We evaluated construct validity, responsiveness, and utility of change indicators of the Dutch-Flemish PROMIS adult v1.0 item banks for Depression and Anxiety administered as computerized adaptive... Show moreWe evaluated construct validity, responsiveness, and utility of change indicators of the Dutch-Flemish PROMIS adult v1.0 item banks for Depression and Anxiety administered as computerized adaptive test (CAT). Specifically, the CATs were compared to the Brief Symptom Inventory (BSI) using pre- and re-test data of adult patients treated for common mental disorders (N = 400; median pre-to-re-test interval = 215 days). Construct validity was evaluated with Pearson's correlations and Cohen's ds; responsiveness with Pearson's correlations and pre-post effect sizes (ES); utility of change indicators with kappa coefficients and percentages of (dis)agreement. The results showed that the PROMIS CATs measure similar constructs as matching BSI scales. Under the assumption of measuring similar constructs, the CAT and BSI Depression scales were similarly responsive. For the Anxiety scales, we found a higher responsiveness for CAT (ES = 0.64) compared to the BSI (ES = 0.50). Finally, both CATs categorized the change scores of more patients as changed compared to matching BSI scales, indicating that the PROMIS CATs may be more able to detect actual change than the BSI. Based on these findings, the PROMIS CATs may be considered a modest improvement over matching BSI scales as tools for reviewing treatment progress with patients. We discuss several additional differences between the PROMIS CATs and the BSI to help test users choose instruments. These differences include the adopted measurement theory (Item Response Theory vs. Classical Test Theory), the mode of administration (CAT vs. fixed items), and the area of application (universal vs. predominantly clinical). (PsycInfo Database Record (c) 2022 APA, all rights reserved). Show less
Flens, G.; Terwee, C.B.; Smits, N.; Williams, G.; Spinhoven, P.; Roorda, L.D.; Beurs, E. de 2021
Public Significance Statement This study suggests that PROMIS CATs for Depression and Anxiety measure the same constructs as matching BSI subscales in Dutch adult patients treated for common mental... Show morePublic Significance Statement This study suggests that PROMIS CATs for Depression and Anxiety measure the same constructs as matching BSI subscales in Dutch adult patients treated for common mental disorders, are at least as able to detect change over time, and categorize the change scores of more patients as actually changed. Based on these findings, the PROMIS CATs may be considered a modest improvement as tools for reviewing treatment progress with patients.We evaluated construct validity, responsiveness, and utility of change indicators of the Dutch-Flemish PROMIS adult v1.0 item banks for Depression and Anxiety administered as computerized adaptive test (CAT). Specifically, the CATs were compared to the Brief Symptom Inventory (BSI) using pre- and re-test data of adult patients treated for common mental disorders (N = 400; median pre-to-re-test interval = 215 days). Construct validity was evaluated with Pearson's correlations and Cohen's ds; responsiveness with Pearson's correlations and pre-post effect sizes (ES); utility of change indicators with kappa coefficients and percentages of (dis)agreement. The results showed that the PROMIS CATs measure similar constructs as matching BSI scales. Under the assumption of measuring similar constructs, the CAT and BSI Depression scales were similarly responsive. For the Anxiety scales, we found a higher responsiveness for CAT (ES = 0.64) compared to the BSI (ES = 0.50). Finally, both CATs categorized the change scores of more patients as changed compared to matching BSI scales, indicating that the PROMIS CATs may be more able to detect actual change than the BSI. Based on these findings, the PROMIS CATs may be considered a modest improvement over matching BSI scales as tools for reviewing treatment progress with patients. We discuss several additional differences between the PROMIS CATs and the BSI to help test users choose instruments. These differences include the adopted measurement theory (Item Response Theory vs. Classical Test Theory), the mode of administration (CAT vs. fixed items), and the area of application (universal vs. predominantly clinical). Show less
Flens, G.; Smits, N.; Terwee, C.B.l.; Pijck, L.; Spinhoven, P.; Beurs, E. de 2021
We investigated longitudinal measurement invariance in the Dutch-Flemish PROMIS adult v1.0 item banks for Depression and Anxiety using two clinical samples with mood and anxiety disorders (n = 640... Show moreWe investigated longitudinal measurement invariance in the Dutch-Flemish PROMIS adult v1.0 item banks for Depression and Anxiety using two clinical samples with mood and anxiety disorders (n = 640 and n = 528, respectively). Factor analysis was used to evaluate whether the item banks were sufficiently unidimensional at two test-occasions and whether the measured constructs remained the same over time. The results indicated that the item banks were sufficiently unidimensional, but the thresholds and residual variances of the constructs changed over time. However, using tentative rules of thumb, these invariance violations did not substantially affect the endorsement of a specific response category of a specific item at a specific test-occasion. Furthermore, the impact on the mean latent change scores of the item banks remained below the proposed cutoff value for substantial bias. These findings suggest that the invariance violations lacked practical significance for test-users, meaning that the item banks provide sufficiently invariant latent factor scores for use in clinical practice. Show less
Flens, G.; Smits, N.; Terwee, C.B.; Pijck, L.; Spinhoven, P.; Beurs, E. de 2019
We investigated longitudinal measurement invariance in the Dutch-Flemish PROMIS adult v1.0 item banks for Depression and Anxiety using two clinical samples with mood and anxiety disorders (n = 640... Show moreWe investigated longitudinal measurement invariance in the Dutch-Flemish PROMIS adult v1.0 item banks for Depression and Anxiety using two clinical samples with mood and anxiety disorders (n = 640 and n = 528, respectively). Factor analysis was used to evaluate whether the item banks were sufficiently unidimensional at two test-occasions and whether the measured constructs remained the same over time. The results indicated that the item banks were sufficiently unidimensional, but the thresholds and residual variances of the constructs changed over time. However, using tentative rules of thumb, these invariance violations did not substantially affect the endorsement of a specific response category of a specific item at a specific test-occasion. Furthermore, the impact on the mean latent change scores of the item banks remained below the proposed cutoff value for substantial bias. These findings suggest that the invariance violations lacked practical significance for test-users, meaning that the item banks provide sufficiently invariant latent factor scores for use in clinical practice. Show less
Flens, G.; Smits, N.; Terwee, C.B.; Dekker, J.; Huijbrechts, I.; Spinhoven, P.; Beurs, E. de 2019
We used the Dutch-Flemish version of the USA PROMIS adult V1.0 item bank for Anxiety as input for developing a computerized adaptive test (CAT) to measure the entire latent anxiety continuum. First... Show moreWe used the Dutch-Flemish version of the USA PROMIS adult V1.0 item bank for Anxiety as input for developing a computerized adaptive test (CAT) to measure the entire latent anxiety continuum. First, psychometric analysis of a combined clinical and general population sample (N = 2,010) showed that the 29-item bank has psychometric properties that are required for a CAT administration. Second, a post hoc CAT simulation showed efficient and highly precise measurement, with an average number of 8.64 items for the clinical sample, and 9.48 items for the general population sample. Furthermore, the accuracy of our CAT version was highly similar to that of the full item bank administration, both in final score estimates and in distinguishing clinical subjects from persons without a mental health disorder. We discuss the future directions and limitations of CAT development with the Dutch-Flemish version of the PROMIS Anxiety item bank. Show less
Flens, G.; Smits, N.; Terwee, C.B.; Dekker, J.; Huijbrechts, I.; Spinhoven, P.; Beurs, E. de 2019
Pseudomonas baetica strain a390T is the type strain of this recently described species and here we present its high-contiguity draft genome. To celebrate the 16th International Conference on... Show morePseudomonas baetica strain a390T is the type strain of this recently described species and here we present its high-contiguity draft genome. To celebrate the 16th International Conference on Pseudomonas, the genome of P. baetica strain a390T was sequenced using a unique combination of Ion Torrent semiconductor and Oxford Nanopore methods as part of a collaborative community-led project. The use of high-quality Ion Torrent sequences with long Nanopore reads gave rapid, high-contiguity and -quality, 16-contig genome sequence. Whole genome phylogenetic analysis places P. baetica within the P. koreensis Glade of the P. fluorescens group. Comparison of the main genomic features of P. baetica with a variety of other Pseudomonas spp. suggests that it is a highly adaptable organism, typical of the genus. This strain was originally isolated from the liver of a diseased wedge sole fish, and genotypic and phenotypic analyses show that it is tolerant to osmotic stress and to oxytetracycline. Show less
Fokkema, M.; Smits, N.; Zeileis, A. Hothorn T.; Kelderman, H. 2018
Identification of subgroups of patients for whom treatment A is more effective than treatment B, and vice versa, is of key importance to the development of personalized medicine. Tree-based... Show moreIdentification of subgroups of patients for whom treatment A is more effective than treatment B, and vice versa, is of key importance to the development of personalized medicine. Tree-based algorithms are helpful tools for the detection of such interactions, but none of the available algorithms allow for taking into account clustered or nested dataset structures, which are particularly common in psychological research. Therefore, we propose the generalized linear mixed-effects model tree (GLMM tree) algorithm, which allows for the detection of treatment-subgroup interactions, while accounting for the clustered structure of a dataset. The algorithm uses model-based recursive partitioning to detect treatment-subgroup interactions, and a GLMM to estimate the random-effects parameters. In a simulation study, GLMM trees show higher accuracy in recovering treatment-subgroup interactions, higher predictive accuracy, and lower type II error rates than linear-model-based recursive partitioning and mixed-effects regression trees. Also, GLMM trees show somewhat higher predictive accuracy than linear mixed-effects models with pre-specified interaction effects, on average. We illustrate the application of GLMM trees on an individual patient-level data meta-analysis on treatments for depression. We conclude that GLMM trees are a promising exploratory tool for the detection of treatment-subgroup interactions in clustered datasets. Show less
Fokkema, M.; Smits, N.; Zeileis, A.; Hothorn, T.; Kelderman, H. 2017
Identification of subgroups of patients for whom treatment A is more effective than treatment B, and vice versa, is of key importance to the development of personalized medicine. Tree-based... Show moreIdentification of subgroups of patients for whom treatment A is more effective than treatment B, and vice versa, is of key importance to the development of personalized medicine. Tree-based algorithms are helpful tools for the detection of such interactions, but none of the available algorithms allow for taking into account clustered or nested dataset structures, which are particularly common in psychological research. Therefore, we propose the generalized linear mixed-effects model tree (GLMM tree) algorithm, which allows for the detection of treatment-subgroup interactions, while accounting for the clustered structure of a dataset. The algorithm uses model-based recursive partitioning to detect treatment-subgroup interactions, and a GLMM to estimate the random-effects parameters. In a simulation study, GLMM trees show higher accuracy in recovering treatment-subgroup interactions, higher predictive accuracy, and lower type II error rates than linear-model-based recursive partitioning and mixed-effects regression trees. Also, GLMM trees show somewhat higher predictive accuracy than linear mixed-effects models with pre-specified interaction effects, on average. We illustrate the application of GLMM trees on an individual patient-level data meta-analysis on treatments for depression. We conclude that GLMM trees are a promising exploratory tool for the detection of treatment-subgroup interactions in clustered datasets. Show less
Flens, G.; Smits, N.; Terwee, C.B.; Dekker, J.; Huijbrechts, I.; Beurs, E. de 2017
In clinical assessment, efficient screeners are needed to ensure low respondent burden. In thisarticle, Stochastic Curtailment (SC), a method for efficient computerized testing for... Show moreIn clinical assessment, efficient screeners are needed to ensure low respondent burden. In thisarticle, Stochastic Curtailment (SC), a method for efficient computerized testing for classificationinto two classes for observable outcomes, was extended to three classes. In a post hocsimulation study using the item scores on the Center for Epidemiologic Studies–DepressionScale (CES-D) of a large sample, three versions of SC, SC via Empirical Proportions (SC-EP),SC via Simple Ordinal Regression (SC-SOR), and SC via Multiple Ordinal Regression (SC-MOR)were compared at both respondent burden and classification accuracy. All methods wereapplied under the regular item order of the CES-D and under an ordering that was optimal interms of the predictive power of the items. Under the regular item ordering, the three methodswere equally accurate, but SC-SOR and SC-MOR needed less items. Under the optimalordering, additional gains in efficiency were found, but SC-MOR suffered from capitalization onchance substantially. It was concluded that SC-SOR is an efficient and accurate method for clinicalscreening. Strengths and weaknesses of the methods are discussed. Show less
Flens, G.; Smits, N.; Carlier, I.; Hemert, A.M. van; Beurs, E. de 2015