However, there is no reliable way to determine which treatment is likely to be the most effective for an individual patient.
A recent study that used machine learning techniques to address this problem has been published.
A research team from Yale University used clinical data from the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) trial in the U.S.
I served as an investigator in the STAR*D and am happy to see this database still in use.
In the current study, the research team used machine learning with a group of 164 pre-treatment variables. From this group of variables, 25 were identified as providing predictive value of response/non-response to treatment with a standard antidepressant drug citalopram.
Clinical predictors of non-response included:
- High baseline depression severity scores
- Presence of psychomotor agitation at baseline
- Reduced energy ratings at baseline (fatigue)
Predictors of depression remission included:
- Current employment
- Higher level of education
- Lower scores on depression insight
The research team was able to build a machine learning model that showed a 63% sensitivity and 66% specificity in prediction response to citalopram. This was statistically greater than random (chance) prediction.
Addition support for their model was gained by replication in a second study of citalopram in depression.
This is an important and exciting finding that suggests low-cost symptom biomarkers may aid in the treatment selection for depression.
You can access the abstract of this important work by clicking HERE or by clicking on the PMID link in the citation below.
Follow the author on Twitter WRY999 HERE.
Photo of sunset on Captiva Island, Florida is from my personal files.
Chekroud AM, Zotti RJ, Shehzad Z, Gueorguieva R, Johnson MK, Trivedi MH, Cannon TD, Krystal JH, & Corlett PR (2016). Cross-trial prediction of treatment outcome in depression: a machine learning approach. The lancet. Psychiatry, 3 (3), 243-50 PMID: 26803397
No comments:
Post a Comment