Data in Online Fox Insight Surveys Similar to Traditional Efforts in Parkinson’s, Study Finds

Marisa Wexler, MS avatar

by Marisa Wexler, MS |

Share this article:

Share article via email
Parkinson's Foundation online offers

People with Parkinson’s disease taking part in the Fox Insight study, where they self-report details about their illness, share similar clinical and demographic features with patients brought through more traditional ways into large data studies, researchers report.

This finding suggests that groups recruited exclusively online can be useful to Parkinson’s research — although they may have drawbacks that need to be addressed.

The study, “Comparison of an Online-Only Parkinson’s Disease Research Cohort to Cohorts Assessed In Person,” was published in the Journal of Parkinson’s Disease

The Fox Insight (FI) study, launched by the Michael J. Fox Foundation, aims to collect large amounts of self-reported data about the health experiences of adults with and without Parkinson’s. Healthy people are serving as a control group.

This study’s recruitment is done online — a departure from more traditional models that usually require participants to travel to a hospital or academic center for interviews and/or exams.

As such, meaningful differences might exist between those willing or able to participate in online and in-person trials.

Researchers compared individuals with a self-reported Parkinson’s diagnosis in the FI study to those taking part in three other traditional (in-person) studies: the Parkinson’s Progression Markers Initiative (PPMI), the Parkinson’s Disease Biomarker Program (PDBM), and the idiopathic PD LRRK2 Consortium (iPDLC). With this last study group, only individuals without LRRK2 mutations — one of the most common genetic causes of Parkinson’s — were included.

Comparisons were done involving a total of 12,654 people participating in FI study, 422 in PPMI, 700 in PDBP, and 508 in iPDLC. (To date, more than 43,450 people are enrolled in Fox Insight. Researchers here looked at those whose data seemed consistent with an established diagnosis, including reporting of results from tests like the Movement Disorders Society-Unified Parkinson’s Disease Rating scale (MDS-UPDRS) II.)

“In general,” wrote the researchers, “we found that the FI cohort is comparable to the other cohorts.”

For example, no significant differences in any recorded motor symptoms were seen between the FI and the PDBP or iPDLC patient groups (cohorts).

Some motor symptoms, however, were more common among people in the FI than PPMI studies. These included difficulty turning in bed (47.22% vs. 25.59%), problems with walking and balance (59.45% vs. 36.73%), and freezing of gait (19.28% vs. 4.74%). The researchers noted that comparisons here were limited to 4,072 people from Fox Insight, all diagnosed in two or fewer years, because the PPMI study only included people diagnosed within two years of enrollment.

Non-motor symptoms also tended to be similar between the FI and other cohorts. Again, however, a few differences were noted.

For instance, the groups reported differences in the frequency of REM sleep behavior disorder, but these frequencies also varied depending on what metric was used in the different datasets. As such, it’s not clear to what extent this reflects actual differences and to what extent it’s just differences in the taking of measurements.

Indeed, FI’s reliance on self-reporting was a consistently addressed limitation. For example, in terms of assessed cardiovascular risk factors, the FI cohort reported significantly higher rates of high blood pressure (43.3% vs. 27.49%) and high cholesterol (47.28% vs. 21.8%) than was seen among people in the PPMI cohort.

“Self-reported medical history, and especially cardiovascular history, is particularly prone to inaccurate reporting,” the researchers wrote, adding “the accuracy of self-reported PD” is an issue “we are currently not able to directly assess.”

But, “it is [also] possible that the FI cohort allows for participation of individuals with medical comorbidities that might otherwise be excluded from other, in-person observational studies” or because of more advanced disease states.

In other words, such differences in data collected might be due to inaccurate self-reporting, or it might be reflective of actual differences between these patient populations. Future studies may be necessary to know for sure.

Likewise, a similar frequency in the use of medications like levodopa — a mainstay of Parkinson’s treatment — was seen among these various study cohorts, but use of other medications — like dopamine agonists — were significantly more common in the PDBP than FI cohorts (49.07% vs. 32.08%).

This finding “could indicate either differences in prescribing trends, or rather an under-reporting of use of non-levodopa agents in FI,” the researchers wrote.

It should also be noted that, in some regards, similarities seen between the FI and other patient groups weren’t necessarily ideal. “PD research studies are lacking in gender, ethnic, and racial diversity, with a significant over-representation of males and whites,” the researchers said.

Although the proportion of females was comparatively high in the FI cohort (45.56%), the vast majority (96.85%) were white, as were more than 90% of people taking part in the other Parkinson’s data studies. Most, more than four-fifths, had also completed at least some schooling beyond high school.

This suggests a continued need for efforts to increase diversity, particularly ethnic and socioeconomic diversity, in these studies, since a representative sample is, by definition, necessary to get a realistic understanding of the disease. Efforts aimed at increasing diversity are underway in the FI study.

Despite potential limitations, this work suggests that the online recruitment and self-reported data of FI patient group is of relevance to Parkinson’s, and its data reasonably similar to patient data gathered through in-person studies.

“The results of this analysis indicate that research utilizing data from an online-only self-reported PD cohort is feasible, and that a cohort assembled in this way has features that are broadly similar to cohorts recruited in person,” the researchers wrote.

“By following a PD population unprecedented in size and scope, the FI cohort can therefore serve to further our understanding of the frequency, severity, and progression of symptoms and quality of life in PD,” they concluded.