Graeme Hawthorne1, Richard H Osborne, Anne Taylor, Jan Sansoni. 1. Department of Psychiatry, Royal Melbourne Hospital, The University of Melbourne, Level 1 North, Main Building, Grattan St, Parkville, VIC, 3052, Australia. graemeeh@unimelb.edu.au
Abstract
BACKGROUND: The SF36 Version 2 (SF36V2) is a revision of the SF36 Version 1, and is a widely used health status measure. It is important that guidelines for interpreting scores are available. METHOD: A population sample of Australians (n = 3015) weighted to achieve representativeness was administered the SF36V2. Comparisons between published US weights and sample derived weights were made, and Australian population norms computed and presented. MAJOR FINDINGS: Significant differences were observed on 7/8 scales and on the mental health summary scale. Possible causes of these findings may include different sampling and data collection procedures, demographic characteristics, differences in data collection time (1998 vs. 2004), differences in health status or differences in cultural perception of the meaning of health. Australian population norms by age cohort, gender and health status are reported by T-score as recommended by the instrument developers. Additionally, the proportions of cases within T-score deciles are presented and show there are important data distribution issues. PRINCIPAL CONCLUSIONS: The procedures reported here may be used by other researchers where local effects are suspected. The population norms presented may be of interest. There are statistical artefacts associated with T-scores that have implications for how SF36V2 data are analysed and interpreted.
BACKGROUND: The SF36 Version 2 (SF36V2) is a revision of the SF36 Version 1, and is a widely used health status measure. It is important that guidelines for interpreting scores are available. METHOD: A population sample of Australians (n = 3015) weighted to achieve representativeness was administered the SF36V2. Comparisons between published US weights and sample derived weights were made, and Australian population norms computed and presented. MAJOR FINDINGS: Significant differences were observed on 7/8 scales and on the mental health summary scale. Possible causes of these findings may include different sampling and data collection procedures, demographic characteristics, differences in data collection time (1998 vs. 2004), differences in health status or differences in cultural perception of the meaning of health. Australian population norms by age cohort, gender and health status are reported by T-score as recommended by the instrument developers. Additionally, the proportions of cases within T-score deciles are presented and show there are important data distribution issues. PRINCIPAL CONCLUSIONS: The procedures reported here may be used by other researchers where local effects are suspected. The population norms presented may be of interest. There are statistical artefacts associated with T-scores that have implications for how SF36V2 data are analysed and interpreted.
Authors: S D Keller; J E Ware; B Gandek; N K Aaronson; J Alonso; G Apolone; J B Bjorner; J Brazier; M Bullinger; S Fukuhara; S Kaasa; A Leplège; R W Sanson-Fisher; M Sullivan; S Wood-Dauphinee Journal: J Clin Epidemiol Date: 1998-11 Impact factor: 6.437
Authors: Brennan M R Spiegel; Zobair M Younossi; Ron D Hays; Dennis Revicki; Sean Robbins; Fasiha Kanwal Journal: Hepatology Date: 2005-04 Impact factor: 17.425
Authors: Kimberly E Alexander; Suzanne Chambers; Amanda B Spurdle; Jyotsna Batra; Felicity Lose; Tracy A O'Mara; Robert A Gardiner; Joanne F Aitken; Judith A Clements; Mary-Anne Kedda; Monika Janda Journal: Qual Life Res Date: 2015-02-28 Impact factor: 4.147
Authors: Sergeui Pakhomov; Nilay Shah; Penny Hanson; Saranya Balasubramaniam; Steven A Smith; Steven Allan Smith Journal: AMIA Annu Symp Proc Date: 2008-11-06