Purpose: To develop a Survival-based Quality of Life Assessment (SQLA) process based on clinically relevant tradeoffs, for use in supporting bedside decisions.
Method: Three small studies were done. (1) 6 electronic cigarette users (vapers) evaluated quality of life with monocular blindness, binocular blindness, current health, and forced return to smoking in a survival curve trading task comparable to a time trade-off or standard gamble. Subjects answered 2 comprehension questions and then selected an indifference curve from 7 rainbow color-coded survival curves representing alternatives to a status-quo black dashed curve. (2) A convenience sample of 17 subjects ranked 4 potential mixed graphic and screen interfaces for SQLA assessments, including an introduction, two text choices and (A) two separate curves, (B) superimposed dashed and solid curves, (C) the dashed interface with text labels, or (D) superimposed curves distinguished by pastel colors matched to choices. (3) A procedure was developed for calculating a fixed discount rate and a fixed utility for a health state from the results of two straight-line survival curve trading tasks.
Result: (1) Vapers answered 83% of comprehension questions correctly, with no errors off by more than one curve. None reversed the expected preference of monocular over binocular blindness. However, vapers gave and cogently defended the usual range of preferences, some valuing blindness close to perfect health, while others valued it close to quick death. Most vapers matched smoking to the same curve they chose for blindness, and the remainder matched it to the next better curve. (2) Colors was significantly preferred to labels, which was preferred to dashes and separate (p<0.001). Colors was the first or second choice for all subjects. (3) Given a linear survival curve with maximum survival time T2 in a health state with utility u, if T1 is the maximum survival time in the equivalent perfect health survival curve, and the discount rate is x, then u = (((1/(1+x))^T1 )*1/T1 – (ln(1/(1+x)) + 1/T1))/(((1/(1+x))^T2 )*1/T2 – (ln(1/(1+x)) + 1/T2)))) If x is unknown, performing two assessments of one health state using different T2 levels permits calculation of x and u.
Conclusion: SQLA tasks are feasible. A human interface comprising superimposed, color-coded survival curves is preferable to some alternatives. Utility and discount rate can be distinguished, if assumptions about their stability hold.
See more of: The 33rd Annual Meeting of the Society for Medical Decision Making