The following is a summary of “Risk Calculation in the Medication Arm of the Ocular Hypertension Treatment Study,” published in the June 2023 issue of Opthalmology by Leshno et al.
For a study, researchers aimed to assess the predictive accuracy of the ocular hypertension study (OHTS) calculator among treated patients in the OHTS medication arm. They conducted a retrospective secondary analysis after a randomized clinical trial. Study included individuals from the OHTS assigned to the medication arm, with complete baseline data in both eyes (n=726).
They determine the generalizability of the OHTS calculator to the OHTS medication arm; the hazard ratios (HR) of the medication group in OHTS were compared to the HR used in the OHTS calculator, employing the Z-test statistic. The performance of the OHTS Calculator was assessed twice within the OHTS medication group, considering both the untreated baseline intraocular pressure (IOP) and the average treated IOP during the initial 24 months as variables.
The model’s performance was evaluated by assessing its accuracy in predicting the risk of reaching an OHTS primary open-angle glaucoma (POAG) endpoint. This was done through the calibration chi-square and discrimination between participants who developed POAG and those who did not.
They observed no significant difference between the HRs of the OHTS medication arm and the HRs used in the OHTS calculator for untreated OHTN based on data. The OHTS calculator prediction model demonstrated good predictive accuracy when utilizing the mean-treated IOP in the medication group. However, its predictive accuracy was lower when using the untreated baseline IOP. The model exhibited good discrimination with treated IOP (c-statistic = 0.77) compared to the reported performance of the OHTS calculator in the OHTS observation group.
They concluded that applying the OHTS Calculator to treated OHTN patients is feasible and conducting repeat risk calculations after initiating IOP reduction can offer valuable insights to guide disease management.