Ana-Maria Orbai, MD, on Validating the PROMIS 29 Profile for PsA
Dr Orbai discusses her project to validate the PROMIS 29 profile for use in assessing several disease domains among patients with psoriatic arthritis, which she spoke about at ACR Convergence 2021.
Ana-Maria Orbai, MD, is an assistant professor of medicine and director of the Psoriatic Arthritis Program at the Johns Hopkins School of Medicine in Baltimore, Maryland.
TRANSCRIPT:
Welcome to another podcast from the Rheumatology and Arthritis Learning Network. I'm your moderator, Rebecca Mashaw. Today, I'm delighted to be speaking with Dr. Ana-Maria Orbai, assistant professor of medicine and director of the Psoriatic Arthritis Program in the Division of Rheumatology at the Johns Hopkins University School of Medicine.
She's going to be talking to us about one of her poster presentations from the recent American College of Rheumatology convergence meeting. Thank you for joining us today, Dr. Orbai.
Dr. Ana-Maria Orbai: My pleasure. Thank you for the opportunity to speak about our research.
Rheumatology and Arthritis Learning Network: This particular abstract concerns validating the PROMIS 29 profile for patients with psoriatic arthritis. Would you start off by explaining why this needed to be done?
Dr. Orbai: Pain, fatigue, physical function, social participation, mood, and sleep are important to assessing psoriatic arthritis because they commonly lead to decreased quality of life in this disease. These domains can be efficiently measured using the PROMIS 29 profile, which is a generic health measure, patient-reported, which hasn't been validated yet in psoriatic arthritis.
To briefly describe PROMIS 29, it is a generic health-related quality-of-life instrument that assesses 7 health domains with 4 items each, and then has 1 additional pain intensity scale. These domains are anxiety, depression, fatigue, pain interference, physical function, sleep disturbance, and satisfaction with social participation.
RALN: How did the profile perform among this patient population? What did you discover?
Dr. Orbai: Our hypothesis was that PROMIS 29 domains would correlate nicely with similar domains on traditional measures that we used to have in clinical trials such as the SF-36 instrument, the Health Assessment Questionnaire disability index, and FACIT Fatigue.
We aimed to demonstrate construct validity of PROMIS 29. The methods used to assess this were assessing correlations with similar and related domains on traditional measures. We found very high correlations between PROMIS 29 physical function and SF-36 physical function, in the range of 0.87, as well as correlations with the other similar domains.
Data from the clinical trial, with these high correlations among similar and related domains on various measures, supports our hypothesis that the construct validity is good, meaning the PROMIS 29 profile is measuring what we want it to measure.
We then further looked at correlations of individual PROMIS 29 domains with FACIT Fatigue, which is an instrument that measures fatigue and has been validated in psoriatic arthritis. As expected, we found very high correlations with fatigue, in the range of 0.88. That's very good, close to 0.9.
Then related domains—social participation, pain interference, physical function—also showed very high correlation with FACIT Fatigue, as expected.
The third legacy measure that we used to demonstrate the validity of the PROMIS 29 domains is the Health Assessment Questionnaire disability index, which measures disability, the inverse of physical function. As expected, we found high inverse correlation with PROMIS 19 physical function and related domains.
These data support construct validity of the PROMIS 29 profile. Then we went a bit further to see whether the PROMIS 29 profile also performs well in distinct subgroups. We took the patient global measure and divided patients into quartiles.
The first quartile, the patient's doing very well on the patient global, with subsequently patients who are doing worse. The PROMIS 29 profile, as expected, had worse scores on the scale of patient global. People who are doing great in the first patient global quartile had PROMIS scores close to the general population mean.
As they're worsening their own assessment of psoriatic arthritis, their PROMIS scores also worsened, showing us that this tool differentiates between patients who are doing well or not doing well, which is one of the main questions during monitoring treatment and monitoring whether we are at treatment goals in the psoriatic arthritis clinic.
Our next question was, "Is the instrument reliable?" If patients don't change in their clinical status, is the PROMIS 29 profile stable as well? This is a very important measurement property. The clinical trial was not specifically designed for test-retest reliability, so the way we performed this analysis was we took patients who did not change on their patient global assessment at week 16 and at week 24.
Then we calculated interclass correlation coefficients. These were acceptable— 0.87 for physical function and pain interference, 0.82 for sleep disturbance, 0.8 for fatigue, and in the range of 0.8 for the other PROMIS instruments as well.
The third measurement property assessed was longitudinal responsiveness. Again, we looked at the changes in the PROMIS 29 t-scores versus patient-reported change in their global assessment of disease. We noticed that patients in the trial who reported the highest improvement on their patient global assessment at week 24 also had the most significant changes, improvements in their PROMIS 29 t-scores.
These were statistically significant between groups defined by improvement in patient global assessment. To give you an example, patients who improved by 31 to 40 points on the patient global assessment scale experienced much more improvement in the PROMIS 29 profile scores compared to patients who did not improve as much.
Based on these data, we conclude that PROMIS 29 profile showed construct validity in psoriatic arthritis. It showed known groups validity, meaning it can differentiate between groups of patients doing well. It seems to have good test-retest reliability.
It also seems to be responsive with changes in treatment, which is good news because it can probably be used longitudinally to measure disease improvement and worsening in this patient population.
RALN: This is going to be a valuable tool in the future for clinicians who are working with patients with psoriatic arthritis and they want to monitor how these patients are doing, how they're improving, or perhaps not improving, on a particular type of medication regimen or treatment plan.
Dr. Orbai: Absolutely. Also, it's going to be a very practical tool, particularly for physicians who use electronic medical records. The PROMIS 29 profile, as well as more advanced versions of PROMIS questionnaires, like computer-adaptive tests, are implemented in electronic medical records and can be administered to patients even before their clinic visits.
I know many clinic visits have this system implemented where the physician already receives the patient's responses before the clinic visit. Then these can be used for shared decision-making in the clinic. Is my treatment working? Do we have to escalate therapy? In addition to the traditional disease activity measures like the joint counts, entheseal count, dactylitis, the disease manifestations.
RALN: Will this simplify the process of analyzing this information? Instead of having perhaps to do 3 or 4 different kinds of testing, will PROMIS 29 provide physicians with a single tool that could answer a lot of the questions that they need to have answered?
Dr. Orbai: Yes, Rebecca, and this is the great thing about PROMIS 29. Also the display. The patients can nicely see a score display. Their scores, if they are close to the general population mean, are presented in green. Then you have shades of yellow, orange, red if things are worsening on one dimension.
It's very easy to notice where a patient may be experiencing difficulties and then address those things during clinical care.
Rebecca: Thank you very much for explaining all of this to us. It sounds like this PROMIS 29 is well-named because it does show a great deal of promise in helping clinicians have a tool that can answer a lot of questions simply and quickly.
Dr. Orbai: My pleasure. Thank you, Rebecca.