Do self-report and medical record comorbidity data predict longitudinal functional capacity and quality of life health outcomes similarly?
1 College of Human Medicine, Clinical Center Building, Michigan State University, 788 Service Road, Room B329, East Lansing, MI, 48824, USA
2 College of Human Medicine, Institute for Health Care Studies, Michigan State University, 965 Fee Road, Room A134, East Lansing, MI, 48824, USA
3 College of Nursing, Michigan State University, 424A West Fee Hall, East Lansing, MI, 48824, USA
4 PharmNet/i3, 5572 Star Flower Dr, Haslett, MI, 48840, USA
5 Health Services Research, Michigan State University, C203 E. Fee Hall, East Lansing, MI, 48824, USA
Citation and License
BMC Health Services Research 2012, 12:398 doi:10.1186/1472-6963-12-398Published: 14 November 2012
The search for a reliable, valid and cost-effective comorbidity risk adjustment method for outcomes research continues to be a challenge. The most widely used tool, the Charlson Comorbidity Index (CCI) is limited due to frequent missing data in medical records and administrative data. Patient self-report data has the potential to be more complete but has not been widely used. The purpose of this study was to evaluate the performance of the Self-Administered Comorbidity Questionnaire (SCQ) to predict functional capacity, quality of life (QOL) health outcomes compared to CCI medical records data.
An SCQ-score was generated from patient interview, and the CCI score was generated by medical record review for 525 patients hospitalized for Acute Coronary Syndrome (ACS) at baseline, three months and eight months post-discharge. Linear regression models assessed the extent to which there were differences in the ability of comorbidity measures to predict functional capacity (Activity Status Index [ASI] scores) and quality of life (EuroQOL 5D [EQ5D] scores).
The CCI (R2 = 0.245; p = 0.132) did not predict quality of life scores while the SCQ self-report method (R2 = 0.265; p < 0.0005) predicted the EQ5D scores. However, the CCI was almost as good as the SCQ for predicting the ASI scores at three and six months and performed slightly better in predicting ASI at eight-month follow up (R2 = 0.370; p < 0.0005 vs. R2 = 0.358; p < 0.0005) respectively. Only age, gender, family income and Center for Epidemiologic Studies-Depression (CESD) scores showed significant association with both measures in predicting QOL and functional capacity.
Although our model R-squares were fairly low, these results show that the self-report SCQ index is a good alternative method to predict QOL health outcomes when compared to a CCI medical record score. Both measures predicted physical functioning similarly. This suggests that patient self-reported comorbidity data can be used for predicting physical functional capacity and QOL and can serve as a reliable risk adjustment measure. Self-report comorbidity data may provide a cost-effective alternative method for risk adjustment in clinical research, health policy and organizational improvement analyses.
Clinical Trials.gov NCT00416026