Do Smartphones Collect Better Clinical Data Than Paper-and-Pencil Tests?
Quick Links
Potential disease-modifying therapies are entering trials for Parkinson’s disease even as researchers need better biomarkers to track clinical improvement. Some believe that smartphones might hold the key. At the 13th International Conference on Alzheimer’s and Parkinson’s Diseases, held March 29 to April 2 in Vienna, scientists from Roche and Prothena discussed how they are using these devices to collect detailed clinical data from trial participants. The two companies collaborate on trials of Prothena’s anti-α-synuclein antibody PRX002 (see related conference story). In Vienna, Michael Lindemann of Roche Research & Early Development presented Phase 1 data from smartphones that suggested they more sensitively and accurately reflect small clinical changes than do traditional measures. Crucially, smartphone data closely correlated with the validated clinical measure for PD, the Movement Disorder Society Unified Parkinson’s Disease Rating Scale (MDS-UPDRS).
“This study shows you can use sensor data to correctly represent clinical severity,” Lindemann said. He believes these data are sensitive enough to track both disease progression and treatment effects. Potentially, the data could distinguish people who progress quickly from those whose disease advances slowly. Such information would help researchers stratify trials and better gauge whether treatments are working, Lindemann noted.
Currently, most PD trials rely on the MDS-UPDRS as the primary outcome measure. However, because the UPDRS is only assessed in the clinic, this scale cannot capture the day-to-day fluctuations in function that torment people with PD, and scientists have been searching for alternatives. Some groups are touting the benefits of technology such as smartphones to collect continuous data on neurodegenerative disease and paint a fuller picture of daily variability than an occasional clinic visit allows (see Dec 2012 news). This idea is starting to take hold in Parkinson’s disease, perhaps because, being a movement disorder, it is particularly amenable to the type of data easily collected by smartphone. Various projects have already begun to gather observational data from PD cohorts in this way (Mar 2016 news; PD Smartphone Data Challenge).
In Vienna, Lindemann described how the PRX002 Phase 1b trial used smartphones to monitor symptoms. These devices contain a gyroscope and accelerometer that enable them to detect movement of various types. Participants in the three highest dose cohorts were offered phones that they carried with them throughout the trial. This was optional, but all 44 participants in these cohorts decided to do it, Lindemann noted. The phones collected data for the full 24 weeks of the study. In addition to passive monitoring of gait and movement by phone, participants were asked to complete six specific tasks every day that measured their balance, walking, dexterity, postural tremor, resting tremor, and steadiness of voice.
Participants performed these tasks quite faithfully, Lindemann noted. At the beginning of the trial, the cohort as a whole collected 75 percent of the requested data points, and by the end of the six months, completed 50 percent of the requested tests. Notably, 90 percent of the participants performed the tasks at least once every four days, providing regular data points for analysis. This level of adherence is much higher than that seen in observational studies. Lindemann ascribes this partly to the personal relationship participants developed with the study center staff, and to the degree they trusted that their data would be used well and contribute to advancing PD treatments. He added that the research team also worked hard to make the apps easy to use and the daily tasks not unduly burdensome.
Others agree. Lara Mangravite of Sage Bionetworks, Seattle, ran a previous observational study. She said that trial participants typically feel a strong sense of commitment to provide as much data as possible to support development of new therapeutic agents for their disease. “Patients who provide mobile data collection within the context of a clinical trial consistently demonstrate greater adherence than those using mobile data collection as part of observational trials that are performed remotely,” she wrote to Alzforum.
Importantly, data collected from the smartphones correlated well with MDS-UPDRS scores, Lindemann said. However, smartphone data did not always match up with UPDRS categories of the same name. Balance detected by smartphone, for example, did not correlate that well with the “postural instability” category of the UPDRS, which is measured by the physician tugging patients backward and seeing how well they can regain their footing. Instead, the smartphone balance measure matched up almost perfectly with the “posture” category of the UPDRS, which assesses static balance. The smartphone dexterity test, which involves tapping spots on the screen, correlated with “dressing” in the UPDRS scale, because this task involves fine motor coordination as people manage buttons and zippers.
While scores from the UPDRS and smartphone tests matched, the latter included additional details that filled in the clinical picture, Lindemann said. Patients who scored a zero on the MDS-UPDRS scale for resting tremor, meaning the clinician did not see a tremor, in fact did have a tremor that their smartphone picked up. When patients self-assessed their own tremors, the results better matched their smartphone data than their UPDRS score, indicating that the phone more faithfully reflected the patient’s experience than did the clinician. Lindemann noted that in these cases, the patient’s tremor might not be present during the 15-minute office visit, but might come out at other times.
Moreover, smartphones were more sensitive at detecting small changes, Lindemann said. For example, one patient who started on a standard symptomatic therapy a month before the end of the trial showed distinct motor improvement on the high-frequency smartphone monitoring data, but the pattern was quite noisy. The effects were also subtle, amounting to about a 10 percent improvement in the dexterity measure. Sampling data from just two timepoints, as baseline and follow-up clinic visits would have done, likely would have missed this narrow improvement, Lindemann said.
The data indicate that smartphone technology is feasible to use in a trial, and can detect subtle and clinically meaningful motor impairments, Lindemann concluded. “Remote patient monitoring will transform clinical research,” he predicted, adding that smartphone monitoring will be included in the Phase 2 PRX002 trial.—Madolyn Bowman Rogers
References
Therapeutics Citations
News Citations
- α-Synuclein Antibodies Enter Phase 2, Sans Biomarker
- Tech Revolution: Monitoring and the Power of Real-Time Data
- Mobile Phone App for Parkinson’s Patients Tests New Model for Data Sharing
External Citations
Further Reading
Annotate
To make an annotation you must Login or Register.
Comments
No Available Comments
Make a Comment
To make a comment you must login or register.