[Skip to Content]
[Skip to Content Landing]
Views 366
Citations 0
Invited Commentary
April 12, 2018

Value of Primary Data in Cost-effectiveness Analyses

Author Affiliations
  • 1Bernard and Shirlee Brown Glaucoma Research Laboratory, Edward S. Harkness Eye Institute, Department of Ophthalmology, Columbia University Medical Center, New York, New York
JAMA Ophthalmol. Published online April 12, 2018. doi:10.1001/jamaophthalmol.2018.0653

Increasing health care costs have spurred the growth of cost-effectiveness analysis (CEA), which is a method of analysis that compares the value of health care gains relative with a comparator. Clinicians, policy makers, payers, and patients have a vested interest in ensuring the publication of high-quality economic evaluations of novel diagnostic or therapeutic approaches. In this issue of JAMA Ophthalmology, Stroupe et al1 have conducted a randomized clinical trial–based cost-effectiveness evaluation of a visual rehabilitation program. The authors report that, compared with low-vision services, vision rehabilitation in a US Department of Veterans Affairs system delivers improved functional visual outcomes at a similar cost.

The publication is timely because last year JAMA reported updated cost-effectiveness guidelines from the Second Panel on Cost Effectiveness in Health and Medicine.2 The original guidelines,3 published 20 years earlier, created a set of standard methods to improve the comparability and quality of CEA studies. Although the guidelines are broad, a few points to highlight are as follows. Costs should be estimated from a societal perspective, meaning that the effects of an intervention on all costs should be considered, not only the direct cost of the intervention. Costs and benefits should be considered over a patient’s lifetime; future events not realized during the study period should be estimated through modeling, and future costs and benefits should be discounted at an equal rate. Lastly, health-related quality of life, measured as a quality-adjusted life-year (QALY), should be a common measure of effectiveness. A QALY is a composite “utility” measure that covers multiple health domains combined into a single overall preference-based score. It incorporates morbidity as well as mortality and is measured on a scale ranging from “perfect health” (1.0) to “death” (0). Health state preferences may be obtained directly,4 which is labor- and resource-intensive, or more commonly obtained through multiattribute questionnaires, such as the EQ-5D (EuroQol Research Foundation).5 The use of QALYs as a common denominator allows interventions to be compared across medical disciplines and thereby provides an additional tool for health policy experts to better allocate resources to the most valuable services.

In this study, Stroupe and colleagues1 chose to present separate cost and effectiveness analyses rather than cost relative to QALYs. Their economic model is based on results from the Low Vision Intervention Trial II randomized clinical trial, the primary outcome of which was comparison of the changes in reading ability on a validated questionnaire (“logit value”) at baseline and 4 months later in a case and control group. Because to our knowledge there are no published studies that convert logit values to QALYs, and because there are no commonly accepted values for “cost per logit gained,” this is a reasonable decision. This decision also highlights an important point—the selection of a modeling approach in CEA relies heavily on the extent of the input data available. At this time, multiattribute questionnaires, such as the EQ-5D, are rarely used in ophthalmic clinical trials because generic health measures lack the specificity that is required in specialty practice. Conversely, vision-specific questionnaires, such as the Veterans Affairs Low-Vision Visual Functioning Questionnaire used in this study, are highly focused and not intended to measure overall well-being. Furthermore, efforts to estimate health state utilities by mapping the National Eye Institute Visual Function Questionnaire to EQ-5D scores6 have shown poor correlations. By extension, clinical measures of visual function7 have had limited success at directly predicting EQ-5D scores. Consequently, the QALY as a general health state measure is problematic for ophthalmic cost-effectiveness researchers.

The authors should be congratulated on evaluating direct costs as well as costs from the patient and caregiver perspective, an element frequently excluded from CEA. The original panel recommended that costs be evaluated from a societal perspective, which would include both direct and indirect costs from multiple perspectives (ie, payer, patients, caregivers, and employers) to understand the full effects of a health intervention. (The second panel more recently recommended that CEAs report 2 reference case analyses, 1 based on a health care sector perspective and 1 on a societal perspective.2) Despite the panel’s recommendation, most CEAs fail to use the societal perspective. The primary reason for this is the practical difficulties in estimating these costs given the absence of primary source data. For example, there are limited high-quality data on costs related to “wages lost” in a population that is retired or has visual impairment. Likewise, the authors chose to estimate the visual gains and the associated costs of the interventions during the trial period and did not choose to project their findings to lifetime gains and costs. Because it is not currently clear to what extent visual rehabilitation outcomes improve overall health and reduce costs, a lifetime model would require multiple assumptions that would necessitate an in-depth sensitivity analyses. However, limiting the analyses to the study period and using a narrower perspective likely undervalues the intervention because the downstream benefits of the visual rehabilitation program are unrealized in this model.

Overall, the authors have justified the cost of the visual rehabilitation program in the US Department of Veterans Affairs system. However, while CEAs provide an additional tool to ensure that eye care clinicians provide high-value care, the scope of findings from CEAs will continue to be limited until we have better primary cost data and can better quantify vision in our general health state measure.

Back to top
Article Information

Corresponding Author: Dana M. Blumberg, MD, MPH, Edward S. Harkness Eye Institute, Department of Ophthalmology, Columbia University Medical Center, 635 W 165th St, New York, NY 10032 (dmb2196@cumc.columbia.edu).

Published Online: April 12, 2018. doi:10.1001/jamaophthalmol.2018.0653

Conflict of Interest Disclosures: The author has completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.

Stroupe  KT, Stelmack  JA, Tang  XC,  et al.  Economic evaluation of low-vision rehabilitation for veterans with macular diseases in the US Department of Veterans Affairs [published online April 12, 2018].  JAMA Ophthalmol. doi:10.1001/jamaophthalmol.2018.0797Google Scholar
Sanders  GD, Neumann  PJ, Basu  A,  et al.  Recommendations for conduct, methodological practices, and reporting of cost-effectiveness analyses: second panel on cost-effectiveness in health and medicine.  JAMA. 2016;316(10):1093-1103.PubMedGoogle ScholarCrossref
Weinstein  MC, Siegel  JE, Gold  MR, Kamlet  MS, Russell  LB.  Recommendations of the panel on cost-effectiveness in health and medicine.  JAMA. 1996;276(15):1253-1258.PubMedGoogle ScholarCrossref
Torrance  GW, Thomas  WH, Sackett  DL.  A utility maximization model for evaluation of health care programs.  Health Serv Res. 1972;7(2):118-133.PubMedGoogle Scholar
EuroQol Group.  EuroQol–a new facility for the measurement of health-related quality of life.  Health Policy. 1990;16(3):199-208.PubMedGoogle ScholarCrossref
Kay  S, Ferreira  A.  Mapping the 25-item National Eye Institute Visual Functioning Questionnaire (NEI VFQ-25) to EQ-5D utility scores.  Ophthalmic Epidemiol. 2014;21(2):66-78.PubMedGoogle ScholarCrossref
Browne  C, Brazier  J, Carlton  J, Alavi  Y, Jofre-Bonet  M.  Estimating quality-adjusted life years from patient-reported visual functioning.  Eye (Lond). 2012;26(10):1295-1301.PubMedGoogle ScholarCrossref