Evaluation of the stages of completion and scoring of the Patient Generated Index (PGI) in patients with rheumatic diseases.

Abstract

To evaluate the stages of completion and approaches to scoring the PGI for reliability, validity and responsiveness. Participants of inpatient rehabilitation or self-management programmes completed the closed PGI with the same areas at 1 year as baseline. Test-retest reliability, validity and responsiveness were assessed for area scores (stage one), points (stage two) and methods of scoring the PGI. One hundred and forty-five patients participated, and 118 (81 %) completed the PGI correctly. Test-retest intraclass correlations were over 0.90 for area scores (stage two) and were 0.87 and 0.86 for final PGI scores with and without the sixth "rest of life" box. Individual area scores had the highest correlations with those for instruments assessing similar constructs; those for the area "rest of life" were lower. Compared to scores based on the sum of the stage two areas, PGI scores had higher correlations of a moderate level with those for patient-reported instruments widely used within rheumatology. Correlations were of a similar level with and without the sixth "rest of life" area, and those based on baseline points at follow-up were highest. The PGI had higher SRMs than the other instruments at 1 year, the highest being for PGI scores based on baseline points. The fully closed version of the PGI, which uses baseline areas and baseline stage three points at follow-up, is most appropriate for assessing outcomes within healthcare evaluation. The sixth "rest of life" area has poorer measurement properties, and its removal does not adversely affect the measurement properties of the PGI.