S Basu1. 1. Department of Ophthalmology, William Harvey Hospital, Kennington Road, Ashford, Kent TN24 0LZ, UK. sbasu2003@doctors.net.uk
Abstract
PURPOSE: This study evaluates the agreement between the calculations performed by IOLMaster and as performed according to the published methods, for three formulas, Holladay I, Hoffer Q, and SRK-T. METHOD: Hypothetical and identical data were used as input variables for calculations by IOLMaster and by a computer program written in visual basic to perform calculations with the three formulas. Results were compared and statistical significance carried out. RESULTS: Significant disagreement was found using Hoffer-Q formula between the results shown by the study program and IOLMaster, while identical results were obtained using Holladay I and SRK-T formulas. For an average eye (AL 23.45 mm and K 45D, ACD constant 6.15, vertex 12 mm), the mean difference in the predicted postoperative (postop) SE between the study-set and IOLMaster dataset was 0.13 D (+/-0.03 D), which is a measure for the systematic error (relative bias). In such eyes, the mean predicted postop ACD in IOLMaster dataset was 0.11 mm (+/-0.04 mm) less than in the study-set, and the data spread of the predicted postop ACD was much higher (+/-0.042 mm) indicating more discrepancy in predicting ACD with IOLMaster. CONCLUSION: When using Hoffer-Q formula, the mean predicted postop SE with IOLMaster was 0.13 D less than the computation in this study, that is, IOLMaster was showing a consistently myopic overestimation, or, a mean undercorrection of necessary intraocular lens power by 0.13 D, or, a residual hyperopic error of 0.13 D. This translates to a consistently more anterior estimation (0.11 mm mean) of predicted postop ACD by IOLMaster, as compared with theoretical analysis.
PURPOSE: This study evaluates the agreement between the calculations performed by IOLMaster and as performed according to the published methods, for three formulas, Holladay I, Hoffer Q, and SRK-T. METHOD: Hypothetical and identical data were used as input variables for calculations by IOLMaster and by a computer program written in visual basic to perform calculations with the three formulas. Results were compared and statistical significance carried out. RESULTS: Significant disagreement was found using Hoffer-Q formula between the results shown by the study program and IOLMaster, while identical results were obtained using Holladay I and SRK-T formulas. For an average eye (AL 23.45 mm and K 45D, ACD constant 6.15, vertex 12 mm), the mean difference in the predicted postoperative (postop) SE between the study-set and IOLMaster dataset was 0.13 D (+/-0.03 D), which is a measure for the systematic error (relative bias). In such eyes, the mean predicted postop ACD in IOLMaster dataset was 0.11 mm (+/-0.04 mm) less than in the study-set, and the data spread of the predicted postop ACD was much higher (+/-0.042 mm) indicating more discrepancy in predicting ACD with IOLMaster. CONCLUSION: When using Hoffer-Q formula, the mean predicted postop SE with IOLMaster was 0.13 D less than the computation in this study, that is, IOLMaster was showing a consistently myopic overestimation, or, a mean undercorrection of necessary intraocular lens power by 0.13 D, or, a residual hyperopic error of 0.13 D. This translates to a consistently more anterior estimation (0.11 mm mean) of predicted postop ACD by IOLMaster, as compared with theoretical analysis.