Follow Datanami:
September 21, 2016

SAS Analytics Helps Pennsylvania Use Data to Protect Patients From Medical Errors

CARY, N.C., Sept. 21 — The Pennsylvania Patient Safety Authority (PSA) is applying SAS Visual Analytics to millions of medical error reports, unearthing new information about medical error trends and causes. Besides revealing new strategies to help protect commonwealth patients, PSA’s program is also saving hundreds of thousands of dollars in database design and staff time.

PSA collects medical error reports from more than 1,200 licensed health care facilities to populate the Patient Safety Reporting System (PA-PSRS). The PA-PSRS database contains more than 2.5 million medical error reports, each with over 200 data and text fields detailing both serious events and near-miss errors. “Serious events” result in actual physical harm to the patient. A “near miss” records a medical error that did not reach or harm the patient. Overall, it adds up to more than 500 million data fields and billions of text field characters.

“It is one thing to develop one of the nation’s largest and best-known statewide patient safety databases. It’s quite another to turn that data into lifesaving information,” said Howard Newstadt, PSA Finance Director and Chief Information Officer.

SAS Visual Analytics is the backbone for PSA’s Visual Analytics Reporting System (PSA VA), through which PSA queries the massive PA-PSRS database and easily reports findings via the web.

“The effect of this project has been transformational,” said Newstadt.

PSA VA is a finalist in the National Association of State Chief Information Officers (NASCIO) State IT Recognition Awards in the Emerging and Innovative Technologies category.

Immediate results 

In its earliest efforts, PSA VA identified previously unrecognized data relationships. One initial test led researchers to further explore how gender may influence specific types of medical errors.

First, PSA VA filtered factors that could skew reported data, such as age range (women live longer) and gender-specific language or care areas. It quickly became clear that certain serious events involving adverse drug reactions were more likely among women, while skin integrity errors (cuts and abrasions) were more likely among men.

A better understanding of factors influencing patient safety events can inform changes to procedures, policies or cultures, affecting clinical risk assessments for certain types of patients, reducing medical errors and patient harm in Pennsylvania and nationally.

Data visualization leads to savings

Analysts use PA-PSRS data in research and to author journal articles, and patient safety liaisons work directly with health care facility staff to improve patient safety and reduce medical errors.

Previously, custom data requests flowed through a single data analyst who constructed complex queries to extract answers. With SAS Visual Analytics, PSA staff build their own customized, web-based data queries and analytic reports. Now, field and analytic staff with little or no IT backgrounds develop their own novel reports, access them over the web and share them remotely with co-workers. Easier, simpler reporting saves PSA staff countless hours.

Analysis validates PSA’s efforts

Pennsylvania’s Medical Care Availability and Reduction of Error Act of 2002 tasked PSA with eliminating medical errors by identifying problems and implementing solutions that promote patient safety.

One of the first reports produced through PSA VA revealed a 45 percent decline in “high harm” events over 10 years, from 2005 through 2014.

“This outcome supports the work of PSA and the intent of the MCARE Act,” said Newstadt. “It also suggests a positive effect on people’s lives, alleviating suffering and in fact preventing death. The monetary savings alone associated with this decline are measured in tens of millions of dollars, if not more.”


Source: SAS

Datanami