Limits...
Toward personalizing treatment for depression: predicting diagnosis and severity.

Huang SH, LePendu P, Iyer SV, Tai-Seale M, Carrell D, Shah NH - J Am Med Inform Assoc (2014)

Bottom Line: In particular, depressed patients exhibit largely unpredictable responses to treatment.The models use commonly available data on diagnosis, medication, and clinical progress notes, making them easily portable.The ability to automatically determine severity can facilitate assembly of large patient cohorts with similar severity from multiple sites, which may enable elucidation of the moderators of treatment response in the future.

View Article: PubMed Central - PubMed

Affiliation: Stanford Center for Biomedical Informatics Research, Stanford University, Stanford, California, USA.

Show MeSH

Related in: MedlinePlus

Receiver operating characteristic (ROC) curves for the model's performance on test data restricted to three cutoff points.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4215055&req=5

AMIAJNL2014002733F2: Receiver operating characteristic (ROC) curves for the model's performance on test data restricted to three cutoff points.

Mentions: Using data for depressed patients up until the time of first diagnosis, the model for predicting diagnosis achieved an area under the receiver operating characteristic (AUC) curve of 0.800 (95% CI 0.784 to 0.815). This cutoff corresponds with a timely prediction. The AUC is 0.712 (95% CI 0.695 to 0.729) for the 6-month cutoff and 0.701 (95% CI 0.684 to 0.718) for the 1-year cutoff. In total, 553 out of the 10 600 features were used in this model. Figure 2 shows the receiver operating characteristic (ROC) curves for these three cutoffs, which display the full range of sensitivity and specificity values achieved by the model. For example, the ROC curves show that at 90% specificity (ie, 10% false positive rate), the model is able to predict a diagnosis of depression at approximately 50% sensitivity at the time of diagnosis and at 25% sensitivity 12 months prior to the time of diagnosis.


Toward personalizing treatment for depression: predicting diagnosis and severity.

Huang SH, LePendu P, Iyer SV, Tai-Seale M, Carrell D, Shah NH - J Am Med Inform Assoc (2014)

Receiver operating characteristic (ROC) curves for the model's performance on test data restricted to three cutoff points.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4215055&req=5

AMIAJNL2014002733F2: Receiver operating characteristic (ROC) curves for the model's performance on test data restricted to three cutoff points.
Mentions: Using data for depressed patients up until the time of first diagnosis, the model for predicting diagnosis achieved an area under the receiver operating characteristic (AUC) curve of 0.800 (95% CI 0.784 to 0.815). This cutoff corresponds with a timely prediction. The AUC is 0.712 (95% CI 0.695 to 0.729) for the 6-month cutoff and 0.701 (95% CI 0.684 to 0.718) for the 1-year cutoff. In total, 553 out of the 10 600 features were used in this model. Figure 2 shows the receiver operating characteristic (ROC) curves for these three cutoffs, which display the full range of sensitivity and specificity values achieved by the model. For example, the ROC curves show that at 90% specificity (ie, 10% false positive rate), the model is able to predict a diagnosis of depression at approximately 50% sensitivity at the time of diagnosis and at 25% sensitivity 12 months prior to the time of diagnosis.

Bottom Line: In particular, depressed patients exhibit largely unpredictable responses to treatment.The models use commonly available data on diagnosis, medication, and clinical progress notes, making them easily portable.The ability to automatically determine severity can facilitate assembly of large patient cohorts with similar severity from multiple sites, which may enable elucidation of the moderators of treatment response in the future.

View Article: PubMed Central - PubMed

Affiliation: Stanford Center for Biomedical Informatics Research, Stanford University, Stanford, California, USA.

Show MeSH
Related in: MedlinePlus