Purchase this article with an account.
D.J. Covert; A Decision Analysis for Laser in situ Keratomileusis (LASIK) in Patients With Myopia and Astigmatism . Invest. Ophthalmol. Vis. Sci. 2006;47(13):527.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Decision analysis is a technique within evidence based medicine used to maximize patient utility by integrating clinical outcomes data and patient preferences. Its strength comes from the implementation of patient preferences not only for treatment outcomes, but also for potential complications. It is useful for a range of clinical settings including elective surgery. Among elective procedures, refractive surgery is uniquely elective: potential refractive surgery patients are acutely sensitive to the possible risks in comparison to anticipated benefits. The goal of this study is to use standard decision analysis methods to develop and test a decision tree to assist patients in deciding whether to undergo LASIK.
Decision trees of varying levels of complexity were constructed. A final model was selected that minimizes complexity but preserves clinical usefulness. Separate versions were constructed for moderate and high myopic astigmatism. The 2002 AAO report on the safety and efficacy of LASIK provided the bulk of the probability estimates in the model. The model was tested with preferences from hypothetical patients with varying risk tolerances and baseline attitudes toward glasses and contact lenses. Subsequently, sensitivity analysis was performed by varying the outcome probability estimates within clinically appropriate ranges.
A model with ten outcomes was developed that included eight possible LASIK outcomes vs. eyeglasses vs. contact lenses. For LASIK outcomes, 4 final visual acuity classes were defined: UCVA 20/20 or better, 20/25 to 20/40, worse than 20/40, and worsened BCVA. Potentially bothersome side effects were grouped together for simplicity. The LASIK option maximized patient utility for a wide range of hypothetical patient preferences, and the model was robust under reasonable changes in the probability estimates.
Although the main model constructed included data from the infancy of LASIK and did not incorporate more recent LASIK advancements such as wavefront mapping and IntraLase, patient utility over a broad range of hypothetical patient preferences was maximized by choosing LASIK. It is likely that with the newer modalities currently available, the model would predict that an even broader population of patients would maximize their utility by undergoing LASIK. Limitations include the simplification of complications and the inherent assumptions of the model, whereas strengths of this technique include the insights it provides toward maximizing patient utility for groups of similar patients considering LASIK and the ability to easily incorporate newer clinical data.
This PDF is available to Subscribers Only