PROMIS PF correlates with HOOS, JR in patients with hip pain

Abstract

Background

Patient-reported outcome measures (PROMs) are increasingly integrated into reporting requirements tied to reimbursement. There may be advantages to computer adaptive tests that apply to many different anatomical regions and diseases, provided that important information is not lost.

Questions

1) Does the Patient-Reported Outcomes Measurement Information System Physical Function (PROMIS PF) computer adaptive test correlate with the Hip injury and Osteoarthritis Outcome Score for Joint Replacement (HOOS, JR: a hip-specific PROM); 2) Is there any difference in the amount of variation explained by various factors (e.g. age, BMI, presence of concomitant knee pain) for both measures?

Methods

In this prospective, cross-sectional study of 213 patients, we assessed the Pearson correlation of PROMIS PF and HOOS, JR. To investigate the variation explained by various patient-level factors, we constructed two multivariable linear regression models.

Results

We found a large correlation between PROMIS PF and HOOS, JR (r 0.58, P < 0.001). Disabled or unemployed status was independently associated with both lower PROMIS PF and HOOS, JR scores (regression coefficient [β] −3.4; 95% confidence interval [CI] −5.8 to −1.0; P = 0.006 and β −11; 95% CI -17 to −5.0; P < 0.001, respectively). Private rather than public insurance was associated with both higher PROMIS PF and HOOS, JR scores (β 4.5; 95% CI 2.2 to 6.8; P < 0.001 and β 6.4; 95% CI 0.49 to 12; P = 0.034, respectively). No floor or ceiling effects were observed for PROMIS PF. HOOS, JR scores showed 4.2% floor and 0.5% ceiling effect.

Conclusions

This study adds to the evidence that general measures of physical limitations may provide similar information as joint- or region-specific measures.

Level of evidence

Level III.

1. Introduction

Patient-reported outcome measures (PROMs) quantify symptom intensity and magnitude of limitations. As such, PROMs may be a good measure of what matters to patients. Consequently, they are increasingly emphasized in alternative payment models rewarding “value” delivered to patients by clinicians.

Disease-specific and joint-specific PROMs are widely used, including the Western Ontario and McMaster Universities Arthritis Index (WOMAC), the Oxford Hip Score, and the Hip injury and Osteoarthritis Outcome Score (HOOS). The 6-item HOOS Joint Replacement (HOOS, JR) questionnaire^1,2 has largely replaced the original 40-item HOOS. The HOOS, JR is part of the Comprehensive Care for Joint Replacement program, along with general measures such as PROMIS Global-10.^3,4 The short-form HOOS, JR reduces survey burden, but there are disadvantages to its joint- and side-specific nature. For example, a patient with bilateral hip osteoarthritis and unilateral knee osteoarthritis would complete three surveys. For longitudinal tracking, three surveys would be required at each follow-up time point. On the other hand, a measure like Patient-Reported Outcomes Measurement Information System Physical Function (PROMIS PF) covers all joints on both sides.

The PROMIS⁵ initiative supported by the National Institutes of Health (NIH) developed freely available measures translated into multiple languages and addressing a broad cross-section of medicine. They are developed based on item response theory, resulting in computer-adaptive tests (CAT) that can be completed in an average of 4–5 questions with little or no ceiling and floor effects.^6,7 The efficiency of CATs, and the practicality of a single survey for physical function inclusive of any pathology, increase the appeal of PROMIS PF CAT.

This study assessed: 1) the correlation between PROMIS PF and HOOS, JR; 2) differences in the amount of variation explained by other clinical or patient variables (e.g. age, BMI, presence of concomitant knee pain) for each of these measures; 3) the strength of the correlation between the two measures accounting for confounder effects; and 4) floor and ceiling effects for both measures.

2. Materials and methods

2.1. Study design

After institutional review board approval of this cross-sectional study, we recruited adult English- or Spanish-speaking patients presenting to an urban outpatient arthroplasty clinic with hip pain. Research assistants in conjunction with clinicians identified eligible patients with atraumatic hip pain of a native joint. Patients under 18 years of age, cognitively impaired, non-English or Spanish speaking, with traumatic hip injury etiology, or with prior hip replacement in the affected joint were excluded. After description of the study, completion of the questionnaires implied informed consent. While specific diagnoses were not reported, the vast majority of patients had hip osteoarthritis. Questionnaires were completed in a private exam room on an electronic tablet using the Research Electronic Data Capture (REDCap) secure web-based application.⁸

2.2. Outcome measures

Patients completed a demographic survey, PROMIS PF, and HOOS, JR in this order. Demographic questions queried for age in years, gender, race, employment status, type of insurance in addition to self-reported height and weight for the purpose of calculating body mass index (BMI), and affected lower extremity joint(s).

PROMIS PF is a CAT that quantifies magnitude of physical limitations. The output is in the form of T-scores indexed to population averages, where a score of 50 defines the population mean with a standard deviation of 10. Higher scores indicate better physical function.⁹

HOOS, JR is a 6-item measure featuring answer choices on a 5-point Likert scale ranging from “None” to “Extreme” in relation to specific activities such as rising from sitting or going up stairs. The summed raw score scales to an interval score that ranges from 0 to 100, wherein a score of 0 suggests total hip disability and a score of 100 indicates perfect hip health.²

2.3. Survey timings

Duration of each PROM (i.e. seconds to complete all items) was computed for both measures. PROMIS PF survey duration was calculable for all patients by subtracting the REDCap timestamp start time for the following survey (HOOS, JR) from the beginning timestamp of the PROMIS PF. HOOS, JR completion time, on the other hand, could only be calculated for those patients filling out a second HOOS, JR for the other side (indicating bilateral hip pain) or those with knee pain who filled out a knee disability score. Only time spent on the first HOOS, JR was tabulated. For both measures, any survey durations >60 s were excluded from calculations since prolonged item completion time signifies filling errors in clinic (e.g. patient is interrupted by the medical assistant). To truly compare completion time, we only used time data of patients for which we were able to calculate both PROMIS PF and HOOS, JR completion time (N = 53; 23%).

2.4. Study population

A total of 231 patients consented to participate in this study. Eighteen (7.8%) patients were excluded for not completing the full sequence of surveys related to overlapping clinical and research time constraints. The final analysis included 213 (92%) patients. Mean patient age was 56 ± 13 years old (range 18–89), with 140 (66%) women, predominately Caucasian (138; 65%; Table 1). All patients had hip pain including 52 patients (24%) with bilateral symptoms. Concomitant knee pain was present in 91 (43%) patients. Mean PROMIS PF score was 38 ± 7.6 and mean HOOS, JR score was 47 ± 20.

2.5. Statistical analysis

Histogram plots of the outcome variables showed a normal distribution. Continuous variables were presented as mean ± standard deviation (SD) and discrete data as proportions. Bivariate analyses using Pearson's correlation and unpaired Student's t-tests were performed to determine the relationships of the explanatory variables with PROMIS PF and HOOS, JR and to understand the correlation between PROMIS PF and HOOS, JR. For patients with bilateral hip pain, the worse HOOS, JR score was used in the analyses. We created four separate multivariable linear regression models to (1) assess factors independently associated with both outcome measures, and (2) to assess if the correlation between both outcome measures remained as strong in a multivariable model that includes other variables. We included all explanatory variables with P < 0.10 on bivariate analyses (Appendix 1) in our final models to see variation explained in both models by the same factors, indicated by R-squared and adjusted R-squared. We also calculated floor (lowest) and ceiling (highest) effect for both measures to assess censoring. We considered P < 0.05 significant.

We powered on our secondary hypothesis since this normally yields more participants. An a priori power calculation indicated that a sample of 168 subjects would provide 90% statistical power, with alpha set at 0.05, for a regression with five predictors if one of the demographic factors would account for 15% of the variability in either PROMIS PF or HOOS, JR, and our complete model would account for 20% of the overall variability. In order to account for discontinuation of the study and to include enough patients with both unilateral and bilateral hip pain we enrolled 25–30% more patients (total of 230).

3. Results

3.1. Correlation PROMIS PF and HOOS, JR

There was a large correlation between PROMIS PF and HOOS, JR (r 0.58; P < 0.001; Fig. 1).

3.2. Factors associated with PROMIS PF

Accounting for potential interaction of variables using multivariable analysis, worse physical function (i.e. lower PROMIS PF) was independently associated with being unemployed or disabled (regression coefficient [β] −3.4; 95% confidence interval [CI] −5.8 to −1.0; P = 0.006; Table 2); private health insurance was independently associated with better physical function (β 4.5; 95% CI 2.2 to 6.8; P < 0.001).

3.3. Factors associated with HOOS, JR

Accounting for potential interaction of variables using multivariable analysis, a higher BMI (β −0.48; 95% CI -0.79 to −0.18; P = 0.002) and an employment status of unemployed, disabled, or other as opposed to employed or retired (β −11; 95% CI -17 to −5.0; P < 0.001) were independently associated worse hip condition (i.e. lower HOOS, JR score; Table 2). Having private insurance as opposed to Medicare, Medicaid, or other was independently associated with better hip condition (β 6.4; 95% CI 0.49 to 12; P = 0.034).

3.4. Correlation between PROMIS PF and HOOS, JR controlling for other factors

Private insurance (β 3.3; 95% CI 1.2 to 5.4; P = 0.002) and higher HOOS, JR scores (β 0.19; 95% CI 0.14 to 0.24; P < 0.001) emerged as the independent variables associated with higher PROMIS PF scores (Table 3). BMI (β −0.38; 95% CI -0.65 to −0.11; P = 0.006) and work status of unemployed, disabled, or other (β −7.0; 95% CI -12 to −1.5; P = 0.013) were independently associated with lower HOOS, JR scores. Higher PROMIS PF scores were associated with higher HOOS, JR scores (β 1.2; 95% CI 0.90 to 1.5; P < 0.001).

3.5. Data distributions

No floor or ceiling effects were observed for PROMIS PF. HOOS, JR scores showed 4.2% floor and 0.5% ceiling effect (Table 2).

HOOS, JR is a fixed 6 questions, whereas PROMIS PF can range from 4 to 12 questions based on patient answers given, and most patients received 4 questions (4.1 ± 0.7).

Patients completed HOOS, JR in an average of 31 ± 13.3 s compared to 38 ± 11.8 s for PROMIS PF (P < 0.001; Table 2).

4. Discussion

As the clinical use of PROMs becomes increasingly widespread and tied to clinician reimbursement, there may be advantages to using fewer, simpler questionnaires that apply to several diseases and anatomical regions. PROMIS PF measures physical limitations, accounting for all musculoskeletal pathology along with psychological and social determinants of health. It also has the advantage of being adaptive which can obviate redundant questions based on previous responses. PROMIS PF may be less onerous and more efficient in populations vulnerable to survey fatigue with pain in multiple joints requiring repeated joint-specific symptom appraisal.

The strong correlation between PROMIS PF and HOOS, JR scores suggests that PROMIS PF may effectively capture the influence of hip pathology on physical limitations. Our results are in line with those of Padilla and colleagues who also found a strong correlation (r 0.60) between PROMIS PF short forms and HOOS, JR.¹⁴

Variation in PROMIS PF and HOOS, JR scores is accounted for by similar variables. The observation that BMI has a small independent influence on HOOS, JR scores, but not on PROMIS PF scores might reflect that larger body mass is more detrimental to the hip joint than to physical ability overall, but it could also be spurious or associated with other, unmeasured variables. It is notable that concomitant knee arthritis was not significant in the final multivariable models for both HOOS, JR and PROMIS PF scores. However, absence of concomitant knee pain signified higher scores in bivariate analysis, which agrees with the higher WOMAC scores for hip arthritis patients without additional knee arthritis shown by Juhakoski and colleagues.¹⁰ Interestingly, bilateral hip pain was likewise significant in bivariate analysis for both PROMIS PF and HOOS, JR, but failed to maintain significance in both multivariate models. The finding that socioeconomic factors (employment and insurance status) accounted for variation in scores of both measures is consistent with findings of studies using the WOMAC and Stanford Health Assessment Questionnaire Disability Index in symptomatic hip osteoarthritis patients.¹¹ In patients with upper extremity problems, both PROMIS PF and PROMIS Upper Extremity scores were significantly lower for unemployed or disabled patients.¹² Other studies found that comparable patients with public insurance have worse PROM scores both pre-operatively and post-operatively.¹³

While additional seconds were required to fill out PROMIS PF compared to HOOS, JR, on average, patients taking PROMIS PF faced fewer questions (4.1 versus the fixed 6 of HOOS, JR). It is possible that webpage-loading time may account for the longer time to complete the CAT. Usually CATs are completed more quickly than longer disease-specific measures in the upper extremity ^15,16 and trauma.¹⁷ The lack of PROMIS PF floor or ceiling effects, in contrast to those of HOOS, JR, matches findings from prior studies and is another advantage of a CAT.18, 19, 20

These results should be considered in the context of the following limitations. This was a single institution study with the potential for spectrum bias. Patients presenting to this urban adult reconstructive specialty clinic may not be representative of degenerative joint disease patients overall across the US population. Response bias is also a consideration, since patients voluntarily completed the questionnaires comprising this study. A few patients were tired of completing questionnaires since they had already completed several as part of their routine intake forms. Our cross-sectional study design cannot evaluate responsiveness to treatment. In one study of nearly 1,000 patients seeking joint-related care, PROMIS PF was more responsive to improvements in pathophysiology after hip or knee joint replacement than the HOOS, JR and KOOS, JR.²¹The HOOS, JR had the lowest effect size.²¹ Another study found that a general health measure (SF-36) was as responsive as a joint specific measure (the Oxford Hip Score) to changes in joint health after total hip arthroplasty.²²

5. Conclusion

The clinical utility of PROMs may be higher when they can be compared across populations of patients with pathophysiology at different regions or of different disease processes (e.g. osteonecrosis vs. arthritis).23, 24, 25 Health care institutions are looking for an efficient, parsimonious PROM menu that applies across the variety of conditions their clinicians treat. Simplification of the delivery process for these tools may curb costs for smaller practices and institutions who are newly establishing PROM collection, or re-evaluating measures used. Our results—consistent with other studies—that PROMIS PF measures similar information to the HOOS, JR supports its use for patients with hip disease, and is consistent with an efficient strategy for PRO collection as part of delivering value-based care.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Appendix A. Supplementary data

The following is the Supplementary data to this article: