PURPOSE: To determine whether the choice of intraocular lens (IOL) power formula improves IOL power predictions and whether personalized constants within the IOL power formula are critical factors in improving refractive predictions after combined penetrating keratoplasty, cataract extraction, and IOL implantation. METHODS: Records of 46 patients who had the triple procedure between January 1988 and December 1992 were evaluated using the SRK II, SRK/T, Holladay, and Hoffer Q formulas to predict the postoperative spherical equivalent refractions for implanted lens power. Calculations were carried out with and without the use of personalized constants. The predictive accuracy of each formula was assessed by comparing the actual postoperative spherical equivalent refractive error with that predicted by the formulas. The predictive error and the distribution of predictive errors were used to assess predictive accuracy. RESULTS: There was no difference in the mean absolute predictive errors and the distribution of predictive errors for the four formulas evaluated (P > .05). The use of personalized formula constants significantly reduced the mean absolute predictive error for the SRK II, SRK/T, and Holladay formulas (P < .05) and approached significance for the Hoffer Q formula. CONCLUSION: The findings suggest that the choice of IOL power formula does not affect IOL power predictions in the corneal triple procedure; however, personalized constants within a formula appears to be a critical factor in improving postoperative refractive predictions.
PURPOSE: To determine whether the choice of intraocular lens (IOL) power formula improves IOL power predictions and whether personalized constants within the IOL power formula are critical factors in improving refractive predictions after combined penetrating keratoplasty, cataract extraction, and IOL implantation. METHODS: Records of 46 patients who had the triple procedure between January 1988 and December 1992 were evaluated using the SRK II, SRK/T, Holladay, and Hoffer Q formulas to predict the postoperative spherical equivalent refractions for implanted lens power. Calculations were carried out with and without the use of personalized constants. The predictive accuracy of each formula was assessed by comparing the actual postoperative spherical equivalent refractive error with that predicted by the formulas. The predictive error and the distribution of predictive errors were used to assess predictive accuracy. RESULTS: There was no difference in the mean absolute predictive errors and the distribution of predictive errors for the four formulas evaluated (P > .05). The use of personalized formula constants significantly reduced the mean absolute predictive error for the SRK II, SRK/T, and Holladay formulas (P < .05) and approached significance for the Hoffer Q formula. CONCLUSION: The findings suggest that the choice of IOL power formula does not affect IOL power predictions in the corneal triple procedure; however, personalized constants within a formula appears to be a critical factor in improving postoperative refractive predictions.