Gillian Penney, Dawn Kernaghan and Victoria Brace
Severe maternal morbidity - the Scottish experience 2003 to 2005
For over 50 years, clinicians in the UK have used measures of maternal mortality as a means of monitoring the quality of maternity services. In addition, detailed study of the circumstances of individual cases has provided material for professional learning. Over the past decade, various groups internationally have investigated the rate of severe maternal morbidity, or ‘near misses’ as a complementary marker of standards of care1-7. The theory underlying this approach is described by Pattinson5: ‘the sequence from good health to death in a pregnant woman is a clinical insult, followed by a systemic inflammatory response, organ failure and finally death. A near miss would be those women with organ dysfunction who survive’. By viewing pregnancy and its potential outcomes as a continuum, beginning at normal pregnancy and concluding with maternal death, the number which can be studied meaningfully can be increased by examining the group of outcomes closest to death.
The Scottish Confidential Audit of Severe Maternal Morbidity began as a pilot study in 2001 and our growing experience has been reported in the last two Why Mothers Die Reports. Since 2003, data have been collected on a consistent national basis on 14 categories of severe maternal morbidity based on criteria originally described by Mantel et al8. We are now in a position to provide data for the triennium 2003-2005, to complement the maternal mortality data for the same period presented in the body of this Report.
Methods
As previously described2,9, each month, every consultant-led maternity unit in Scotland reports the number of women meeting one or more of the 14 agreed definitions to the central office of the Scottish Programme for Clinical Effectiveness in Reproductive Health. The categories and definitions currently in use are summarised in Table 20.1. A minimal dataset on each case is collected, comprising: a unique identifier; age; date of event; and limited clinical information to verify that the case definitions are being met. These monthly returns are collated centrally and used to calculate national and unit-level rates of severe morbidity events.
In addition, case assessment proformas relating to the most common categories of severe morbidity events (major obstetric haemorrhage and eclampsia) have been developed. These proformas comprise both condition-specific (ie assessing adherence to national guidance) and general (ie root cause analysis) sections. These national proformas are used by local clinical risk management teams during assessment of cases of major obstetric haemorrhage (2003 to 2005) or eclampsia (from 2004). They serve to guide local teams through a systematic and structured assessment of each case. Risk management teams are required to make an overall assessment of quality of care using definitions of suboptimal care similar to those used by the Confidential Enquiries into Maternal Deaths. They are also required to formulate an action plan. The completed proformas are collated centrally in order to identify recurrent themes and draw generalisable lessons for Scotland as a whole. Thus, both case ascertainment (permitting the calculation of rates of events) and case assessment (permitting the learning of clinical lessons) take place.
Table 20.1 Inclusion criteria used in 2003-05.
| Code | Category | Definition |
|---|---|---|
| 1 | Major obstetric haemorrhage | Estimated blood loss ≥2,500ml, or transfused five or more units of blood, or received treatment for coagulopathy (fresh frozen plasma, cryoprecipitate, platelets). (Includes ectopic pregnancy meeting these criteria). |
| 2 | Eclampsia | Seizure associated with antepartum, intrapartum or postpartum symptoms and signs of pre-eclampsia. |
| 3 | Renal or Liver dysfunction | Acute onset of biochemical disturbance, urea > 15mmol/l, creatinine >400mmol/l, aspartate aminotransferase / alanine aminotransferase >200u/l. |
| 4 | Cardiac arrest | No detectable major pulse. |
| 5 | Pulmonary oedema | Clinically diagnosed pulmonary oedema associated with acute breathlessness and 02 saturation < 95%, requiring 02, diuretics or ventilation. |
| 6 | Acute respiratory dysfunction | Requiring intubation or ventilation for >60 minutes (not including duration of general anaesthetic). |
| 7 | Coma | Including diabetic coma. Unconscious >12 hours. |
| 8 | Cerebro-vascular event | Stroke, cerebral/cerebellar haemorrhage or infarction, subarachnoid haemorrhage, dural venous sinus thrombosis. |
| 9 | Status epilepticus | Unremitting seizures in a patient with known epilepsy. |
| 10 | Anaphylactic shock | An allergic reaction resulting in collapse with severe hypotension, difficulty breathing and swelling/rash. |
| 11 | Septicaemic shock | Shock (systolic blood pressure <80 mm/Hg) in association with infection. No other cause for decreased blood pressure. Pulse of 120 beats per minute or more. |
| 12 | Anaesthetic problem | Aspiration, failed intubation, high spinal or epidural anaesthetic. |
| 13 | Massive pulmonary embolism | Increased respiratory rate (>20 per minute), tachycardia, hypotension. Diagnosed as ‘high’ probability on V/Q scan or positive spiral chest CT scan. Treated by heparin, thrombolysis or embolectomy. |
| 14 | Intensive care or coronary care admission |
Unit equipped to ventilate adults. Admission for one of the above problems or for any other reason. Include admissions to Coronary Care Units. |
Results
During the 2003 to 2005 triennium, 845 women were reported as meeting one or more of our defined inclusion criteria. Using a denominator of 159,223 maternities, the rate of severe maternal morbidity during the triennium was 5.3 with a 95% confidence interval from 5.0 to 5.7 per 1000 maternities). During the triennium, 15 Direct or Indirect maternal deaths were reported to the Confidential Enquiry into Maternal Deaths in Scotland; giving a ‘near-miss to death ratio’ of 56:1, with a 95% CI from 34:1 to 100:1.
Many women met the definitions for more than one severe morbidity category. For example, some suffered major haemorrhage and were admitted to intensive care. Thus, during the triennium the 845 women who were reported to us experienced a total of 1,135 events meeting our inclusion criteria. Rates of the 14 individual categories of severe morbidity are summarised in Table 20.2. Major obstetric haemorrhage was the most numerous category occurring in 582 women during the triennium. Thus major haemorrhage occurred in 69% of all 845 women with severe maternal morbidity.
Table 20.2 Numbers and rates per 1,000 maternities of individual categories of severe maternal morbidity; Scotland: 2003-05.
| Number of cases |
Rate per
1,000 maternities |
95 per cent CI | ||
|---|---|---|---|---|
| 1. Major Obstetric Haemorrhage | 582 | 3.66 | 3.37 | 3.96 |
| 2. Eclampsia | 55 | 0.35 | 0.26 | 0.45 |
| 3. Renal or liver dysfunction | 60 | 0.38 | 0.29 | 0.48 |
| 4. Cardiac arrest | 8 | 0.05 | 0.03 | 0.09 |
| 5. Pulmonary oedema | 42 | 0.26 | 0.20 | 0.36 |
| 6. Acute respiratory dysfunction | 58 | 0.36 | 0.28 | 0.47 |
| 7. Coma | 0 | 0.00 | 0.00 | 0.02 |
| 8. Cerebro-vascular event | 15 | 0.09 | 0.06 | 0.15 |
| 9. Status epilepticus | 2 | 0.01 | 0.00 | 0.04 |
| 10. Anaphylactic shock | 5 | 0.03 | 0.01 | 0.07 |
| 11. Septicaemic shock | 20 | 0.13 | 0.08 | 0.19 |
| 12. Anaesthetic problem | 26 | 0.16 | 0.11 | 0.24 |
| 13. Massive pulmonary embolism | 14 | 0.09 | 0.05 | 0.15 |
| 14. Intensive care or coronary care admission |
248 | 1.56 | 1.38 | 1.76 |
| All women included | 845 | 5.31 | 4.96 | 5.68 |
Overall Scottish rates of severe maternal morbidity in individual years from 2003 to 2005 are summarised in Table 20.3. The increase in rate from 4.6 per 1,000 in 2004 to 6.1 per 1,000 in 2005 is greater than would be expected by chance. The data in Table 20.3 suggest that this increase in severe maternal morbidity is almost entirely accounted for by an increase in major obstetric haemorrhage. The increase in rate of major haemorrhage from 3.2 per 1,000 in 2004 to 4.4 per 1,000 in 2005 was also statistically significant. In contrast, the increase in the rate of severe maternal morbidity if cases of haemorrhage are excluded is compatible with random variation. In commenting on these apparent increases between 2004 and 2005, it must be noted that there is no consistent upward trend over the three years studied to date.
Table 20.3 Numbers and rates of severe maternal morbidity (overall and from haemorrhage and non-haemorrhage causes) in individual years: Scotland: 2003 to 2005.
| Year | Number of maternities |
Number of women |
Rate per 1000 maternities |
95 per cent CI | |
|---|---|---|---|---|---|
Women with severe maternal morbidity |
|||||
| 2003 | 51,902 | 270 | 5.2 | 4.6 | 5.8 |
| 2004 | 53,502 | 246 | 4.6 | 4.0 | 5.2 |
| 2005 | 53,819 | 329 | 6.1 | 5.5 | 6.8 |
| Change 2004 to 2005 | 1.5 | 0.6 | 2.4 | ||
Major obstetric haemorrhage |
|||||
| 2003 | 51,902 | 176 | 3.4 | 2.9 | 3.9 |
| 2004 | 53,502 | 171 | 3.2 | 2.8 | 3.7 |
| 2005 | 53,819 | 235 | 4.4 | 3.8 | 5.0 |
| Change 2004 to 2005 | 1.2 | 0.4 | 1.9 | ||
Non-haemorrhage severe morbidity |
|||||
| 2003 | 51,902 | 94 | 1.8 | 1.5 | 2.2 |
| 2004 | 53,502 | 75 | 1.4 | 1.1 | 1.8 |
| 2005 | 53,819 | 94 | 1.7 | 1.4 | 2.1 |
| Change 2004 to 2005 | 0.3 | -0.1 | 0.8 | ||
Detailed case assessments by hospital risk management teams, using the national proforma, have been conducted on cases of major obstetric haemorrhage throughout the triennium. Forms were returned for 517 of the 582 cases, but in 14 cases the space for the local assessment of sub-optimal care were left blank. Overall assessments of quality of care for the 503 cases for which information was returned are summarised in Table 20.4. The majority of cases (65%) were considered to be well-managed with only 3% of cases being judged as receiving major sub-optimal care. Similar detailed assessments of cases of eclampsia have been conducted since 2004. Again, only a small minority of cases were judged as receiving major sub-optimal care.
Table 20.4 Overall assessments of sub-optimal care in 503 cases of major obstetric haemorrhage.
Category |
Number | Percentage |
|---|---|---|
| 1 = Appropriate care, well managed. | 327 | 65 |
|
2 = Incidental sub-optimal care – |
114 | 23 |
|
3 = Minor sub-optimal care – |
49 | 10 |
| 4 = Major sub-optimal care – Different management might have been expected to result in a more favourable outcome. The management of this case contributed significantly to the morbidity of this patient. |
13 | 3 |
Discussion
The Scottish Confidential Audit of Severe Maternal Morbidity has collected data prospectively using 14 well-defined, consistent categories of morbidity for the triennium (2003-2005), contemporaneous with this maternal death enquiry. The audit is funded by NHS Quality Improvement Scotland. Continuous data collection under the auspices of a national agency, has two principal advantages. Firstly, aggregation of data over several years means that meaningful national rates of rare outcomes such as eclampsia can be calculated. Secondly, consistent methods mean that trends over time in overall rates of severe morbidity can be examined.
The causes of severe maternal morbidity found in our study differ from the causes of maternal death found in the Confidential Enquiries. Major haemorrhage accounted for 69% of our cases of severe morbidity but for only 11% of Direct maternal deaths in the UK in the 2003 to 2005 triennium. In contrast, venous thromboembolism was the principal cause of maternal death (28% of cases), but accounted for under 2% of our cases of severe morbidity. Thus, the pattern of morbidity and mortality appears to differ from the continuum described by Pattinson5, with some clinical insults (e.g. major haemorrhage) being more amenable to alteration by prompt and appropriate treatment than others.
The approach used by the Scottish Confidential Audit relies on accurate and complete case identification at local hospital level. There have always been concerns that incident reporting systems under-estimate the true level of reportable events. In order to improve reproducibility, we have a designated member of staff within each unit with responsibility for reporting and we regularly update these participants. We would anticipate that with increasing familiarity with the study, case ascertainment would improve. Improved ascertainment may explain the apparent rise in the rate of severe morbidity.
Another explanation is that a rise in major obstetric haemorrhage has resulted from changes in the obstetric population: increasing numbers of mothers with complex medical conditions, increasing age at childbirth, increasing number of multiple pregnancies following assisted reproduction, and increasing number of caesarean sections with subsequent placenta praevia and accreta. Data from Information Services of NHS Scotland10 show that the national rate of emergency caesarean section remained constant at 15.4% over the three years of this study but the rate of elective Caesarean section rose year on year, from 8.8% in 2003 to 9.5% in 2005 (Chi-squared test for trend, p<0.0001). The relationship between changes in the obstetric population and in obstetric practice and maternal morbidity requires continuing monitoring.
Although the number of women suffering major haemorrhage increased from 2004 to 2005, there was no concomitant rise in the number of women admitted to an intensive care unit. This suggests that high dependency facilities within labour wards are absorbing this rise. This pattern of care lends support to the continued use of definitions based on pathophysiological features, rather than aspects of clinical management such as admission to an intensive care unit, which may underestimate the burden of morbidity in our population.
In general, cases of major haemorrhage were considered to be well-managed, with only 3% judged to have major sub-optimal care. This is in marked contrast to assessments of cases of maternal death; in the 2003 to 2005 triennium, 10 of 17 cases (59%) of maternal deaths due to haemorrhage were judged by the Enquiry assessors to have major substandard care. These figures suggest that our cases of severe maternal morbidity should genuinely be thought of as ‘great saves’, rather than ‘near misses’: cases where appropriate care has prevented progression of morbidity to mortality.
Currently, the Scottish Confidential Audit of Severe Maternal Morbidity can only provide theories to explain any changes in morbidity in maternity units across Scotland. Although we cannot make broad assumptions based on a rising rate in a single year, it does highlight the need for continuing monitoring. Changes in the obstetric population and in obstetric practice result in changes to the risks of childbirth for both mother and baby; continuous monitoring of mortality and morbidity represents an essential element of reflective clinical practice.