Section - Quick Takes

04. Predicting Conversion to Psychosis Using Machine Learning

Published on June 1, 2024 Certification expiration date: June 1, 2027

Oliver Freudenreich, M.D.

Co-director of the MGH Psychosis Clinical and Research Program Associate Professor of Psychiatry Harvard Medical School - Massachusetts General Hospital

Key Points

The study highlights the use of machine learning algorithms in predicting psychosis conversion in high-risk or prodromal patients, with a focus on the North American Prodrome Longitudinal Study (NAPLS) risk calculator.
The study found that the random forest algorithm performed best, achieving a prediction accuracy of 90%, sensitivity of 75%, and specificity of 85%. Key predictors included delusions and suspiciousness.
Despite promising results, the clinical utility of these calculators remains uncertain, as the models require validation in independent datasets and clinicians must still rely on their judgment in evaluating patients.

To unlock this audio you need to become a Silver, Gold, Silver extended or Gold extended Member.

Check Our CME Program

Downloads for Offline Access

Download Audio File (MP3)

Text version

Hi from Boston, Massachusetts. I’m Oliver Freudenreich, or Dr. F, representing the Psychopharmacology Institute.

In today’s Quick Take, I’ll discuss the conversion to psychosis in high-risk or possibly prodromal patients, prompted by a recent brief research letter published in the American Journal of Psychiatry ‘s November issue. This research focuses on risk prediction in psychiatry. Consider an adolescent falling behind academically, withdrawn, and exhibiting subtle paranoia. These symptoms—academic failure, social withdrawal, and attenuated psychosis—should alert clinicians to the potential prodromal phase of schizophrenia. However, literature indicates that the general conversion risk from a clinical high-risk state to full-blown psychosis is about 20%–30% over 3 years. The challenge is to better ascertain an individual’s conversion risk.

After 2 decades of research, we now have large longitudinal datasets. These have facilitated the development of risk calculators for conversion, similar to the Framingham risk calculator for heart disease. The most commonly used calculator,

Unlock this Quick Take and earn 0.75 CMEs
Become a Silver, Gold, Silver extended or Gold extended Member.

Get Instant Access

Already have an account? Sign in

Next Section

05. Iloperidone: A Potential Treatment Option for Bipolar Mania

Learning Objectives:

After completing this activity, the learner will be able to:

Explain the mechanism of action of muscarinic agonists in treating psychosis and how they differ from current antidopaminergic antipsychotics.
Analyze one's own biases and decision-making processes in the context of restraint use, considering the potential impact on patient outcomes.
Compare and contrast the potential outcomes and treatment implications of accurately diagnosing substance-induced psychosis vs mislabeling it as schizophrenia.
Understand the role of machine learning algorithms in improving the prediction of psychosis conversion in high-risk or prodromal patients.
Assess the overall risk–benefit profile of iloperidone for individual patients with bipolar mania.

Original Release Date: June 1, 2024

Expiration Date: June 1, 2027

Experts: Scott Beach, M.D., Paul Zarkowski, M.D., Oliver Freudenreich, M.D.

Medical Editor: Lorena Rodriguez, M.D.

Relevant Financial Disclosures:

Oliver Freudenreich declares the following interests:

- Alkermes: Research grant, consultant honoraria

- Janssen: Research grant, consultant honoraria

- Otsuka: Research grant

- Karuna: Research grant, consultant honoraria

- Neurocrine: Consultant honoraria

- Vida: Consultant honoraria

- American Psychiatric Association: Consultant honoraria

- Medscape: Honoraria

- Elsevier: Honoraria

- Wolters-Kluwer: Royalties

- UpToDate: Royalties, honoraria

All of the relevant financial relationships listed above have been mitigated by Medical Academy and the Psychopharmacology Institute.

None of the other faculty, planners, and reviewers for this educational activity have relevant financial relationships to disclose during the last 24 months with ineligible companies whose primary business is producing, marketing, selling, re-selling, or distributing healthcare products used by or on patients.

Contact Information: For questions regarding the content or access to this activity, contact us at support@psychopharmacologyinstitute.com

Instructions for Participation and Credit:

Participants must complete the activity online during the valid credit period that is noted above.

Follow these steps to earn CME credit:

View the required educational content provided on this course page.
Complete the Post Activity Evaluation for providing the necessary feedback for continuing accreditation purposes and for the development of future activities. NOTE: Completing the Post Activity Evaluation after the quiz is required to receive the earned credit.
Download your certificate.

Accreditation Statement

This activity has been planned and implemented in accordance with the accreditation requirements and policies of the Accreditation Council for Continuing Medical Education through the joint providership of Medical Academy LLC and the Psychopharmacology Institute. Medical Academy is accredited by the ACCME to provide continuing medical education for physicians

Credit Designation Statement

Medical Academy designates this enduring activity for a maximum of 0.75 AMA PRA Category 1 credit(s)™. Physicians should claim only the credit commensurate with the extent of their participation in the activity.