Measuring physical and mental health using the SF-12: implications for community surveys of mental health

Aust N Z J Psychiatry. 2006 Sep;40(9):797-803. doi: 10.1080/j.1440-1614.2006.01886.x.

Abstract

Objective: The effects of using different approaches to scoring the SF-12 summary scales of physical and mental health were examined with a view to informing the design and interpretation of community-based survey research.

Method: Data from a population-based study of 7485 participants in three cohorts aged 20-24, 40-44 and 60-64 years were used to examine relationships among measures of physical and mental health calculated from the same items using the SF-12 and RAND-12 approaches to scoring, and other measures of chronic physical conditions and psychological distress.

Results: A measure of physical health constructed using the RAND-12 scoring showed a monotonic negative association with psychological distress as measured by the Goldberg depression and anxiety scales. However, a non-monotonic association was evident in the relationship between SF-12 physical health scores and distress, with very high SF-12 physical health scores corresponding with high levels of distress. These relationships highlight difficulties in interpretation that can arise when using the SF-12 summary scales in some analytical contexts.

Conclusions: It is recommended that community surveys that measure physical and mental functioning using the SF-12 items generate summary scores using the RAND-12 protocol in addition to the SF-12 approach. In general, researchers should be wary of using factor scores based on orthogonal rotation, which assumes that measures are uncorrelated, to represent constructs that have an actual association.

Publication types

  • Comparative Study
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Age Factors
  • Data Collection*
  • Factor Analysis, Statistical
  • Female
  • Health Status Indicators*
  • Humans
  • Male
  • Mental Health / statistics & numerical data*
  • Middle Aged
  • Psychometrics
  • Reproducibility of Results
  • Sickness Impact Profile*
  • Statistics as Topic
  • Surveys and Questionnaires