Study: Patient Safety Indicators Need Refinement
Efforts to identify hospital-based patient safety problems have culminated in a list of indicators known as PSIs that initially were intended to help to detect adverse events and prompt quality improvements.
Now, however, PSIs are emerging as standards by which the public will be able to compare hospitals – a move that two VA Boston Healthcare System researchers, both professors at the BU School of Public Health (BUSPH), find problematic.
In a paper published in the Journal of American College of Surgeons, Amy Rosen and Ann Borzecki, professors of health policy and management at BUSPH, examined the validity of the PSIs, or Patient Safety Indicators, in a study of Veterans Administration hospitals.
What they found was that while some of the indicators had strong predictive value, for others, the predictive validity was relatively low. The criterion validity was questionable because of coding or clinical errors — sometimes because the medical conditions flagged by the indicators were present before the patients were hospitalized.
“Many of the true positives had patient-related risk factors that may have increased the likelihood of experiencing a PSI,” the study said. For example, many patients with pressure ulcers showing as “true positives” had them prior to admission to the hospital. On the other hand, despite the relatively high predictive value of indicators such as Accidental Puncture or Laceration, many of those “true positive” cases turned out to be minor injuries with no clinical significance, the authors found.
Rosen and Borzecki suggested several ways to improve the validity of the indicators, including introducing an “admission type code” that would distinguish between elective and emergency admissions, and changing coding for some of the indicators to address inaccuracies and limitations.
“Working with providers to improve the content and accuracy of medical record documentation would also help improve validity,” they said.
Overall, the review – titled “Validating the Patient Safety Indicators in the Veterans Health Administration: Are they Ready for Prime Time?” — suggested that the PSI indicators were not yet reliable enough to represent solid information about hospital performance.
Give those problems, “We recommend that the PSIs should continue to be used primarily for screening and case-finding, and not for public reporting or pay-for-performance,” the authors said, while noting some exceptions.
They recommended that surgeons find ways to monitor and improve medical record-keeping and the accuracy of coding at their hospitals, in part by actively participating in educating those who do coding.
“The PSIs should ultimately be used to detect real complications that are caused by quality of care problems, in order to improve surgical care and enable [federal regulators] and other organizations to legitimately adopt them for public reporting and pay-for-performance,” the study concludes.
Submitted by Lisa Chedekel