Patient-Reported Outcome, Biodata and Process Data to Evaluate Dialysis Tolerability (ProDial)

Hemodialysis treatment affects less than 1% of the population, yet is responsible for 5 to 10% of overall health costs in developed countries. Furthermore, despite its necessity as a life-sustaining procedure, mortality among renal disease patients is still 5-10 times greater compared to the general population. This is because currently, hemodialysis treatment is not individually tailored to patients but based on non- individual assessment of biochemical and procedural factors. The proposed project intends to change that, and aims to provide an individual, dynamic hemodialysis treatment plan based on aforementioned non-personal factors combined with monitored, unique personal patient data (e.g., continuous heart rate variability), individual physical activity (step-count), sleep quality and dialysis recovery time (movement analysis). Furthermore, it takes into account a patients self-reported perception of its treatment, including the treatments impact on the patients quality of life. To this end, the project incorporates the innovative, anonymized and non-trackable recording of personalized data, including wearable personal devices and sensors. Furthermore, (mobile) self-reporting applications will be created and made available to patients in order to record their perception of the hemodialysis treatment. These data are stored in an integrated, searchable and personalized data storage, where they will be analyzed using the latest bio- statistical and machine-learning techniques in order to find associations between the referenced data categories (self-reported outcome, biodata, machine data) in order to develop personalized, real-time adaptable hemodialysis treatment procedures.

Every person is different, yet in haemodialysis treatment, these differences are - for the most part - not taken into account. Instead, patient haemodialysis treatment is largely driven by non-individual assessment of biochemical and procedural dialysis measures. As a result, the experience is more unpleasant for most people than it needs to be. ProDial aims to change this. In the ProDial project, we strive to develop a software system that stores and links data from several sources (electronic patient-reported outcomes, dialysis devices and personalized health measurement devices such as smart watches) and with those data, determines the best dialysis settings and procedure. With electronic patient-reported outcomes, we capture a patient's self-reported tolerability outcome, e.g., how the patient perceives his own recovery, by having him or her fill out a questionnaire pertaining data such as recovery time, dialysis fatigue, and sleep quality. With personalized health meters, we record a patient's activity level in the days after dialysis treatment, with data points such as continuous heart rate variability, step-count, sleep quality and movement time. Finally, we also capture procedural data recorded by dialysis machines so that they can be correlated with treatment outcome and tolerability. To uniformly store and link these recorded heterogeneous data, in a persistent data storage across borders, languages, and different healthcare systems, a specialized medical data standard called Fast Healthcare Interoperability Resources (FHIR) was chosen. Using this FHIR standard, localized versions of questionnaires for patient-reported outcomes were created, so that patients from all study centres (Spain, Germany and Poland) could submit their data in the study site's primary language. Furthermore, code systems such as ICD-10, LOINC, and SNOMED-CT were used for coding of data in FHIR where suited, in order to allow more uniform processing of the data. Using the data gathered during the study trial, the selected data fields were assessed for correlation on the self-reported tolerability, and those with significant impact were selected for model construction. An evaluation of the model is currently in progress.