Elsevier

The Lancet

Volume 355, Issue 9208, 18 March 2000, Pages 1004-1007
The Lancet

Public Health
Mortality rates after surgery for congenital heart defects in children and surgeons' performance

https://doi.org/10.1016/S0140-6736(00)90001-1Get rights and content

Summary

Background

A public inquiry into surgery for paediatric congenital heart defects in Bristol, UK, underscored the need for reliable data on overall mortality rates, which would allow assessment of individual surgeons' performance. We aimed to gather and report such data for 1 year to provide information for clinicians, researchers, policy makers, and the general public.

Methods

We collected data on all operations (1378) for congenital heart defects done by 11 surgeons in five departments in the UK between April 1, 1997, and March 31, 1998. These operations represented about 36% of all operations done in the UK during that time. Clearly defined criteria were agreed to classify operations into subgroups.

Findings

The overal mortality rate for all operations was 4·0% (95% Cl 3·0–5·2). No deaths occurred for 67 arterial-switch operations. Mortality rates for coarctation, ventricular septal defect, atrioventricular septal defect, Fallot, and truncus arteriosus operations were 1·1%, 0·6%, 3·6%, 2·3%, and 28·6%, respectively. Although overall mortality rates between surgeons varied (1·6–6·9%), no surgeon's were higher than the 95% Cl. The numbers of operations done by individual surgeons were small, which led to wide confidence intervals and made the detection of differences in performance difficult.

Interpretation

The participating departments seemed to reach high standards of care for children with congenital heart defects, although more data would be needed to assess performance of individual surgeons. The development of quality standards will be difficult because of the complexity of defects, the different types of operations, and few patients in each subgroup. Collection of larger sets of data for more patients and centres are needed.

Introduction

A public inquiry into surgery for congenital heart defects in Bristol, UK, has underscored the need for reliable information about results in paediatric cardiac surgery, which could be analysed, published, and made available for the assessment of surgical performance. Few data have been reported on overall mortality rates for paediatric cardiac surgery.1, 2 Published data are mainly from centres achieving outstanding results and refer to selected types of operations.3, 4

The development of a standard system for collection of data on congenital heart defects in the UK and abroad has been slow. Differences in nomenclature of diagnoses and operations, difficulties associated with coding of complex lesions, concern about reliability and validation of data, and lack of resources have been the main hindrances.5, 6 Some, but not all, centres have used commercially available databases or their own databases to monitor surgical performance.

In 1997, five UK centres of paediatric cardiac surgery with established data-collection procedures (Glasgow, London [Great Ormond Street Hospital for Sick Children], Leeds, Newcastle, and Southampton) agreed to participate in this study. They were able to provide results of all operations for congenital heart defects for 1 year. Our main goal was to provide basic data and make them available to clinicians, researchers, policy makers, and the general public. We assessed also variability between and within centres and methods for assessment of mortality.

Section snippets

Centres and procedures

The five invited centres all used electronic systems for data collection; had databases that allowed them to report mortality rates for all open, closed, and six selected marker operations; agreed on definitions of operations to be used in the study, had one consultant who would collect data or closely supervised collection' and gave permission that data specific to each surgeon and centre could be published.

Each centre agreed to provide results of operations in two agegroups—infants (≤1 year)

Findings for operations

1378 operations were performed, with 55 (4%) deaths, during the study. The total number represents about 36% of all procedures for paediatric congenital heart defects done in the UK during that time. Closed and open heart surgery carried a higher risk for infants than for older children. Overall mortality for closed procedures was low; six (1·7%) of all 360 children died (table 1). For open heart surgery 49 (4·8%) of 1018 children died. For five of the marker operations, the technical

Differences between centres and surgeons

Since the numbers of open and closed operations done in each centre, except centre five, were small, the 95% CI were wide. Mortality rates for open operations ranged from 2·2% to 7·4%, and for closed operations from 0% to 3·8% (table 3). The variations in mortality rates for the six marker operations are shown in table 4. There were substantial apparent differences between centres for some operations—for example, in centre one, one death occurred after Fallot's tetralogy repair, but gave a

Study quality

The strength of this study is in the care taken by all participants to collect the data as completely and as accurately as possible. We believe that the information obtained can contribute to the current discussion about assessment of surgical performance.

We recognise that our methods bear scrutiny and, as is common in observational studies, there are undoubtedly weaknesses. The duration of the study and sample size were not based on statistical power calculations but were dictated by the

Implications

Our results show that the participating departments reached a high standard of care in surgery for congenital heart defects. The overall mortality rate was 4%. Analysis of the marker operations suggested that even technically difficult operations, such as repair of Fallot's tetralogy or arterial switch, can carry low risks. The only operation associated with a high mortality was truncus arteriosus. Since, however, the numbers of operations were small and 95% CI wide, the results should be

Standards and guidelines

Use of standards would facilitate quality control in paediatric cardiac surgery. Preparation of standards could proceed in several stages. The data needed for setting them should be collected purposely over a long time, preferably from all departments in the country. Standards for all operations and for specific operations, should be based on reliable data. A system for collecting such data needs to be organised and supervised.6 Before standards are agreed, it is important to discuss how they

Conclusion

Most children with congenital heart defects can now undergo operations with low risk of mortality. Although the number of cases we studied was large in absolute terms, creating subgroups by diagnosis and operation gave small datasets, which may have made detection of systematic differences in performance between centres, or between surgeons, difficult. Surgeons' performance should not be assessed based on results achieved for one operation only.

We thank M R de Leval, M J Elliott, A Hasan, M P

References (19)

There are more references available in the full text version of this article.

Cited by (0)

View full text