Abstract
PURPOSE: The purpose of this study was to validate the use of a mixture of strategies (Swedish Interactive Thresholding Algorithm [SITA] Fast and Standard) when conducting the mean total deviation (mTD) trend analysis. METHODS: Data were collected retrospectively from 128 eyes of 95 patients with open-angle glaucoma with at least 5 SITA Fast visual field (VF) followed by 5 SITA Standard VFs. First, the mTD of a fifth future VF in 5.3 years was predicted using various numbers (from 1-2 to 1-4) of SITA Standard VFs. Similar prediction calculations were performed simultaneously using the SITA Fast VFs before the SITA Standard VFs. Furthermore, this calculation was iterated by adjusting for the difference between the SITA Fast and SITA Standard. RESULTS: Without SITA Fast VFs, the means of prediction errors were 10.40 and 4.44 decibels (dB) with 2 and 3 SITA Standard VFs, respectively. These values significantly decreased to 2.83 and 2.40 dB with SITA Fast VFs without adjusting for the difference between SITA Fast and Standard VFs. The values after adjusting for the difference between SITA Fast and Standard VFs were 2.88 and 2.39 dB, respectively, which were not significantly different. With 4 SITA Standard VFs, the prediction errors varied between 2.12 and 2.34 dB across the 3 methods; however, they were not significantly different from each other. CONCLUSIONS: It is useful to add information from prior SITA Fast VFs when conducting mTD trend analysis with SITA Standard VFs. TRANSLATIONAL RELEVANCE: It is clinically beneficial to include both SITA Standard and Fast VFs in mTD trend analysis.