CoChat
Sign Up
Cited by 50Treatment of Major DepressionMental Health Treatment and AccessElectroconvulsive Therapy Studies

Finding treatment-resistant depression in real-world data: How a data-driven approach compares with expert-based heuristics

Open access
Cepeda, M. Soledad, Reps, Jenna, Fife, Daniel, Blacketer, Clair, Stang, Paul, Ryan, Patrick·2018·Depression and Anxiety
DOI 10.1002/da.22705PMC PMC5873404
Open this paper in CoChat — free accountOpen PDF

Abstract

BACKGROUND: Depression that does not respond to antidepressants is treatment-resistant depression (TRD). TRD definitions include assessments of treatment response, dose and duration, and implementing these definitions in claims databases can be challenging. We built a data-driven TRD definition and evaluated its performance. METHODS: We included adults with depression, ≥1 antidepressant, and no diagnosis of mania, dementia, or psychosis. Subjects were stratified into those with and without proxy for TRD. Proxies for TRD were electroconvulsive therapy, deep brain, or vagus nerve stimulation. The index date for subjects with proxy for TRD was the procedure date, and for subjects without, the date of a randomly selected visit. We used three databases. We fit decision tree predictive models. We included number of distinct antidepressants, with and without adequate doses and duration, number of antipsychotics and psychotherapies, and expert-based definitions, 3, 6, and 12 months before index date. To assess performance, we calculated area under the curve (AUC) and transportability. RESULTS: We analyzed 33,336 subjects with no proxy for TRD, and 3,566 with the proxy. Number of antidepressants and antipsychotics were selected in all periods. The best model was at 12 months with an AUC = 0.81. The rule transported well and states that a subject with ≥1 antipsychotic or ≥3 antidepressants in the last year has TRD. Applying this rule, 15.8% of subjects treated for depression had TRD. CONCLUSION: The definition that best discriminates between subjects with and without TRD considers number of distinct antidepressants (≥3) or antipsychotics (≥1) in the last year.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1002/da.22705

Cites

27 of 27 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • The algorithmically-derived decision-tree rule in Cepeda et al. (2018) performed more accurately in internal validation than any of the five decision rules defined by expert psychiatrists (F-1 = 0.44 compared to F-1′s = 0.39-0.42) and held up in external validation samples totaling 14 845 patients from alternate insurance databases (AUCs = 0.78-0.79).

    “The authors found that their algorithmically-derived decision-tree rule performed more accurately in internal validation than any of the five decision rules defined by expert psychiatrists (F-1 = 0.44 compared to F-1′s = 0.39–0.42) and held up in external validation samples totaling 14 845 patients from alternate insurance databases (AUCs = 0.78–0.79).”

    From Irremediability in psychiatric euthanasia: examining the objective standard (2023) · cites this paper as Cepeda et al., 2018

    Split into 2 atoms — 1 confirmed, 1 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

    Experimental — accuracy not yet measured

  • One study used insurance claims of 22 057 patients to predict which patients would receive neuromodulation after trying an antidepressant in the past year.

    “One study used insurance claims of 22 057 patients to predict which patients would receive neuromodulation after trying an antidepressant in the past year (Cepeda et al., 2018).”

    From Irremediability in psychiatric euthanasia: examining the objective standard (2023) · cites this paper as Cepeda et al., 2018

    The Results table lists 2,076 patients with the treatment-resistant-depression proxy and 19,981 without it in the CCAE claims database, totaling 22,057. The Target population and Analysis sections confirm a prior-year antidepressant requirement and a predictive model, but 22,057 was one database cohort, not the study-wide sample. The model distinguished patients with a neuromodulation procedure code near the index visit from those without one; it did not predict future receipt of neuromodulation. (LLM, conf=0.87)

    Experimental — accuracy not yet measured

  • A third set of two studies used patient records to predict treatment-resistance.

    “A third set of two studies used patient records to predict treatment-resistance (Cepeda et al., 2018; Perlis et al., 2012).”

    From Irremediability in psychiatric euthanasia: examining the objective standard (2023) · cites this paper as Cepeda et al., 2018

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

Flagged first, then by citations

2 of the 29 references couldn’t be matched to a known paper.

Research works printed but not linked: 1

  1. Powers, D. M. W. (2011). Evaluation: From precision, recall and f‐measure to ROC, informedness, markedness& correlation. Journal of Machine Learning Technologies, 2(1), 37.

Cited by

Sorted by citations

Related

Authors
Cepeda, M. S., Reps, J., Fife, D., Blacketer, C., Stang, P., & Ryan, P.
Affiliations
Janssen Research and Development; Titusville NJ USA
Corresponding author
Cepeda, M. Soledad
ORCID
Venue
Depression and Anxiety
Publisher
Wiley
Published
2017-12-15
Volume
35
Issue
3
Pages
220-228
ISSN
1091-4269 (ISSN-L)
DOI
10.1002/da.22705
PMC
PMC5873404
OpenAlex
W2771774637
Cited by
50
References
29
Citations by year
2026: 3 · 2025: 5 · 2024: 5 · 2023: 3 · 2022: 7 · 2021: 8 · 2020: 5 · 2019: 8 · 2018: 5 · 2017: 1
FWCI
3.63
Type
journal article
Language
en
Open access
Licence
Locations
  • Depression and Anxiety (publishedVersion, cc-by-nc)
  • PubMed Central (publishedVersion, cc-by-nc)
  • PubMed (publishedVersion)
Topics
  • Health Sciences › Medicine › Pharmacology › Treatment of Major Depression
  • Social Sciences › Psychology › Social Psychology › Mental Health Treatment and Access
  • Health Sciences › Medicine › Psychiatry and Mental health › Electroconvulsive Therapy Studies
Keywords
treatment-resistant depression, antidepressant non-response, real-world data, claims data, decision tree, antipsychotic drugs, transportability, electroconvulsive therapy, deep brain stimulation, vagus nerve stimulation
MeSH terms
Heuristics, Adult, Aged, Antidepressive Agents, Female, Humans, Male, Middle Aged, Antipsychotic Agents, Depressive Disorder, Treatment-Resistant

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review