Introduction In the UK, the confidential enquiry into maternal death recommended that a modified early warning score (MEWS) should be used for all obstetric admissions to improve early recognition of sick women.

Methods The authors adapted the standard MEWS taking into account the physiological changes of pregnancy and introduced a combined observation/early warning chart with parameters colour coded to highlight abnormal readings and management escalation policy for abnormal scores. Compliance with this was variable so the authors started bedside training of staff. One day a month, the observation charts of all patients on the ward were checked for compliance with our guidelines.

Results The charts of 694 women were audited. Compliance with recording MEWS was good, but repeating it at appropriate intervals, as specified in our guidelines improved from 40% to 100%, over the audit period as did recording three times daily until discharge from 23% to 86%. Compliance with recording pulse, blood pressure and temperature was consistently >95%, but other parameters less so. By the end of the audit period compliance for all parameters was 100% from: respiratory rate 73%, urine output 32%, oxygen saturation 74% and conscious level 11%.

5.8% of women had MEWS >3, our cut off for escalation, 80% of these was reviewed within 30 min by medical staff.

Conclusion Full compliance with recording and charting basic observations and MEWS can be achieved by using a bedside teaching approach and audit. More work is needed to see whether using MEWS in addition to basic observations improves outcomes.