Table 1.

Summary of study characteristics

Authors	Year	Country	Study design	Sample characteristics	Outcomes measured	Medication	Genes
Studies with three comparators
Jürgens et al.26	2020	Denmark	Single-blind RCT Three arms: pharmacogenetics versus SCM versus TAU Assessment points: baseline and 12 months	Total n = 161. Pharmacogenetics (n = 84) versus TAU (n = 77) Age, median (years): 42 versus 42 Sex (% female): 43 versus 46 Ethnicity: N/A Medication: N/A Diagnosis (%): paranoid schizophrenia, 72 versus 65; schizotypal disorder, 20 versus 20; persistent delusional disorders, 3 versus 2; acute and transient psychotic disorders, 1 versus 2; schizoaffective disorders, 2 versus 8	Antipsychotic drug persistence, that is, days to first modification of the initial treatment Number of drug and dose changes Adverse drug response (UKU) Symptom severity (using SAPS) Compliance (ROMI)	Antipsychotics	CYP2D6 and CYPC19
Studies with two comparators
Kang et al.24	2023	China	Double-blind RCT Two arms: pharmacogenetics versus TAU Assessment points: baseline and 6 weeks	Total n = 210 Pharmacogenetics (n = 113) versus TAU (n = 97) Age, median (years): 29.9 versus 28.3 Sex (%): male, 100 Ethnicity (%): Han Chinese, 100 Medication: quetiapine, 26.5 versus 19.6; risperidone, 31.0 versus 45.4; olanzapine, 13.3 versus 11.3; aripiprazole, 15.0 versus 8.2; ziprasidone, 0.9 versus 3.1; paliperidone, 6.2 versus 7.2; clozapine, 1.8 versus 2.1; amisulpride, 5.3 versus 3.1. Diagnosis (%): schizophrenia, 100	Symptom severity (using PANSS) ADRs	Antipsychotics	CYP1A2, CYP2D6, CYP3A4, DRD2, EPM2A, HTR1A, HTR2A, HTR2C, MC4R, RGS4 and SH2B1
Herbild et al.28	2013	Denmark	Double-blind RCT Three arms: pharmacogenetics versus extensive clinical monitoring versus TAU Time horizon: 1 year	Total n = 207 Pharmacogenetics (n = 103) versus TAU (n = 104) Age, mean (years): 41 versus 42 Sex (% female): 45 versus 44 Ethnicity: N/A Medication: N/A Diagnosis (%): schizophrenia, 74 versus 71; schizotypal disorders, 24 versus 21; other disorders, 5 versus 12	Pharmaceutical costs Hospitalization costs	Antipsychotics	CYP2D6 and CYP2C19
Arranz et al.27	2019	Spain	Double-blind RCT Two arms: pharmacogenetics versus TAU Assessment points: baseline and 12 weeks	Total n = 290. Pharmacogenetics (n = 123) versus TAU (n = 167) Age, median (years): 46.1 versus 48.7 Sex (% female): 48.8 versus 43.7 Ethnicity: N/A Medication (%): clozapine, 35 versus 52.7; risperidone, 13 versus 12; olanzapine, 20.3 versus 8.4; paliperidone, 13 versus 13; aripiprazole, 5.7 versus 7.8; quetiapine 8.9 versus 3; ziprasidone, 0.8 versus 1.2; trifluoperazine, 0.8 versus 0.6; haloperidol, 0.8 versus 0.6; asenapine, 0.8 versus 0.6; pimozide, 0.8 versus 0. Diagnosis (%): schizophrenia, 86 versus 69; schizoaffective, 5 versus 4; delusional disorder, 9 versus 27	Symptom severity (using PANSS) ADRs (using UKU)	Antipsychotics	CYP2D6, CYPC19, CYP1A2 and CYP3A5
Arranz et al.21	2022	Spain	Prospective observational study Two arms: pharmacogenetics versus TAU Assessment points: baseline and 4 months	Total n = 104. Pharmacogenetics (treatment resistant) (n = 42) versus TAU (n = 62) Age, mean (years): 18.79 versus 13.83 Sex (% female): 26% versus 8% Ethnicity: N/A Medication (%): antipsychotics, 67 versus 32; antidepressants, 48 versus 11; anxiolytics, anticonvulsants and others, 26 versus 56; no current medication, 7 versus 0 Diagnosis: 100% autism spectrum disorder	Symptom severity (using CGI and CGAS)	Antipsychotics, antidepressants, anxiolytics and anticonvulsants	CYP1A2, CYP2C19, CYP2D6 and SLC6A4
Studies with one comparator
Carrascal-Laso et al.22	2020	Spain	Retrospective observational study One arm: pharmacogenetics only Time horizon: 3 years	Total sample (n = 188) offered pharmacogenetics test. Age, median (years): 47 Sex (% female): 37.8% Ethnicity: N/A Medication (%): N/A Diagnosis (%): dementia, 0.53; substance-related disorder 6.38; schizophrenia, 67.02; persistent delusional disorder, 1.06; brief and acute psychotic disorder, 0.53; schizoaffective disorder, 6.92; bipolar disorder, 13.30; major depressive disorder, 0.53; specific personality disorder, 1.06; mixed personality disorder, 0.53; intellectual disability, 1.06	Mean daily dose Polytherapy cases	Antipsychotics	CYP1A2, CYP2B6, CYP2C9, CYP2C19, CYP2D6, CYP3A5 and ABCB1
Carrascal-Laso et al.23	2021	Spain	Retrospective observational study One arm: pharmacogenetics only Time horizon: 3 years	Total sample (n = 188) offered pharmacogenetics test. Age, median (years): 47 Sex (% female): 37.8% Ethnicity: N/A Medication (%): N/A Diagnosis (%): dementia, 0.53; substance-related disorder 6.38; schizophrenia, 67.02; persistent delusional disorder, 1.06; brief and acute psychotic disorder, 0.53; schizoaffective disorder, 6.92; bipolar disorder, 13.30; major depressive disorder, 0.53; specific personality disorder, 1.06; mixed personality disorder, 0.53; intellectual disability, 1.06	Pharmaceutical costs Hospitalization costs	Antipsychotics	CYP1A2, CYP2B6, CYP2C9, CYP2C19, CYP2D6, CYP3A5 and ABCB1
Walden et al.25	2019	Canada	Prospective observational study One arm: pharmacogenetics only Assessment points: baseline, 6 weeks and 12 weeks	Total sample (n = 80) offered pharmacogenetics test. Age, mean (years), 43 Sex (% female): 43.8% Ethnicity/race (% of participants): European Caucasian 68.8%, African 3.8%, Asian 3.8%, others 12.5%, mixed 11.3% Medication (%): antipsychotics, 47.5; antidepressants, 23.8; anxiolytics, 7.5; antipsychotics and antidepressants, 11.3; antipsychotics, antidepressants, and anxiolytics, 6.3; no medication, 3.8 Diagnosis: schizophrenia/schizoaffective, 53.8%; anxiety/depression, 40%, others, 6.3%	Physician’s opinions (using PIP-FQ) ADRs (UKU)	Antidepressants, anxiolytics and antipsychotics	CYP2D6 and CYP2C19
Markov/decision models
Ninomiya et al.31	2022	United Kingdom	Decision tree with Markov model Third-party payer perspective Two arms: pharmacogenetics versus TAU Time horizon: 10 years	The target population was adult men and women with treatment-resistant schizophrenia	ICER	Clozapine	SLCO1B3–SCLO1B7, HLA-DQB1 and HLA-B
Girardin et al.35	2019	United States	Decision tree with semi-Markovian model Third-party payer perspective Three arms: (1) PGx-guided clozapine treatment with ANCM for patients who test positive for one or both alleles, (2) PGx-guided clozapine treatment for patients who test negative or alternative antipsychotics for patients who test positive, (3) TAU. Time horizon: 3 years	The target population was adult men and women with treatment-resistant schizophrenia	ICER	Clozapine	HLA-DQB1 and HLA-B
Kurylev et al.29,	2018	Russia	Decision tree Three arms: (1) PGx in 100% of patients, (2) PGx in 30% of patients, (3) TAU. Time horizon: N/A	The target population was patients diagnosed with paranoid schizophrenia	Hospitalization costs Medication costs	Antipsychotics	CYP2D6
Rejon-Parrilla et al.32,	2014	United Kingdom	Decision tree with Markov model Healthcare provider perspective (NHS) Two arms: (1) traditional dosing, (2) pharmacogenetic testing Time horizon: 2 years	The target population was previously untreated patients newly diagnosed with schizophrenia, aged 25	Incremental cost-effectiveness	Risperidone	CYP2D6
Perlis et al.34,	2005	United States	Decision tree with Markov model Societal perspective Three arms: (1) no PGx test, clozapine as first-line treatment, (2) PGx testing, clozapine as first line if they test positive for or third line if the test negative, (3) no PGx testing, clozapine as third line. Time horizon: lifetime	The target population was a 30-year-old patient with schizophrenia	Incremental cost-effectiveness	Clozapine	N/A

N/A indicates that information was not reported in the original article. ANCM, absolute neutrophil count monitoring; CGAS, Children’s Global Assessment Scale; PGx, pharmacogenetics; SCM, structural clinical monitoring.