INTRODUCTION AND OBJECTIVES: An increasing number of parents and practitioners use the Internet for health-related purposes, and an increasing number of predictive models are available on the Internet for predicting spontaneous resolution rates for children with vesico-ureteral reflux (VUR). We sought to determine whether currently available Internet-based calculators of VUR resolution produce systematically different results.

METHODS: Using a systematic Internet search, we identified 3 Internet-based calculators of spontaneous resolution rates for children with VUR: two academic and one industry-affiliated. We generated a random cohort of 100 hypothetical patients with a wide range of clinical characteristics and entered the data on each patient into each calculator. We then compared the results in terms of both mean predicted resolution probability and in terms of their discriminatory ability (i.e. the number of hypothetical patients whose probability estimates came in above or below a given threshold value), and assessed this number a across a range of potential threshold values. Statistical analysis was performed using ANOVA or Fisher’s exact tests.

RESULTS: The mean predicted resolution probability for the academic calculators were 41% and 36%, while that of the industry-affiliated calculator was 33% (p=0.02). For some patients, the calculators produced markedly different probabilities of spontaneous resolution, in some cases ranging from 24% to 89% for the same patient. At thresholds of greater than 5%, 10%, and 25% probability of spontaneous resolution, the calculators significantly diverged from one another regarding which (and how many) patients would be above the threshold (all p<0.0001).

CONCLUSIONS: Internet-based calculators of VUR resolution significantly differ from one another in terms of their predicted probabilities of spontaneous resolution. For certain patients, particularly those with a lower probability of resolution, these differences have the potential to significantly influence clinical decision-making.

Source of Funding: AHRQ T32 grant number HS00063