S78 Do Tuberculosis Cases Managed By Clinicians With Average Annual Caseloads Below 10 Have Poorer Treatment Outcomes?

Abstract

Introduction and objectives The 2007 Department of Health Tuberculosis Toolkit advises that clinicians should not be solely managing tuberculosis (TB) cases if their average caseload is less than 10 per year. A systematic evaluation of whether these guidelines are being followed, and how effective such a threshold is, has not been undertaken in the UK. Methods All UK TB cases notified 2003–2011 were extracted from Public Health England’s Enhanced Tuberculosis Surveillance system. Mean caseload for each clinician was calculated over the preceding year and three years by using case notification date. 12 month TB treatment outcomes were categorised as unfavourable or good/neutral.1,2 Cases without clinician information and resistant to rifampicin were excluded, the latter due to UK recommendations on the length of treatment. The proportion of cases managed by clinicians with a caseload under 10 was analysed, then random effects logistic regression utilised to determine the relationship between caseload and treatment outcomes, adjusting for clustering by clinician and confounding. Results 74,550 TB cases were notified 2003–11. The proportion of TB cases seen by a clinician who had a low caseload (less than 10) in the preceding year declined gradually 2004–11 (42 to 28%), with no apparent acceleration post-Toolkit. Univariate modelling demonstrated very strong evidence of increased odds of an unfavourable treatment outcome among cases seen by a clinician who had a low caseload over the preceding three years (cluster-specific odds ratio 1.23 (95% confidence interval 1.14–1.33), p-value Conclusions Our analysis indicates that TB cases managed by clinicians with a mean caseload of under 10 over the preceding three years have increased odds of an unfavourable treatment outcome. However other factors, such as the number of clinicians managing TB cases in the hospital overall, should potentially be taken into account for policy recommendations. References Ditah, Thorax 2008;63:440–446 Anderson, Euro Surveill. 2013;18(40):pii=20601