Optimizing the ability of the Hamilton Depression Rating Scale to discriminate across levels of severity and between antidepressants and placebos
Bibliographic record
Abstract
Efforts to improve the Hamilton Rating Scale for Depression (HRSD) have included shortening the scale by selecting the best performing items, lengthening the scale by assessing additional symptoms, modifying the format and scoring of existing items, and developing structured interview guides for administration. We defined item performance exclusively in terms of the ability of items to discriminate differences among levels of depressive severity which has not be used to guide any revisions of the HRSD conducted to date. Two techniques derived from item response theory were used to improve the ability of the HRSD to discriminate among individuals with different degrees of depressive severity. Item response curves were used to quantify the ability of items to discriminate among individual differences in depressive severity, on the basis of which the most discriminating items were selected. Maximum likelihood estimates were used to compute an optimal depressive severity score, using all items, but which weighted highly discriminating items more so than items that did not discriminate well. The utility of each method was evaluated by comparing a subset of optimally discriminating items and maximum likelihood estimates of depressive severity to the Maier Philipp subscale of the HRSD, in terms of how well scales discriminate treatment effects. Effect sizes for overall change in depression severity as well as effect sizes differentiating response to treatment versus placebo were evaluated in a sample of 491 patients receiving fluoxetine and 494 patients receiving placebo. Results of analyses identified a new subset of items (IRT-6), selected on the basis of their ability to discriminate among differences in depressive severity, that accounted for more variance in full-scale HRSD scores and was better at detecting change in illness severity than the Maier Philipp subscale of the HRSD. The IRT-6 subscale was equally good as the Maier Philipp subscale in differentiating treatment from placebo response. No evidence supporting the benefits of using maximum likelihood estimates to develop optimally performing subscales was found. Implications of the results are discussed in terms of strategies for optimizing the assessment of change in overall depression severity as well as differentiating treatment response.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".