Reliability, Validity, and Health Issues Arising From Questionnaires Used to Measure Psychosocial and Organizational Work Factors (POWFs) Among Hospital Nurses: A Critical Review

Abstract

This systematic review assesses the validity of epidemiological questionnaires used to measure psychosocial and organizational work factors (POWFs) in nurses. Of the 632 articles published between 1980 and July 2008 identified in this review, 108 provide some data concerning analysis of the intrinsic characteristics of such instruments (content validity or conceptual basis, reliability, validation of internal construction) and their external validity with respect to health aspects (concurrent validity and predictive validity). Psychometric properties of generalist questionnaires validated among blue collar or white collar workers were also assessed in the nurse population. The Job Content Questionnaire (JCQ), because of its longevity and reputation, was the generalist questionnaire most used among this population. Although its structure often raises questions in the nurse population, its dimensions (mainly the control one) have been shown to be predictive of some health outcomes measured with “objective” indicators concerning absenteeism, injuries, and musculoskeletal disorders. Effort Reward Imbalance (ERI), which has a structure more stable among the nurse population, has shown concurrent validity in terms of intent to leave the nursing profession. No questionnaire specifically designed for nurses can claim to satisfy all of the recommendations in terms of internal validity. Nevertheless, the Practice Environment Scale–Nursing Work Index (PES–NWI) seems to be one of the most promising instruments because of its appropriateness (content validity), its structure, which has a rather good fit (construct validity), its ability to discriminate magnet hospitals like other NWI derivates (discriminant validity), and it has also been associated in cross-sectional studies with health outcomes, especially nurses’ self-assessed mental health but also with patients’ health outcomes objectively assessed (concurrent validity). However, elements for predictive validity are still lacking with NWI derivates. The Discussion provides recommendations for measuring POWFs, encompassing the use of external validated measurements.