until discharge or death and, for perinatal outcome, until 7 days after birth. The number of live births and obstetric admissions in the same period was taken from the labour room and ward registers. Outcome indicators Two WHO outcome indicators were calculated from the severe maternal outcome cohort [7]. The maternal near-miss to mortality ratio was the number of near-miss cases for each maternal death. The mortality index was maternal deaths / (near-miss cases + maternal deaths) × 100. The mortality index was also calculated for each underlying cause; a higher index was taken to mean poorer care for that condition. Perinatal mortality was the number of stillbirths and early neonatal deaths per 100 babies born at 20 weeks or more. Statistical analysis Data were entered in Microsoft Excel and analysed with IBM SPSS Statistics. Categorical variables are presented as numbers and percentages and continuous variables as mean and standard deviation (SD) or median and interquartile range (IQR). The 95% confidence interval (CI) of the mortality index was calculated with the Wilson method. Near-miss cases and maternal deaths were compared with the chi-square test, or the Fisher exact test when an expected count was below 5, and with the independent t test or Mann-Whitney U test for continuous variables. A p value below 0.05 was considered significant.

Severe maternal outcomes and indicators

A total of 120 women had a severe maternal outcome: 105 near-miss cases and 15 maternal deaths. The near-miss to mortality ratio was 7.0:1, and the mortality index was 12.5% (95% CI 7.7 to 19.6) (Table 1).

Table 1. WHO outcome indicators of the severe maternal outcome cohort

Indicators defined according to the WHO near-miss approach. CI: confidence interval (Wilson method).

Characteristics of the women

The mean age of the women was 26.7 ± 5.4 years (range 18 to 43), and 19.2% were younger than 20 or older than 35 years. Half lived in rural areas, 14.2% were primigravidae and 45.0% had fewer than four antenatal visits. Most women (86, 71.7%) were referred from another facility; the median distance travelled was 50 km (IQR 21.8 to 91.8), and half travelled more than 50 km. The mean haemoglobin at admission was 9.2 ± 2.0 g/dL, and 10.8% had haemoglobin below 7 g/dL. The event occurred after delivery or during labour in 83 women (69.2%), before delivery in 27 (22.5%) and after abortion or ectopic pregnancy in 10 (8.3%). Among those who delivered, 43 had a caesarean and 57 a vaginal delivery, and 20 women were not delivered. Women who died more often had their event before delivery (53.3% vs 18.1%, p = 0.005); the overall distribution of timing differed between the groups (p = 0.007). Age, residence, parity, antenatal visits, referral status, distance travelled, mode of delivery and haemoglobin at admission did not differ significantly between near-miss cases and maternal deaths (Table 2).

Table 2. Characteristics of women with maternal near miss and maternal death

Values are n (%) unless stated otherwise. SD: standard deviation; IQR: interquartile range. Chi-square or Fisher exact test for categorical variables; t test or Mann-Whitney U test for continuous variables.

Causes of severe maternal outcome

Obstetric haemorrhage was the leading underlying cause (34 women, 28.3%), followed by hypertensive disorders (22, 18.3%), pregnancy-related sepsis (20, 16.7%), severe anaemia (13, 10.8%), ruptured uterus (12, 10.0%), pregnancy with abortive outcome (10, 8.3%) and other indirect causes (9, 7.5%). Among the 15 deaths, haemorrhage caused 4, and hypertensive disorders, sepsis and ruptured uterus 3 each. The mortality index was highest for ruptured uterus (25.0%), followed by sepsis (15.0%) and hypertensive disorders (13.6%). There were no deaths after abortion or ectopic pregnancy (Table 3, Figure 1).

Table 3. Underlying causes of severe maternal outcome and cause-specific mortality index

Mortality index = maternal deaths / (near-miss cases + maternal deaths) × 100.

Figure 1. Underlying causes of severe maternal outcome in 120 women, showing near-miss cases, maternal deaths and the cause-specific mortality index (MI).

WHO near-miss criteria

Clinical criteria were met by 73.3% of women, laboratory criteria by 57.5% and management criteria by 52.5%; 57.5% met criteria in at least two of these groups. The commonest markers were shock (48.3%), thrombocytopenia below 50,000/µL (24.2%), continuous vasoactive drugs (24.2%) and creatinine of 3.5 mg/dL or more (17.5%). Sixty-nine women (57.5%) received blood, and 19 (15.8%) received 5 or more units of red cells. Nine women (7.5%) needed hysterectomy, and 95 (79.2%) were admitted to the intensive care unit. Maternal deaths more often met management criteria (80.0% vs 48.6%, p = 0.023) and more often needed ventilation (33.3% vs 12.4%, p = 0.049). Cardiopulmonary resuscitation was needed only in women who died (4 of 15) (Table 4).

Table 4. WHO near-miss criteria met by near-miss cases and maternal deaths

A woman could meet more than one criterion. Chi-square or Fisher exact test.

Delays and perinatal outcome

At least one delay was found in 92 women (76.7%). Phase 2 delay (reaching the hospital) occurred in 40.8%, phase 1 delay (deciding to seek care) in 40.0% and phase 3 delay (receiving care after arrival) in 20.8%. The frequency of delays did not differ between near-miss cases and maternal deaths. Of 100 babies born at 20 weeks or more, 79 were born alive and 21 were stillborn; 6 died in the first week, giving a perinatal mortality of 27.0%. Perinatal mortality was 23.6% among near-miss cases and 54.5% among maternal deaths (p = 0.064). Sixty-nine babies (69.0%) were born preterm (Table 5).

]

Table 5. Delays in care and perinatal outcome among women with a severe maternal outcome

A woman could have more than one type of delay. Percentages for babies are of babies born at 20 weeks or more in each group. Perinatal mortality = stillbirths + early neonatal deaths. Fisher exact test.

This prospective study applied the WHO criteria to 120 women with a severe maternal outcome at a teaching hospital that receives referrals from Jalgaon district. One woman died for every seven who survived a near miss, and the mortality index was 12.5%. Haemorrhage, hypertensive disorders and sepsis caused nearly two-thirds of severe maternal outcomes, most women were referred, and three in four had a delay in care. The mortality index reflects how well a hospital manages women who reach it with a life-threatening condition. The index of 12.5% in this study is lower than the 14.9% reported from Manipal [21], 28.1% from Vadodara [20] and 29.07% from Yavatmal in rural Maharashtra [19]. In north India, Pandey and colleagues recorded 633 near-miss cases and 247 deaths in a tertiary hospital, a mortality index of 28.1% [23]. The present index is close to the 13.1% in a district hospital in Rwanda [24] and the 12.9% case fatality rate in a rural referral hospital in Tanzania [25]. The near-miss to mortality ratio of 7:1 is higher than the 5.6:1 in Manipal [21] and 3.43:1 in Yavatmal [19], and close to the 9:1 in six public hospitals in Baghdad [26]. A higher ratio means that more women with organ dysfunction survive. Even so, in the WHO Multicountry Survey, mortality in countries with high maternal mortality was two to three times higher than expected for the severity of illness [10], and each death in this series needs individual review. Haemorrhage was the leading cause, as in most Indian and international studies. It caused 44.2% of near-miss cases in Manipal [21] and 45.7% in north India [23], and hypertensive disorders caused 53.4% in two Mumbai hospitals [22]. Anaemia and hepatitis were also important in rural Maharashtra [19], and in southern India about half of maternal deaths were from indirect causes, mainly infections [27]. The present pattern of haemorrhage, hypertensive disorders and sepsis agrees with these reports and with global data [4]. Ruptured uterus had the highest mortality index (25.0%). It usually follows obstructed or poorly supervised labour, and it shows the cost of delayed referral from peripheral facilities. Caesarean complications and uterine rupture have also been important causes of near miss in Tanzania [25,28]. Sepsis had the second highest index, which points to late presentation and the need for early antibiotics and source control. No woman died after abortion or ectopic pregnancy, which suggests that these conditions were managed promptly once the women reached hospital. Women who died more often had their event before delivery. Antepartum haemorrhage, eclampsia and sepsis before delivery leave less time to stabilise the woman and add the risks of emergency delivery. Women who died more often met management criteria and needed ventilation, which reflects a greater severity of organ failure. Shock was the commonest marker, as in the WHO Multicountry Survey, where cardiovascular, respiratory and coagulation dysfunction were the most frequent organ dysfunctions [10], and in the Vellore series [27]. Delays were common. In the Mumbai study, 68.2% of women with near miss had found no treatment at lower-level facilities before admission [22]. In Brazil, delays became more frequent as severity increased [15]. In the present study, 76.7% of women had at least one delay, mostly in deciding to seek care or in reaching the hospital. These delays depend on awareness, transport and the referral system, while phase 3 delays depend on the hospital itself [14]. Almost three in four women were referred, and half travelled more than 50 km. Women reaching a referral hospital in shock or after eclampsia have little reserve left, so earlier recognition and stabilisation at the first facility matter as much as care at the referral centre. Studies in Tanzania and Rwanda found gaps in simple interventions, such as oxytocin for postpartum haemorrhage and antibiotics at caesarean section [24,25], while a study in rural Uttarakhand found high coverage of oxytocin and magnesium sulphate [29]. Near miss also affects the baby. Perinatal mortality was 27.0%, and more than half the babies of women who died did not survive the first week. Among near-miss cases in Vadodara, 35% of births were stillbirths and 42% were preterm [20], and in Brazil 19.5% of pregnancies ended in stillbirth and 7.7% in neonatal death [30]. These results confirm that a severe maternal outcome is also a severe perinatal outcome. The WHO criteria were practical in this setting because the hospital has laboratory and intensive care services. In facilities with limited laboratory support, the organ dysfunction criteria may underestimate near miss [12], and even in high-income settings disease-based criteria remain useful [13]. Near-miss review is feasible in Indian facilities and helps identify avoidable factors [17,18]. Regular review of each near-miss case and maternal death, with feedback to referring facilities, should become routine [16]. Strengths and limitations The study was prospective, used the standard WHO definition and criteria, and included consecutive women, so the indicators can be compared with those of other facilities. Delays were recorded with a standard model, and perinatal outcome was followed. The study has limitations. It was done in one hospital, and as a referral centre it does not reflect the frequency of near miss in the district population. Hospital totals of live births were not part of this analysis, so the near-miss ratio, severe maternal outcome ratio and maternal mortality ratio per live birth were not calculated. Women who died at home, on the way or in other facilities were not captured. The number of maternal deaths was small, which limits the power of comparisons between near-miss cases and deaths. Information on delays was partly taken from relatives and may be subject to recall bias. Some laboratory criteria depend on tests that may not have been done in every woman, so some near-miss cases may have been missed.

At this referral hospital, one woman died for every seven who survived a maternal near miss, and the mortality index was 12.5%. Haemorrhage, hypertensive disorders and sepsis were the main causes, ruptured uterus carried the highest risk of death, and three in four women had a delay in care. More than a quarter of their babies died. Early recognition and referral, stabilisation before transfer, better management of haemorrhage and hypertensive disorders at first-level facilities, and regular facility-based near-miss review can reduce severe maternal outcomes in the district.

DECLARATIONS

Acknowledgement: The author thanks the residents and nursing staff of the Department of Obstetrics and Gynaecology, Dr. Ulhas Patil Medical College & Hospital, Jalgaon, for their help with case identification and data collection.

Funding: None.

Conflict of interest: The author declares no conflict of interest.

Author contribution: Dr. Ashok Bajirao Mahankale designed the study, collected and analysed the data, wrote the manuscript and approved the final version.

Data availability: The data are available from the corresponding author on reasonable request.