Page 166 - Exploring the Potential of Self-Monitoring Kidney Function After Transplantation - Céline van Lint

164
Chapter 7
during the whole study period. Although the number of face-to-face visits was significantly lower for our intervention group, the reduction in number of visits was less than anticipated. Patients therefore continued to visit the outpatient clinic relatively often, potentially resulting in a lower perceived need to perform (all requested) measurements.
The reliability of patient-reported data
For self-measured values to be clinically useful, they need to be reported accurately. In chapter 6, we showed that approximately 90% of both creatinine and blood pressure measurements was registered correctly in the SMSS. This percentage corresponds to what has previously been described for patient- reported blood pressure[51, 53] and anticoagulation[54], but is much higher than has been observed for patient-reported levels of blood glucose[55-57]. In case of non-correspondence between measured and actually registered values, the values that were registered in the SMSS were significantly lower than those actually measured. This suggests that patients select, alter or add values in such a way that their creatinine profile looks more positive. This corresponds to what has been found in a population of patients self-monitoring International Normalized Ratio (INR), where the measurements that fell within the desired range were significantly higher when using patient-reported data compared to data stored in the device[35]. For patients with diabetes or hypertension, it was found that inaccurate reporting increased with increasing levels of blood glucose[56] or blood pressure[53]. Why patients report values that look better than the values they actually measured or add non-existent measurements has not yet been fully clarified. For diabetes, it has been suggested that patients report false glucose levels due to a feeling of guilt for not having achieved glycaemic goals[57] or add phantoms values in an attempt to fill up logbooks and satisfy their healthcare providers[55]. Both situations seem to represent an attempt to be a ‘good’ patient. However, altering and selecting data that is not representative of the actual clinical situation or adding phantom values in any case may be dangerous. This can lead to suboptimal treatment and, eventually, to worsened patient outcomes[53, 55]. Kendrick and colleagues have indeed shown that women with pregnancy-derived diabetes received suboptimal treatment due to a large difference between their reported glucose values and what they had actually measured[57]. It also seems to work the other way around: diabetic patients who were more reliable in their reporting had a significantly better glycaemic control[55]. This is probably due to a clinicians’ ability to adjust therapy more precisely if measurements are reported accurately. To rule out the possibility of incorrect reporting, other authors have already recommended the use of devices that can transfer data automatically [54-56, 58].