Annex A > Chapter 19 - Statistics Relating to the Clinical Performance of Paediatric Cardiac Surgical Services in Bristol Compared with Other Specialist Centres during the Period 1984 to 1995 > The Bristol surgeons' formal written comments on the Experts' Overview Report and the Experts' responses > Assessments of the Experts' Overview Report
<< previous | next >>
Assessments of the Experts' Overview Report
212 The Inquiry commissioned two experts in statistics to assess the Overview Report: Professor Stephen Gallivan Director of the Clinical Operational Research Unit at University College London; and Professor Robert Curnow, Emeritus Professor at the University of Reading and former President of the Royal Statistical Society. Professor Gallivan commented:
`In order to investigate the robustness of their findings, the authors have carried out an extensive programme of "sensitivity analysis".
`... In all cases, even with the final artificially optimistic scenario, analysis still indicated strong evidence for substantial excess mortality at Bristol.
`In view of the evidence from this sensitivity analysis, one can only conclude that the analytical approach adopted was indeed statistically robust.' [293]
`No major errors or ambiguities were detected in the report.' [294]
`... to deny the divergence of outcome, one would need to believe that HES and [UK]CSR both gave grossly inaccurate estimates for many of the 11 mortality rates, or alternatively, that case load and mortality at Bristol has been hugely misrepresented. In the opinion of this reviewer, this is highly improbable.' [295]
`Given such strong evidence that outcomes at Bristol were indeed divergent, analysis has been carried out to examine whether there were intrinsic factors, other than case mix, that could mitigate such findings.
`... It can be seen that none of these factors provides mitigation for the level of excess mortality estimated and provides further reassurance that the findings are not a statistical artefact. Again, this suggests that the analysis provides reliable and valid evidence of divergence of the outcomes at Bristol.' [296]
`In my view, the scrutiny of the various data sources, their comparison and synthesis; the data analyses, and the reporting of the conclusions from the analyses are of the highest professional standard. The analyses have been exhaustive and have been based on appropriate and up-to-date methodology. In their interpretative role, the authors have steered a careful route, balancing the possible misinterpretations of the available data because of biases in the collection and collation stages with the need to derive robust conclusions where this is possible. The authors have studied the sensitivity of their conclusions and recommendations to the inadequacies of the data. This has been done in terms of both possible biases in the data and the representation of the inevitable random variation in the outcome measures. In no place could I identify areas where the authors had been other than cautious in their findings.
`I can therefore confirm that I believe the authors have fulfilled their remit. The analytical approach is statistically robust and fit for purpose. I have found no errors or ambiguities of a statistical nature. I believe the overall conclusions to be reliable and valid.' [297]
<< previous | next >> | back to top
Footnotes
[293] INQ 0045 0070 - 0071 ; `Peer Review Report - Overview of statistical evidence', October 2000, Professor Gallivan
[294] INQ 0045 0071; `Peer Review Report - Overview of statistical evidence', October 2000, Professor Gallivan
[295] INQ 0045 0073; `Peer Review Report - Overview of statistical evidence', October 2000, Professor Gallivan (emphasis in original)
[296] INQ 0045 0074; `Peer Review Report - Overview of statistical evidence', October 2000, Professor Gallivan
[297] INQ 0045 0067 Professor Robert Curnow