Comparing the Cancer-Specific EORTC-8D and the Generic EQ-5D-3L

There is a wealth of literature that questions the sensitivity of generic preference-based measures (PBMs) of health. To address this a number of condition-specific PBMs have been developed, although to date their use in decision-making is limited. If CSPBMs are to be more widely adopted, then evidence of their performance is required. In this recently published open access paper OHE’s Paula Lorgelly (with colleagues from Cambridge and Sheffield) assess the validity, responsiveness and sensitivity of a cancer-specific preference-based measure, the EORTC-8D, relative to the generic PBM, the EQ-5D-3L.

The analysis utilises Cancer 2015 a large-scale prospective longitudinal population-based molecular cancer cohort study from Victoria, Australia. In addition to clinical data and patient socio-demographics and familial history and participants self-completed patient reported outcome measures (PROMs) at baseline and various follow-up points depending on how advanced the cancer was. The PROMs included the EORTC QLQ-C30 which gives EORTC-8D values and the EQ-5D-3L. The analysis of baseline values used a complete case sample of 1678 patients; while 1157 patients had at least one follow-up point (some had up to five follow-ups) for whom QALYs were estimated.  The average follow-up was 434 days, the longest follow-up (in this slice of the dataset) was 1146 days.

Descriptive analysis found that the mean baseline health state value for the EORTC-8D was higher (0.829) than for the EQ-5D-3L (0.748). This may be a function of the EORTC-8D having a higher ‘floor’, and the lowest possible health state value is 0.292 compared with the EQ-5D-3L floor of -0.594. The assessment of convergent validity found that the various dimensions in each instrument and instrument scores were strongly correlated, while the analysis of content validity found few ceiling effects. Despite this, the agreement between the instruments was poor, with considerable variation in values for those with lower baseline health related quality of life.

The condition-specific QALYs estimated using the EORTC-8D were significantly higher (0.909) than those derived from the EQ-5D-3L over time (0.860) although the difference was small (0.049, p<0.001). Both the generic and condition-specific QALYs were found to be similarly sensitive to a number of patient and disease characteristics. When specifically considering the difference in QALYs (condition-specific minus generic QALY estimates) multivariate regression analysis found that the variation in baseline health state values had a large influence on the difference in QALYs.

The research team concluded that given the variability in sensitivity between the baseline values and the QALY estimations means researchers and decision makers are advised to be cautious if using the instruments interchangeably.  The marginally higher QALY estimates may be appealing but further research is required to confirm if these higher estimates offer a more accurate reflection of health related quality of life gains.

Reference:

Lorgelly, P.K., Doble, B., Rowen, D. Brazier, J., 2016. Condition-specific or generic preference-based measures in oncology? A comparison of the EORTC-8D and the EQ-5D-3L. Quality of Life Research

Posted in EQ-5D and PROMs, Research | Tagged External publications