Estimating the Incidence of Acute Infectious Intestinal Disease in the Community in the UK: A Retrospective Telephone Survey
Authors:
Laura Viviani, Mike van der Es, Lisa Irvine, Clarence C. Tam, Laura C. Rodrigues, Kathryn A. Jackson, Sarah J. O’Brien, Paul R. Hunter, IID2 Study Executive Committee
Abstract:
Objectives:
To estimate the burden of intestinal infectious disease (IID) in the UK and determine whether disease burden estimations using a retrospective study design differ from those using a prospective study design.
Design/Setting:
A retrospective telephone survey undertaken in each of the four countries comprising the United Kingdom. Participants were randomly asked about illness either in the past 7 or 28 days.
Participants:
14,813 individuals for all of whom we had a legible recording of their agreement to participate.
Outcomes:
Self-reported IID, defined as loose stools or clinically significant vomiting lasting less than two weeks, in the absence of a known non-infectious cause.
Results:
The rate of self-reported IID varied substantially depending on whether asked for illness in the previous 7 or 28 days. After standardising for age and sex, and adjusting for the number of interviews completed each month and the relative size of each UK country, the estimated rate of IID in the 7-day recall group was 1,530 cases per 1,000 person-years (95% CI: 1135-2113), while in the 28-day recall group it was 533 cases per 1,000 person-years (95% CI: 377-778). There was no significant variation in rates between the four countries. Rates in this study were also higher than in a related prospective study undertaken at the same time.
Conclusions:
The estimated burden of disease from IID varied dramatically depending on study design. Retrospective studies of IID give higher estimates of disease burden than prospective studies. Of retrospective studies longer recall periods give lower estimated rates than studies with short recall periods. Caution needs to be exercised when comparing studies of self-reported IID as small changes in study design or case definition can markedly affect estimated rates.
Journal:
PLOS One