Introducing maternal morbidity audit in the Netherlands


Dr J van Dillen, Department of Obstetrics, Leiden University Medical Centre, Albinusdreef 2, Leiden, The Netherlands. Email


Please cite this paper as: van Dillen J, Mesman J, Zwart J, Bloemenkamp K, van Roosmalen J. Introducing maternal morbidity audit in the Netherlands. BJOG 2010;117:416–421.

Objective  To describe the panel audit process and to identify substandard care in selected women from a nationwide prospective cohort study into severe acute maternal morbidity (SAMM) in the Netherlands.

Design  Prospective audit of selected women with SAMM.

Setting  Eight audit meetings held throughout the Netherlands.

Population  All pregnant women in the Netherlands.

Methods  Before each meeting, SAMM details of selected women were sent to all panel members for individual assessment by completing an audit form. During a subsequent plenary meeting, findings were discussed and substandard care factors as judged by the majority of assessors were scored.

Main outcome measures  Incidence of substandard care and recommendations for improving the quality of care.

Results  Substandard care was identified in 53 of 67 women (79%). Specific recommendations were formulated concerning the procedure of audit and concerning local as well as national management guidelines.

Conclusion  Our findings reflect SAMM in the Netherlands and substandard care is present in four out of five women. Ongoing audit of women with SAMM is promoted both at local and national level.


Maternal mortality has traditionally been used as an important indicator of health care, making comparison over time and between services possible. Detailed assessment of individual women through audit by the Confidential Enquiry into Maternal Deaths in the United Kingdom has been acknowledged as a major contributor to the decline of maternal deaths in the UK over the past 50 years. Other countries have followed this example, among them South Africa and the Netherlands. Nowadays, maternal mortality in high-income countries is too rare to be used as a sensitive marker for the quality of services. Therefore, severe acute maternal morbidity (SAMM) has been introduced.1–6 SAMM complicates at least 0.71% of all pregnancies in the Netherlands, and should be considered as a new indicator of the quality of obstetric care next to maternal mortality.7 Auditing SAMM to identify substandard care has generally been accepted as complementary to maternal death reviews.8 In this study we describe the introduction of SAMM audits in the Netherlands with the primary aim to illustrate the panel audit process and to identify substandard care.


This study was part of the nationwide prospective cohort study into SAMM in the Netherlands, called ‘LEMMoN’. The LEMMoN study was conducted to assess incidence, case fatality rates (CFRs), risk factors and substandard care overall and for different subgroups of SAMM. All 98 hospitals (100%) with a maternity unit in the Netherlands participated and from 1 August 2004 until 1 August 2006, a total of 371 021 deliveries were included. SAMM was classified according to disease-specific and management-based criteria and categorised into five groups (Figure 1). Detailed methods are described previously.7 From 2004 onwards, eight audits have been organised throughout the Netherlands during which 67 incidents of SAMM (2.6% of all SAMM occurrences) were assessed (Table 1). Audits included regionally or nationally selected incidents of SAMM and included audits with specific topics: eclampsia, major obstetric haemorrhage (MOH) and SAMM after delivery under primary care (Table 1). The pilot audit was initiated to evaluate if audit on the collected data was possible, and second to evaluate if the women included in the LEMMoN study were considered ‘true SAMM’ by the panel members. This pilot included all 23 women with SAMM in two hospitals during the first 10 months of the study, of whom 14 were eventually selected for discussion during the panel meeting.9 Since then, we applied initial selection and discussed all the women during the plenary meetings. During an in-depth eclampsia audit in Delft, nationally selected women were discussed without the presence of medical staff (consultants, midwives or registrars) involved with these women. It was noted that this left many in-depth questions unanswered and therefore, two additional audit meetings were held with the involved staff present. These three audits are presented here as one. To ensure enough time for in-depth discussion, a maximum of 12–14 women were selected for each panel meeting. Concerning the primary-care audit of MOH, we restricted the number of women by auditing only severe cases of SAMM assuming that the recommendations formulated could also apply to less severe cases. Therefore, for the primary-care audit of MOH, women were eligible when eight or more units of blood were transfused, and the woman was either admitted to intensive care or had undergone major surgery or arterial embolisation to stop the haemorrhage.

Figure 1.

 Inclusion criteria for severe acute maternal morbidity (SAMM).

Table 1.   Selected characteristics from seven SAMM audit meetings
LocationDateInclusion criteriaSAMM (n)Assessors (n)Substandard care* (%)
  1. MOH, major obstetric haemorrhage; SAMM, severe acute maternal morbidity.

  2. *Substandard care by majority of the assessors after group discussion.

The Hague06/2005Local (pilot: all cases)141786
Groningen03/2006Regional (severe cases)122375
Leiderdorp09/2006Regional (severe cases)121367
Delft/Zwolle/Amsterdam02/2007National (eclampsia)12892
Utrecht10/2008Primary care: National (eclampsia)81863
11/2008National (severe MOH)92489

For each audit, panel members were persons selected from the LEMMoN advisory board and the national Maternal Mortality Committee, as well as local healthcare workers involved. The number of assessors and the individuals selected varied by audit (n = 8–23, Table 1), but panel members were chosen in such a way that each audit included staff from university as well as non-university hospitals. Furthermore, for each audit, members from different specialties (mainly obstetricians, midwives and internal-medicine specialists) were selected with special attention to including members with experience in the audit process. Almost all panel members were involved with patient care in obstetrics, while some were involved within national organisations (obstetrics and gynaecology or midwifery) and were invited to attend because of their experience in national programmes and protocols.

Each panel meeting considered eight to 14 women. Anonymised notes from the LEMMoN database, selected by one member of the LEMMoN audit team (J.Z.), were sent to the panel members and included discharge letters, details from delivery, operation notes, laboratory results and a summary of file notes. Each panel member was requested to perform individual assessment of the patient notes using a standardised audit form used by the Maternal Mortality Committee. Substandard care was identified at the level of the woman, the care provider or the healthcare system (15 items). In the case of eclampsia or MOH, additional substandard care items concerning specific management of these complications were scored. Individual scores were collected before each plenary meeting. During the plenary meeting, incidences of SAMM were discussed and subsequently assessed for substandard care using the same audit form. However, now additional information from the original patient file and—for some audit meetings—clarification by the involved care provider were available. Final decision of substandard care, which is used in this report, was reached by consensus. Substandard care was assumed if the majority of assessors judged this to be the case. Substandard care was primarily identified if care deviated from national guidelines. If national guidelines were not available, local protocols, best available evidence or expert consensus were used.


Pilot audit

During a pilot audit 23 women with SAMM were selected in two teaching hospitals in The Hague and these were assessed by 17 audit members.9 Individual assessment of patient notes was judged to be possible in 16 women (69.6%), with 18 women classified as ‘true SAMM’ (78.3%) and identification of substandard care during individual assessment in ten women (43.5%). Of five women not classified as ‘true SAMM’ by the panel members, three where included because of MOH with transfusion of 4 units of red blood cells and two women were admitted to intensive care units for observation because of pre-eclampsia and mild peripartum cardiomyopathy. Fourteen women were subsequently selected by the panel members for plenary discussion with additional information from the original patient file. Of these, 12 women were classified as having ‘true SAMM’ (85.7%) and substandard care was judged to be present in 12 (85.7%) of them. In one woman, lack of information as the result of poor records was judged to be substandard care. The pilot audit concluded that the LEMMoN study identified SAMM in the Netherlands and that substandard care analysis was possible in most women, especially if care providers and patient file were made available during the panel meeting. In addition to substandard care analysis, recommendations were made concerning future LEMMoN audits (Table 2).

Table 2.   Recommendations from selected SAMM audit meetings
  1. MOH, major obstetric haemorrhage; SAMM, severe acute maternal morbidity.

GeneralAdditional information with patient records is often necessary for effective audit
Improve record keeping, especially concerning timing of interventions
Improve treatment guidelines concerning pre-eclampsia and MOH, for primary as well as secondary care
Eclampsia10Improve adequate treatment of hypertension
Improve adequate seizure prophylaxis
Primary-care MOHReduce the delay in reaching the hospital by timely referral (if placenta not delivered after 30 minutes)
Importance of intravenous access and initiation of resuscitation before transport to hospital
Discussion about the need and feasibility for misoprostol® at primary-care level
Discussion concerning emergency transport and acceptance of home delivery in areas where referral to secondary care might result in delay
Need of delivery on ground floor because of regulations for emergency transport employees restricting them to carry women downstairs
Repeated consultation from secondary-care provider for suspected pre-eclampsia should lead to referral and continued secondary care, irrespective of whether woman fits criteria
Primary-care eclampsiaStandard measuring of blood pressure is indicated 2 hours after delivery or before leaving the woman after home delivery

All audits

Of 358 874 births during the study period, 2552 incidents of SAMM were included in LEMMoN (7.1 per 1000 births). Of 67 women with SAMM discussed during the panel meetings, substandard care was judged to be present by the majority of assessors for 53 women (79.1%). From five of the audits, including 53 (74.6%) of the women with SAMM, more detailed scoring of substandard care items was available. From a total of 17 430 possible substandard care items (number of assessors × number of cases × 15 scoring items) 1223 (7.0%) were scored. Only 73 (6.0%) were identified at the level of the woman, 933 (76.3%) at the level of the care providers and 217 (17.7%) at the level of the healthcare system (Table 3).

Table 3.   Substandard care items and their contribution during five severe acute maternal morbidity (SAMM) audit meetings
  1. *Only for primary-care audits. percentage for total substandard care items.

Delay in consulting doctor433.5
Refusal of medical help or advice151.2
Language barrier151.2
Inadequate antenatal care927.5
Delay in recognition of symptoms/signs1139.2
Delay in referral to obstetrician1219.9
Inadequate risk selection*413.4
Inadequate antenatal care887.2
Delay in recognition of symptoms/signs18114.8
Delay in treatment after diagnosis25520.9
Delay in referral to tertiary-care centre352.9
Other consultant70.6
Delay in consulting obstetrician70.6
Healthcare system21717.7
Home birth influenced outcome1038.4
Birth in general hospital influenced outcome766.2
Quality of transport influenced outcome383.1

Of 358 874 births, 145 703 (40.6%) were under the responsibility of primary-care givers and 213 171 (59.4%) delivered in hospital under the responsibility of a gynaecologist.7 Of the 145 703 births handled by primary-care givers, 113 404 were home births (31.6% of total deliveries in the Netherlands) and 32 299 were hospital deliveries under primary care. Of 2552 incidents of SAMM, 227 (1.6 per 1000) were included after delivery under the responsibility of primary-care provider, 154 (1.4 per 1000) were included after home birth and 73 after delivery under the responsibility of primary-care in hospital (2.3 per 1000). During two audit sessions (one concerning MOH and one concerning eclampsia), 17 of these women with SAMM after delivery under primary care (7.5%) were assessed.

Primary-care audits

From 1606 SAMM linked to MOH, 140 (8.7%) were included after home delivery. Nine women (6.4%) met the criteria for MOH after delivery under primary care mentioned in the Methods section and were assessed by 24 panel members. Substandard care was judged to be present by the majority of the assessors in eight women (88.9%) and inadequate risk selection was judged to be present by the majority of the assessors in four women (44.4%). From a total of 4410 possible substandard care items (number of assessors × number of women × 21 scoring items), 387 (8.8%) were recorded: 134 (34.6%) were at the level of the primary-care provider and 72 (18.6%) concerned the management of MOH irrespective of the level of care. Specific recommendations were made concerning more stringent risk selection, delay in reaching the hospital and timing of referral (Table 2).

From 239 women with SAMM included because of eclampsia or severe haemolysis, elevated liver enzymes and low platelet count (HELLP), all eight (3.3%) women for whom delivery had been under primary care were assessed by 18 panel members. Substandard care was judged to be present by the majority of the assessors in five women (62.5%). Inadequate risk selection was identified by a minority of the assessors in four women (ranging from 16.7% to 44.4% of assessors). From a total of 2940 possible substandard care items (number of assessors × number of cases × 21 scoring items), 221 (7.5%) were recorded: 69 (31.2%) were at the level of the primary-care provider and 62 (28.1%) concerned the management of eclampsia irrespective of the level of care. Specific recommendations were made concerning the diagnosis and management of pre-eclampsia (Table 2).


During eight audit meetings in the Netherlands, 67 women with SAMM were assessed and substandard care was identified in almost four out of five women. Substandard care was judged to be present at the level of the woman and the level of the healthcare system but mainly at the level of the care provider. For optimal substandard care analysis, additional information from the original patient files was often required. Without this information, substandard care was often suspected but not demonstrated. However, even with the complete patient file available for assessment, substandard care analysis was not always possible. The lack of information as a result of inadequate record keeping can also be regarded as substandard care. During the panel meetings, with availability of original patient file and discussion among panel members, the identification of substandard care increased. Although this pattern was consistent throughout all audits, the magnitude of the increase in substandard care identification during the pilot audit (from 43% after individual audit to 86% after group audit) has not been seen during successive audits (data not shown). This might reflect a learning curve for audit. The earlier reported lower incidence of substandard care in the LEMMoN study (61.9%) is the result of the inclusion of individual audit results in that report compared with rates after group audit in the current report (79.1%).7

Only 2.6% of all incidents of SAMM have been audited. Although women were purposefully selected to include women with severe SAMM, we think the recommendations can be used to improve general obstetric care also in women with less severe problems. We specifically and purposefully used an open audit instead of criterion-based audit during these sessions. Although more time consuming for the assessors and more subjective (dependent on the assessors’ experience and expertise), this kind of audit also identified substandard care in unforeseen areas. Examples are the discussions concerning the use of misoprostol in primary care and the management guidelines used by emergency transport employees. The recommendations from these open audits can be formulated in management guidelines, which in turn can assist in future criterion-based audits (for example the use of magnesium sulphate in the prevention of eclampsia), which will then be more objective.10

The incidence of SAMM as the result of eclampsia in the Netherlands is markedly increased compared with other Western countries.10 Substandard care was identified in most incidences of SAMM, mainly at the level of the care providers and often because of inadequate treatment of hypertension and inadequate seizure prophylaxis. As for maternal death as the result of hypertensive disease in pregnancy, in 26 (96%) out of 27 maternal deaths of this type occurring in the Netherlands between 2000 and 2004, substandard care factors were present.11 In 2005, the national guideline ‘Hypertensive disorders in pregnancy’ of the Dutch Society of Obstetrics and Gynaecology has been adjusted and multiple papers and presentations have been given informing obstetricians concerning this issue.12 However, the guideline and its implementation can still be improved.11 Future repeated assessment of SAMM in the Netherlands could demonstrate the impact of these efforts.

Half of all SAMM concern MOH.7 Obstetric haemorrhage is the third direct cause of maternal death in the Netherlands with CFR of 1 in 201, compared with CFR of 1 in 53 for all incidents of SAMM. The relatively low CFR for MOH reflects the quality of blood supply in the Netherlands with women having received up to 50 units of blood. Hence, half of the women with SAMM because of MOH (n = 811) received more than 4 units of blood. From these figures it is clear that MOH is an important contributor to SAMM and not so much to maternal death. Where this might result in an attitude of acceptance towards morbidity, the risk of blood transfusion, especially during the reproductive period, should not be neglected. Audit revealed that there is ample room for improvement in the management of MOH, especially by reducing delays in diagnosis and treatment. Therefore, regular skills training in obstetric emergencies like MOH should be implemented in any unit.6,13 The Managing Obstetric Emergencies and Trauma course was introduced in the Netherlands in 2003 and it is encouraged during this national training to initiate regular local multidisciplinary skill trainings. A recent questionnaire indicated that at least 29% of Dutch obstetric units have regular skills trainings and 22% are in the process of organising these trainings (personal communication).

The lower risk for SAMM after delivery under the responsibility of the primary-care giver (relative risk 0.1; 95% confidence interval 0.1–0.2) seems to reflect the proper functioning of the Dutch system of risk selection.7 However, also among these women, substandard care was judged to be present in the majority of women with SAMM. Furthermore, inadequate risk selection in cases leading to severe MOH was present in almost half of women. The definition ‘retained placenta’ is used when the placenta has not been delivered within 1 hour after the birth of the baby.14 In the Netherlands, women are therefore only referred to secondary care 1 hour after delivery in the absence of severe bleeding (>1000 ml). For term pregnancy (which applies to all deliveries under primary care), however, the duration of the third stage of labour is under 15 minutes for 90% of deliveries.15 Therefore, we recommend earlier referral to secondary care in the case of retained placenta, especially after home delivery because of delay in reaching secondary care, as mentioned in Table 2.

Concerning audit in general, although the effect of critical incident audit has not been proven in randomised controlled trials, it is clear that morbidity and mortality reviews do more good than harm.16 Critical incident audit both monitors the quality of services and is a resource for professional learning.13,17 The openness in provision of data and participation during these audits in the Netherlands is encouraging because it is thought to reflect quality awareness. Continuing local audit of women with eclampsia and MOH have already been implemented in the national quality assurance programme to improve management and local guidelines. In addition to these national initiatives, auditing SAMM at local or regional level should be encouraged to improve the quality of obstetric care. In the Netherlands, however, obstetric audit is relatively new. After the 2004 results from the EURO-PERISTAT project (, which indicated that Dutch perinatal mortality rates ranked unfavourably compared with other European countries, many measures have been taken to improve the quality of perinatal care. The most important are the initiation of the nationwide perinatal audit, better prenatal screening and the introduction of preconception care.18 The national perinatal audit programme includes training of audit members at regional and local levels. In the near future, more healthcare workers will be familiar with obstetric audit and it is envisaged that the tradition of audit, as in the UK, will eventually also be reached in Dutch obstetric health care.

Disclosure of interest

We declare that we have no conflict of interest.

Contribution to authorship

J.vD. organised the pilot audit, performed the statistical analysis and wrote the manuscript. J.M. organised the primary-care audits and assisted with the manuscript. J.J.Z. conducted the LEMMoN project, supervised enrolment, data collection and management, and assisted with a version of the manuscript. K.B. was involved in the design of the project and provided feedback on an earlier version of the manuscript. J.vR. conceived the LEMMoN project, being project leader, and provided feedback on an earlier version of the manuscript. J.vR. will act as study guarantor. All authors gave final approval of the manuscript. J.vD. and J.J.Z. are members of the LEMMoN audit team and K.B. and J.vR. are members of the LEMMoN expert panel.

Details of ethics approval

The study was centrally approved by the medical ethics committee of Leiden University Medical Centre (P04-020; 8 March 2004).


The study was supported by the Dutch Organisation for Health Research (ZonMw; grant 3610.0024) and the Matty Brand Foundation. The funding sources had no role in study design, data collection, data analysis, data interpretation or writing of the report.


We would like to thank all panel members for their time and effort during the different audit meetings. Furthermore, we would like to thank M. Smit, midwife, and A. Dijkman, gynaecologist, for sharing their data concerning the skill’s training questionnaire. Finally, we would like to thank the Netherlands Perinatal Registry for national reference data on delivery.