Using Item Response Theory to Identify Responders to Treatment: Examples with the Patient-Reported Outcomes Measurement Information System (PROMIS®) Physical Function Scale and Emotional Distress Composite.
Ontology highlight
ABSTRACT:
SUBMITTER: Hays RD
PROVIDER: S-EPMC8437927 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:BACKGROUND:Psychometric theory offers a range of tests that can be used as supportive evidence of both validity and reliability of instruments aimed at measuring patient-reported outcomes (PRO). The aim of this paper is to illustrate psychometric tests within the Classical Test Theory (CTT) framework, comprising indices that are frequently applied to assess item- and scale-level psychometric properties of PRO instruments. METHODS:Using data on the PROMIS Depression Item Bank, typical CTT indices for the assessment of psychometric properties are illustrated, including content validity, item-level data exploration, reliability, and construct validity, particularly confirmatory factor analysis, to test the unidimensionality assumption underlying the item bank. Analyses are carried out on an original item set of 51 depression items, the final (official) PROMIS Depression Item Bank consisting of 28 items, and an 8-item short form. RESULTS:The analyses reported provide an informative illustration on how item- and scale-level reliability and validity statistics can be used to assess the psychometric quality of a PRO instrument. The results illustrate how the reported statistics can be used for item selection from an item pool (here: 51 items). Both the (final) 28-item bank and the 8-item short form show good psychometric properties supporting the high quality of individual items and the unidimensionality assumption of the item bank. CONCLUSIONS:It is our hope that our illustration of CTT methods, in conjunction with two companion papers illustrating modern test theory methods, will help researchers to confidently apply a range of statistical tests to evaluate item- and scale-level psychometric performance of PRO instruments.
| S-EPMC6663945 | biostudies-other
Project description:BACKGROUND:The Patient-Reported Outcomes Measurement Information System® (PROMIS) includes a Physical Function (PF) item bank and an Upper Extremity (UE) item bank, which is composed of a subset of items from the PF bank. The UE item bank has few items and known ceiling effects. Therefore, this study aimed to expand the item bank to assess a wider range of functioning. With the additional content, other psychometric properties-improved content validity, item bank depth, range of measurement, and score reliability-were also evaluated. We convened an expert panel to review potential items, and then conducted psychometric analyses on both extant and newly-collected data. RESULTS:Expert focus groups reviewed the PF item bank for items that were "sufficiently" related to upper extremity functioning for inclusion in the expanded UE item bank. The candidate item bank was quantitatively evaluated in a new sample of 600 people. The final items were calibrated in an aggregated dataset (n?=?11,635) from two existing datasets, and the newly collected sample. The original UE item bank included 15 items. After expert review and quantitative evaluation, 31 items were added. The combined 46 items were calibrated using item response theory (IRT). Then computer adaptive tests (CATs) were simulated based off of the psychometric results. These indicated that the new UE item bank has an extended measurement range compared to the original version. CONCLUSIONS:The expanded PROMIS UE item bank assesses a wider range of upper extremity functioning compared to the initial UE item bank. However, ceiling effects remain a concern for unimpaired groups. The new UE item bank is recommended for individuals with known or suspected upper extremity limitations.
| S-EPMC6879697 | biostudies-literature
Project description:Composite assessments aim to combine different aspects of a disease in a single score and are utilized in a variety of therapeutic areas. The data arising from these evaluations are inherently discrete with distinct statistical properties. This tutorial presents the framework of the item response theory (IRT) for the analysis of this data type in a pharmacometric context. The article considers both conceptual (terms and assumptions) and practical questions (modeling software, data requirements, and model building).
| S-EPMC5915608 | biostudies-literature
Project description:PurposeThis study aimed to validate the PROMIS Pediatric item bank v2.0 Peer Relationships and compare reliability of the full item bank to its short form, computerized adaptive test (CAT) and the social functioning (SF) subscale of the Pediatric Quality of Life Inventory (PedsQL™).MethodsChildren aged 8-18 (n = 1327), representative of the Dutch population completed the Peer Relationships item bank. A graded response model (GRM) was fit to the data. Structural validity was assessed by checking item-fit statistics (S-X2, p < 0.001 = misfit). For construct validity, a moderately strong correlation (> 0.50) was expected between Peer Relationships and the PedsQL SF subscale. Cross-cultural DIF between U.S. and NL was assessed using logistic regression, where an item with McFadden's pseudo R2 > 0.02 was considered to have DIF. Percentage of participants reliably measured was assessed using the standard error of measurement (SEM) < 0.32 as a criterion (reliability of 0.90). Relative efficiency ((1-SEM2)/nitems) was calculated to compare how well the instruments performed relative to the amount of items administered.ResultsIn total, 527 (response rate: 39.7%) children completed the PROMIS v2.0 Peer Relationships item bank (nitems = 15) and the PedsQL™ (nitems = 23). Structural validity of the Peer Relationships item bank was sufficient, but one item displayed misfit in the GRM model (S-X2 < 0.001); 5152R1r ("I played alone and kept to myself"). The item 733R1r ("I was a good friend") was the only item that displayed cross-cultural DIF (R2 = 0.0253). The item bank correlated moderately high (r = 0.61) with the PedsQL SF subscale Reliable measurements were obtained at the population mean and > 2SD in the clinically relevant direction. CAT outperformed all other measures in efficiency. Mean T-score of the Dutch general population was 46.9(SD 9.5).ConclusionThe pediatric PROMIS Peer Relationships item bank was successfully validated for use within the Dutch population and reference data are now available.
| S-EPMC8233291 | biostudies-literature
Project description:Two item banks for substance use were developed as part of the Patient-Reported Outcomes Measurement Information System (PROMIS(®)): severity of substance use and positive appeal of substance use.Qualitative item analysis (including focus groups, cognitive interviewing, expert review, and item revision) reduced an initial pool of more than 5300 items for substance use to 119 items included in field testing. Items were written in a first-person, past-tense format, with 5 response options reflecting frequency or severity. Both 30-day and 3-month time frames were tested. The calibration sample of 1336 respondents included 875 individuals from the general population (ascertained through an internet panel) and 461 patients from addiction treatment centers participating in the National Drug Abuse Treatment Clinical Trials Network.Final banks of 37 and 18 items were calibrated for severity of substance use and positive appeal of substance use, respectively, using the two-parameter graded response model from item response theory (IRT). Initial calibrations were similar for the 30-day and 3-month time frames, and final calibrations used data combined across the time frames, making the items applicable with either interval. Seven-item static short forms were also developed from each item bank.Test information curves showed that the PROMIS item banks provided substantial information in a broad range of severity, making them suitable for treatment, observational, and epidemiological research in both clinical and community settings.
| S-EPMC4633351 | biostudies-literature
Project description:This is the first study of the measurement equivalence of the Patient Reported Outcomes Measurement Information System® (PROMIS®) Anxiety short forms in a large ethnically diverse sample. The psychometric properties and differential item functioning (DIF) were examined across different racial/ethnic, educational, age, gender and language groups. METHODS:These data are from individuals selected from cancer registries in the United States. For the analyses of race/ethnicity the reference group was non-Hispanic Whites (n = 2,263), the studied groups were non-Hispanic Blacks (n = 1,117), Hispanics (n = 1,043) and Asians/Pacific Islanders (n = 907). Within the Hispanic subsample, there were 335 interviews conducted in Spanish and 703 in English. The 11 anxiety items were from the PROMIS emotional disturbance item bank. DIF hypotheses were generated by content experts who rated whether or not they expected DIF to be present, and the direction of the DIF with respect to several comparison groups. The primary method used for DIF detection was the Wald test for examination of group differences in item response theory (IRT) item parameters accompanied by magnitude measures. Expected item scores were examined as measures of magnitude. The method used for quantification of the difference in the average expected item scores was the non-compensatory DIF (NCDIF) index. DIF impact was examined using expected scale score functions. Additionally, precision and reliabilities were examined using several methods. RESULTS:Although not hypothesized to show DIF for Asians/Pacific Islanders, every item evidenced DIF by at least one method. Two items showed DIF of higher magnitude for Asians/Pacific Islanders vs. Whites: "Many situations made me worry" and "I felt anxious". However, the magnitude of DIF was small and the NCDIF statistics were not above threshold. The impact of DIF was negligible. For education, six items were identified with consistent DIF across methods: fearful, anxious, worried, hard to focus, uneasy and tense. However, the NCDIF was not above threshold and the impact of DIF on the scale was trivial. No items showed high magnitude DIF for gender. Two items showed slightly higher magnitude for age (although not above the cutoff): worried and fearful. The scale level impact was trivial. Only one item showed DIF with the Wald test after the Bonferroni correction for the language comparisons: "I felt fearful". Two additional items were flagged in sensitivity analyses after Bonferroni correction, anxious and many situations made me worry. The latter item also showed DIF of higher magnitude, with an NCDIF value (0.144) above threshold. Individual impact was relatively small. CONCLUSIONS:Although many items from the PROMIS short form anxiety measures were flagged with DIF, item level magnitude was low and scale level DIF impact was minimal; however, three items: anxious, worried and many situations made me worry might be singled out for further study. It is concluded that the PROMIS Anxiety short form evidenced good psychometric properties, was relatively invariant across the groups studied, and performed well among ethnically diverse subgroups of Blacks, Hispanic, White non-Hispanic and Asians/Pacific Islanders. In general more research with the Asians/Pacific Islanders group is needed. Further study of subgroups within these broad categories is recommended.
| S-EPMC5479355 | biostudies-literature
Project description:PURPOSE:Health status descriptive systems based on item response theory (IRT), such as the Patient-Reported Outcomes Measurement Information System (PROMIS®), have item banks to measure domains of health. We developed a method to present such banks for health-state valuation. METHODS:We evaluated four different presentation approaches: a single item (1S), 2 items presented separately (2S), 2 items presented together (2T), or 5 items presented together (5T). We evaluated these four approaches in three PROMIS item banks (depression, physical function, and sleep disturbance). Adult community members valued health-state descriptions using the visual analog scale and standard gamble methods. We compared the approaches by the range of item bank theta scores captured, participants' assessments of difficulty (1 = very easy to 7 = very hard), and exit interviews. RESULTS:Participants (n = 118) ranged in age from 18 to 71; 63% were female and 54% were white. The 1S approach captured the smallest range of theta scores. A monotonic relationship between theta score and mean standard gamble estimate was found with all approaches except 2S. Across all 3 item banks, mean difficulty assessments were 2.35 (1S), 2.69 (2T), 2.78 (5T), and 2.80 (2S). In exit interviews, participants generally found all four approaches similarly meaningful and realistic. CONCLUSIONS:Creating health descriptions by presenting 2 items maximized the range of theta while minimizing difficulty and maintaining a monotonic relationship with utility estimates. We recommend this approach for valuation of IRT-based descriptive systems such as PROMIS.
| S-EPMC6611549 | biostudies-literature
Project description:BACKGROUND:The Dutch-Flemish PROMIS® Upper Extremity (DF-PROMIS-UE) V2.0 item bank was recently developed using Item Response Theory (IRT). Unknown for this bank are: (1) if it is legitimate to calculate IRT-based scores for short forms and Computerized Adaptive Tests (CATs), which requires that the items meet the assumptions of and fit the IRT-model (Graded Response Model [GRM]);(2) if it is legitimate to compare (sub) groups of patients using this measure, which requires measurement invariance; and (3) the precision of the estimated patients' scores for patients with different levels of functioning and compared to legacy measures. Aims were to evaluate (1) the assumptions of and fit to the GRM, (2) measurement invariance and (3) (comparative) precision of the DF-PROMIS-UE v2.0. METHODS:Cross-sectional data were collected in Dutch patients with upper extremity disorders. Assessed were IRT-assumptions (unidimensionality [bi-factor analysis], local independence [residual correlations], monotonicity [coefficient H]), GRM item fit, measurement invariance (absence of Differential Item Functioning [DIF] due to age, gender, center, duration, and location of complaints) and precision (standard error of IRT-based scores across levels of functioning). To study measurement invariance for language [Dutch vs. English], additional US data were used. Legacy instruments were the Disability of the Arm, Shoulder and Hand (DASH), the QuickDASH and the Michigan Hand Questionnaire (MHQ). RESULTS:In total 521 Dutch (mean age?±?SD?=?51?±?17?years, 49% female) and 246 US patients (mean age?±?SD?=?48?±?14?years, 69% female) participated. The DF-PROMIS-UE v2.0 item bank was sufficiently unidimensional (Omega-H?=?0.80, Explained Common Variance?=?0.68), had negligible local dependence (four out of 1035 correlations >?0.20), good monotonicity (H?=?0.63), good GRM fit (no misfitting items) and demonstrated sufficient measurement invariance. Precise estimates (Standard Error?<?3.2) were obtained for most patients (7-item short form, 88.5%; standard CAT, 91.3%; and, fixed 7-item CAT, 87.6%). The DASH displayed better reliability than the DF-PROMIS-UE short form and standard CAT, the QuickDASH displayed comparable reliability. The MHQ-ADL displayed better reliability than the DF-PROMIS-UE short form and standard CAT for T-scores between 28 and 50. For patients with low function, the DF-PROMIS-UE measures performed better. CONCLUSIONS:The DF-PROMIS-UE v2.0 item bank showed sufficient psychometric properties in Dutch patients with UE disorders.
| S-EPMC7077019 | biostudies-literature
Project description:PurposePrevious research indicated that the Patient-Reported Outcomes Measurement Information System (PROMIS®) item bank v2.0 'Ability to Participate in Social Roles and Activities' may miss subdomains of social participation. The purpose of this study was to generate items for these missing subdomains and to evaluate their content validity.MethodsA three-step approach was followed: (1) Item generation for 16 International Classification of Functioning Disability and Health subdomains currently not covered by the item bank; (2) Evaluation of content validity of generated items through expert review (n = 20) and think-aloud interviews with a purposeful sample of people with and without (chronic) health conditions (n = 10), to assess item comprehensibility, relevance, and comprehensiveness; and 3) Item revision based on the results of step 2, in a consensus procedure.ResultsFirst, 48 items were generated. Second, overall, content experts indicated that the generated items were relevant. Furthermore, based on experts' responses, items were simplified and 'participation in social media' was identified as an important additional subdomain of social participation. Additionally, 'participating in various social roles simultaneously' was identified as a missing item. Based on the responses of the interviewed adults items were simplified. Third, in total 17 items, covering 17 subdomains, were proposed to be added to the original item bank.DiscussionThe relevance, comprehensibility and comprehensiveness of the 17 proposed items were supported. Whether the proposed extension of the item bank leads to better psychometric properties of the item bank should be tested in a large-scale field study.
| S-EPMC7561593 | biostudies-literature
Project description:AIMS:The goals of these analyses were to examine the psychometric properties and measurement equivalence of a self-reported cognition measure, the Patient Reported Outcome Measurement Information System® (PROMIS®) Applied Cognition - General Concerns short form. These items are also found in the PROMIS Cognitive Function (version 2) item bank. This scale consists of eight items related to subjective cognitive concerns. Differential item functioning (DIF) analyses of gender, education, race, age, and (Spanish) language were performed using an ethnically diverse sample (n = 5,477) of individuals with cancer. This is the first analysis examining DIF in this item set across ethnic and racial groups. METHODS:DIF hypotheses were derived by asking content experts to indicate whether they posited DIF for each item and to specify the direction. The principal DIF analytic model was item response theory (IRT) using the graded response model for polytomous data, with accompanying Wald tests and measures of magnitude. Sensitivity analyses were conducted using ordinal logistic regression (OLR) with a latent conditioning variable. IRT-based reliability, precision and information indices were estimated. RESULTS:DIF was identified consistently only for the item, brain not working as well as usual. After correction for multiple comparisons, this item showed significant DIF for both the primary and sensitivity analyses. Black respondents and Hispanics in comparison to White non-Hispanic respondents evidenced a lower conditional probability of endorsing the item, brain not working as well as usual. The same pattern was observed for the education grouping variable: as compared to those with a graduate degree, conditioning on overall level of subjective cognitive concerns, those with less than high school education also had a lower probability of endorsing this item. DIF was also observed for age for two items after correction for multiple comparisons for both the IRT and OLR-based models: "I have had to work really hard to pay attention or I would make a mistake" and "I have had trouble shifting back and forth between different activities that require thinking". For both items, conditional on cognitive complaints, older respondents had a higher likelihood than younger respondents of endorsing the item in the cognitive complaints direction. The magnitude and impact of DIF was minimal. The scale showed high precision along much of the subjective cognitive concerns continuum; the overall IRT-based reliability estimate for the total sample was 0.88 and the estimates for subgroups ranged from 0.87 to 0.92. CONCLUSION:Little DIF of high magnitude or impact was observed in the PROMIS Applied Cognition - General Concerns short form item set. One item, "It has seemed like my brain was not working as well as usual" might be singled out for further study. However, in general the short form item set was highly reliable, informative, and invariant across differing race/ethnic, educational, age, gender, and language groups.
| S-EPMC5433382 | biostudies-literature