CoChat
Sign Up
Cited by 75Psychotherapy Techniques and ApplicationsAnxiety, Depression, Psychometrics, Treatment, Cognitive ProcessesTreatment of Major Depression

Precision medicine for long-term depression outcomes using the Personalized Advantage Index approach: cognitive therapy or interpersonal psychotherapy?

Open access
van Bronswijk, Suzanne C., DeRubeis, Robert J., Lemmens, Lotte H. J. M., Peeters, Frenk P. M. L., Keefe, John R., Cohen, Zachary D., Huibers, Marcus J. H.·2021·Psychological Medicine
DOI 10.1017/s0033291719003192PMC PMC7893512
Open this paper in CoChat — free accountOpen PDF

Abstract

Background: Psychotherapies for depression are equally effective on average, but individual responses vary widely. Outcomes can be improved by optimizing treatment selection using multivariate prediction models. A promising approach is the Personalized Advantage Index (PAI) that predicts the optimal treatment for a given individual and the magnitude of the advantage. The current study aimed to extend the PAI to long-term depression outcomes after acute-phase psychotherapy. Methods: Data come from a randomized trial comparing cognitive therapy (CT, n = 76) and interpersonal psychotherapy (IPT, n = 75) for major depressive disorder (MDD). Primary outcome was depression severity, as assessed by the BDI-II, during 17-month follow-up. First, predictors and moderators were selected from 38 pre-treatment variables using a two-step machine learning approach. Second, predictors and moderators were combined into a final model, from which PAI predictions were computed with cross-validation. Long-term PAI predictions were then compared to actual follow-up outcomes and post-treatment PAI predictions. Results: One predictor (parental alcohol abuse) and two moderators (recent life events; childhood maltreatment) were identified. Individuals assigned to their PAI-indicated treatment had lower follow-up depression severity compared to those assigned to their PAI-non-indicated treatment. This difference was significant in two subsets of the overall sample: those whose PAI score was in the upper 60%, and those whose PAI indicated CT, irrespective of magnitude. Long-term predictions did not overlap substantially with predictions for acute benefit. Conclusions: If replicated, long-term PAI predictions could enhance precision medicine by selecting the optimal treatment for a given depressed individual over the long term.

AI snapshot

Written from the paper’s full text.

TL;DR

This study extended the Personalized Advantage Index (PAI) to 17-month follow-up outcomes after acute-phase cognitive therapy (CT) or interpersonal psychotherapy (IPT) for major depressive disorder. Long-term PAI predictions identified parental alcohol abuse as a predictor and recent life events and childhood maltreatment as moderators, with significant PAI effects mainly among CT-indicated individuals and those with stronger predicted advantage.

Key findings

  • One predictor (parental alcohol abuse) and two moderators (number of life events past year; number of childhood trauma events) were selected for long-term outcome. Parental alcohol abuse was associated with higher follow-up BDI-II scores; more life events and more childhood trauma events were associated with lower follow-up BDI-II scores in CT relative to IPT.
  • Overall, individuals assigned to their PAI-indicated treatment had lower follow-up depression severity than those assigned to their PAI-non-indicated treatment, but this difference was not significant (14.5 vs. 17.2, t = 1.39, df = 149, p = 0.17, d = 0.23). Among the top 60% PAI scores, the difference was significant (13.2 vs. 18.2, t = 2.22, df = 89, p = 0.03, d = 0.47).
  • For individuals whose PAI indicated CT, those randomized to CT had lower follow-up BDI-II scores than those randomized to IPT (14.4 vs. 19.8, t = 1.95, df = 85, p = 0.05, d = 0.42); in the top 60% PAI subset, this difference was larger (11.1 vs. 22.3, t = 3.56, df = 45, p < 0.001, d = 1.04). For individuals whose PAI indicated IPT, no significant differences were found (14.7 vs. 13.7, p = 0.67, d = −0.11; top 60%: 15.6 vs. 14.1, p = 0.61, d = −0.16).
  • Long-term PAI scores correlated weakly with post-treatment PAI scores (corr. = 0.33); 46 of 76 (62.2%) long-term CT indications agreed with post-treatment CT indications, and 43 of 58 (74.1%) long-term IPT indications agreed with post-treatment IPT indications.

Methods

  • Data came from a randomized trial comparing CT (n = 76) and IPT (n = 75) for MDD; the current study used the two active conditions (n = 151). Primary outcome was BDI-II depression severity during 17-month follow-up (months 7, 8, 9, 10, 11, 12, and 24), aggregated into an area under the curve and residualized on baseline BDI-II.
  • From 69 pre-treatment variables, 38 were retained after removing 31 highly correlated variables (cor. > 0.70). A two-step machine learning approach selected predictors/moderators: model-based recursive partitioning with random forest, followed by backward elimination across 1,000 bootstrapped samples; variables selected in at least 60% of bootstraps were considered robust.
  • PAI scores were computed from a final regression model using fivefold cross-validation. Utility was evaluated by comparing observed follow-up BDI-II scores for PAI-indicated vs. PAI-non-indicated treatment, overall and in the top 60% PAI subset, and separately for CT- and IPT-indicated individuals. Long-term PAI scores were compared with post-treatment PAI scores in a completer subset (n = 134). A secondary analysis repeated variable selection and model fitting in fivefold held-out samples.

Limitations

  • The paper states: the long-term PAI was not externally validated; although cross-validation was used for final model parameters, variable selection used the full sample. Relevant predictors or moderators may have been omitted despite starting with 69 variables. Participants could seek additional treatment during follow-up, though the authors state this did not significantly affect long-term outcomes. The sample size of 151 may be insufficient according to recent suggestions for multivariate prediction models.
  • Evident from the text: the overall PAI-indicated vs. non-indicated difference was non-significant, and the IPT-indicated comparisons showed no significant benefit. Long-term and post-treatment PAI recommendations overlapped only moderately, which the authors note could pose a clinical dilemma if different treatments are optimal at different stages.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1017/s0033291719003192

Cites

72 of 72 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 49 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    The difference in follow-up depression severity between PAI-indicated and PAI-non-indicated treatment was significant in the subset whose PAI score was in the upper 60%.

    “This difference was significant in two subsets of the overall sample: those whose PAI score was in the upper 60%, and those whose PAI indicated CT, irrespective of magnitude.”
    • Also stated in: § The Personalized Advantage Index · plain
      “Among those with the highest 60% PAI scores, 47 individuals received their PAI-indicated treatment and 44 individuals received their PAI non-indicated treatment. Mean follow-up BDI-II scores differed significantly between these groups (indicated treatment = 13.2, non-indicated treatment = 18.2, t = 2.22, df = 89, p = 0.03), with an effect size estimate of 0.47.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Comparing individuals who received their PAI-indicated treatment with those who received their PAI-non-indicated treatment yields an unbiased estimate of the PAI's utility, without adjustment for baseline variables.
  • Findingown data

    The difference in follow-up depression severity between PAI-indicated and PAI-non-indicated treatment was significant in the subset whose PAI indicated CT, irrespective of magnitude.

    “This difference was significant in two subsets of the overall sample: those whose PAI score was in the upper 60%, and those whose PAI indicated CT, irrespective of magnitude.”
    • Also stated in: § The Personalized Advantage Index · plain
      “for individuals whose PAI indicated CT as the optimal treatment, those who received CT (n = 43) reported lower follow-up BDI-II scores as compared to those who were allocated to IPT (n = 44; indicated treatment = 14.4, non-indicated treatment = 19.8, t = 1.95, df = 85, p = 0.05, Cohen's d = 0.42).”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Comparing individuals who received their PAI-indicated treatment with those who received their PAI-non-indicated treatment yields an unbiased estimate of the PAI's utility, without adjustment for baseline variables.
  • Findingnone given

    Long-term PAI predictions did not overlap substantially with predictions for acute benefit.

    “Long-term predictions did not overlap substantially with predictions for acute benefit.”
    Quote in the paperSubstantial claim
  • Findingown data

    Of the 58 individuals with a long-term PAI indicating IPT, 43 (74.1%) had a post-treatment PAI indicating IPT.

    “Of the 58 individuals with a long-term PAI indicating IPT, 43 (74.1%) had a post-treatment PAI indicating IPT.”
    Quote in the paperNot a substantial claim
  • Findingown data

    The between-treatment difference in follow-up BDI-II scores for life events was significant for individuals with two or more life events.

    “Results of the Johnson–Neyman technique indicated that this between-treatment difference was significant for individuals with two or more life events.”
    Quote in the paperSubstantial claim
  • May combine two claims. 21
  • Classified as probably not a substantial claim. 24

  • May combine two claims.

    Findingown data

    One predictor (parental alcohol abuse) and two moderators (recent life events; childhood maltreatment) were identified.

    “One predictor (parental alcohol abuse) and two moderators (recent life events; childhood maltreatment) were identified.”
    • Also stated in: § Variable selection · plain
      “parental alcohol abuse was identified as a predictor and number of life events past year and number of childhood trauma events were selected as moderators.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Individuals assigned to their PAI-indicated treatment had lower follow-up depression severity than those assigned to their PAI-non-indicated treatment, but the difference was not statistically significant in the overall sample.

    “Individuals assigned to their PAI-indicated treatment had lower follow-up depression severity compared to those assigned to their PAI-non-indicated treatment.”
    • Also stated in: § The Personalized Advantage Index · plain
      “Although the average follow-up BDI-II scores were lower for those who received their indicated treatment, this difference was not significant (indicated treatment = 14.5, non-indicated treatment = 17.2, t = 1.39, df = 149, p = 0.17).”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Comparing individuals who received their PAI-indicated treatment with those who received their PAI-non-indicated treatment yields an unbiased estimate of the PAI's utility, without adjustment for baseline variables.

  • May combine two claims.

    Findingown data

    Of these, three variables were selected in at least 60% of bootstrapped samples: parental alcohol abuse as a predictor and number of life events past year and number of childhood trauma events as moderators.

    “Of these variables, three variables were selected in at least 60% of the bootstrapped samples using the backwards elimination technique: parental alcohol abuse was identified as a predictor and number of life events past year and number of childhood trauma events were selected as moderators.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Long-term PAI scores and post-treatment PAI scores were only weakly correlated (corr. = 0.33), indicating that the magnitude of the predictive advantage was not very consistent between them.

    “The magnitude of the predictive advantage was not very consistent between long-term and post-treatment PAI scores, as indicated by a weak correlation (corr. = 0.33).”
    • Also stated in: § Discussion · plain
      “Individual treatment recommendations and predicted advantages from the long-term PAI scores and the post-treatment were correlated, but only moderately.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    In the secondary analysis, the difference was more pronounced among those with the highest 60% PAI scores (indicated treatment n = 46 = 13.7, non-indicated treatment n = 45 = 19.9, t = 2.33, df = 89, p = 0.02), with an effect size of 0.49.

    “Similar to the primary analysis, this difference was more pronounced among those with the highest 60% PAI scores [mean follow-up BDI-II scores indicated treatment (n = 46) = 13.7, non-indicated treatment (n = 45) = 19.9, t = 2.33, df = 89, p = 0.02], with an effect size of 0.49.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The variable selection identified parental alcohol abuse as a predictor and recent life events and childhood maltreatment as moderators of long-term depression outcome following CT and IPT.

    “Using state-of-the-art variable selection techniques, one predictor (parental alcohol abuse) and two moderators (life events past year and childhood maltreatment) for long-term depression outcome following CT and IPT were identified.”
    • Also stated in: § Predictors and moderators · plain
      “In the current study, we identified parental alcohol abuse as a predictor, and recent life events and childhood maltreatment as moderators of long-term outcome.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The difference in observed depression severity between PAI-indicated and PAI non-indicated treatment was only present in individuals who were recommended to receive CT, whereas no mean differences were found for individuals recommended to receive IPT.

    “Notably, this difference was only present in individuals who were recommended to receive CT, whereas no mean differences were found for individuals recommended to receive IPT.”
    • Also stated in: § Evaluating the long-term PAI · plain
      “Interestingly, further analyses showed that this difference was primarily due to the outcomes observed in individuals whose PAI indicated CT.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    The Personalized Advantage Index (PAI) predicts the optimal treatment for a given individual and the magnitude of the advantage.

    “A promising approach is the Personalized Advantage Index (PAI) that predicts the optimal treatment for a given individual and the magnitude of the advantage.”
    • Also stated in: § Introduction · plain
      “This method not only provides an individual treatment recommendation, it also delivers a quantitative estimate of the predicted advantage of the indicated treatment over the non-indicated treatment(s).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites DeRubeis et al., 2014

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Individuals often go through multiple antidepressant therapies before an effective regimen is identified.

    “In addition, treatment response is highly unpredictable; for example, individuals often go through multiple antidepressant therapies before an effective regimen is identified (Rush et al., 2006).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Rush et al., 2006

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    PAI predictions were computed with cross-validation.

    “Second, predictors and moderators were combined into a final model, from which PAI predictions were computed with cross-validation.”
    • Also stated in: § Statistical analyses · plain
      “Based on this regression model, individual outcome predictions for each treatment were made using a fivefold cross-validation.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Picard and Cook, 1984

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    The PAI method provides an individual treatment recommendation and a quantitative estimate of the predicted advantage of the indicated treatment over the non-indicated treatment.

    “This method not only provides an individual treatment recommendation, it also delivers a quantitative estimate of the predicted advantage of the indicated treatment over the non-indicated treatment(s).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites DeRubeis et al., 2014

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    The PAI approach has been replicated and extended to acute phase CT v. IPT for MDD, continuation CT v. fluoxetine for recurrent MDD, sertraline v. placebo for MDD, trauma-focused CBT and eye movement desensitization for PTSD, and dropout in MDD and PTSD.

    “Since then, the PAI approach has been replicated and extended to acute phase CT v. interpersonal psychotherapy (IPT) for MDD (Huibers et al., 2015), continuation CT v. fluoxetine for recurrent MDD (Vittengl et al., 2017), sertraline v. placebo for MDD (Webb et al., 2019), trauma-focused cognitive behavioral therapy (CBT) and eye movement desensitization for posttraumatic stress disorder (PTSD) (De”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Huibers et al., 2015

      Not checked in the free run.

    • cites Vittengl et al., 2017

      Not checked in the free run.

    • cites Webb et al., 2019

      Not checked in the free run.

  • May combine two claims.

    Study design and conductown data

    Of all participants, 139 (92.1%) had no missing variables and 119 (78.1%) had no missing BDI-II scores.

    “Of all participants, 139 individuals (92.1%) had no missing variables and 119 individuals (78.1%) had no missing BDI-II scores.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data

    To adjust for baseline BDI-II differences between CT and IPT, residuals of a regression with AUC as dependent and baseline BDI-II as independent were used as the outcome.

    “To adjust for this difference, we calculated the residuals of a regression function with the AUC as the dependent variable and the BDI-II at baseline as the independent variable. We used these residuals as the outcome variable for further analyses.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    The imputation approach has been shown to be accurate and comparable to multiple imputation, with lower imputation errors compared to many other imputation methods.

    “This imputation approach has been shown to be accurate and comparable to multiple imputation, with lower imputation errors compared to many other imputation methods (Stekhoven and Bühlmann., 2012; Waljee et al., 2013).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Stekhoven and Bühlmann., 2012

      Not checked in the free run.

    • cites Waljee et al., 2013

      Not checked in the free run.

  • May combine two claims.

    Study design and conductown data

    Missing data were limited: 25 observations of all 38 variables were missing (0.4%) and 164 BDI-II values were missing (15.5%).

    “A total of 25 observations of all 38 variables were missing (0.4%). On the BDI-II (7, 8, 9, 10, 11, 12, and 24 months), 164 values were missing (15.5%).”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data

    Secondary analyses repeated variable selection and model fitting to a fivefold held-out sample to examine overfitting and out-of-sample predictions.

    “To examine if this affected the results, we ran secondary analyses repeating the process of variable selection and model fitting to a fivefold held-out sample creating five separate models.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Participants who received CT had a higher number of comorbid axis II disorders compared to IPT.

    “On average, participants who received CT had a higher number of comorbid axis II disorders and a lower number of axis II traits as compared to IPT (t = 2.00, df = 144, p = 0.047 and t = 2.31, df = 144, p = 0.02 for disorders and traits, respectively).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    CT and IPT have comparable prophylactic effects after treatment termination.

    “CT and IPT are two frequently practiced psychotherapies for MDD and have been shown to be equally effective in the acute phase (Jakobsen et al., 2012; Lemmens et al., 2015) with comparable prophylactic effects after treatment termination (Lemmens et al., 2019).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Lemmens et al., 2019

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    Treatment quality was rated by independent assessors as '(very) good' to 'excellent'.

    “For both CT and IPT, treatment quality was rated by independent assessors as ‘(very) good’ to ‘excellent’ (Lemmens et al., 2015).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Lemmens et al., 2015

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductown data

    A total of 74 individuals had been assigned, by chance, to their PAI-indicated treatment, and 77 received, by chance, their PAI non-indicated treatment.

    “A total of 74 individuals had been assigned, by chance, to their PAI-indicated treatment, and 77 received, by chance, their PAI non-indicated treatment.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    Variables selected in at least 60% of bootstrapped samples were considered robust.

    “If variables were selected in at least 60% of the bootstrapped samples, they were considered robust (Austin and Tu, 2004) and used to build the PAI.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Austin and Tu, 2004

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Participants who received CT had a lower number of axis II traits compared to IPT.

    “On average, participants who received CT had a higher number of comorbid axis II disorders and a lower number of axis II traits as compared to IPT (t = 2.00, df = 144, p = 0.047 and t = 2.31, df = 144, p = 0.02 for disorders and traits, respectively).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    Missing BDI-II outcomes and variables were imputed using a non-parametric random forest approach.

    “Missing BDI-II outcomes and variables were imputed using a non-parametric random forest approach (R package ‘MissForest’, Stekhoven and Bühlmann, 2012).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Stekhoven and Bühlmann, 2012

      Not checked in the free run.

    • Rests on an assumption: Missing BDI-II outcomes and pre-treatment variables are missing at random, so random forest imputation yields unbiased values.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    It is possible that relevant predictors or moderators were not included among the 69 variables used in variable selection.

    “Second, although we began our variable selection with 69 variables, it is still possible that relevant predictors or moderators were not included in our study.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    There is evidence that the association between parental alcohol abuse and depressive mood or mood disorders is mediated by adverse childhood experiences.

    “although there is evidence that this association is mediated by adverse childhood experiences (Anda et al., 2002).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Anda et al., 2002

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    The moderator effects and the differential performance of the PAI for CT v. IPT indicate a specific benefit of CT for a subgroup of individuals who suffered from childhood maltreatment events and recently experienced significant life events, whereas for the remainder of individuals no differential effect was observed.

    “These moderator effects and the differential performance of the PAI for CT v. IPT indicate a specific benefit of CT for a subgroup of individuals who suffered from childhood maltreatment events and recently experienced significant life events, whereas for the remainder of the individuals, no differential effect was observed.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    In a previous study, a tendency was found for individuals with severe negative life events prior to their onset of depression to respond better to IPT than to CBT.

    “In a previous study, a tendency was found for individuals with severe negative life events prior to their onset of depression to respond better to IPT than to CBT.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Bulmash et al., 2009

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given

    Although cross-validation was used to compute the regression parameters of the final model, the full study sample was used for the variable selection procedure.

    “Although we used a cross-validation approach to compute the regression parameters of the final model, we used the full study sample for the variable selection procedure.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductcitation

    Individuals were allowed to seek additional treatment during follow-up, but this did not significantly affect the long-term outcomes.

    “Third, individuals were allowed to seek additional treatment during follow-up. However, this did not significantly affect the long-term outcomes (Lemmens et al., 2019).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Lemmens et al., 2019

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    The number of life events was also identified as one of six moderators of the post-treatment PAI of the same study sample, with lower post-treatment depression severity in CT as compared to IPT.

    “This variable was also identified as one of the six moderators of the post-treatment PAI of the same study sample, with lower post-treatment depression severity in CT as compared to IPT (Huibers et al., 2015).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Huibers et al., 2015

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Research in adult children of alcoholics has reported an association between parental alcohol abuse and depressive mood and mood disorders.

    “This finding is in line with the research in adult children of alcoholics that reported an association between parental alcohol abuse and depressive mood (Kelley et al., 2010; Klostermann et al., 2011), and mood disorders (Cuijpers et al., 1999)”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Kelley et al., 2010

      Not checked in the free run.

    • cites Klostermann et al., 2011

      Not checked in the free run.

    • cites Cuijpers et al., 1999

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    The moderator effects indicate that CT is indicated for individuals with two or more pre-treatment life events and/or one or more events of childhood maltreatment, whereas individuals with one or no life events and no childhood trauma have no indication of a meaningful difference between CT and IPT.

    “One can interpret these moderator effects as follows: when an individual had two or more pre-treatment life events and/or one or more events of childhood maltreatment, CT would be indicated, whereas individuals with one or no life events and no childhood trauma have no indication of a meaningful difference between CT and IPT”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Cohen and DeRubeis, 2018

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    A secondary analysis reran the complete process with five folds, producing five models that estimated the PAIs of individuals whose data were not used in any way to develop the algorithm that yielded the PAIs.

    “To examine potential bias, we did a secondary analysis rerunning the complete process with fivefolds, producing five models that estimated the PAIs of individuals whose data were not used in any way to develop the algorithm that yielded the PAIs.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Findings of that same study indicated that response to treatment in individuals with severe negative life events prior to their depression treatment was superior in both CBT and IPT, relative to antidepressant medication.

    “However, findings of that same study indicated that response to treatment in individuals with severe negative life events prior to their depression treatment was superior in both CBT and IPT, relative to antidepressant medication (Bulmash et al., 2009).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Bulmash et al., 2009

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Differential treatment outcomes for individuals with a history of childhood maltreatment have been described in previous studies.

    “Differential treatment outcomes for individuals with a history of childhood maltreatment have been described in previous studies (Nemeroff et al., 2003; Barbe et al., 2004; Asarnow et al., 2009; Lewis et al., 2010; Harkness et al., 2012).”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Nemeroff et al., 2003

      Not checked in the free run.

    • cites Barbe et al., 2004

      Not checked in the free run.

    • cites Asarnow et al., 2009

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    The findings of Mulder et al. (2017) were not in line with our results: no significant moderators were identified, and personality variables were identified as significant predictors.

    “The findings of this study were not in line with our results: no significant moderators were identified, and personality variables were identified as significant predictors.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Mulder et al., 2017

      Not checked in the free run.

37 of 37

  • The paper relies on this without support.

    Missing BDI-II outcomes and pre-treatment variables are missing at random, so random forest imputation yields unbiased values.

    “Missing BDI-II outcomes and variables were imputed using a non-parametric random forest approach”

  • The paper relies on this without support.

    Comparing individuals who received their PAI-indicated treatment with those who received their PAI-non-indicated treatment yields an unbiased estimate of the PAI's utility, without adjustment for baseline variables.

    “To test the utility of the PAI, actual follow-up outcomes (AUCs) of individuals receiving the PAI-indicated treatment were compared with those of individuals receiving the PAI non-indicated treatment, using t tests.”
2 of 2

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?

  • All personalized intervention studies use predictive models based on multi-subject data.

    “Yet, of note, all of these personalized intervention studies use predictive models based on multi-subject data.”

    From Personalized machine learning guided intervention for optimizing lifestyle behaviors in depression: a pilot study (2026) · cites this paper as [24]

    The Methods (Statistical analyses) describe a personalized treatment model trained on data from multiple participants: each individual's outcomes were predicted using model weights from the other four groups of the sample. The Introduction names other multivariate approaches, but our copy does not establish that all personalized intervention studies use models based on multi-subject data. (LLM, conf=0.83)

    Experimental — accuracy not yet measured

  • In a depression intervention study, patients receiving personalized advantage index (PAI)-indicated treatment (cognitive or interpersonal therapy) had decreased long-term symptom severity compared with those receiving PAI-non-indicated treatment.

    “A depression intervention study that used a personalized advantage index (PAI) to assign either cognitive or interpersonal therapy to individuals showed decreased long-term symptom severity in patients receiving PAI-indicated vs. PAI-non-indicated treatment [24].”

    From Personalized machine learning guided intervention for optimizing lifestyle behaviors in depression: a pilot study (2026) · cites this paper as [24]

    The Results (The Personalized Advantage Index) report lower average long-term BDI-II scores with PAI-indicated treatment, but the overall difference was not statistically significant; it was significant in the highest-60%-PAI subgroup. Participants were assigned to treatment by chance, not assigned using the PAI, as the citing sentence says. (LLM, conf=0.92)

    Experimental — accuracy not yet measured

Self-check

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

  • Incorrect percentage whose post-treatment PAI also indicated CT

    Minor

    Among 76 individuals, 46 is 60.5%, not 62.2%.

    Reported 62.2 · recomputed 60.526316

1 of 1

Cites

Flagged first, then by citations

Research works printed but not linked: 4

  1. First, 1995, Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I)
  2. Beck, 1988, Manual for the Beck Hopelessness Scale
  3. Klerman, 1984, Interpersonal Psychotherapy for Depression
  4. Beck, 1996, Beck Depression Inventory II: Manual

Cited by

Sorted by citations

Related

Authors
van Bronswijk, S. C., DeRubeis, R. J., Lemmens, L. H. J. M., Peeters, F. P. M. L., Keefe, J. R., Cohen, Z. D., & Huibers, M. J. H.
Affiliations
  • University of Pennsylvania (US)
  • Maastricht University (NL)
  • Weill Cornell Medicine (US)
  • University of California, Los Angeles (US)
ORCID
Venue
Psychological Medicine
Publisher
Cambridge University Press (CUP)
Published
2019-11-22
Volume
51
Issue
2
Pages
279-289
ISSN
  • 0033-2917 (ISSN-L)
  • 1469-8978
DOI
10.1017/s0033291719003192
PMC
PMC7893512
OpenAlex
W2990568952
Semantic Scholar
bce67c1841b26a0634fdd0edcdba9cb2b66b06d4
Cited by
75
References
71
Citations by year
2026: 9 · 2025: 14 · 2024: 9 · 2023: 13 · 2022: 15 · 2021: 7 · 2020: 8
FWCI
8.76
Type
journal article
Language
en
Open access
Licence
Locations
  • Psychological Medicine (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Research Publications (Maastricht University) (publishedVersion, cc-by)
  • Utrecht University Repository (Utrecht University) (submittedVersion, other-oa)
  • VU Research Portal (submittedVersion, other-oa)
  • Research Publications (Maastricht University) (submittedVersion, other-oa)
  • Data Archiving and Networked Services (DANS) (submittedVersion, other-oa)
Topics
  • Social Sciences › Psychology › Clinical Psychology › Psychotherapy Techniques and Applications
  • Social Sciences › Psychology › Experimental and Cognitive Psychology › Anxiety, Depression, Psychometrics, Treatment, Cognitive Processes
  • Health Sciences › Medicine › Pharmacology › Treatment of Major Depression
Keywords
precision medicine, major depressive disorder, cognitive therapy, interpersonal psychotherapy, treatment selection, parental alcoholism, child abuse, recent life events, treatment response prediction
MeSH terms
Interpersonal Psychotherapy, Adolescent, Adult, Aged, Depressive Disorder, Major, Major Depressive Disorder, Female, Humans, Male, Middle Aged, Netherlands, Psychiatric Status Rating Scales, Cognitive Behavioral Therapy, Treatment Outcome, Young Adult, Precision Medicine

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.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-09.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-09.

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