A study published online on October 7 by the Journal of the American Medical Association (JAMA) finds that postoperative venous thromboembolism (VTE) rates are likely not a good measure of hospital quality because hospitals that routinely order more imaging studies find more VTE events.
The study’s authors looked at 2010 Hospital Compare and American Hospital Association data from 2,838 hospitals, as well as 2009-2010 Medicare claims data for 954,926 surgical patient discharges from 2,786 hospitals. When all of this data was analyzed, the researchers found that indeed, the more you look the more you find, meaning that hospitals that ordered more VTE imaging studies (duplex ultrasonography, chest computed tomography/magnetic resonance imaging, and ventilation-perfusion scans) had slightly higher risk adjusted VTE event rates.
When the researchers looked at other quality measures, they found that a higher rate of risk adjusted VTE events did not correlate well with other hospital quality measures. The hospitals that scored well on the other quality measures did worse on VTE event rates than hospitals that had lower overall quality measure scores. This indicated to the researchers that because of “surveillance bias,” lower quality hospitals were having fewer post-op VTE events not because they were better but because they were not looking for these events as aggressively.
There is some debate over the right amount imaging for post-op VTE diagnosis. For patients at low risk undergoing certain surgeries, imaging may be unnecessary and even wasteful. However, this study only looked at 11 surgeries (abdominal aortic aneurysm repair, coronary artery bypass graft, craniotomy, colectomy, cystectomy, esophagectomy, gastric bypass, lung resection, pancreatic resection, proctectomy, and total knee arthroplasty) and did not examine what amount of testing was the right level for these procedures. Rather, it was focused just on the validity of post-op VTE rate as a quality measure and found that this was likely not a valid measure, and possibly a problematic one, if hospitals altered their procedures to manipulate this score in their favor at the expense of patient care.
“Surveillance bias likely influences other quality metrics, and the possibility that it might perversely encourage the wrong practices should be carefully considered when quality metrics are developed,” the authors wrote.