Psychological Treatment Symptom Data for 6,464 Patients Over 12 Weeks

6,464 patients undergoing a 12-week psychological treatment program provided self-rated symptom data from three questionnaires over the first 4 weeks. The data was used to compare the predictive performance of traditional sum scores versus psychometrically optimized latent trait scores for post-treatment outcomes. Nils Hentati Isacsson authored this study, which was last updated on March 18, 2026.

Use Cases

• Comparing predictive model performance based on latent trait scores versus traditional sum scores.
• Evaluating the impact of questionnaire length reduction on prediction accuracy for treatment outcomes.
• Testing machine learning models like linear regression, Bayesian ridge regression, and random forest on clinical symptom data.
• Studying the application of Rasch Measurement Theory to improve measurement in predictive research.

Strengths

• Data from 6,464 patients provides a substantial sample size for analysis.
• Includes a direct comparison of two measurement approaches (sum scores vs. latent trait scores) on the same patient cohort.
• Uses multiple imputation for missing data and nested cross-validation for model tuning, suggesting a methodologically sound approach.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The dataset is very small at 19.5 KB, indicating it likely contains summary or analysis results rather than the raw patient-level data.

Provenance

Source

figshare

Collection Method

Self-rated symptom data collected from patients during the first 4 weeks of a psychological treatment program.

Time Range

Data collection covered the first 4 weeks of a 12-week treatment program; specific years are not provided.

Freshness

Last updated 2026-03 07:45:58.

Geography

Spatial coverage is not specified in the provided description.