M W Cooke1, S Wilson, P Bridge. 1. Emergency Medicine, Research Group, University of Warwick, Coventry. mwcooke@emerg-uk.com
Abstract
BACKGROUND: Questionnaires are commonly sent to accident and emergency (A&E) departments to determine common practice and are often extrapolated to best practice. AIMS: To determine if questionnaire based studies have a defined population of A&E departments and whether studies are reproducible. METHODS: All questionnaires in the Journal of Accident and Emergency Medicine were reviewed and assessed for inclusion criteria, departments studied and study design. RESULTS: 30 questionnaires were detected, 22 were postal, six telephone and two did not state method of contact. Sample sizes ranged from 15 to 740 and inclusion of A&E departments was highly variable according to geographical area, size of department or consultant status. Seventeen (54.8%) did not state the source of A&E department listings. Response rates ranged from 55-100%. Only three studies undertook subset analysis according to either size or locality. CONCLUSIONS: Questionnaire of studies A&E departments have poor methodology descriptions, which means that many are not reproducible. Inclusion criteria are highly variable and failure to analyse important subsets may mean that individual departments cannot apply recommendations. Questionnaire studies relating to A&E do not use a consistent well defined population of A&E departments. Information in the studies is usually inadequate to allow them to be repeated.
BACKGROUND: Questionnaires are commonly sent to accident and emergency (A&E) departments to determine common practice and are often extrapolated to best practice. AIMS: To determine if questionnaire based studies have a defined population of A&E departments and whether studies are reproducible. METHODS: All questionnaires in the Journal of Accident and Emergency Medicine were reviewed and assessed for inclusion criteria, departments studied and study design. RESULTS: 30 questionnaires were detected, 22 were postal, six telephone and two did not state method of contact. Sample sizes ranged from 15 to 740 and inclusion of A&E departments was highly variable according to geographical area, size of department or consultant status. Seventeen (54.8%) did not state the source of A&E department listings. Response rates ranged from 55-100%. Only three studies undertook subset analysis according to either size or locality. CONCLUSIONS: Questionnaire of studies A&E departments have poor methodology descriptions, which means that many are not reproducible. Inclusion criteria are highly variable and failure to analyse important subsets may mean that individual departments cannot apply recommendations. Questionnaire studies relating to A&E do not use a consistent well defined population of A&E departments. Information in the studies is usually inadequate to allow them to be repeated.