Page 163 - Exploring the Potential of Self-Monitoring Kidney Function After Transplantation - Céline van Lint

increase of 10% (true positive) in 70% of all cases and a decrease of 10% (true negative) in 67% of all cases. These results were obtained in a laboratory setting. For self-monitoring to be a safe alternative to regular transplant follow-up, it is important to know whether the creatinine device is also sufficiently able to detect deteriorations in kidney function when used by patients. For the detection of clinically relevant increases in creatinine (>10%), we even found a higher level of agreement when self- monitoring was conducted in a real-life setting by patients themselves: in 78% of all relevant cases a similar trend for home-based and laboratory-based creatinine levels was observed (described in chapter 5). The higher level of similarity between home-based and laboratory-based creatinine trends first suggests that patients are able to correctly perform the creatinine measurements at home. Second, it suggests that the accuracy of the creatinine device and test strips has improved over time. The test strips that were used during the laboratory study came from a batch that was manufactured at least 2 years before the batches of strips that were used during the RCT. With the test strips being constantly improved, this could possibly explain the increase of 10% in similarity between home and laboratory measurements and it may indicate that the ability of the StatSensor® to detect changes in kidney function is improving.
However, to make the StatSensor® a more reliable tool for self-monitoring, the further improvement of its analytical performance remains very important as this will automatically improve its’ clinical performance (both for diagnostic and monitoring purposes). Awaiting the improvement of the analytical performance of the StatSensor®, two manoeuvres could offer a provisional solution. First, it is desirable that the number of false negatives is decreased, as it leads to delayed detection of rejection, which is dangerous and should not or hardly occur. If applied to the StatSensor®, one could choose a cut-off percentage that is lower for StatSensor® results than for laboratory measurements. For example, by lowering the StatSensor® cut-off percentage to ≥5%, the number of correctly identified relevant increases in level of creatinine (≥10% as determined by the central laboratory method) would have increased from 70% to 82%. A drawback of this approach is that it would result in an increased number of false positives and, consequently, additional diagnostic interventions. Second, the clinical performance regarding monitoring will improve when the number of measurements is increased. With increasing the frequency of StatSensor® measurements, a more reliable trend will be obtained. As patients perform the measurements themselves, we could advise patients to increase their measurement frequency. With lowering the cut-off and a guideline to measure in a higher frequency, the chances of detecting rejection are increased and theoretically, the number of outpatient visits can be safely reduced.
General discussion 161
7