Using audits to reduce the incidence of uterine rupture in a Malawian district hospital

1. Introduction 

Uterine rupture is a serious complication of labor, causing maternal and fetal morbidity and mortality around the globe, although much more frequently in low-income countries [1], [2]. A WHO systematic review of maternal mortality and morbidity in 2005 showed large differences in the incidence of uterine rupture among countries. In high-income countries the incidence of uterine rupture was approximately 1% in women with a history of cesarean delivery, and less than 1 per 10000 in women with an unscarred uterus [1].

A recent nationwide prospective cohort study in the Netherlands confirmed the rarity of uterine rupture in a high-income country setting. The overall incidence was 6.1 per 10000 deliveries [3], most of which (86.7%) occurred in women with a history of cesarean delivery [4].

In low-income countries, however, uterine rupture was found to be a far more common problem, with overall rates ranging between 0.1% and 1% of all births [1]. The lack of differentiation of uterine rupture in women with and without a history of cesarean delivery is the most important shortcoming of the available data. In studies from Nigeria, Ghana, Ethiopia, and Bangladesh, about 75% of cases of uterine rupture occurred in women with an unscarred uterus, with obstructed labor being the most common cause [5], [6], [7]. Case fatality rates ranged between 1% and 13% and perinatal mortality between 74% and 92% [1]. Few studies have been done in rural areas of Sub-Saharan Africa, and most data are from urban centers, where the incidence of uterine rupture may be lower because services are more readily accessible [6], [8], [9], [10], [11].

In Thyolo district, Malawi, which is the setting of the present study, 23% of women deliver at the district hospital and 32% in public health centers. Adding women who deliver in the mission hospital and private clinics raises the total to 58.7% of deliveries in the district with the help of a skilled birth attendant. Before the onset of this study, uterine rupture occurred approximately once every week in Thyolo District Hospital, but accurate numbers are not available owing to inadequate hospital records (personal communication from district health officer and maternity staff). This reflects the situation of an overburdened health facility where documentation is not a priority [12]. The incidence of uterine rupture is unknown among those who do not reach a health facility [13].

The purpose of the present study was to improve obstetric care and to reduce the incidence of uterine rupture in Thyolo District Hospital. We performed an audit to identify the incidence of ruptured uterus and modifiable care-related factors [14].

2. Materials and methods 

3. Results 

Table 1 shows the characteristics of all 35 cases of uterine rupture during the study period.

Table 1. Characteristics of women with uterine rupture (n=35).

	Characteristics			No. (%)a	Thyolo maternal population, % [95% CI]
	Patient	Age (mean 27.4 y)	<20	3 (9)	31.3b (combined)
			20–25	6 (17)
			25–30	15 (43)	30.7b
			>30	10 (29)	27b
			Unknown	1 (3)	–
		HIV-positive	Yes	9 (26)	23.1c
			No	23 (66)	n/a
			Unknown	3 (9)	n/a
		Previous cesarean delivery	Yes	9 (26)	6.6b
			No	26 (74)	93.4b
	Pregnancy	Attitudes to pregnancy	Wanted and planned	18 (51)	n/a
			Wanted and unplanned	3 (9)	n/a
			Unwanted	8 (23)	n/a
			Unknown	6 (17)	n/a
		Parity	0	1 (3)	n/a
			1–3	22 (63)	n/a
			4	5 (14)	n/a
			≥5	7 (20)	n/a
		Prenatal hemoglobin measurement	Yes	6 (17)	35.3b [29.8–41.1]
			No	25 (71)	n/a
			Unknown	4 (11)	n/a
	Delivery	Obstructed labor	Yes	27 (77)	n/a
			No	8 (23)	n/a
		TBA involved in care during labor	Yes	10 (29)	21.3b [16.8–26.4]
			No	25 (71)	78.7b
		Health center involved	Yes	22 (63)	32b [26.8–37.6]
			No	13 (37)	68b
		Type of operation	Emergency hysterectomy	12 (34)	n/a
			Repair with tubal ligation	10 (29)	n/a
			Repair without tubal ligation	11 (31)	n/a
			Died before theatre	2 (6)	n/a
	Outcome	Maternal mortality		4 (11)	n/a
		Perinatal mortality		29 (83)	n/a

Abbreviations: TBA, traditional birth attendant; n/a, not available.

a Values may not add up to 100% because of rounding. b Data taken from baseline report for The Community-Based Maternal and Newborn Care Learning Programme, March 2008. c Data taken from the Malawi Demographic and Health Survey 2004 for the general female population, not available for maternal population.

Median parity of the women was 3 (range, 0–7), mean age was 27.4years (range, 18–40years), and two-thirds were living below the poverty line of US $2 per day. The HIV-prevalence was 26%. Eleven of 29 pregnancies (38%) were unplanned and 8 (23%) were unwanted.

During the first 3months of study, 16 cases of ruptured uterus occurred in 833 deliveries, an incidence of 19.2 per 1000. Nine of the 16 ruptures occurred after the woman was admitted, 5 in women with an unscarred uterus and 4 in those with a history of cesarean delivery. The other ruptures occurred in unscarred uteri prior to hospital arrival. These numbers have been converted to incidence per 1000 deliveries in Fig. 1.

View Large Image Download to PowerPoint

Fig. 1 Incidence and type of uterine rupture before and after audit in Thyolo District Hospital. Key: Red: ruptures that occurred after hospital admission in previously unscarred uteri; Orange: ruptures that occurred after hospital admission in previously scarred uteri; Yellow: ruptures that occurred before hospital admission in previously unscarred uteri. There were no ruptures in scarred uteri before hospital admission.

During the first 3months of the study, 7 cases were audited by hospital staff, 4 of which ruptured in hospital. In all 4 cases, care at hospital level was assessed to be substandard according to the district health protocols. Of the 3 audited ruptures that occurred outside hospital, the assessors identified substandard hospital management of the complication in 2 cases. An example of one case is described below.

Example

A 25-year-old woman was admitted to Thyolo District Hospital. She was G4P3, 2 of her children were alive, 1 infant had died from malaria. The woman had dropped out from primary school, was HIV-negative and had previously been healthy. All 3 deliveries had been spontaneous. Her pregnancy was wanted and planned. She had attended the prenatal clinic twice during pregnancy; hemoglobin had been measured during the third trimester at 10.8g/dL.

At 2:45 AM she was seen by a nurse/midwife in the labor ward. She reported that heavy labor pains had started at 6:00 PM the day before. She denied that she had visited a traditional birth attendant or a health center before arrival. The midwife examined her and found the cervix to be 9cm dilated with moderate contractions. She artificially ruptured the membranes, the liquor was clear. Fetal heart tones were heard and were regular.

At 6:00 AM, the midwife decided to start oxytocin (2.5U in a pint of 5% glucose at an undocumented rate). By that time the cervix was fully dilated. At 6:45 AM the action line for multigravida on the partograph was crossed and she informed a clinical officer, who came to assess the patient at 9:15 AM, 2.5hours later. By that time the woman had gone into shock and the clinician diagnosed uterine rupture. He consulted the medical officer on duty, who performed an emergency laparotomy. At 10:20 AM a fresh stillbirth weighing 4.4kg was extracted, a large posterior rupture was found and subtotal hysterectomy was performed [27]. The woman was transfused 2 pints of whole blood preoperatively. Postoperatively she recovered well and was discharged on the ninth day post partum.

The case was discussed with an audit panel of hospital staff who unanimously agreed that rupture had been caused by major substandard care in hospital. Identified substandard care factors were: (1) poor assessment of potential risk factors by the midwife: prolonged labor in a multipara; (2) incorrect indication for oxytocin after 12hours of active labor; (3) midwife giving oxytocin without consulting a clinician; (4) inadequate case filing of labor management; (5) failure of the clinical officer to assess the woman in due time (2.5hours); and (6) a delay in starting surgery of almost an hour.

A hospital protocol for augmentation of labor was developed. The midwife and clinician in question were briefed on the findings. The midwife attended training organized on the use of the partograph and the identification of prolonged labor.

Table 2 shows the main reasons for substandard care identified by hospital staff and Table 3 shows the recommendations that were implemented following the audit processes.

Table 2. Important reasons for substandard care identified by hospital staff.

	Hospital	1. Delay in recognizing prolonged labor because of poor follow-up of laboring women and delayed action to findings recorded in the partograph, especially in association with other risk factors for uterine rupture (grand multiparity, previous cesarean delivery) [identified in 4 out of 4 cases that occurred during admission].
		2. Poor case filing of delivery management, especially in case of augmented labor and other risk factors [in 3 out of 4 cases]
		3. Delay in providing adequate treatment after a diagnosis of uterine rupture [in 2 out of 4 cases]. Decision-to-operation interval of all cases combined averaged 96minutes.
		4. Inadequate preoperative resuscitation (timely acquirement of intravenous access with giving of intravenous fluids) [in 2 out of 4 cases]
	Health center	1. Lack of documentation of management [identified in 3 out of 3 cases that occurred before admission in hospital]
		2. Delay in decision to refer [in 2 out of 3 cases]
	Transport	Delay of ambulance to arrive at health center, after health center staff had called for one [in 2 out of 3 cases]

Table 3. Recommendations for improvement of care put forward by hospital staff and implemented toward the end of the first 3months.

	1. Organize a more elaborate audit session with external experts.
	2. Organize teaching by external experts.
	3. Organize training for midwives on the partograph, prolonged labor, and ruptured uterus.
	4. Increase supervision of higher cadres of staff on labor management and proper documentation by lower cadres.
	5. Organize a session for theater staff with the intention to reduce delays and improve surgical care.
	6. Arrange follow-up visits to health centers that had referred cases of uterine rupture.
	7. Organize a meeting between district health management and transport management to decide on solutions to reduce transport delays.
	8. Implementation of the WHO protocol for augmentation of labor in the hospital [20].

After implementation of the audit recommendations, 19 ruptures out of 3099 deliveries occurred over the following 9months, or 6.1 per 1000 deliveries (Fig. 1). This represented a reduction of 68% in the incidence of ruptures compared with the pre-audit period of 3months (OR 0.32; 95% CI, 0.16–0.63). Eleven of these ruptures occurred after admission, 4 of which were uterine scar ruptures. None of the 8 ruptures diagnosed on admission occurred in scarred uteri.

There were no ruptures in scarred uteri before hospital admission. In the 9-month period after implementation of the recommendations, only 2 uterine ruptures were audited, and the focus shifted to other complications. In general, compliance with the current hospital protocols improved.

Four maternal deaths and 29 perinatal deaths yielded a case fatality rate of 11.4% and a perinatal mortality rate of 829 per 1000 live births; 74% of ruptures occurred in unscarred uteri.

4. Discussion 

The present study shows that audit of uterine ruptures is a useful tool to rapidly improve quality of care in a district hospital setting with a very high incidence of uterine rupture. The incidence of rupture reported in the first 3months of the study has not often been reported in the literature [1], [8], [9], [10], [11], [28]. However, it is unclear whether the incidence rate was a continuation of previous high trends in our facility because recording of uterine rupture before the advent of the “4M-study” was inadequate. Case fatality and perinatal mortality rates in our cohort were comparable with other studies in resource-poor settings [1], [5]. The number of deliveries per month increased throughout the 1-year study period, probably because of simultaneous initiatives to encourage hospital deliveries.

According to hospital staff, uterine rupture has long been a neglected problem in Thyolo district, yet the reasons for its persistently high incidence have never been systematically described. The high prevalence of malnutrition in the district increases the baseline risk of complications of labor due to stunted growth and smaller pelvic size [29], [30], [31]. Poor contraceptive use (22%) by women in the district [31], and a high incidence in this study (31%) of ruptures in unwanted pregnancies, confirms the importance of promoting family planning in Thyolo [1].

Thaddeus and Maine [13] have classified delays leading to maternal mortality into 3 phases: delays at home before a decision is made to seek health care; delays in reaching the appropriate level of care; and delays in adequate management after arrival at the health facility. Using this approach, a number of factors can be identified that may lead to the high incidence of uterine rupture in Thyolo district.

Factors that may delay seeking treatment include a high female illiteracy rate (51.6%) [32], low social status [21], [22], [33], and cultural norms that make it more acceptable for women to deliver at home than in a health facility [33]. Delays in reaching the appropriate level of care may be due to travel costs and deficiencies in transport and referral systems [13], [32], [33], and factors delaying adequate management at the health facility are the shortage of qualified health staff [16], [32], the performance of available staff [14], and the scarcity of blood for transfusion [34]. We hope that addressing these shortcomings and thus reducing delays in referral and obstetric management will also reduce delays in women in labor seeking health care [35].

One reason why audit may be effective is that health workers are stimulated to provide better care when they know that they are being monitored, independent of the educational value of audit sessions or the implemented policy changes [36]. This “watchdog function” may be especially important in settings where legal consequences for clinical mistakes are uncommon [14], [37], [38]. However, adverse effects of audit on staff motivation and cohesion have also been reported [14], [39]. Although we did not specifically study the perceptions of audit by our staff, our experience certainly does not confirm these adverse effects. Audits in our hospital were well attended and much appreciated by all cadres of staff. The organizers of audit sessions tried to ensure open and constructive discussions, and to present audit as a learning experience, rather than as an opportunity to blame one another. Self-criticism from different cadres of staff was encouraged and respected.

Second, audit may enhance continuous professional development in cadres of staff that have had relatively little formal obstetric training owing to short training programs designed to rapidly increase numbers of health staff [38]. This is especially relevant in areas such as our district, with a shortage of human resources that requires more technically complicated work to be done by lower cadres of staff [16], [38].

A third reason for the success of the Thyolo audit may be that audit sessions were also attended by district health managers. Their participation in the audit process made them aware of some of the challenges in the functioning of the district's health system and allowed them to take appropriate action.

One important factor in substandard care, which has proven difficult to address, is the limited availability of blood. On average, women in our cohort received just over 1 pint of blood—inadequate to correct severe anemia. In one of the 4 maternal deaths, lack of blood for transfusion was a major contributing factor.

Audit is an inexpensive and simple intervention, requiring little technology [14], [37], [38], and is therefore an ideal method for health workers in resource-poor settings. By improving facility care through audit, health workers have the potential to help meet the targets set by Millennium Development Goals 4 (the reduction of child mortality) and 5 (the reduction of maternal mortality) [40]. A consequence of improved facility care could be more births attended by skilled health workers, as it becomes a more attractive option for pregnant women in the district [13], [29], [34], [40].

The improvement of hospital care, however, should not be the only step in attempting to reduce the incidence of ruptured uterus in any given population. Community sensitization to influence health-seeking behavior of women and guardians is also necessary [40], [41], [42]. More specifically, health education provided to women and traditional birth attendants on danger signs during pregnancy and delivery, and encouraging women in the community to prepare for birth by making a birth plan, can be effective measures, which we seek to implement in the district [43].

It is the duty of the district management team to ensure access to safe delivery for all women in the area it serves [44]. Increased access to safe obstetric care is an essential element of the right to health [45], and any occurrence of uterine rupture due to obstructed labor, especially when rupture occurs within a health facility, is a violation of this right.