Psymatik: A New Online Tool That Sorts Antidepressant or Antipsychotic Side Effects

Published on: 

Investigators of a new meta-analysis designed the Psymatik Treatment Optimizer, an online tool which synthesizes side-effects in a database and can personalize individual user side effect concerns.

A new meta-analysis broke down the side effects of antidepressants and antipsychotic medication to see how they affect the quality of life and functioning in treated patients.

The study, led by Toby Pillinger, PhD, of the institute of psychiatry, psychology, and neuroscience at King’s College London, examined 14 articles containing 65 meta-analyses of individual side-effects associated with antidepressants and antipsychotics.

After reviewing 19 guidelines, the team found 7 provided ordinal data. Antipsychotic data was pulled out of 5 studies (11 meta-analyses and 594 patients) and 4 guidelines. Antidepressant data was pulled out of 3 guidelines.

The databases shared the “clinical dilemma” related with handling side effects. Patients often prefer one medication over another in fear of a side effect. For example, one patient might avoid taking a medication that increases the risk of weight gain, while another patient might avoid a medication due to the risk of akathisia.

That’s where the Psymatik Treatment Optimizer comes in, created by the team using the Dijango tool. The system’s purpose is to “synthesize the side-effect databases with individual user-defined concern weights.”

“To our knowledge, we have compiled the most comprehensive side-effect databases for antipsychotics and antidepressants to date, incorporating meta-analytic data for more than 65,000 people, alongside a synthesis of 7 national or international guidelines for 69 drugs,” the investigators wrote. “Furthermore, for the first time to our knowledge, meta-analytic data were complemented by imputed results that were based on ordinal ranking scores, providing a compromise between the superior accuracy of meta-analytic results and broader drug coverage of guidelines.”

The Pysmatik makes it easier for individuals to know what the best medication for them would be. The system is complete with a side-effect menu and a side-effect concern rating, incorporated as a slider, ranging from 0 (a minor concern) to 100 (a major concern).

Pillinger and his colleagues put into the system 5851 pairwise comparisons for antidepressants and 5142 pairwise comparisons for antipsychotics. With that data, Psymatik ranks medication in order of preference for the individual user.

“The Psymatik Treatment Optimizer is not intended to dictate prescribing decisions directly, but to facilitate informed discussions between the patient and prescriber about appropriate treatment,” the team wrote. “As such, this application holds promise as a tool for shared decision making within psychiatry.”

The database will help make it easier to pinpoint what side effect is in what medication, which is important since in high-income countries, ≥17% of adults are prescribed antidepressants and 2% are prescribed antipsychotics. Among those treated, 75% have side-effects.

Side-effects of antidepressants and antipsychotics ultimately affect a patient’s life because they increase risks for diseases and death, as well as perpetuating a stigma. Patients who are unhappy with their prescription’s side effects sometimes switch their prescription. Nearly 1 in 10 patients with a new episode of depression switch their antidepressant. On the other hand, 73% of patients with first-episode psychosis switch their antipsychotic.

Switching a medication hurts a patient’s quality of life. Patients who switch antidepressants lose productivity at work by 70% and miss twice as many days of work compared to their colleagues who do not switch medication.

The team had 2 methodological stages to create side effect databases.

The first stage was a systematic umbrella review of published meta-analyses. The team looked at data up to June 26, 2023 through Embross, PsychINFO, and MEDLINe, finding meta-analyses that ranked antidepressants when used as monotherapy.

The team studied side effects that are common among antipsychotics or antidepressants. The side-effect databases included meta-analytic, ordinal, and imputed data. Both the meta-analytic and ordinal outcomes were normalized to provide values between 0 and 1. The team examined the meta-analysis that examined the largest number of drugs.

The second stage of the study was creating the Pysmatik web application.

In the meta-analysis, the investigators noticed a consistent grouping of antipsychotics that contrasted side-effects. For instance, partial agonists like ripiprazole; brexpiprazole; cariprazine, lurasidone; and ziprasidone had a low risk for weight gain and metabolic dysregulation. Yet these same drugs showed an increased risk of akathisia. Though other older drugs that showed an increased risk of akathisia—haloperidol; flupentixol; zuclopenthixol; fluphenazine; and pimozide—also had a low risk for weight gain but an increased risk of Parkinsonism and hyperprolactinaemia.

The investigators found side-effect patterns in antidepressant classes, discovering selective serotonin reuptake inhibitors like fluoxetine and sertraline often had gastrointestinal and sexual side-effects and a high risk of hyponatremia, but a lower risk of anticholinergic effects and sedation.

On the other hand, Tricyclic antidepressants like amitriptyline and trimipramine were associated with the increased risk of weight gain; sedation; orthostatic hypotension; sexual dysfunction; and anticholinergic effects, with a lower risk of insomnia or agitation and gastrointestinal effects.

Ultimately, Psymatik is intended to serve as a resource for clinicians.

“The Psymatik Treatment Optimizer should be implemented and tested to establish to what degree the tool could form a relevant part of future psychiatric practice,” the team wrote. “However, the utility of Psymatik is reliant on database accuracy, and we have identified that meta-analytic side-effect data for many psychiatric treatments are either absent or insufficient, which needs to be addressed by future studies.”

Pillinger and colleagues said the databases will be updated when new adverse effect data become available, “with meta-analytic data replacing ordinal data.”


Pillinger T, Howes OD, Correll CU, et al. Antidepressant and antipsychotic side-effects and personalised prescribing: a systematic review and digital tool development [published online ahead of print, 2023 Sep 26]. Lancet Psychiatry. 2023;S2215-0366(23)00262-6. doi:10.1016/S2215-0366(23)00262-6