A novel excel sheet conversion tool from Snellen fraction to LogMAR including ‘counting fingers’, ‘hand movement’, ‘light perception’ and ‘no light perception’ and focused review of literature of low visual acuity reference values

Visual acuity (VA) is paramount to outcome measures in ophthalmic studies. The Early Treatment Diabetic Retinopathy Study (ETDRS) chart is internationally recognized as 'gold standarnumerical imputation for low vision stad' for VA measurement in clinical trials and population surveys [ICO 1984 + 2002 reports]. It utilizes the logarithm of the minimum angle of resolution (LogMAR); a continuous measure of VA with a linear scale. In contrast, Snellen fraction uses a geometric progression, which is inherently skewed, however, remains commonly reported in clinical practice. When performing statistical analysis, non-conversion of Snellen fraction to LogMAR results in erroneous results and misrepresentative statistical analyses (Holladay 1997; Holladay 2004; Yu & Afifi 2009). Therefore, it becomes necessary to convert Snellen fraction to LogMAR accurately when using retrospective data (Holladay 1997; Holladay 2004). Online conversion tables and calculators are available albeit impractical; they are tedious for large databases and do not account for additional or missed letters. For additional or missing letters, the most common methodology employs ± 0.02 LogMAR per missing/gained letter as per Ferris et al. (1982). Furthermore, in patients with poor vision, the resolving visual acuities' ordinal scale of 'counting fingers' (CF), 'hand movement' (HM), 'light perception' (LP) and 'no light perception' (NLP) are employed. These values commonly are quantified and converted to a numerical form permitting statistical analysis (Grover et al. 1999; Holladay 2004; Schulze-Bonsel et al. 2006; Bach 2007; Lange et al. 2009; Day et al. 2015). Multiple authors have recommended or used various conversions which are summarized in Table 1. NLP would intuitively equal 0.0 decimal VA, which is non-quantifiable in LogMAR units. Whilst CF and HM are measurable with acceptable test-retest repeatability, LP and NLP are not (Schulze-Bonsel et al. 2006). Schulze and Bach (co-author for Schulze-Bonsel et al. (2006)) propose that three lines on a LogMAR chart are generally clinically significant (Schulze-Bonsel et al. 2006), consequently assigning LP and NLP numerical imputations of 0.3 and 0.6 LogMAR units below HM, respectively (Bach 2007). The National Ophthalmology Database (NOD) has published many papers for numerous cases in the UK, consistently using the reference values found in Table 1 for their publications. These referenced values have been utilized in other large case-series publications (Chu et al. 2016). 0.014 (decVA) 1.85 LogMAR 0.005 (decVA) 2.30 LogMAR 433 On: 06/06/20 246 On 06/06/20 National ophthalmology database. Values consistently used in every publication from this group to allow comparison Reason for reference values, not stated. 112 On 06/06/20 144 On 06/06/20 498 On 06/06/20 Microsoft Excel® (Microsoft Corporation, Redmond, WA) is a spreadsheet program that is widely used for data collection and basic analysis. A recent publication by colleagues outlined a method using Excel to convert Snellen fraction to LogMAR (Tiew et al. 2020). This method, however, is limited in that it does not allow direct entry of the Snellen fraction into one cell. For example, '6/9-2' would need to be entered across three different cells ('6', '9', '-2'). Moreover, this method does not allow for conversion of non-numerical values such as CF, HM, LP and NLP. We demonstrate a user-friendly and efficient Excel sheet® method, enabling conversion of Snellen fraction and non-numerical VA values to LogMAR. This formula converts one cell of Excel to another cell and can be copy pasted and dragged down a column to convert large data of Snellen fraction rapidly and effectively including low VA, to LogMAR. Our Excel formula has several advantages including: A downloadable Excel Sheet (Appendix S1) will enable customization of the formula in terms of reference LogMAR for CF, HM, LP and NLP. It allows re-labelling, for example LP to PL, as well as editing the starting reference cell (currently A2). By default, the formula utilizes the United Kingdom NOD values (Table 1; Day et al. 2015). Figure 1 outlines more detailed instructions and illustrates use. We exercise caution in performing statistical analysis particularly with VA < HM. As no author has reliably and reproducibly converted LP and NLP, these are numerical imputations to aid quantification in database analysis. Consequently, we recommend avoiding parametric analysis and utilizing non-parametric analyses to interpret values as ordinal data if converting to LogMAR for statistical analysis. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.