advertisement

Editors Selection IGR 10-1

Response

Comment by William Steward on:

20389 Cost-effectiveness of treating ocular hypertension, Stewart WC; Stewart JA; Nassar QJ et al., Ophthalmology, 2008; 115: 94-98

Find related abstracts

It is most gratifying that the reviewers found acceptance of the concept of outcome studies in ophthalmology. Economic outcome studies have only recently become published in the ophthalmic literature despite their prior common use in systemic medicine. Economic outcomes data have the advantage of providing cost/efficacy information to assist the practitioner with their therapeutic decisions.

However, several difficulties exist in using economic outcomes data clinically. Firstly, outcomes research evaluates efficacy from a cost, as well as the health, view point. However, cost considerations may not always be appropriate clinically. Secondly, much of the economic data is based on assumptions, either from the literature or from physician interviews that limit a model's ability to fit any disease process perfectly. Thirdly, few standards exist for economic outcome study methodology. This lack of standards may lead to conflicting preferred methodology among individual investigators. Lastly, a danger exists that government or private payers will use outcomes data to limit physician and patient therapeutic choices. Consequently, as stated in our article, and emphasized by several reviewers, economic outcomes data cannot replace the practitioner's clinical judgment in making therapeutic decisions. Accordingly, these data must be used with caution by both physicians and payers.

Our own model had its own set of needed assumptions and limitations. We based our utilities on the work of Tengs and Wallace who created utilities for visual acuity versus those specifically for glaucoma by Kobelt or Jampel.^2,3,4When our model was created these two glaucoma based utility studies were not known to us. However, after careful review by the publishing journal the utilities were not changed for several reasons. First, this was an ocular hypertension (OHT) model and glaucoma utilities, as useful as they are, did not generally apply to our study. Only approximately 4% of patients progressed to glaucoma in our model.⁵ Second, for the small percent of patients who progressed, and for whom the glaucoma utilities would apply, the differences between the Tengs and Wallace values and the glaucoma utilities were small. We compared internally two separate glaucoma Markov models with the Kobelt utilities versus those by Tengs and Wallace. We found no subjective and minimal objective differences in the results. Consequently, the use of the glaucoma based utilities in our model would have had minimal influence on our findings and the current results are valid. We used the OHTS and EMGT as a basis for our treatment decisions. One comment notes that the treatment guidelines in these studies were 'tight'. Unfortunately, there are few standards for how to treat OHT. Consequently, although these two excellent trials were not necessarily a perfect clinical fit for most routine clinical practices, they provided a basis of treatment decisions that are generally accepted in the literature. We chose to evaluate costs for our sensitivity analyses. This is traditional for sensitivity analyses testing to show the robustness of the Markov model from a cost standpoint.⁶ However, to the reviewer's point such sensitivity analysis is not always limited to cost parameters. We hoped in publishing this study that it would provide a basis of discussion to assist in incorporating economic outcomes data into treatment decision parameters within OHT and glaucoma on an ap-propriate clinical basis. As several of the comments recognized, much further work is needed in outcomes research to standardize outcomes procedures, gain acceptance within the medical community and learn how to use the results in an appropriate clinical fashion.

• ← Previous
• Next →