Abstract
Objectives
To study the test-retest reliability and measurement error, construct validity, responsiveness, interpretability, and floor/ceiling effects of a Patient-Reported Outcomes Measurement Information System (PROMIS) short form designed to measure physical function in geriatric rehabilitation patients (PROMIS-PF-GR).
Design
Prospective cohort.
Setting and Participants
Inpatient geriatric rehabilitation patients.
Methods
We evaluated the test-retest reliability by re-administering PROMIS-PF-GR 3 to 5 days after the admission measurement. The intraclass correlation coefficient (ICC) was calculated to determine test-retest reliability; an ICC of ≥0.70 was considered sufficient. Measurement error was established by calculating the standard error of measurement and smallest detectable change. Construct validity and responsiveness were determined by testing a priori formulated hypotheses (criterion: ≥75% hypothesis not rejected). Interpretability was evaluated by calculating the minimal important change using predictive modeling and a global rating as criterion for change. Floor/ceiling effects were established by calculating the percentage patients with the minimum/maximum raw score (criterion: ≤15%) at admission and discharge.
Results
A total of 207 patients participated in the study [mean ± standard deviation age (80 ± 8.3 years), 58% female]. More than one-half of patients (56%) reported to be improved during rehabilitation. The ICC was 0.79 (95% confidence interval 0.70–0.84), the standardized error of measurement was 3.8, and the smallest detectable change 10.6. None of the 4 hypotheses for construct validity were rejected; 2 out of 5 hypotheses for responsiveness were rejected. The minimal important change was 8.0 (95% confidence interval 4.1–12.5). No floor/ceiling effects were found.
Conclusions and Implications
The PROMIS-PG-GR showed sufficient test-retest reliability, measurement error, and construct validity. We did not find sufficient evidence for responsiveness, which may be due to the unexplained weak correlation between the PROMIS change score and the Global Rating Scale. We still recommend the use the PROMIS-PG-GR for measuring self-reported physical function in geriatric rehabilitation.
Keywords
One of the major goals of geriatric rehabilitation is to restore physical functioning or enhance residual functional capability in frail and older patients with disabilities.
1
,2
Ideally, treatment progress should be measured and evaluated not only from the perspective of the professional but also from the perspective of the patient. The patient's perspective gives insight in the rehabilitation progress, which is considered meaningful for patients and could help professionals focus on the patient's specific desired health outcomes. Patient-Reported Outcome Measures (PROMs) can be used for this purpose and help to monitor improvement in health and well-being and improve shared decision making.3
, 4
, 5
, 6
, 7
Multiple PROMs are available to measure physical function in geriatric patients.
8
However, many PROMs have developmental and psychometric shortcomings, such as the absence of a clearly described measurement aim, lack of patient involvement in the development process, and insufficient responsiveness.8
Moreover, most have not been developed and validated by modern psychometric methods like Item Response Theory. To overcome these shortcomings a new high quality PROM based on the Patient-Reported Outcomes Measurement Information System (PROMIS) was recently developed.9
PROMIS is an innovative psychometrically sound and clinically meaningful measurement system of item banks for measuring domains such as fatigue, anxiety, and physical function across a wide range of medical conditions.
10
, 11
, 12
Each bank consists of a set of items (questions) with responses (answers) that measure the same domain (construct, in this case physical function) and whose item parameters (the item locations on the underlying metric and discriminative ability) have been established using Item Response Theory analysis.10
, 11
, 12
PROMIS item banks can be used to create custom short forms, consisting of a fixed subset of relevant items from an item bank. We have chosen to develop a new custom-made PROMIS short form for geriatric rehabilitation rather than using one of the existing PROMIS short forms, as geriatric rehabilitation patients have relatively low physical function levels while the existing short forms generally speaking have been developed to cover the whole range of physical function. This implies that some items in the existing PROMIS short forms are irrelevant for the geriatric rehabilitation population and that these short forms lack a sufficient number of items addressing lower levels of physical function. Advantages of custom short forms are that they can be tailored to the expected level of functioning of the target population at issue, in this case geriatric rehabilitation patients. This means that a custom short form can measure the level of physical function of these patients more precisely than the existing short forms.
10
We developed and validated a custom PROMIS short form from the PROMIS physical function item bank with the involvement of experienced health professionals and geriatric rehabilitation patients: the 24-item PROMIS Physical Function Geriatric Rehabilitation (PROMIS-PF-GR) short form.
9
The PROMIS-PF-GR previously showed sufficient content validity, structural validity, measurement invariance, and internal consistency.9
The objective of the current study was to test additional psychometric properties of the PROMIS-PF-GR in geriatric rehabilitation patients: test-retest reliability, measurement error, construct validity, responsiveness, and interpretability.Methods
The study was approved by the Medical Ethics Review Board of the institution (number FWA00017598).
Setting and Participants
The PROMIS-PF-GR was tested in 11 nursing homes throughout the Netherlands in 2016–2017. To be eligible for participation, nursing homes had to have specialized geriatric rehabilitation wards. All new patients admitted for geriatric inpatient rehabilitation were eligible and were recruited by a research assistant. Exclusion criteria were decision incompetent patients, as rated by the attending physician; patients who did not master the Dutch language, and patients who did not provide informed consent.
Procedure
Three measurements were performed: in the week of admission, 3 to 5 days after the first measurement, and in the week of discharge. At admission we collected demographic data and presence and status of a caregiver, and administered the updated Charlson Comorbidity Index, the Mini-Mental State Examination (MMSE), the Barthel Index, and the PROMIS-PF-GR. Three to 5 days after the first measurement, we repeated the PROMIS-PF-GR. At the week of discharge, we collected data about the discharge destination, the Barthel Index, and PROMIS-PF-GR, and administered a Global Ratings Scale. The attending physician provided information on comorbidity of the participating patients from their charts. The attending nurses filled in the Barthel Index. The PROMIS-PF-GR was filled in by the patients or with help of a research assistant. All other information was provided by the patients.
Measures
We used the updated Charlson Comorbidity Index, a widely used measure of burden of disease. It assesses the presence of 12 comorbidities and predicts mortality.
13
,14
A total score is obtained by assigning a weighted score to each comorbidity, yielding a maximal score of 24 points. Higher scores indicate higher risk of mortality.The MMSE is a widely used screening test for cognitive limitations. It consists of 11 questions, grouped in 7 categories, measuring aspects of cognitive functioning.
15
The 7 categories are added into a total score ranging from 0 to 30. Higher scores indicate a higher level of cognitive functioning. The MMSE has sufficient diagnostic accuracy. Several cut-off points for the MMSE are used as indicator of cognitive dysfunction. We applied a score below 25 as cut-off as this cut-off has a high negative predictive value for detecting cognitive dysfunction.16
The Barthel Index is an observational instrument that measures independent functioning in 10 basic activities of daily life activities, such as eating, mobility, and grooming.
17
The total score ranges from 0 to 20. Higher scores indicate a higher level of independence in activities of daily life functioning. Its psychometric properties are sufficient for geriatric rehabilitation.18
The PROMIS-PF-GR measures self-reported physical function, its content was compiled by geriatric rehabilitation patients and experts from the PROMIS physical function item bank.
9
It consists of 24 items on a 5-point Likert scale and takes between 4 and 7 minutes to complete and can be found in Supplementary Table 1. Four items address the degree to which the patients’ health limits physical activities such as climbing stairs and taking a shower (“not at all” to “cannot do”). The other 20 items address the level of difficulty in carrying out activities such as walking, dressing, and going to the toilet (“without any difficulty” to “unable to do that”). When patients were not able to complete the PROMIS-PF-GR, a research assistant read each item aloud and the patient verbally expressed his/her answer, which the research assistant filled in.9
Raw PROMIS-PF-GR scores can be converted to T-scores, using Health Measures Scoring Service powered by the US Assessment Center.
19
The mean T-score of the general (US) population is 50 with a standard deviation (SD) of 10. A score between 30 and 40 represents moderate limitations and below 30 severe limitations in physical function.Assessment Center.
http://www.assessmentcenter.net
Date accessed: February 26, 2021
20
,Health measures.
http://www.healthmeasures.net/score-and-interpret/interpret-scores/promis/promis-score-cut-points
Date accessed: February 26, 2021
21
The PROMIS-PF-GR has sufficient content and structural validity for measuring self-reported physical function during geriatric rehabilitation.- Rothrock N.E.
- Cook K.F.
- O’Connor M.
- et al.
Establishing clinically-relevant terms and severity thresholds for Patient-Reported Outcomes Measurement Information System (PROMIS) measures of physical function, cognitive function, and sleep disturbance in people with cancer using standard setting.
Qual of Life Res. 2019; 28: 3355-3362
9
Results from the original developmental study suggest that PROMIS-PG-GR can even be used in geriatric rehabilitation patients with mild cognitive functioning.9
The Global Rating Scale (GRS) is a single question that asks the patient to what extent his/her physical function has changed during the rehabilitation period, with 5 response options: (1) much deteriorated; (2) somewhat deteriorated; (3) no change; (4) somewhat improved; and (5) much improved.
Analysis
Test-retest reliability and measurement error
Reliability and measurement error address the precision of the measure. Reliability refers to the proportion of the total variance in the measurement which is due to “true” differences between patients, whereas measurement error refers to the systematic and random error of a patient's score that is not attributed to “true” changes.
22
Intraclass correlation coefficients (ICC, absolute agreement, 2-way random effects model) were calculated as a measure of reliability. The following formulas were used to calculate the ICCs, the standard error of measurement (SEM), and smallest detectable change (SDC)
23
:(1)
(2)
(3)
An ICC of ≥0.70 was considered sufficient. Ideally, the smallest detectable change should be smaller than the minimal important change in order to be at least 95% sure that changes as large as the minimal important change are not measurement error.
Construct validity
Construct validity addresses whether the measure assesses the intended construct. This was evaluated by testing a priori formulated hypotheses. We formulated 4 hypotheses based on psychometric and clinical assumptions (Table 1). Construct validity was considered to be sufficient when at least 75% of the hypotheses would not be rejected.
Responsiveness
Responsiveness addresses whether the measure is able to detect change over time in the construct (physical function) being measured.
22
Responsiveness was determined by testing a priori formulated hypotheses. We formulated 5 hypotheses based on psychometric and clinical assumptions (Table 1). Responsiveness was considered to be sufficient when at least 75% of the hypotheses would not be rejected.Three groups were created based on the responses to the GRS. Patients who indicated that they were much improved were labeled as “improved,” patients who indicated that they were somewhat improved, not changed, or somewhat deteriorated were labeled as “not changed,” and patients who indicated that they were much deteriorated were labeled as “deteriorated.”
23
Next, effect sizes and standardized response means were calculated and compared for the “improved” and “not changed” group. The data from the “deteriorated” group was not used because we were interested in minimal important improvement.23
We calculated the mean change in PROMIS-PF-GR T-scores between baseline and discharge divided by the SD of the baseline score for the effect size, and divided the SD of the change score for and standardized response means.24
Correlations were calculated by means of Pearson correlations between PROMIS-PF-GR and Barthel Index at admission, discharge, and changes scores. In addition, the correlation was determined between PROMIS-PF-GR at admission and discharge. We set a threshold of R >0.3 for moderate correlations (different constructs; observed vs self-reported physical function) and R > 0.5 for strong correlations (same construct; both self-reported physical function).22
,23
Finally, we plotted the receiver operating characteristics curve with the GRS as anchor and determined the area under the curve to assess the ability to discriminate between “improved” and “not changed” patients. An area under the curve of 0.7 was considered evidence for responsiveness.25
Interpretability
To evaluate the interpretability of PROMIS-PF-GR change scores, we calculated the minimal important change using predictive modeling.
26
,27
The predictive modeling method uses logistic regression analysis to determine the relationship between the change score and group membership (improved vs not changed). The minimal important change is equal to the score that corresponds to a likelihood ratio of 1, which is the change score for which the post-test probability of belonging to the improved group is equal to the pre-test probability. The minimal important change was adjusted for the proportion of improved patients and non-parametric bootstrapping (1000 samples) was used to calculate the 95% confidence interval (CI).26
,27
If the minimal important change is lower than the smallest detectable change, the probability that a change as large as the minimal important change (MIC) is not due to measurement error can be determined by calculating the Z-score with formula 4:(4)
Floor/Ceiling effects
Floor/ceiling effects addresses the proportion of respondents with the lowest/highest score on a measure. We calculated the percentage of patients with the lowest (floor) and the highest possible (ceiling) raw summary score both at admission and discharge. We considered a floor/ceiling effect present, if the percentage was ≥15%.
28
The minimal important change was determined using R in the R package pROC (R project for statistical computing, v 3.4.3; R Foundation for Statistical Computing, Vienna, Austria). All other analysis were performed using SPSS (IBM SPSS Statistics for Window v 22.0; IBM Corp, Armonk, NY). The sample size for the whole psychometric study required a minimal of 200 patients.
23
Results
Participants
A total of 207 patients participated in the study; no patients were excluded for the study. Test-retest reliability could be determined in 207 patients. Thirty-nine patients were lost to follow-up because they were readmitted to the hospital or deceased. Data at discharge was available for 168 (81%) patients. There were no missing items scores.
Table 2 summarizes patient characteristics and PROMIS-PF-GR scores at admission and discharge. The mean age of the participants was 80 years, and 58% were female. The main reasons for admission to rehabilitation were trauma, including fractures (27.1%), elective total joint replacement (15.9%), stroke (15.5%), and miscellaneous (40.6%). The latter group consists of patients admitted for pulmonary, cardiac, and oncological rehabilitation and patients in need for rehabilitation after infection, postintensive care, major surgery or kidney failure. At admission, average T-score was 26 (SD 8), 70% of the patients had moderate functional limitations (PROMIS-PF-GR T-score between 30 and 40), and 23% had severe limitations (PROMIS-PF-GR T-score below 30). At discharge, the average T-score was 35 (SD 8), 54% had moderate limitations in physical function, and 28% had severe limitations. The average length of stay was 41 days .
Table 1Hypotheses Testing
| Construct Validity | Result |
|---|---|
| 1. Correlation between BI and PROMIS-PF-GR at admission: Pearson R > 0.3 | R = 0.58 |
| 2. Correlation between BI and PROMIS-PF-GR at discharge: Pearson R > 0.3 | R = 0.63 |
| 3. The mean change score of the PROMIS-PF-GR of patients returning home is at least 5 T-score points higher (better) than the mean change score of patients who are not able to return to home. | Not home: 9.3 Home: 3.7 |
| 4. Correlation between PROMIS-PF-GR admission and discharge: Pearson R > 0.5 | R= 0.6 |
| Responsiveness | Result |
| 5. The change scores of the PROMIS-PF-GR correlate with the change scores of the BI with Pearson R > 0.3 | R= 0.31 |
| 6. ES and SRM <0.2 for the “no change” patient group | ES = 0.7 SRM = 1.0 |
| 7. ES and SRM ≥ 0.5 for the “improved” patient group | ES = 1.2 SRM = 1.4 |
| 8. The mean change score of the PROMIS-PF-GR correlates with the raw GRS-score with Pearson R > 0.5. | R = 0.32 |
| 9. Area under the (ROC) curve “improved” vs “no change” group ≥0.7 | AUC = 0.7 |
AUC, area under the curve; BI, Barthel Index; ES, effect size; SRM, standardized response means.
Test-retest reliability and measurement error
The ICC was 0.79 (95% CI 0.70–0.84) indicating sufficient test-retest reliability. The standard error of measurement was 3.8, and the smallest detectable change 10.6 T-score points.
Construct validity
The correlation between the Barthel Index and PROMIS-PF-GR was 0.58 at admission (hypothesis 1 not rejected) and 0.63 at discharge (hypothesis 2 not rejected). The mean change in PROMIS-PF-GR T-score for patients who were able to go home was 9.3 and for those who were not able to go home 3.7 (hypothesis 3 not rejected). The correlation between the PROMIS-PF-GR at admission and discharge was 0.6 (hypothesis 4 not rejected) .
Table 2Patient Characteristics (n = 207)
| Sex, n (%) | |
| Female | 120 (58%) |
| Male | 87 (42%) |
| Age, mean (SD; range) | 80 (8.3; 61–95) |
| Highest educational level, n (%) | |
| Low | 38 (18%) |
| Medium | 114 (55%) |
| High | 53 (26%) |
| Missing | 2 (1%) |
| Primary diagnosis, n (%) | |
| Trauma | 56 (27.1%) |
| Elective joint replacements | 33 (15.9%) |
| Stroke | 32 (15.5%) |
| Amputation | 2 (1.0%) |
| Miscellaneous | 84 (40.6%) |
| Updated Charlson Comorbidity Index, mean (SD; range) | 1.5 (1.8; 0–9) |
| MMSE, mean (SD; range) | 25 (4; 9–30) |
| Length of stay (mean d, SD) | 41 (26) |
| PROMIS-PF-GR T-score, mean (SD; range) | |
| At admission (n = 207) | 26 (8; 11–51) |
| At discharge (n = 168) | 35 (8; 13–58) |
| Change | 9 (7; −6 to 33) |
| Moderate limitation physical function (n; %) | |
| At admission (n = 207) | 47 (22.7%) |
| At discharge (n = 168) | 90 (53.6%) |
| Severe limitation physical function (n; %) | |
| At admission (n = 207) | 146 (70.5%) |
| At discharge (n = 168) | 47 (28.0%) |
| Barthel Index (SD; range) | |
| At admission (n = 207) | 57.5 (22.5; 0–100) |
| At discharge (n = 168) | 77.5 (25.0; 0–100) |
| Change | 21 (22.5; −95 to 80) |
| GRS at discharge categories, n (%) | |
| Deteriorated | 1 (0.6) |
| No change | 73 (43.5) |
| Improved | 94 (56.0) |
| Missing (n = 39) | |
| T-score change between admission and discharge per GRS category | |
| Deteriorated | −1.0 |
| No change | 6.2 |
| Improved | 10.5 |
∗ T-scores are based on US item parameters.
† PROMIS physical function reference scores.
20
Health measures.
http://www.healthmeasures.net/score-and-interpret/interpret-scores/promis/promis-score-cut-points
Date accessed: February 26, 2021
Responsiveness
The correlation between the PROMIS-PF-GR and Barthel Index changes scores was 0.31 (hypothesis 5 not rejected). The effect size and standardized response means were 0.7 and 1.0, respectively, for the “no change” group (hypothesis 6 rejected), and 1.2 and 1.4 for the “improved” group (hypothesis 7 not rejected). The correlation between PROMIS-PF-GR change score and the GRS-score was 0.32 (hypothesis 8 rejected). The area under the curve was 0.7 (hypothesis 9 not rejected).
Interpretability
The minimal important change could be assessed in 167 patients and 56% of the patients reported being “improved” during rehabilitation and their change score was 8.0 T-score points (95% CI 4.1–12.5). This is smaller than the smallest detectable change (10.6 T-score points) and corresponded with a Z-score of 1.49. This indicates that if a patient changes 10.6 T-score points, there is 95% chance that this change reflects a real change, and not measurement error or lack of measurement precision. We found a minimal important change of 8.0 T-score points. An additional calculation showed that there was an 86% chance that a change of 8.0 T-score points reflects a real change.
23
We performed a sensitivity analysis in which we added the “somewhat changed” patients to the improved group and found a minimal important change of 6.4 T-score points (95% CI 2.8–6.8). There was a 72% chance that a change of 6.4 T-score points reflects a real change.Floor/Ceiling effects
One percent of the patients had the lowest raw score, and no patients had the highest raw score at admission. No patients had the lowest raw score and 1.8% had the highest raw score at discharge.
Discussion
The PROMIS-PF-GR has sufficient test-retest reliability and construct validity, but no convincing evidence for responsiveness was found. The smallest detectable change (10.6 T-score points) was slightly larger than the minimal important change (8.0 T-score points), which implies that when a patient improves 8.0 points or more, we can be at least 86% sure that this change is not measurement error.
Other studies found sufficient reliability for other PROMIS physical function short forms in orthopaedic trauma patients,
29
orthopedic foot and ankle surgery patients30
and patients with rheumatoid arthritis.31
Only 1 of the studies determined the standard error of measurement (3.8) and smallest detectable change (10.6) and found similar results as our study.30
Construct validity was also tested for other PROMIS physical function short forms. Several PROMIS short forms (PROMIS PF SF4a,8a,10a) were compared with different legacy instruments and showed high correlation for patients in spine clinics,
32
upper extremity trauma patients,29
and older patients in the emergency department.33
We did not find sufficient evidence for responsiveness of the PROMIS-PF-GR as only 3 (instead of 4 required) of the 5 hypotheses were confirmed. The 2 hypotheses that were rejected both addressed the relation between PROMIS-PF-GR and the anchor, the GRS. A surprising high effect size of 0.7 for patients who considered themselves “not changed” and a low correlation was found between the change in PROMIS-PF-GR and the GRS. The PROMIS-PF-GR may have overestimated the real change; however, we think it is more likely that our anchor question, the GRS, was not valid. Perhaps it was difficult for these patients (considering an average age of 80 years and an average MMSE of 25) to recall their level of physical function at admission, weeks after being admitted to geriatric rehabilitation. Other studies found more support for responsiveness of different PROMIS-PF short forms. Two studies tested the responsiveness the PROMIS-PF 20a in patients with rheumatoid arthritis and found the PROMIS-PF 20a to be more responsive than the Short Form (36) Health Survey and Health Assessment Questionnaire.
34
,35
Whal et al found that change scores of the PROMIS-PF 10a were significantly associated with changes in rheumatoid arthritis disease activity.- Hays R.D.
- Spritzer K.L.
- Fries J.F.
- Krishnan E.
Responsiveness and minimally important difference for the patient-reported outcomes measurement information system (PROMIS) 20-item physical functioning short form in a prospective observational study of rheumatoid arthritis.
Ann Rheum Dis. 2015; 74: 104-107
36
High effect sizes (>0.80) were found for PROMIS-PF 8a after treatment for lower29
and upper extremity trauma.37
Finally, Stephan et al found an effect size of 0.80 in the PROMIS-PF 4a in foot and ankle patients after surgery (as expected), and correlations between change scores on PROMIS-PF 4a and the Foot Function Index disability subscale of 0.48 and with global treatment outcome of 0.45 (whereas >0.50 was expected).30
A limitation is that most of these studies used distribution-based methods instead of the preferred method of hypotheses testing to determine responsiveness.22
,38
,39
We found a minimal important change value of 8 points. Two studies determined the minimal important change of the PROMIS-PF-10a in patients with osteoarthritis and rheumatoid arthritis using multiple anchor-based methods and found a range of 1.9–2.2.
35
,- Hays R.D.
- Spritzer K.L.
- Fries J.F.
- Krishnan E.
Responsiveness and minimally important difference for the patient-reported outcomes measurement information system (PROMIS) 20-item physical functioning short form in a prospective observational study of rheumatoid arthritis.
Ann Rheum Dis. 2015; 74: 104-107
40
The minimal important change of PROMIS-PF 4a in orthopaedic foot and ankle surgery patients was estimated at 4.6 points.30
It is important to note that none of these studies reported a correlation score between the anchor and the PROMIS change score. A recent study by Devji et al showed the importance of a good correlation between the anchor and the PROM change score.41
We used an conservative approach to determine the minimal important change because patients who reported “somewhat improved” were included in the “no change” group and this could have led to an overestimation. A sensitivity analysis indeed showed a lower minimal important change of 6.4 T-score points if the “somewhat changed” patients were added to the improved group. However, using this cut-off of 6.4 points as an indicator of change increases the probability that “change” is the result of measurement error instead of real change. We found no floor or ceiling effects for the PROMIS-PF-GR at admission or discharge, in accordance with other studies.30
,34
,40
,42
Limitations
First, test-retest-reliability was tested by re-administering the PROMIS-PF-GR 3 to 5 days after the original completion. A potential disadvantage of this short follow-up period is a higher risk of recall bias. However, a longer follow-up period would result in a higher chance of changes in physical function, especially in patients who rehabilitated quickly. Second, in 4 out of the 5 responsiveness hypotheses the GRS was used. A potential problem is that this global rating relies on the correct recall of retrospective function,
43
,44
which may be problematic in an older frail population. Third, the use of only 1 physical comparison test, namely the Barthel Index, for testing construct validity and responsiveness could be considered a limitation.Conclusions and Implications
The PROMIS-PG-GR showed sufficient construct validity and test-retest reliability and that important changes can be distinguished from measurement error with a high certainty. We did not find sufficient evidence for responsiveness. This may, however, be due to lack of validity of the GRS that was used as an anchor.
Clinical Implication and Future Studies
Despite some uncertainty about the responsiveness of the PROMIS-PF-GR, we consider the results of this study solid enough to recommend its application in geriatric rehabilitation to assess progress of self-reported physical function during geriatric rehabilitation. Further research should focus on responsiveness and on how PROMIS-PF-GR can be used to promote shared decision making. In addition, future research might evaluate whether the PROMIS-PF-GR could be shortened without compromising its content validity. This could be done either with an updated short form with less items or by the use of computer adaptive testing (CAT). Theoretically, PROMIS-PF-GR gives insight in the patient's perspective of rehabilitation progress, and this could be a starting point for shared decision making and goal setting in geriatric rehabilitation. PROMIS-PF-GR was developed and validated for use in geriatric rehabilitation, however, it may also be useful in other older or geriatric patients in settings outside of geriatric rehabilitation. Further research on the psychometric properties in of the PROMIS-PF-GR in other patient populations is recommended.
Appendix
Supplementary Table 1PROMIS v 2.0 PROMIS-PF-GR
†
Terms and conditions of use can be found on https://www.healthmeasures.net/2-uncategorised/285-agree-to-the-healthmeasures-terms-of-use. The full version PROMIS-PF-GR can be requested from C.B. Terwee ([email protected]).
| Item | Content |
|---|---|
| PFA5 | Does your health now limit you in lifting or carrying groceries? |
| PFA17 | Are you able to reach into a high cupboard? |
| PFA23 | Are you able to go for a walk of at least 15 minutes? |
| PFA30 | Are you able to step up and down curbs? |
| PFA37 | Are you able to stand for short periods of time? |
| PFA45 | Are you able to get out of bed into a chair? |
| PFA47 | Are you able to pull on trousers? |
| PFA52 | Are you able to tie your shoelaces? |
| PFA53 | Are you able to run errands and shop? |
| PFA54 | Are you able to button your shirt? |
| PFB3 | Does your health now limit you in putting a trash bag outside? |
| PFB11 | Are you able to wash dishes, pots, and utensils by hand while standing at a sink? |
| PFB17 | Are you able to put on and take off your socks? |
| PFB30 | Are you able to open a new milk carton? |
| PFB36 | Are you able to put on a pullover sweater? |
| PFB48 | Does your health now limit you in taking a shower? |
| PFC6r1 | Are you able to walk a block (about 100 m) on flat ground? |
| PFC37 | Does your health now limit you in climbing one flight of stairs? |
| PFC39 | Are you able to stand without losing your balance for several minutes? |
| PFC45r1 | Are you able to sit on and get up from the toilet? |
| PFC47 | Are you able to be out of bed most of the day? |
| PFC51 | Are you able to wipe yourself after using the toilet? |
| PFC52 | Are you able to turn from side to side in bed? |
| PFC53 | Are you able to get in and out of bed? |
∗ The items of PROMIS-PF-GR were originally selected from the PROMIS item bank PF v 1.2, however, these items are also included in the PROMIS PF bank v 2.0. For this reason we designate PROMIS-PF-GR also as v 2.0.
† Terms and conditions of use can be found on https://www.healthmeasures.net/2-uncategorised/285-agree-to-the-healthmeasures-terms-of-use. The full version PROMIS-PF-GR can be requested from C.B. Terwee ([email protected]).
References
- Inpatient rehabilitation specifically designed for geriatric patients: Systematic review and meta-analysis of randomised controlled trials.BMJ. 2010; 340: c1718
- Geriatric rehabilitation: State of the art and future priorities.J Am Med Dir Assoc. 2019; 20: 396-398
- Clinical outcome assessments: Conceptual foundation-Report of the ISPOR clinical outcomes assessment: Emerging Good Practices for Outcomes Research Task Force.Value Health. 2015; 18: 741-752
- Montreal accord on patient-reported Outcomes (PROs) use series-paper 4: Patient-reported outcomes can inform clinical decision making in chronic care.J Clin Epidemiol. 2017; 89: 136-141
- How do patient reported outcome measures (PROMs) support clinician-patient communication and patient care? A realist synthesis.J Patient Rep Outcomes. 2018; 2: 42
- The impact of measuring patient-reported outcomes in clinical practice: A systematic review of the literature.Qual Life Res. 2008; 17: 179-193
- Patient reported outcome measures could help transform healthcare.BMJ. 2013; 346: f167
- PROactive consortium. A comprehensive systematic review of the development process of 104 patient-reported outcomes (PROs) for physical activity in chronically ill and elderly people.Health Qual Life Outcomes. 2011; 9: 116
- Development of a Patient-Reported Outcomes Measurement Information System (PROMIS) short form for measuring physical function in geriatric rehabilitation patients.Qual Life Res. 2020; 29: 2563-2572
- The Patient-Reported Outcomes Measurement Information System (PROMIS): Progress of an NIH roadmap cooperative group during its first two years.Med Care. 2007; 45: S3-S11
- The future of outcomes measurement: Item banking, tailored short-forms, and computerized adaptive assessment.Qual Life Res. 2007; 16: 133-141
- The Patient-Reported Outcomes Measurement Information System (PROMIS) developed and tested its first wave of adult self-reported health outcome item banks: 2005–2008.J Clin Epidemiol. 2010; 63: 1179-1194
- A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation.J Chronic Dis. 1987; 40: 373-383
- Updating and validating the Charlson comorbidity index and score for risk adjustment in hospital discharge abstracts using data from 6 countries.Am J Epidemiol. 2011; 173: 676-682
- “Mini-mental State” A practical method for grading the cognitive state of patients for the clinician.J Psychiatr Res. 1975; 12: 189-198
- Cognitive tests to detect dementia: A systematic review and meta-analysis.JAMA Intern Med. 2015; 175: 1450-1458
- Functional evaluation: The Barthel index.Md State Med J. 1965; 14: 61-65
- Measurement properties of the Barthel Index in geriatric rehabilitation.J Am Med Dir Assoc. 2019; 20: 420-425
- Assessment Center.(Available at:)http://www.assessmentcenter.netDate accessed: February 26, 2021
- Health measures.(Available at:)http://www.healthmeasures.net/score-and-interpret/interpret-scores/promis/promis-score-cut-pointsDate accessed: February 26, 2021
- Establishing clinically-relevant terms and severity thresholds for Patient-Reported Outcomes Measurement Information System (PROMIS) measures of physical function, cognitive function, and sleep disturbance in people with cancer using standard setting.Qual of Life Res. 2019; 28: 3355-3362
- The COSMIN study reached international consensus on taxonomy, terminology, and definitions of measurement properties for health-related patient-reported outcomes.J Clin Epidemiol. 2010; 63: 737-745
- Measurement in Medicine.Cambridge University Press, New York2011
- Effect sizes for interpreting change in health status.Med Care. 1989; 27: 178-189
- Assessing the responsiveness of functional scales to clinical change: An analogy to diagnostic test performance.J Chronic Dis. 1986; 39: 897-906
- Minimal important change (MIC) based on a predictive modeling approach was more precise than MIC based on ROC analysis.J Clin Epidemiol. 2015; 68: 1388-1396
- The anchor-based minimal important change, based on receiver operating characteristic analysis or predictive modeling, may need to be adjusted for the proportion of improved patients.J Clin Epidemiol. 2017; 83: 90-100
- Quality criteria were proposed for measurement properties of health status questionnaires.J Clin Epidemiol. 2007; 60: 34-42
- Validation of PROMIS Physical Function Instruments in patients with an orthopaedic trauma to a lower extremity.J Orthop Trauma. 2019; 33: 377-383
- Measurement properties of PROMIS short forms for pain and function in orthopedic foot and ankle surgery patients.Qual Life Res. 2019; 28: 2821-2829
- Reliability and validity of Patient-Reported Outcomes Measurement Information System short forms in ankylosing spondylitis.J Rheumatol. 2020; 47: 1182-1188
- Correlation between the Oswestry disability index and the 4-item short forms for physical function and pain interference from PROMIS.J Neurosurg Spine. 2019; 9: 1-6
- PROMIS physical function 10-item short form for older adults in an emergency setting.J Gerontol A Biol Sci Med Sci. 2020; 75: 1418-1423
- Improved responsiveness and reduced sample size requirements of PROMIS physical function scales with item response theory.Arthritis Res Ther. 2011; 13: R147
- Responsiveness and minimally important difference for the patient-reported outcomes measurement information system (PROMIS) 20-item physical functioning short form in a prospective observational study of rheumatoid arthritis.Ann Rheum Dis. 2015; 74: 104-107
- Validity and responsiveness of a 10-Item Patient-Reported Measure of Physical Function in a rheumatoid arthritis clinic population.Art Care Res. 2017; 69: 338-346
- Longitudinal validation of the PROMIS physical function item bank in upper extremity trauma.J Orthop Trauma. 2017; 31: e321-e326
- Responsiveness of the PROMIS and FAAM instruments in foot and ankle orthopedic population.Foot Ankle Int. 2019; 40: 56-64
- COSMIN risk of bias checklist for systematic reviews of patient-reported outcome measures.Qual Life Res. 2018; 27: 1171-1179
- Responsiveness and minimally important differences for 4 Patient-Reported Outcomes Measurement Information System short forms: Physical function, pain interference, depression, and anxiety in knee osteoarthritis.J Pain. 2017; 18: 1096-1110
- Evaluating the credibility of anchor based estimates of minimal important differences for patient reported outcomes: Instrument development and reliability study.BMJ. 2020; 4369: m1714
- Patient-Reported Outcomes Measurement Information System (PROMIS) instruments among individuals with symptomatic knee osteoarthritis: A cross-sectional study of floor/ceiling effects and construct validity.BMC Musculoskelet Disord. 2015; 16: 253
- Clinimetrics Corner: The Minimal Clinically Important Change Score (MCID): A necessary pretense.J Man Manip Ther. 2008; 16: E82-E83
- Clinimetrics corner: A closer look at the minimal clinically important difference (MCID).J Man Manip Ther. 2012; 20: 160-166
Article info
Publication history
Published online: February 25, 2021
Footnotes
Informed consent was obtained from all individual participants included in the study. Decision incompetent patients (as rated by the attending physician) were excluded from the study, in such we considered all included patients to be competent to sign informed consent.
C.B. Terwee is president of the PROMIS Health Organization. L.D. Roorda and C.B. Terwee are members of the Dutch-Flemish PROMIS group. All authors have no financial or nonfinancial conflicts of interest.
This research did not receive any funding from agencies in the public, commercial, or not-for-profit sectors.
Identification
Copyright
© 2021 The Authors. Published by Elsevier Inc. on behalf of AMDA - The Society for Post-Acute and Long-Term Care Medicine.
User license
Creative Commons Attribution (CC BY 4.0) | How you can reuse 
Elsevier's open access license policy
Creative Commons Attribution (CC BY 4.0)
Permitted
- Read, print & download
- Redistribute or republish the final article
- Text & data mine
- Translate the article
- Reuse portions or extracts from the article in other works
- Sell or re-use for commercial purposes
Elsevier's open access license policy
