Multicentre criterion based audit of the management of induced abortion in ScotlandBMJ 1994; 309 doi: https://doi.org/10.1136/bmj.309.6946.15 (Published 02 July 1994) Cite this as: BMJ 1994;309:15
- G C Penney,
- A Glasier,
- A Templeton
- Department of Obstetrics and Gynaecology, Aberdeen Maternity20Hospital, Aberdeen AB9 2ZD Lothian Health Board, Family Planning and Well Woman Services, Edinburgh EH4 1NL
- Correspondence to: Dr Penney.
Objectives: To assess and improve the quality of care provided to women undergoing induced abortion. Design — Two rounds of prospective, criterion based case note review audit.
Setting: Ten NHS gynaecology units throughout Scotland.
Subjects: 2004 patient episodes of abortion care identified consecutively during two rounds of audit. The first round comprised 967 cases and the second round 1037.
Interventions: Dissemination of results from the first round of audit and recommendations for change20in the form of a written report and at postgraduate meetings in participating hospitals.
Main outcome measures: Improvements in quality of care as assessed against 16 previously agreed criteria, both overall across the 10 study hospitals and within individual hospitals.
Results: Overall, four significant improvements occurred: increased availability of early medical abortion, decreased utilisation of surgical abortion at very early gestation, increased use of mifepristone priming before second trimester medical abortion, and increased provision of follow up. At the individual hospital level 42 of 150 elements of care studied20were “close to optimal” at the time of the first round of audit, rising to 54 at the second round (NS). A total of 31 significant improvements in individual elements of care occurred, but 11 significant deteriorations also occurred (at the P<0.05 level).
Conclusions: The prospective multicentre audit proved feasible aqnd achieved the aims of any form of audit in terms of identifying deficiencies and variations in care. The audit results promoted objective review of local abortion services in participating hospitals. At least for some elements of care in some hospitals significant improvements were detectable.
Induced abortion is one of the commonest components of the Scottish NHS gynaecological workload, accounting for around 11 000 procedures per year.1 Abortion care was therefore considered to be a particularly appropriate topic for medical audit and was recently addressed by means of a criterion based approach in the gynaecology audit project in Scotland.2
A list of criteria for good quality care was agreed by a combination of objective review of contemporary medical publications, panel discussions, and postal survey of all consultant gynaecologists in Scotland (response rate 92%), as described.3 The 16 criteria which were addressed in the multicentre audit are listed in the box.
Ten hospitals throughout Scotland representing 10 different health board areas and employing around half of all consultant gynaecologists in Scotland were included. During each of two audit periods an audit assistant (with a medical secretarial background) in each hospital identified all patients undergoing induced abortion. Mothers undergoing abortion because of fetal abnormality were excluded as they were thought to raise different care issues. Data relating to the agreed criteria were collected in a standardised manner in all 10 hospitals by transfer of information from case notes on to a review document. Each audit assistant was trained by the research fellow (GCP) in data transfer. Data from all hospitals were entered into a purpose designed database by using Paradox (Borland) software on an IBM compatible personal computer. Statistical significance testing by X2 analysis (unless otherwise indicated) was performed with Instat 2 (Graphpad) software.
After the first audit period results and recommendations for change were disseminated as a written report mailed to all consultant gynaecologists. In addition, in seven of the 10 hospitals a postgraduate meeting was held, allowing a further opportunity to highlight local issues.
A data collection period of six months (beginning in October 1992) was available for the first round of audit. It was planned to review 100 consecutive cases managed in each hospital during this time. In the larger hospitals 100 cases were identified during as little as six weeks, whereas in one of the smaller hospitals only 67 cases presented during the whole six months. Thus in total 967 cases were studied. Owing to constraints of the project timetable and funding data collection for the second round of audit was restricted to a two month period (beginning in August 1993). All cases managed in each hospital during this time were identified and reviewed. The mean number of cases per hospital was 104 (range 20 to 266), and in total 1037 cases were studied.
At the time of the second round of audit four significant overall improvements across the 10 hospitals were detected:
The use of medical abortion for women at <9 weeks' gestation (criterion 8) rose from 39 to 516 (7.6%) to 172 of 541 (31.8%) (P<0.0001)
The inappropriate use of surgical abortion in women at <7 weeks' gestation (criterion 9) decreased from 68 to 85 (80.0%) to 56 of 98 (57.1%) (P=0.0017)
The use of mifepristone cervical priming before midtrimester medical abortion (criterion 12) increased from 15 of 64 women (23.4%) to 64 of 102 (62.7%) (P<0.0001)
The recording of a follow up arrangement in the case notes (criterion 15) increased from 52% to 69% of cases (Wilcoxon signed rank test, P=0.037), and the advising of follow up within the recommended interval of two weeks after abortion also increased (from 5% to 32%; Wilcoxon signed rank test, P=0.0645).
There were no overall deteriorations in relation to any elements of care.
Improvements in individual hospitals
Fifteen of the agreed criteria were addressed on an individual basis in each of the 10 hospitals. (Criterion 9 was omitted in view of the small numbers of patients presenting at <=6 weeks in any one hospital.) Thus 150 individual elements of care were assessed. Results from the first round of audit are summarised in table I. Of these 150 elements of care, 42 (28.0%) were judged to be “close to optimal” (arbitrarily defined as >=90% of cases meeting the criterion, or (for criterion 5) <=25% of cases undergoing ultrasound scanning, (for criteria 8 and 12) mifepristone being available).
Performance was particularly good in relation to recording of Rh status (nine hospitals achieved >90%) and recording of patient's contraceptive plan (eight hospitals achieved >90%). No hospital achieved close to optimal performance in relation to seeing patients within five days of referral, recording of cervical smear history, providing contraceptive supplies, or suggesting follow up within two weeks of the abortion.
The results from the second round of audit are summarised in table II. Fifty four elements of care (36.0%) were judged to be close to optimal. This increase over the first round failed to reach significance (P=0.1732). However, major improvements occurred in relation to 31 individual elements of care. These comprised the introduction of a new service (either early medical abortion or the use of mifepristone for second trimester medical abortion) in six instances or significant (P<0.05) improvements in the proportion of cases meeting a given criterion in 25 instances. The most common significant improvements were in relation to recording of cervical smear history, recording of a follow up plan, and advising early follow up (each occurred in five hospitals).
Eleven significant deteriorations in the proportion of cases meeting a given criterion also occurred. The most common deterioration was in relation to the proportion of cases seen by a gynaecologist within five days of referral (three hospitals).
There was no significant difference between the number of improvements occurring in the seven hospitals in which a postgraduate meeting was held (23 improvements out of a possible 105) and the three hospitals in which results and recommendations were disseminated as a report only (eight out of a possible 45) (P=0.6632).
The multicentre audit proved feasible and achieved good cooperation among gynaecologists. While acknowledging the theoretical advantages of clinical rather than medical audit we perceived that this project engendered a strong sense of ownership as a result of being a confidential single discipline exercise initiated by the Scottish branch of the Royal College of Obstetricians and Gynaecologists, of which all participants were members.
The first round of audit enabled variations and deficiencies in care to be identified and undoubtedly prompted open and objective discussions about local abortion services in participating hospitals. It might be argued that some of the changes detected in the second round were unrelated to the audit feedback exercise. However, one of the most dramatic improvements (the increase in mifepristone priming before midtrimester abortion) occurred at a time unrelated to new publications on the topic and before the granting of a product licence. Medical audit can never be the sole stimulus to change; it acts alongside other educational initiatives. In this instance, however, we believe the audit was instrumental in changing practice.
It is clear that in order to maximise the benefits from the national investment in clinical audit attention must be directed at identifying the most effective strategies for disseminating audit results and for implementing change. Russell and Grimshaw have provided suggestions for the effective dissemination and implementation of clinical practice guidelines4 (to which audit recommendations are closely related).5 They suggest that dissemination is best accomplished by “specific educational intervention” rather than simply “mailing targeted groups.” However, in this study no more improvements occurred in those hospitals where a postgraduate meeting was held specifically to address the audit findings than in those where a postal report only was provided.
Implementation entails encouraging clinicians to adopt the audit recommendations. The only implementation strategy employed in this study was forewarning participants that a second round of audit was imminent. A more effective strategy might have been to incorporate recommendations in a structured patient record document for use while managing abortion referrals, a technique which has proved successful in implementing guidelines for the management of infertility.6
The modest success of this national audit has confirmed an interest and willingness to participate in medical audit among Scottish gynaecologists. Refinements in audit methodology are required to translate this good will into substantial improvements in patient care.
We acknowledge the cooperation of the following consultant gynaecologists/audit assistants who cordinated the audit in their respective hospitals NHS Trust, P B Terry/J Mowat, J Fryers; Edinburgh Royal Infirmary, G E Smart/K Duncan; Ninewells Hospital, Dundee, N Patel/W Barrie; J Watson; Southern General Hospital, Glasgow, M Carty/C Doak; Ayrshire Central Hospital, S Prigg/J, Morland, A-M Garvey, J Howie, C McIntyre; Raigmore Hospital, Inverness, M Hulse/M Gordon; Victoria Hospital, Kirkcaldy, M Hill/M Telfer; Dumfries and Galloway Infirmary, G Gordon/A Graham; Inverclyde Royal Hospital, Greenock, L Cassidy/U Stevenson; Hairmyres Hospital, East Kilbride, J Gran/M McCartney, Elaine Stirton computerised the data and typed the manuscript. Daphne Russell provided statistical advice. The gynaecology audit project in Scotland is founded by a grant from the Clinical Resource and Audit Group of the Scottish Office of the Home and Health Department (but the views expressed are ours and not necessarily those of the Clinical Resource and Audit Group or the department).