Eyeworld

EW RESIDENTS 64 November 2015 current (pilot) study is large, testing the hypothesis that LCS may cause more CME than PCS would require approximately four times the num- ber of eyes than that in the current study. Most single-site studies may regrettably suffer from inadequate power and subsequent difficulty interpreting results. In addition, in ideal experi- mental design, both the patients and the providers would be masked to treatment (randomized, masked design). This would minimize any observer bias that might influence the visual assessments, decision to order an SD-OCT, or the appropriate follow-up interval. Similarly, the patients would be randomized and masked to the treatment interven- tion. In this study, patients who met inclusion criteria were offered the laser-assisted option, but only those who chose to pay extra were allo- Review continued from page 63 cated to the laser treatment group. The extra cost may affect a patient's reporting of symptoms, which could, in turn, influence the practi- tioner's decision making. However, this non-randomized grouping does reflect how treatment decisions are made in the real world, and is per- haps more clinically relevant, even if it is not experimentally optimal. The journal club is an excellent format to explore these important aspects of experimental design and the inherent limitations of clinical research. Understanding what con- clusions from published studies are supported by the data and which are speculative based on a study design that cannot adequately address the hypothesis is a critical skill. As the medical system is moving more to- ward "evidence-based medicine," it is imperative that physicians under- stand the difference between find- ings in a study and the conclusions drawn. As an example, retrospective reviews of clinical outcomes are gen- erally useful for hypothesis generat- ing, but not for hypothesis testing due to lack of treatment consistency and changing practice patterns over time. Prospective studies are far less prone to outcomes biases, but even in the present study, we observed that the software upgrades changed the treatment parameters during the recruitment period. Further, as the current study demonstrates, trans- lating the study results to clinical practice is more likely when the study uses clinically relevant indica- tors in the experimental design (i.e., obtaining SD-OCT only when there is reduced visual acuity at the 3–4 week visit). The current study is an excellent example of how real-world outcomes can inform our under- standing of the choices we make in clinical practice and how to begin to assess what those decisions might mean for our patients. EW References 1. Abell RG, Kerr NM, Vote BJ. Toward zero effective phacoemulsification time using femtosecond laser pretreatment. Ophthalmology 2013;120:942–948. 2. Mayer WJ, Klaproth OK, Hengerer FH, Kohnen T. Impact of crystalline lens opacification on effective phacoemulsification time in femtosecond laser-assisted cataract surgery. Am J Ophthalmol 2014;157:426–432. 3. Ewe SYP, Oakley C, Abell RG, Allen PL, Vote BJ. Cystoid macular edema after femtosecond-laser as- sisted versus phacoemulsification cataract surgery – a prospective comparative cohort case series. J Cataract Refract Surg 2015;41. Article in press. 4. Henderson BA, Kim JY, Ament CS, Ferrufino -Ponce ZK, Grabowska A, Cremers SL. Clinical pseudophakic cystoid macular edema. Risk factors for development and duration after treatment. J Cataract Refract Surg 2007;33:1550–1558. 5. Loewenstein A, Zur D. Postsurgical cystoid macu- lar edema. Dev Ophthalmol 2010; 47:148–159. 6. Melberg NS, Olk RJ. Corticosteroid-induced ocular hypertension in the treatment of aphakic or pseudophakic cystoid macular edema. Ophthalmol- ogy. 1993;100(2):164–7. 7. Levitz L, Reich J, Roberts TV, Lawless M. Incidence of cystoid macular edema: femto- second laser-assisted cataract surgery versus manual cataract surgery. J Cataract Refract Surg 2015;41:683–6. Contact information Culican: Culican@vision.wustl.edu

• Cover
• _GoBack
• _GoBack (1)