The Study

Mobile health (mHealth) technology has advanced at a breakneck pace as wireless access has become pervasive. We wanted to test this powerful technology for clinical trials by equipping trial volunteers with sensors, wearables, and apps. In other words, we instrumented them.

To experience the opportunity first hand, we collaborated with a prominent clinical research center and sponsored a study.

Welcome to MOVE-2014

“A Pilot Open Label Clinical Trial to Evaluate the Combined Impact of Two Mobile Health Products on Health Outcomes in Overweight Adults With Type 2 Diabetes”

We equipped patients with a smartphone (iPhone) and an activity tracker (Fitbit Flex).

We provided them with a mobile app to capture diary and quality of life data (Patient Cloud).

Over an eight-week period, we tracked the effect of sensors, wearables and apps on health outcomes in overweight people with Type 2 Diabetes.


The Results

Here's what we found...

Bottom Line

MOVE-2014 demonstrated that studies can successfully instrument patients in a clinical trial.

View additional information on the data collected.

Data Quality

Higher Visibility of Data Led to Higher Data Quality

Outliers may indicate device quality issues.

Data Quality:
Sleep Efficiency Example

One subject had perfect sleep quality nearly 45% of the time, an outlier that was determined to be implausible.

View formula used Download the chart

Data Quality:
Steps Walked Example

Another subject walked close to 35,000 steps in a day, but that outlier made sense given the subject's data patterns.

Download the chart

Patient Compliance

Compliance is an Important Data Quality Check

The study experienced high compliance with patients wearing the activity tracker. However, patients were not always compliant setting their device to sleep mode.

Patient Compliance:
Fitbit Wear Compliance

Subjects wore their devices over 90% of the time.

View formula used Download the chart

Patient Compliance:
Sleep Mode Compliance

However, subjects struggled to put their devices into sleep mode consistently.

Download the chart

Patient Compliance:
Questionnaire Compliance

Subjects were highly compliant with completion of questionnaires via a mobile app.

Download the chart


mHealth Data Yields Many Possible Clinical Insights

By reviewing the data across all sources (investigators, patients and sensors), some intriguing possible correlations exist (e.g., between pain and activity, between fatigue and activity).

Relationship Between Pain, Activity and Weight Loss

Subjects who reported the highest levels of pain walked the least, and ultimately displayed lower decrease in BMI than other subjects.

Download the chart

Cohort of Subjects with Successful Performance on Study Endpoints

We identified a cluster of subjects who lowered both blood glucose and BMI in order to understand whether phenotypic (behavioral) characteristics could help explain this promising response.

View additional information Download the chart

Implications for eClinical:

mHealth data offers a new class of data which is:

  • Objective
  • eSource
  • Real time
  • Real world
  • Remote
  • Continuous

It's relatively easy to collect mHealth data, but doing so in a regulatory compliant way can be challenging.

Visualization tools, along with data science expertise, are necessary to derive clinical meaning from the data.

Patient-reported outcomes go hand-in-hand with mHealth. There's a lot to learn about sensor/wearable data, so pairing it with subjective patient self-assessments can facilitate understanding.