Terrence Steyer, M.D.
Chair, Department of Family Medicine
G.S. Hammond, T.T. Aoki (1992)
The DIMS was designed at the University of California, Davis to measure longitudinal changes in health status of both type 1 and type 2 diabetes patients during clinical trials. The instrument is comprised of 44 items covering five subscales: diabetes-specific symptoms, nonspecific symptoms, well-being, diabetes-related morale, and social role fulfillment. It is self-administered and takes 15-20 minutes to complete. Four items correlated poorly with the overall score and should be dropped from future applications. For missing items, the value for the average of all responses for that item was assigned. Internal reliability of the five subscales were fair to good, with Cronbach’s ? ranging from 0.60 to 0.85. While no true test-retest analysis was conducted, surveys completed by patients a second time at least one month later had good to excellent correlations for the five subscores (0.60 to 0.78, p<0.001). Principle component analysis was conducted to verify the appropriateness of the five subscales. As a measure of the measure’s validity, the total score and four of the five subscores (except for social role) correlated significantly with HbA1c levels. However, DIMS scores were not significantly correlated with a diabetes complications index, and positive correlations with age and sex did occur (Garratt et al. 2002).
DIMS has not been specifically tested in an elderly population.
DIMS has not been specifically tested in a minority population.
No subsequent studies using this survey tool have been reported.
In a review of health outcome measures for diabetes Garratt et al. (2002) state that since the development of DIMS did not employ patient input, it is not as strong a measure as some other tools. The survey has fair to good internal reliability for the five subscales, but even the authors state that some uncertainty exists in interpretation and meaningfulness of the subscales (Hammond and Aoki 1992). For now, they suggest, only the total score should be used as an overall indicator of diabetes impact. The subscores are most useful in longitudinal studies as effective individual measures comparing treatments.
The DIMS is published in full as an appendix to Hammond and Aoki (1992). A copy of the questionnaire, instructions, and a scoring key are available from the authors.
Thomas T. Aoki
University of California, Davis
Division of Endocrinology
4301 X Street, Building FOLB II-C
Sacramento CA, 95816
Garratt, A.M., L. Schmidt, and R. Fitzpatrick (2002) Patient-assessed health outcome measures for diabetes: a structured review. Diabetic Med., 19, 1-11.
Hammond, G.S. and T.T. Aoki (1992) Measurement of Health Status in diabetic patients. Diabetes Care 15(4), 469-477.