Use of SF-36 and SF-12 health status measures: a quantitative comparison for groups versus individual patients

Med Care. 2001 Aug;39(8):867-78. doi: 10.1097/00005650-200108000-00012.


Background: The extent to which SF-12 scores reflect SF-36 scores has not been well studied.

Objectives: One purpose was to compare the sensitivity to change of the SF-36 Physical Function sub-score, and the Physical Component Summary Scores (PCS) of the SF-36 and SF-12 on patients with low back pain (LBP). A second purpose was to determine if the SF-12 could serve as a surrogate measure for the SF-36 when making decisions about individual patients.

Subjects: The sample consisted of 101 consecutive patients.

Measures: SF-36 questionnaires were completed by patients at both initial and discharge examinations. SF-12 scores were calculated from the completed SF-36 questionnaires. Therapists' judgments of whether patients were judged to have returned to premorbid function served as the construct for meaningful clinical change.

Analysis: Receiver Operating Characteristic (ROC) curve analysis and repeated measures MANCOVA were used to assess sensitivity to change. Linear regression and 95% prediction bands described the extent to which SF-12 scores predict individual SF-36 scores.

Results: No significant differences were found between the ROC curve areas for the Physical Function sub-scale, the PCS-36 and PCS-12. No significant differences were found for the comparison of change scores between PF-36, PCS-36 and PCS-12 scores.

Conclusion: The findings suggest that Physical Function sub-scores, SF-36 and SF-12 PCS scores are equally sensitive to change. SF-12 PCS scores do not adequately predict SF-36 PCS scores for individual patients. The PCS-12 should probably not be used to make judgments about the health status of individual patients with LBP.

Publication types

  • Comparative Study

MeSH terms

  • Activities of Daily Living / classification*
  • Adult
  • Aged
  • Decision Making
  • Female
  • Health Status Indicators*
  • Humans
  • Linear Models
  • Low Back Pain / diagnosis*
  • Low Back Pain / rehabilitation*
  • Male
  • Middle Aged
  • Multivariate Analysis
  • Pain Measurement / methods*
  • ROC Curve
  • Surveys and Questionnaires*