Objective
The purpose of this study was to identify factors that contributed to severe maternal morbidity, defined by admission of pregnant women and women in the postpartum period to the intensive care unit (ICU) from 2010-2011 at Auckland City Hospital (ACH), a tertiary hospital that delivers 7500 women/year, and to determine potentially avoidable morbidity with the use of local multidisciplinary review.
Study Design
All admissions of pregnant women and women in the postpartum period (to 6 weeks) to the ICU at ACH from 2010-2011 were identified from hospital databases. Case notes were summarized and discussed by a multidisciplinary team. The presence of contributory factors and potentially avoidable morbidity were determined by consensus with a tool that was developed by the New Zealand Perinatal and Maternal Mortality Review Committee for the review of maternal and perinatal deaths. Specific recommendations for clinical management were identified by the multidisciplinary group.
Results
Nine pregnant women and 33 women in the postpartum period were admitted to the ICU from 2010-2011. Contributory factors were identified in 30 cases (71%); 20 cases (48%) were considered to be potentially avoidable; personnel factors were the most commonly identified avoidable causes. Specific recommendations that resulted from the study included the need for the development of guidelines for puerperal sepsis, improved planning for women at known risk of postpartum hemorrhage, enhanced supervision of junior staff, and enhanced communication through multidisciplinary meetings.
Conclusion
Forty-eight percent of severe maternal morbidity, which was defined as admission to the ICU at ACH from 2010-2011, was considered to be potentially avoidable by a local multidisciplinary review team; priorities were identified for improvement of local maternity services.
The maternal mortality ratio in New Zealand was 17.8 of 100,000 births in the 5 years of 2006-2010, ranging between 12.3 and 24.7 in individual years. This compares with rates of 11.8 of 100,000 in the United States and 11.4 of 100,000 in the United Kingdom.
However, in high-income countries, the rarity of maternal death limits its utility as a robust measure of the quality of a maternity system and provides limited information about risk factors for specific pregnancy complications. Severe acute maternal morbidity (SAMM) has been proposed as an alternative measure; in some countries, a review of SAMM cases has been used to identify quality improvement priorities.
Published reviews of SAMM from different countries and centers have used a variety of methods and tools, although most publications focus on system, provider, and patient contributory factors. In New Zealand, a single center pilot review of admissions to the intensive care unit (ICU) has been reported with the methods of Geller et al. The LEMMoN study from the Netherlands defined substandard care at each level of the maternity system as “care that deviated from national or local guidelines, local protocols, best available evidence or expert consensus,” and substandard care was considered to be present in 79% of all cases. However, this study did not further determine whether major morbidity was considered preventable. Other models that have determined preventable provider and system factors have reported that as many as one-third of cases were preventable. In Scotland, all maternity centers audit severe maternal morbidity and contribute the findings to an annual national audit.
The NZ Perinatal and Maternal Mortality Review Committee (PMMRC) has developed a tool for measuring contributory factors and potentially avoidable maternal death. The tool includes subheadings for contributory factors under organizational and management, personnel, technology and equipment, environment, and barriers to access/engagement with care.
The primary aims of this study were to apply the PMMRC tool to SAMM cases, which were defined as admission of pregnant women and women in the postpartum period to the ICU at Auckland City Hospital (ACH), to determine the incidence of contributory factors and potentially avoidable maternal morbidity and to identify recommendations that could address these issues.
Materials and Methods
ACH is a large tertiary-level teaching hospital, with tertiary level maternity and neonatal units with approximately 7500 deliveries annually and 1000 neonates admitted to the neonatal ICU. In addition, ACH is the tertiary care provider for a wider catchment birthing population of approximately 17,500 per year. It is also responsible for the provision of the fetal medicine service for the whole of New Zealand. Primary maternity care in New Zealand is provided by community-based lead maternity caregivers (86%) who are predominantly midwives, but may be private obstetricians or general practitioners, and by hospital-based services (14%). Secondary and tertiary care is predominantly provided by hospital-based services.
Study cases were identified by merging the hospital maternity clinical database with the hospital clinical management system database to identify women who gave birth at ACH and who were admitted to the ICU either during pregnancy or in the first 6 weeks after delivery during 2010-2011. The ACH ICU database was also accessed to identify women who delivered in other centers or who had been admitted to the ICU at <20 weeks’ gestation. No cases were excluded from the review.
A ratio of ICU admissions to maternities was calculated for the time period. Maternities were defined as women who delivered at ACH at ≥20 weeks’ gestation.
A multidisciplinary review team was established to represent hospital-based and community midwifery services, hospital-based and private obstetricians, anesthetists, intensive care physicians, and obstetric physicians after invitation to the departmental heads to recommend potential team members.
Each case was summarized on a standard form and timeline by 2 researchers (D.M.A. and L.C.S.). A summary of clinical observations was also prepared on an early warning score template for ease of presentation. Sociodemographic and clinical data were extracted from the patient’s notes. The clinical notes were used to assign an Acute Physiology, and Chronic Health Evaluation II (APACHE II) score (based on the 12 hours before and 12 hours after admission to the ICU), which is a measure of severity of illness at ICU admission (which has a range from 0–71) and is predictive of in-hospital death, although it has not been validated specifically for pregnant women. The summary, timeline, and early warning score record were sent to review team members 1 week before each review meeting.
Contributory factors and potentially avoidable morbidity were assessed with the PMMRC tool that was developed based on the London Protocol and which had been published previously. Contributory factors were defined for this purpose as modifiable components of the health system and issues of quality of care that cover a broad spectrum of management, personnel, and patient responsibility. The tool includes a list of questions to determine whether any of the following issues have been identified: (1) organizational and/or management factors, (2) factors relating to personnel, (3) factors relating to technology and equipment, (4) factors relating to the environment, and (5) barriers to accessing/engaging with care. Within each question, a checklist of 4-11 items is given ( Table 1 ). The presence or absence of a category of contributory factor and the specific items from the checklist was recorded. Multiple contributory factors and items from the checklist could be selected. Potentially avoidable morbidity was considered to have occurred when the absence of any of the contributory factors could have resulted in less severe maternal morbidity and potentially avoided admission to the ICU. The decision about potentially avoidable morbidity was taken formally after the completion of the assessment of the presence of any contributory factors. The team was then asked to decide which contributory factor or factors had led to a case being defined as potentially avoidable and which caregiver groups that they considered responsible for the factors. Specific recommendations for improvements in care were also sought from the team at each review meeting.
| Contributory factors | Contributory factors (n = 42) | Potentially avoidable a (n = 42) | ||
|---|---|---|---|---|
| n | % | n | % | |
| Any contributory factor | 30 | 71 | 20 | 48 |
| Organizational/management factors | 16 | 38 | 12 | 29 |
| Poor organizational arrangements of staff | 4 | 10 | 2 | 5 |
| Inadequate education and training | 4 | 10 | 2 | 5 |
| Lack of policies, protocols, or guidelines | 8 | 19 | 6 | 14 |
| Inadequate numbers of staff | 1 | 2 | 0 | |
| Poor access to senior clinical staff | 5 | 12 | 3 | 7 |
| Inadequate supervision of staff | 3 | 7 | 3 | 7 |
| Lack of communication between services | 4 | 10 | 4 | 10 |
| Failure or delay in emergency response | 4 | 10 | 3 | 7 |
| Delay in procedure (eg, cesarean delivery) | 2 | 5 | 2 | 5 |
| Delayed access to test results or inaccurate results | 2 | 5 | 2 | 5 |
| Other | 3 | 7 | 1 | 2 |
| Personnel factors | 23 | 55 | 17 | 40 |
| Knowledge and skills of staff lacking | 15 | 36 | 13 | 31 |
| Delayed emergency response by staff | 10 | 24 | 10 | 24 |
| Failure to maintain competence | 0 | 0 | ||
| Inadequate communication among staff | 3 | 7 | 3 | 7 |
| Failure to seek help/supervision | 8 | 19 | 6 | 14 |
| Failure to follow recommended best practice | 12 | 29 | 9 | 21 |
| Lack of recognition of complexity or seriousness of condition by caregiver | 19 | 45 | 15 | 36 |
| Other | 0 | 0 | ||
| Technology and equipment factors | 3 | 7 | 3 | 7 |
| Essential equipment not available | 0 | 0 | ||
| Lack of maintenance of equipment | 0 | 0 | ||
| Malfunction/failure of equipment | 1 | 2 | 1 | 2 |
| Failure/lack of information technology | 2 | 5 | 2 | 5 |
| Other | 0 | 0 | ||
| Environmental factors | 0 | 0 | 0 | 0 |
| Geography (eg, long transfer distance) | 0 | 0 | ||
| Administration systems inadequate | 0 | 0 | ||
| Building and design functionality limited clinical response | 0 | 0 | ||
| Other | 0 | 0 | ||
| Barriers to access or engagement with care | 14 | 33 | 4 | 10 |
| Unbooked or late booking for antenatal care, infrequent antenatal visits | 5 | 12 | 1 | 2 |
| Barriers to access/engagement with care related to | ||||
| Substance use | 1 | 2 | 0 | |
| Family violence | 0 | 0 | ||
| Lack of recognition of complexity or seriousness of condition by patient or family | 6 | 14 | 4 | 10 |
| Maternal mental illness | 0 | 0 | ||
| Cultural barriers | 3 | 7 | 0 | |
| Language barriers | 2 | 5 | 0 | |
| Not eligible to access free care | 2 | 5 | 0 | |
| Other | 2 | 5 | 0 | |
a Absence of this contributory factor might have prevented or reduced severity of morbidity
The team members attended a training meeting before they began reviewing the cases, at which 2 cases of SAMM from 2009 were reviewed. Because the process was part of a local review, some reviewers had been involved in the care of the cases that were discussed. Reviewers were advised that they may choose to excuse themselves from such a review or to remain and contribute to the discussion, when appropriate. Rules were sent to each member of the team and at the start of each meeting, members were reminded about confidentiality and respect.
Monthly meetings were organized; each meeting lasted 3-4 hours, with an average of 6 cases presented at each meeting. Time was given for clarifications and discussion, followed by scoring of contributory factors and then assessment of potential avoidability. Consensus was required to determine potential avoidability. The meetings were chaired by 1 of 2 researchers (L.C.S. or D.M.A.), while the other researcher recorded the findings.
At the completion of all case reviews, a final meeting was held to discuss the recommendations to be made as a result of discussion of the cases. These were grouped and summarized for feedback to the service for consideration through the ACH Women’s Health Clinical Governance structure.
All data were entered into an Access database (Microsoft Corporation, Redmond, WA), and data were analyzed with the use of STATA software (version 9; Stata Corporation, College Station, TX).
Ethics approval for this study was given by the Northern X Regional Ethics Committee. Individual consent was not required for review of the clinical records.
Results
In 2010-2011, a total of 42 cases of SAMM were identified at ACH. Nine women were admitted during pregnancy, and 33 women were admitted in the postpartum period. There were no cases of maternal death. Three of the 42 women who were admitted had delivered in other centers. The ratio of SAMM to maternities at ACH was 276 of 100,000. The most common reason for admission was postpartum hemorrhage (n = 14), followed by preexisting medical condition and infection ( Table 2 ). The APACHE II scores of admissions ranged from 2–22 (median, 14; interquartile range, 10–16.8). Sixteen patients (38%) required ventilation during ICU admission. The median length of stay in the ICU was 0.95 days (interquartile range, 0.8–1.5 days).
| Indication | Intensive care unit maternity admissions (N = 42) | Contributory factors | Potentially avoidable | |||
|---|---|---|---|---|---|---|
| n | Column, % | n | Row, % | n | Row, % | |
| Postpartum hemorrhage | 14 | 33 | 10 | 71 | 7 | 50 |
| Preexisting medical condition | 12 | 29 | 5 | 42 | 2 | 17 |
| Infection | 10 | 24 | 10 | 100 | 6 | 60 |
| Obstetric infection | 5 | 12 | 5 | 100 | 4 | 80 |
| Nonobstetric infection | 5 | 12 | 5 | 100 | 2 | 40 |
| Hypertension | 3 | 7 | 3 | 100 | 3 | 100 |
| Overdose (unintentional medicinal) | 2 | 5 | 2 | 100 | 2 | 100 |
| Venous thromboembolism | 1 | 2 | 0 | 0 | 0 | 0 |
The full multidisciplinary review team included 6 obstetricians (1 from private practice), 4 midwives (2 from community midwifery and 2 from hospital-based practice), 1 obstetric physician, 2 intensive care physicians, 1 obstetric anesthetist, and an epidemiologist. There were 5-9 members at each review meeting. The case summaries required on average 3 hours to complete, with a range of 1–9 hours. Ten meetings with a total duration of 30-40 hours were required to review all 42 cases, and the average time to complete multidisciplinary review of a single case was 1 hour.
There were no significant differences in admission to the ICU by ethnicity in 2010-2011. More women who were admitted to the ICU were under the care of the hospital obstetric service (midwifery or medical) at the time of birth compared with community lead maternity caregivers ( Table 3 ).
