Introduction The “Hawthorne Effect” is the phenomenon in which subjects modify practice as a consequence of the knowledge that they are being observed. This is a potential confounder during periods of national endosocopy audit and may result in spuriously improved outcome reporting during audit periods. We aimed to investigate whether the Hawthorne Effect influences colonoscopy practice. We also aimed to ascertain if the national colonscopy audit could result in a change in practice, and whether any such change was maintained.
Methods The Unisoft endoscopy database at Whipps Cross University Hospital was interrogated to determine patient demographics, sedation rates, quality of bowel preparation, diagnoses and therapeutic interventions during 5 2-week time periods; The national colonoscopy audit period (t), t−1 year, t−2 weeks, t+2 weeks and t+3 months. Results were compared to determine whether there was a statistically significant difference in measurable indices of clinical practice that may be due to the Hawthorne Effect. Time periods following the audit period were included to establish whether there was any evidence of a “washout period” of improved outcomes following the national audit—that is, if the process of observed audit results in a lasting improvement in clinical practice. The null hypothesis was suggested that all periods would be similar, and tested to a 95% confidence level.
Results Colonoscopies performed during the national colonoscopy audit period (t) were compared with 2-week periods t−1 year, t−2 weeks, t+2 weeks and t+3 months. Similar numbers of procedures were carried out during the five time periods. Basic patient demographics were similar, as were the numbers of male and female patients. No statistically significant differences were found in the sedative dose, ceacal or TI intubation rates between the audit period and any other time period. Moreover, polyp detection and retrieval was likewise also not statistically significantly different when the four time periods were compared with the fortnight of the national colonscopy audit. Small differences were noted in the colonscopists assessment of bowel preparation—there was more likely be a comment on poor bowel preparation during the audit period than any of the other time periods.
Conclusion Data from Whipps Cross University Hospital demonstrate that observation of colonoscopists during the recent BSG national colonoscopy audit does not alter significantly the clinical practice or interpretation of findings when compared to time periods before or after the audit period. This validates the national colonoscopy audt findings; the data are indeed a true reflection of “normal” colonsocopy practice—colonoscopists are apparently not affected by the “Hawthorne Effect”.
Competing interests None declared.