Article Text

PDF
Diagnostic performance dashboards: tracking diagnostic errors using big data
  1. Ketan K Mane1,
  2. Kevin B Rubenstein1,
  3. Najlla Nassery2,
  4. Adam L Sharp3,
  5. Ejaz A Shamim4,
  6. Navdeep S Sangha5,
  7. Ahmed Hassoon6,
  8. Mehdi Fanai7,
  9. Zheyu Wang8,
  10. David E Newman-Toker7
  1. 1 Mid-Atlantic Permanente Research Institute, Kaiser Permanente, Mid-Atlantic States, Rockville, Maryland, USA
  2. 2 Department of General Internal Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA
  3. 3 Department of Emergency Medicine, Kaiser Permanente, Southern California, Los Angeles, California, USA
  4. 4 Department of Neurology, Kaiser Permanente, Mid-Atlantic States, Rockville, Maryland, USA
  5. 5 Department of Neurology, Kaiser Permanente, Southern California, Los Angeles, California, USA
  6. 6 Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, USA
  7. 7 Department of Neurology, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA
  8. 8 Department of Oncology, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA
  1. Correspondence to Dr David E Newman-Toker, Department of Neurology, Johns Hopkins University School of Medicine, Baltimore, MD 21218, USA; toker{at}jhu.edu

Statistics from Altmetric.com

In their 2015 report, Improving Diagnosis in Health Care, the National Academy of Medicine asserted that most individuals ‘will experience at least one diagnostic error in their lifetime, sometimes with devastating consequences’1 and ‘improving the diagnostic process is not only possible, but it also represents a moral, professional, and public health imperative.’1 A key barrier to eliminating diagnostic errors is the lack of operational measures to track diagnostic performance.2 3

Novel approaches using ‘big data’ to identify statistically meaningful patterns offer unique opportunities to operationalise measurement of diagnostic errors and misdiagnosis-related harms.4 In a study of ~190 000 US inpatient stroke admissions, the authors found missed opportunities to diagnose stroke early were often linked to clinical presentations with dizziness or vertigo.5 A graphical temporal profile analysis of treat-and-release emergency department (ED) visits showed an exponential increase in visit frequency in the days before stroke admission, establishing these as likely misdiagnoses.5

We operationalised this approach by constructing a diagnostic performance dashboard to monitor diagnostic quality and safety. Kaiser Permanente-Mid Atlantic Permanente Medical Group (KP) and the Johns Hopkins University School of Medicine (JHM) partnered to build a learning ecosystem using visual analytics tools. Visual analytics combines expert knowledge with machine computational power for smart data exploration.6 Visual representations allow users to see the big picture and visually explore relevant data. Interactive data discovery supports ‘slice-and-dice’ operations with data drill-through capabilities, enabling exploratory data mining, hypothesis testing and decision making.

Leveraging the exploratory data …

View Full Text

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles