C Jenkinson1, R Layte. 1. Health Services Research Unit, University of Oxford, Radcliffe Infirmary, UK.
Abstract
OBJECTIVES: The 36 item short form health survey (SF-36) has proved to be of use in a variety of settings where a short generic health measure of patient-assessed outcome is required. This measure can provide an eight dimension profile of health status, and two summary scores assessing physical function and mental well-being. The developers of the SF-36 in America have developed algorithms to yield the two summary component scores in a questionnaire containing only one-third of the original 36 items, the SF-12. This paper documents the construction of the UK SF-12 summary measures from a large-scale dataset from the UK in which the SF-36, together with other questions on health and lifestyles, was sent to randomly selected members of the population. Using these data we attempt here to replicate the findings of the SF-36 developers in the UK setting, and then to assess the use of SF-12 summary scores in a variety of clinical conditions. METHODS: Factor analytical methods were used to derive the weights used to construct the physical and mental component scales from the SF-36. Regression methods were used to weight the 12 items recommended by the developers to construct the SF-12 physical and mental component scores. This analysis was undertaken on a large community sample (n = 9332), and then the results of the SF-36 and SF-12 were compared across diverse patient groups (Parkinson's disease, congestive heart failure, sleep apnoea, benign prostatic hypertrophy). RESULTS: Factor analysis of the SF-36 produced a two factor solution. The factor loadings were used to weight the physical component summary score (PCS-36) and mental component summary score (MCS-36). Results gained from the use of these measures were compared with results gained from the PCS-12 and MCS-12, and were found to be highly correlated (PCS: rho = 0.94, p < 0.001; MCS: rho = 0.96, p < 0.001), and produce remarkably similar results, both in the community sample and across a variety of patient groups. CONCLUSIONS: The SF-12 is able to produce the two summary scales originally developed from the SF-36 with considerable accuracy and yet with far less respondent burden. Consequently, the SF-12 may be an instrument of choice where a short generic measure providing summary information on physical and mental health status is required.
OBJECTIVES: The 36 item short form health survey (SF-36) has proved to be of use in a variety of settings where a short generic health measure of patient-assessed outcome is required. This measure can provide an eight dimension profile of health status, and two summary scores assessing physical function and mental well-being. The developers of the SF-36 in America have developed algorithms to yield the two summary component scores in a questionnaire containing only one-third of the original 36 items, the SF-12. This paper documents the construction of the UK SF-12 summary measures from a large-scale dataset from the UK in which the SF-36, together with other questions on health and lifestyles, was sent to randomly selected members of the population. Using these data we attempt here to replicate the findings of the SF-36 developers in the UK setting, and then to assess the use of SF-12 summary scores in a variety of clinical conditions. METHODS: Factor analytical methods were used to derive the weights used to construct the physical and mental component scales from the SF-36. Regression methods were used to weight the 12 items recommended by the developers to construct the SF-12 physical and mental component scores. This analysis was undertaken on a large community sample (n = 9332), and then the results of the SF-36 and SF-12 were compared across diverse patient groups (Parkinson's disease, congestive heart failure, sleep apnoea, benign prostatic hypertrophy). RESULTS: Factor analysis of the SF-36 produced a two factor solution. The factor loadings were used to weight the physical component summary score (PCS-36) and mental component summary score (MCS-36). Results gained from the use of these measures were compared with results gained from the PCS-12 and MCS-12, and were found to be highly correlated (PCS: rho = 0.94, p < 0.001; MCS: rho = 0.96, p < 0.001), and produce remarkably similar results, both in the community sample and across a variety of patient groups. CONCLUSIONS: The SF-12 is able to produce the two summary scales originally developed from the SF-36 with considerable accuracy and yet with far less respondent burden. Consequently, the SF-12 may be an instrument of choice where a short generic measure providing summary information on physical and mental health status is required.
Authors: Patrick M Pullicino; Leslie A McClure; Virginia J Howard; Virginia G Wadley; Monika M Safford; James F Meschia; Aaron Anderson; George Howard; Elsayed Z Soliman Journal: J Stroke Cerebrovasc Dis Date: 2011-12-03 Impact factor: 2.136
Authors: Rudolf W Poolman; Marc F Swiontkowski; Jeremy C T Fairbank; Emil H Schemitsch; Sheila Sprague; Henrica C W de Vet Journal: J Bone Joint Surg Am Date: 2009-05 Impact factor: 5.284
Authors: Cati G Brown-Johnson; Janine K Cataldo; Nicholas Orozco; Nadra E Lisha; Norval J Hickman; Judith J Prochaska Journal: Am J Addict Date: 2015-05-01
Authors: Alexandra Koreli; George Briassoulis; Michail Sideris; Anastas Philalithis; Savvas Papagrigoriadis Journal: In Vivo Date: 2021 Mar-Apr Impact factor: 2.155