Estimates of Minimally Important Differences and Patient Acceptable Symptom State in Five Patient-Reported Outcomes Measurement Information System Short-Forms Among Individuals with SLE

Patricia Katz^1,2, Evo Alemao³, Jayanti Mukherjee³ and Kaleb Michaud^2,4, ¹Forward/National Data Bank for Rheumatic Diseases, Wichita, KS, ²University of California San Francisco, San Francisco, CA, ³Bristol-Myers Squibb, Princeton, NJ, ⁴Rheumatology, University of Nebraska Medical Center, Omaha, NE

Meeting: 2018 ACR/ARHP Annual Meeting

Keywords: patient-reported outcome measures and systemic lupus erythematosus (SLE)

Session Information

Date: Monday, October 22, 2018

Title: Patient Outcomes, Preferences, and Attitudes Poster I: Patient-Reported Outcomes

Session Type: ACR Poster Session B

Session Time: 9:00AM-11:00AM

Background/Purpose: As the NIH PROMIS (Patient-Reported Outcomes Measurement Information System) measures are used more widely, information is needed to enhance interpretation. Minimally important differences (MIDs; estimate of clinical significance of change; “feeling better/worse”) and patient acceptable symptom state (PASS; state at which patients consider symptoms to be acceptable; “feeling good”) are important components in interpretation. We estimated MIDs and PASS for five PROMIS domains in SLE.

Methods: Data were from the Forward/National Data Bank for Rheumatic Diseases. Participants complete questionnaires every 6 months. In July 2015 to July 2017 (five administrations), 4-item short-forms for 5 PROMIS domains were added: Physical Function, Fatigue, Sleep Disturbance, Pain Interference and Satisfaction with Social Roles. PROMIS scales were scored to derive T-scores scaled to population means of 50 and SD of 10. Changes were calculated for consecutive administrations, yielding 4 change periods. MID estimates were calculated using both anchor- and distribution-based methods.¹ Anchors were comparisons of each domain and overall health to 6 months before (rated as much better, somewhat better, neither better nor worse, somewhat worse, much worse).² Domain-specific comparisons (e.g., “compared to 6 months ago, is your fatigue now…?”) were asked only once; overall health comparisons were asked in all questionnaires. PASS was estimated as the 75^th percentile positive score of those who stated their current health was acceptable.³

Results: The number of respondents ranged from 389 – 462 in the 5 administrations. Table 1 shows characteristics of respondents from one administration. Tables 2 and 3 show MID and PASS estimates, respectively.

Conclusion: MIDs for PROMIS scales in SLE appear to be similar to those reported elsewhere (±2 points).⁴ PASS estimates have not been examined previously for PROMIS. PASS estimates for PROMIS T-scores are 0.5 to 1 standard deviation better than the population mean of 50. Such information will improve interpretation of PROMIS scores and changes in those scores.

References:

Revicki D et al. J Clin Epidemiol 2008;61:102
Bellamy N et al. Arthritis Care Res 2015;67:972
Tubach F et al. Arthritis Care Res 2012;64:1699
Hays RD et al. Ann Rheum Dis 2015;74:104

Table 1. Characteristics of Respondents to July 2017 Questionnaire (N=397)
Sociodemographic
Age, years	60.6±12.4
Female, %	94.5
Race, White, %	81.8
Health characteristics
SLE duration, years	24.2±12.5
Physician-confirmed SLE diagnosis, %	73.0
Rheumatic Disease Comorbidity Index	2.7±2.0
HAQ score	0.93±0.70
Fatigue, NRS^†	4.4±3.0
Sleep problems, NRS^†	4.3±3.1
Pain, NRS^‡	3.7±2.9
Lupus activity, NRS^§	2.5±2.5
PROMIS^® scores
Physical function	42.7±9.6
Fatigue	55.2±11.5
Pain interference	56.8±9.9
Sleep disturbance	53.0±9.6
Satisfaction with social roles	47.4±10.3
Values are mean ±SD unless otherwise noted ^†Scale: 0=no problem, 10=severe problem ^‡Scale: 0=no pain, 10=severe pain ^§Scale: 0=no activity, 10=severe activity NRS=numeric rating scale; PROMIS^®=Patient-Reported Outcomes Measurement Information System

Table 2. MID Analyses

Anchor-based analysis

Distribution-based analysis

ΔPROMIS^®

SE of measurement

0.35 SD

MID best estimate

PROMIS^® scale

Compared with 6 months before

Better*

Worse*

Mean

Range

Mean

Range

Better

Worse

Fatigue

−2.2

1.8

1.44

1.38–1.52

2.51

2.41–2.67

−2

Overall health^†

−2.4

1.8

Pain Interference

Pain interference

−2.1

1.7

2.20

2.19–2.22

3.45

3.43–3.47

−2

Pain^†

−1.9

1.4

Overall health^†

−1.4

1.8

Sleep disturbance

Sleep

−3.0

2.7

4.14

4.00–4.31

3.26

3.13–3.37

−3

Overall health^†

−2.6

1.0

Satisfaction with social roles

Social functioning

0.8

−2.1

2.01

1.95–2.06

3.51

3.41–3.61

−2

Overall health^†

1.2

−2.0

Physical function

Function^†

1.1

−1.2

2.34

2.32–2.35

3.34

3.31–3.64

−2

Overall health^†

1.1

−1.4

*Change shown for “somewhat worse” and “somewhat better” groups. “Much worse” and “much better” groups were excluded, as per Bellamy²

^†ΔPROMIS^® and standardized response mean were averaged over four change periods

Δ=change; MID=minimally important difference; PROMIS^®=Patient-Reported Outcomes Measurement Information System

Table 3. PASS Estimates
	Based on “last 48 hours” question	Based on satisfaction with health question*	Best estimate of PASS
Fatigue^†	46	47.6	47
Pain interference^†	41.6	50.6	42
Sleep disturbance^†	46.2	47.1	47
Satisfaction with social roles	64.2	53.3	60
Physical function	56.9	49.8	55
Population mean ±SD T-scores are 50±10 *Scores averaged over four change periods ^†Lower scores reflect better health status. Otherwise, higher scores reflect better health status PASS=patient acceptable symptom state

Disclosure: P. Katz, Bristol-Myers Squibb, 2; E. Alemao, Bristol-Myers Squibb, 1, 3; J. Mukherjee, Bristol-Myers Squibb, 1, 3; K. Michaud, University of Nebraska Medical Center and FORWARD, The National Databank for Rheumatic Diseases, 3,Rheumatology Research Foundation and Pfizer, 2.

To cite this abstract in AMA style:

Katz P, Alemao E, Mukherjee J, Michaud K. Estimates of Minimally Important Differences and Patient Acceptable Symptom State in Five Patient-Reported Outcomes Measurement Information System Short-Forms Among Individuals with SLE [abstract]. Arthritis Rheumatol. 2018; 70 (suppl 9). https://acrabstracts.org/abstract/estimates-of-minimally-important-differences-and-patient-acceptable-symptom-state-in-five-patient-reported-outcomes-measurement-information-system-short-forms-among-individuals-with-sle/. Accessed .

« Back to 2018 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/estimates-of-minimally-important-differences-and-patient-acceptable-symptom-state-in-five-patient-reported-outcomes-measurement-information-system-short-forms-among-individuals-with-sle/