The benefits of PROMIS
As health care shifts its focus from volume to value, use of performance measures, such as patient-reported outcomes, has become increasingly important. "The question isn't whether or not we need to collect patient-reported outcomes, but rather 'Which measures do we use to collect them?'" said Mark S. Vrahas, MD, Robert W. Lovett Professor of Orthopaedic Surgery, Orthopaedic Trauma Initiative, Harvard Medical School.
Dr. Vrahas made his comments during the 2016 AAOS Now Forum, "How Good an Orthopaedist Are You? Risk Stratification & Quality-of-Care Reporting," held following the AAOS Annual Meeting in Orlando, Fla.
The PROMIS initiative
In 2004, the National Institutes of Health (NIH) initiated the Patient Reported Outcomes Measurement Information System (PROMIS) project. According to Dr. Vrahas, the project's goal was to develop standard patient-reported outcome measures across important health domains. More than 60,000 clinicians, researchers, and patients have contributed data to the project.
"There are several good reasons for orthopaedic surgeons to use PROMIS measures," said Dr. Vrahas. "To begin with, these measures are supported by the NIH. They are also based on current measurement science, which offers advantages over legacy measures."
Dr. Vrahas explained there has been a paradigm shift within psychometric testing from classical test theory—the science on which legacy measures were based—to item-response theory (IRT). "With classical test theory, the entire test is validated as a whole. IRT, however, validates each question independently," he said.
PROMIS measures include computer-adaptive tests (CATs), based on IRT, as well as static fixed-length short forms. The CATs use the responses from previous questions to select other questions, thereby improving test precision. Physicians can also create their own custom short-form by selecting questions from a large bank of items. As each item is calibrated, any combination of items produces a score on the same metric, allowing for comparisons between measures and patients.
"Most of the NIH-developed CATs produce a very high level of precision with just four to six questions. Because they contain fewer questions, they also reduce patient burden, which ultimately means they cost less to use," said Dr. Vrahas. In a 2015 study he co-authored, the PROMIS Physical Function (PF) CAT was compared to three traditional physical function measures in elderly patients with proximal humerus fractures. The PROMIS PF CAT and the traditional measures yielded similar patient assessments of upper extremity function. "In addition, it took the patients—who averaged 75 years of age—less than 2 minutes to complete the PROMIS test on a tablet computer," he said.
The results of similar studies on spine, foot and ankle, and upper extremity disorders support Dr. Vrahas' findings. "In each of the studies, the PROMIS PF CAT was faster to complete and correlated well with legacy physical function measures. Moreover, all the authors felt the PROMIS PF CAT could replace the legacy measures," he said.
In addition, there were fewer measurement floor and ceiling effects, making it more likely that a difference in outcomes would be detected when differences actually exist. This can improve quality of care as well as research.
Most importantly, according to Dr. Vrahas, PROMIS measures can be improved over time as more people use and study them. He noted that 30 new questions will be added to the PROMIS PF item bank later this year to improve measurement at the very high end of function (eg, athletes). As the underlying metric is unchanged, results from studies using this expanded measure can be compared with existing data.
To enhance comparability with other patient-reported outcome measures, the NIH funded a program to convert legacy measure scores to the PROMIS metric (www.PROsettastone.org). PROMIS measures are also very accessible, Dr. Vrahas noted. PROMIS short forms available in portable document format (PDF) can be downloaded from www.healthmeasures.net, while PROMIS CATs are embedded in various data collection platforms. "If your hospital wants to use PROMIS CATs, they can contact the PROMIS network for an Application Program Interface and put them behind their firewall," he added.
Dr. Vrahas concluded, "When it comes to patient-reported outcome measures, I think it makes sense that we all collect the same thing. That's why I am a big advocate of the PROMIS measures."
Dr. Vrahas' disclosure information can be accessed at www.aaos.org/disclosure
Maureen Leahy is assistant managing editor of AAOS Now. She can be reached at firstname.lastname@example.org
- PROMIS was initiated to develop improved methods for measuring patient-reported outcomes.
- PROMIS measures include static fixed-length short-form tests and CATs.
- Compared to legacy measures, the PROMIS PF CAT has been shown to improve accuracy and reduce patient burden.
- Additional benefits of PROMIS measures include comparability and accessibility.