Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

Purpose: Cost-utility analysis is the preferred method of economic evaluation to support health policy decision-making in most developed nations. Utility estimation is based upon the assumption that a single universal construct, ?health related quality of life? (HRQoL) is measured. However, this assumption has never been tested empirically. If there are important aspects of health that are not captured by the HRQoL construct, the effectiveness of interventions directed towards treatment of those aspects might be underestimated. We hypothesize that vision related diseases might be a family of conditions falling into this category by testing whether vision-related quality of life is a construct that is distinct from health-related quality of life. We test this hypothesis using two instruments: the SF-36 to measure HRQoL, and the NEI Vision Function Questionnaire (NEI-VFQ) to measure vision-related quality of life (VRQoL). Methods: Between June 1998 and October 1999 443 patients from 18 ophthalmic practices were interviewed: diabetic retinopathy (DR) =59, glaucoma=99, macular degeneration (AMD)=44, cataract=132, correctable refractive error (RE)=109. The SF-36 and NEI-VFQ were administered. The SF-36 is widely used for assessment of HRQoL, and has been shown to be valid and reliable for this purpose. The NEI-VFQ has also been shown to be reliable and valid for assessment of VRQoL in people with chronic eye disease. The relationship between the two instruments was assessed using exploratory factor analysis (EFA) and variable cluster analysis (VCA). If HRQoL and VRQoL are related constructs, it would be expected that the constructs identified by these instruments would be measured by items from both the SF-36 and NEI-VFQ. In this manner, we will determine if the variables form distinct, overlapping, or integrated constructs. Results: The preliminary results suggest that vision and non-vision related quality of life are indeed two distinct constructs. In exploratory factor analysis no items from the NEI-VFQ loaded on constructs formed by the SF-36 items, or vice-versa. The SF-36 items form multiple constructs that support previous research concerning this instrument; while the NEI-VFQ items form two constructs, visual function and ocular pain. Variable cluster analysis confirms the EFA, with the SF-36 and NEI-VFQ items showing moderate correlation with items from their ?home? instrument, but weak correlation with items from the other instrument.

Conclusions: Our preliminary results provide evidence that VRQoL and HRQoL appear to be two distinct constructs that have modest association. While these analyses are based upon use of functional based measures as opposed to preference based measures such as the standard gamble (SG) or time trade-off (TTO), the results have important implications for the validity of preference based measures for assessment of effectiveness. If the SG or TTO (or instruments based upon these methods, such as the EQ-5D, HUI, or SF-6D) do not adequately measure all aspects of health, interventions addressing poorly measured problems (vision-related problems in this case) may be substantially underestimated. If this is the case, health policy makers relying on cost-effectiveness studies using these instruments might incorrectly reject health programs for treatment of important medical conditions.