2007; 29: 504–506 SHORT COMMUNICATION Validation of the PHEEM instrument in a Danish hospital setting K. ASPEGREN, L. BASTHOLT, K. M. BESTED, T. BONNESEN, E. EJLERSEN, I. FOG, T. HERTEL, T. KODAL, J. LUND, J. S. MADSEN, A. MALCHOW-MØLLER, M. PETERSEN, B. SØRENSEN & L. WERMUTH The Postgraduate Deanery and Office of Medical Education, Faculty of Medicine, University of Southern Denmark, Odense, Denmark Abstract The Postgraduate Hospital Educational Environment Measure (PHEEM) has been translated into Danish and then validated with good internal consistency by 342 Danish junior and senior hospital doctors. Four of the 40 items are culturally dependent in the Danish hospital setting. Factor analysis demonstrated that seven items are interconnected. This information can be used to shorten the instrument by perhaps another three items. Introduction The education of junior doctors is an important aspect of hospital output. For it to be effective, the hospital must be a ‘learning organization’, possessing qualities such as informa- tion sharing, worker participation and innovation, using failure events to improve work processes (Argyris 1999). ‘The establishment of a supportive learning-oriented culture is of outmost significance in creating competent physicians’ (Hoff et al. 2004). To facilitate hospital development towards this goal, measurement instruments of good quality directed towards situated learning in hospitals are needed. The subject has been comprehensively reviewed by Roff (2005). The Postgraduate Hospital Educational Environment Measure (PHEEM) is a 40-item instrument to evaluate junior hospital doctors’ learning environment developed in the UK (Roff et al. 2005). We were interested in using it in Denmark for quality work in the early phases of postgraduate training. However, due to possible differences in work conditions for junior doctors and in hospital culture between UK and Denmark, validation in a Danish hospital setting was necessary before regular use. We were also interested in investigating whether a shorter version of the questionnaire could be made without losing important aspects of the original 40-item version, because short questionnaires stand a better chance of being answered by those asked (Streiner & Norman 1995). Methods To ensure content and meaning, the questionnaire was first translated from English into Danish by a Danish doctor proficient in English and then retranslated into English by a professional translator. An English doctor, working in Denmark, proficient in the Danish language, compared the two versions. Some minor differences in meaning between the two versions were adjusted. This adjusted Danish version was sent to senior doctors (consultants and doctors in administrative posts) and junior doctors (pre-registration house officers and doctors in specialist training programmes) in a broad selection of hospital departments (see Table 1). Selection of departments was non-randomized but aimed at cover specialities with many junior doctors in training. Doctors were asked to rate the importance of each item on a five-grade Likert-type scale (from Not important to Very important). Data were analysed by the SAS statistical program for internal consistency using Cronbach’s alpha and factor analysis using Varimax rotation. Factors were chosen using Kaiser–Guttman criterion (Eigenvalue >1.00). Only factors with a loading value larger than 0.60 were included. In the following sections, we use the numbering of items from the original publication of PHEEM (Roff et al. 2005). Results Complete questionnaire ratings were obtained from 159 senior and 183 junior doctors. Mean participation rate was 55% of approached doctors and is specified by specialities in Table 1. Compliances with each item varied between 100% and 97% and are given in Table 2. Internal consistency of data was high (Cronbach’s alpha 0.93). Mean rating and standard deviation for each of the 40 items are given in Table 2. Items numbers 1, 9, 11 and 26 were found to have relatively low ratings. Many responders also wrote comments to these items, considering them irrelevant in a Danish hospital setting. Therefore, we exclude them from the Danish version of the PHEEM. Factor analysis showed that seven items were interconnected (nos 10, 15, 22, 28, 31, 33, and 38), explaining 8.6% of the total variance. Correspondence: Professor Knut Aspegren, Office of Medical Education, Faculty of Medicine, University of Southern Denmark Winsløwparken 19, DK-5000 Odense C, Denmark. Email: knut.aspegren@post.tele.dk 504 ISSN 0142–159X print/ISSN 1466–187X online/07/050504–3 ß 2007 Informa UK Ltd. DOI: 10.1080/01421590701477357 Med Teach Downloaded from informahealthcare.com by University of Southern Denmark on 09/30/14 For personal use only.