Global web ALERT button

Big Data for Rehabilitation

 Click the arrow at the bottom of the video to navigate between topics or click to view the video in a new window.

Notes

Recorded at the American Society of Neuroradiology Annual Meeting on November 10-11, 2016. Speakers: Steven Cramer, MD; Liam Johnson, PhD; Sook-Lei Liew, PhD, OTR/L; Keith Lohse, PhD; Kenneth Ottenbacher, PhD, OTR.

A persistent challenge in rehabilitation research is the vast heterogeneity within clinical populations. This inter-individual variability makes it difficult to establish significance and reliably replicate findings of rehabilitation studies across smaller sample sizes. Large, diverse datasets have the potential to drive rehabilitation research by providing the greater statistical power needed for evaluating clinical hypotheses and validating findings from smaller studies. However, collecting, organizing, and analyzing large amounts of data has limitations and considerations. Here, we present current applications of ‘big data’ approaches for rehabilitation research across collections of behavioral, neuroimaging, and clinical outcomes data.

Learning Module Notes Modules

  1. What is Big Data? (0:00-2:34)
    • Specifically for biomedical research
    • Issues: data collection, storage, and management 
  2. What are examples of Big Data? (2:34 – 9:05)
    • Cooperative groups, complex data sets, genetics, digital data
    • Measuring neuro function still is difficult to measure vs. structural changes (the reason why EEG is promising (7:25) Telehealth: efficient, inexpensive, will create huge streams of data 
  3. Why “big data?” (9:05 – 10:22)
    • New techniques, high variance in disease expression, greater statistical power, increased complexity of clinical practice, increased regulatory demands/oversight
    • Because we can peruse big data now, we can peruse tackling these other issues 
  4. Possible Benefits with a Big Data approach include: (10:22 – 11:32)
    • Ask questions with greater granularity, personalized medicine, greater efficacy of science research, health care policy will be based on more precise knowledge, greater sensitivity to outliers, understand complex interactions.
  5. Possible Risks with Big Data approach: (11:23-11:53)
    • Inaccurate data, some features of data sets becoming increasingly opaque, the quantity of data too large to be useful, expense, data security breach.
  6. Promises, Pitfalls, future Potential – Large Data Definitions & Resources for Rehab by Ken Ottenbacher (11:53-35:30)
    • Intro & Overview: Intro Challenges & Opportunities CLDR: Center for Lage Data Research & Data Sharing
    • Data Discovery
    • Challenges: overwhelming datasets
    • Opportunities: Advances in data, policy questions, clinical questions to be answered with this data,
      • Large Data Analytics is an Evolving Field & What is “Big Data”?
      • Big data, data science, and data discovery is a growing field
      • BD2K: NIH Big Data to Knowledge Grant (focus: facilitate broad use of data, analysis methods and software, enhance training, centers of excellence)
      • Characteristics of Big Data
      • Large Data vs Big Data
      • Opportunities with Big Data
      • Wearable devices, motion tech to produce data
      • PRORI’s PCORnet: Health systems data on patients within a network
      • Health care reform: ACA, payment structures, hospital readmission as a quality indicator
      • Example: Readmission in Post-Acute Care
      • CLDR: Education & Training, Data Directory – large data sets of interest to people in the rehab field, Pilot Project Program – opportunity to explore a large data set study, data sharing and archiving studies, Visiting Scholars – visit consortium to get support & help with data, Archiving and Sharing Data, Resources
  7. Keith Lohse: Information Architecture in Rehab Trials (35:30-59:00)
    • Intro 36:06 – 53:52 Presentation Aims:
    • Theory & Method & Clinical Interpretation – Ontology
    • Motor Rehab for Stroke
    • How can we use existing data to address current questions in RS and inform research decisions?
    • Conclusions and Q&A
  8. Sook-Lei Liew: ENIGMA for Neurorehab: Large-Scale Meta-Analysis Approach to Modeling, Neuroimaging, Genetics, and Behavior (59:00-1:22:39)
    • The precision medicine initiative
    • Application to Neuroimaging 
  9. Liam Johnson: High-Quality Data in a Clinical Trial (1:22:39-1:41:00)
    • Snapshot of AVERT Trial data & complexity of managing and analyzing the data
    • Insight into lessons learned from running multi-site/continent clinical trial
    • Highlight growing collaborative efforts in neuro rehab
  10. Q&A
  Learning Modules