Johns Hopkins researchers use big data analytics to target diagnostic errors, improve quality 

Reducing diagnostic errors is a crucial component of improving care quality, but current methods of monitoring such mistakes may be time consuming. Now, researchers at Johns Hopkins have developed a strategy that uses big data to speed up the process. 

David Newman-Toker, M.D., director of the Center for Diagnostic Excellence at Johns Hopkins' Armstrong Institute for Patient Safety and Quality, and his team developed a new approach called SPADE (Symptom-Disease Pair Analysis of Diagnostic Error) to allow providers to harness databases instead of having staff members comb medical records for more information, according to a study published in BMJ Quality & Safety. 

SPADE uses statistical analyses to find and flag patterns that can predict diagnostic errors. It mines available databases for common symptoms that lead patients to visit a doctor and then compares those data with diseases that are often misdiagnosed.  

For example, SPADE would illustrate how frequently patients who visit a provider with dizziness are sent home because doctors believe the condition is minor, when they had actually suffered a stroke, Newman-Toker said in an announcement. Having these data available would improve outcomes and could refine quality measures to better reflect patients' concerns. 

"Being able to do that using big data is an important innovation for diagnostic quality and safety," he said. 

RELATED: 'Diagnostic calibration' to blame for medical errors 

Diagnostic errors are an industry-wide issue. A recent Mayo Clinic study suggested that more than 20% of patients are misdiagnosed by their primary care doctors. Physicians can mitigate the risk by seeking out a second opinion or taking a "time out" prior to surgery to ensure the correct patient is on the operating table. 

SPADE's approach is likely to be most effective with acute or subacute conditions for which a misdiagnosis could have serious consequences: hospitalization, disability or death within six months. Newman-Toker said it would likely be effective for the "big three" conditions that lead to death or disability after a diagnostic error: infections, cancer or vascular events. 

Taking a new approach to finding and preventing diagnostic errors will take time and support from leadership, but Newman-Toker sees SPADE's analyses as data that could be part of providers' public quality reporting in the near future, he said. 

"Patients will have the opportunity, for the first time, to see how their hospital is performing on diagnostics and ask themselves, 'Do I want to choose a hospital that has fewer misdiagnosis-related deaths?'" he said. "And that is a step toward patient empowerment in diagnostics that has never existed before."