Page 138 - Exploring the Potential of Self-Monitoring Kidney Function After Transplantation - Céline van Lint
P. 138

 136
Chapter 6
measurements in the device and registrations in the SMSS were selected. Total number of measurements performed and registered on these days was 3963. Figure 4 gives an overview of the reliability of the registered data, showing that 87% of these values was registered correctly.
Non-corresponding registrations
In 515 cases (13%) of all creatinine registrations, the registered value did not correspond to the value that was measured on that day. In 174 cases, we could determine the origin of the difference (e.g. wrong combination of date and measured value, typo, rounding off). The remaining 341 registrations (9% of all registered values) were used for further analysis. Median number of non-corresponding values per patient was 3 (IQR 8). Eleven patients (25%) made no mistakes at all, while another eleven patients made more than 10 mistakes. Two patients had an extremely high number of non- corresponding registrations. The first one had 52 non-corresponding registrations, half of which were found to be exactly 10, 20, 30 or 40 μmol/l lower than what was actually measured. In total 83% of his non-corresponding entries was lower than what was measured. The other patient registered 92 non- corresponding values, which were lower than the actual measured values in 93% of his cases. In case of non-correspondence, the difference between measured and registered ranged from 1 to 73 mmol/l with a median of 9 mmol/l (IQR 13). The non-corresponding registrations were significantly lower than the actual measured ones: 123 mmol/l (SD 28) versus 130 mmol (SD 33), respectively (t = 8.7, p=<.001).
Phantom values
In total 93 phantom values were found, which was 2% of all registered values. Thirteen patients (30%) registered at least one phantom value, six patients (14%) registered seven or more. Twenty phantom values resembled the measurements of surrounding days, which would suggest these phantom values were only registered to adhere to the registration protocol. However, this appeared not to be the case as 16 of these 20 resembling phantom values were registered by a single patient who would already have been overly adherent without these phantom values. Three patients registered creatinine values during months where no measurements were logged in the device. For example, one patient quit measuring in February, but registered three measurements during March and April. Phantom values were significantly lower than actually measured ones, respectively 107 (SD 26) and 123 (15) mmol/l (t =3.9, p=< .001).
  



























































































   136   137   138   139   140