Skip to main content

Table 5 Simulation analysis – effect of alternative assumptions for number of weight readings submitted

From: Assessing the reliability of self-reported weight for the management of heart failure: application of fraud detection methods to a randomised trial of telemonitoring

  Assignment of EDP patients (%) Assignment of NEDP patients (%) Proportion of people in A1-A4 with NEDP (%)
A0 A1-A4 A0 A1-A4
10 42.6 1.9 3.5 12.9 97.5
20 96.0 4.7 8.4 34.1 97.6
40 99.8 0.6 4.4 17.4 99.4
60 100.0 0.3 4.5 16.9 99.7
80 100.0 0.3 6.7 25.9 99.8
162 (from Table 1) 100.0 0.0 5.9 23.5 100.0
180 100.0 0.0 6.0 22.6 100.0
  1. Note: Assumes 15% of patients have EDP, 30% of the time, with weights that are uniformly distributed on {0, 1, 2, 3, and 4} modulo 5. Based on 10,000 replications of the simulation experiment