Dataset Information

Machine-Learning for Prescription Patterns: Random Forest in the Prediction of Dose and Number of Antipsychotics Prescribed to People with Schizophrenia.

ABSTRACT:

Objective

We aimed to predict antipsychotic prescription patterns for people with schizophrenia using machine learning (ML) algorithms.

Methods

In a cross-sectional design, a sample of community mental health service users (SUs; n = 368) with a primary diagnosis of schizophrenia was randomly selected. Socio-demographic and clinical features, including the number, total dose, and route of administration of the antipsychotic treatment were recorded. Information about the number and the length of psychiatric hospitalization was retrieved. Ordinary Least Square (OLS) regression and ML algorithms (i.e., random forest [RF], supported vector machine, K-nearest neighborhood, and Naïve Bayes) were used to estimate the predictors of total antipsychotic dosage and prescription of antipsychotic polytherapy (APP).

Results

The strongest predictor of the total dose was APP. The number of Community Mental Health Centers (CMHC) contacts was the most important predictor of APP and, with APP omitted, of dosage. Treatment with anticholinergics predicted APP, emphasizing the strong correlation between APP and higher antipsychotic dose. RF performed better than OLS regression and the other ML algorithms in predicting both antipsychotic dose (root square mean error = 0.70, R2 = 0.31) and APP (area under the receiving operator curve = 0.66, true positive rate = 0.41, and true negative rate = 0.78).

Conclusion

APP is associated with the prescription of higher total doses of antipsychotics. Frequent attenders at CMHCs, and SUs recently hospitalized are often treated with APP and higher doses of antipsychotics. Future prospective studies incorporating standardized clinical assessments for both psychopathological severity and treatment efficacy are needed to confirm these findings.

SUBMITTER: Marchi M

PROVIDER: S-EPMC9329108 | biostudies-literature | 2022 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Machine-Learning for Prescription Patterns: Random Forest in the Prediction of Dose and Number of Antipsychotics Prescribed to People with Schizophrenia.

Marchi Mattia M Galli Giacomo G Fiore Gianluca G Mackinnon Andrew A Mattei Giorgio G Starace Fabrizio F Galeazzi Gian M GM

Clinical psychopharmacology and neuroscience : the official scientific journal of the Korean College of Neuropsychopharmacology 20220801 3

<h4>Objective</h4>We aimed to predict antipsychotic prescription patterns for people with schizophrenia using machine learning (ML) algorithms.<h4>Methods</h4>In a cross-sectional design, a sample of community mental health service users (SUs; n = 368) with a primary diagnosis of schizophrenia was randomly selected. Socio-demographic and clinical features, including the number, total dose, and route of administration of the antipsychotic treatment were recorded. Information about the number and ...[more]

PMID: 35879029

Similar Datasets

Project description:BackgroundPeople with schizophrenia-spectrum disorders (SSD) often have high levels of obesity and poor cardiometabolic health. Certain types of antipsychotics have been shown to contribute towards weight gain and there is some equivocal evidence that obesity is related to poor health-related quality of life (HRQoL) in people with SSD. It is also still uncertain if antipsychotic polypharmacy/higher doses of antipsychotics are linked with HRQoL and/or increased risk of obesity/Cardiovascular Disease (CVD). Therefore, this study aimed to examine potential relationships between prescribed antipsychotic medication regimens, cardiometabolic health risks and HRQoL in community-based Chinese people with SSD.MethodThis cross-sectional study reports the results of baseline measurements of a random sample of patients in an ongoing controlled trial of physical health intervention for people with severe mental illness. Data from these randomly-selected participants (n = 82) were analysed to calculate 10-year CVD relative-risk (using QRISK®2 score), estimate the prevalence of metabolic syndrome and contextualize patients' prescribed antipsychotics (types, combinations and Daily Defined Dose equivalent). Patients self-reported their HRQoL (SF12v2) and their obesity condition was assessed by waist-circumference and Body Mass Index (BMI).ResultsTwo-thirds of patients had a BMI ≥23 kg/m2, almost half were centrally obese and 29% met the criteria for metabolic syndrome. The individual relative-risk of CVD ranged from 0.62 to 15, and 13% had a moderate-to-high 10-year CVD risk score. Regression models showed that lower physical HRQoL was predicted by higher BMI and lower mental HRQoL. Higher Defined Daily Dose, clozapine, younger age and male gender were found to explain 40% of the variance in CVD relative risk.ConclusionThe findings indicate that cardiometabolic health risks in people with SSD may be more common than those reported in the general Hong Kong population. The results also provide further support for the need to consider antipsychotic polypharmacy and higher doses of antipsychotics as factors that may contribute towards cardiometabolic risk in Chinese patients with SSD. Clinicians in Hong Kong should consider using routine CVD risk screening, and be aware that younger male patients who are taking clozapine and prescribed higher Defined Daily Dose seem to have the highest relative-risk of CVD.Trial registrationClinicaltrials.gov NCT02453217 . Prospectively registered on 19th May 2015.

Project description:Introduction:While patients' perspectives toward pharmacotherapy are expected to be directly influenced by their motivation and understanding of the treatment that they are currently receiving, no study has comprehensively investigated the impact of insight into illness and knowledge for the ongoing pharmacotherapy on the attitude towards drug treatment among patients with schizophrenia. Materials and Methods:One hundred forty-eight Japanese outpatients diagnosed with schizophrenia, according to the International Classification of Diseases 10th edition, were included (mean±SD age, 47.3±12.4 years; 90 men (60.8%)). Attitudes toward antipsychotic treatment and insight into illness were assessed with the Drug Attitude Inventory-10 (DAI-10) and the VAGUS, respectively. In addition, a multiple-choice questionnaire that was designed to examine patients' knowledge about therapeutic effects, types, and implicated neurotransmitters of antipsychotic drugs they were receiving was utilized. Results:The mean±SD of DAI-10 score was 4.7±4.2. The multiple regression analysis found that lower Positive and Negative Syndrome Scale (PANSS) scores, higher VAGUS scores, and longer illness duration were significantly associated with higher DAI-10 scores (?=-0.226, P=0.009; ?=0.250, P=0.008; ?=0.203, P=0.034, respectively). There was a significant difference in the DAI-10 scores between the subjects who gave more accurate answers regarding the effects of their primary antipsychotic and those who did not (mean±SD, 5.57±4.38 vs 4.13±4.04, P=0.043); however, this finding failed to survive the multiple regression analysis. Conclusion:Better insight into illness and treatment, lower illness severity, longer illness duration, and possibly greater knowledge about the therapeutic effects of medications may lead to better attitudes towards pharmacotherapy among patients with schizophrenia, which has an important implication for this typically chronic mental condition requiring long-term antipsychotic treatment to sustain stability.

Project description:Background: A prolonged electrocardiogram (ECG) QT interval is associated with cardiac events and increased mortality. Antipsychotics can prolong the QT interval. The QT interval requires correction (QTc) for heart rate using a formula or QT-nomogram. The QT and QTc can be calculated automatically by the ECG machine or manually; however, machine-measured QT(c) intervals may be inaccurate. Objective: We aimed to investigate the mean QTc and proportion of prolonged QTc intervals in people taking antipsychotic medicines. Methods: We conducted an observational retrospective chart review and data analysis of all consecutive patients taking antipsychotics, with an ECG record, admitted to the psychiatric unit of a large tertiary hospital in Brisbane, Australia, between 1 January 2017 and 30 January 2019. We investigated the mean QTc of people taking antipsychotics to determine differences using (a) machine versus manual QT interval measurement and (b) QTc correction formulae (Bazett, Fridericia, Framingham, Hodges and Rautaharju) and the QT-nomogram. We also determined the number of people with a prolonged QTc using different methods and compared rates of prolonged QTc with antipsychotic monotherapy and polypharmacy. Results: Of 920 included people, the mean (±SD) machine-measured, Bazett-corrected QT interval (recorded from the ECG) was 435 ms (±27), significantly longer (p < 0.001) than the mean manually measured corrected QT intervals with Fridericia 394 ms (±24), Framingham 395 ms (±22), Hodges 398 ms (±22) and Rautaharju 400 ms (±24) formulae. There were significantly more people with a prolonged QTc using machine-measured QT and the Bazett formula (12.0%, 110/920) when compared with manually measured QT and the Fridericia formula (2.2%, 20/920) or QT-nomogram (0.7%, 6/920). Rates of QTc prolongation did not differ between people taking antipsychotic polypharmacy compared with monotherapy. Conclusion: Machine-measured QTc using the Bazett formula overestimates the QTc interval length and number of people with a prolonged QTc, compared with other formulae and the QT-nomogram. We recommend manually measuring the QT and correcting with the Fridericia formula or QT-nomogram prior to modifying antipsychotic therapies.

Project description:Since 2017, we have used IonTorrent NGS platform in our hospital to diagnose and treat cancer. Analyzing variants at each run requires considerable time, and we are still struggling with some variants that appear correct on the metrics at first, but are found to be negative upon further investigation. Can any machine learning algorithm (ML) help us classify NGS variants? This has led us to investigate which ML can fit our NGS data and to develop a tool that can be routinely implemented to help biologists. Currently, one of the greatest challenges in medicine is processing a significant quantity of data. This is particularly true in molecular biology with the advantage of next-generation sequencing (NGS) for profiling and identifying molecular tumors and their treatment. In addition to bioinformatics pipelines, artificial intelligence (AI) can be valuable in helping to analyze mutation variants. Generating sequencing data from patient DNA samples has become easy to perform in clinical trials. However, analyzing the massive quantities of genomic or transcriptomic data and extracting the key biomarkers associated with a clinical response to a specific therapy requires a formidable combination of scientific expertise, biomolecular skills and a panel of bioinformatic and biostatistic tools, in which artificial intelligence is now successful in developing future routine diagnostics. However, cancer genome complexity and technical artifacts make identifying real variants challenging. We present a machine learning method for classifying pathogenic single nucleotide variants (SNVs), single nucleotide polymorphisms (SNPs), multiple nucleotide variants (MNVs), insertions, and deletions detected by NGS from different types of tumor specimens, such as: colorectal, melanoma, lung and glioma cancer. We compared our NGS data to different machine learning algorithms using the k-fold cross-validation method and to neural networks (deep learning) to measure the performance of the different ML algorithms and determine which one is a valid model for confirming NGS variant calls in cancer diagnosis. We trained our machine learning with 70% of our data samples, extracted from our local database (our data structure had 7 parameters: chromosome, position, exon, variant allele frequency, minor allele frequency, coverage and protein description) and validated it with the 30% remaining data. The model offering the best accuracy was chosen and implemented in the NGS analysis routine. Artificial intelligence was developed with the R script language version 3.6.0. We trained our model on 70% of 102,011 variants. Our best error rate (0.22%) was found with random forest machine learning (ntree = 500 and mtry = 4), with an AUC of 0.99. Neural networks achieved some good scores. The final trained model with the neural network achieved an accuracy of 98% and an ROC-AUC of 0.99 with validation data. We tested our RF model to interpret more than 2000 variants from our NGS database: 20 variants were misclassified (error rate < 1%). The errors were nomenclature problems and false positives. After adding false positives to our training database and implementing our RF model routinely, our error rate was always < 0.5%. The RF model shows excellent results for oncosomatic NGS interpretation and can easily be implemented in other molecular biology laboratories. AI is becoming increasingly important in molecular biomedical analysis and can be very helpful in processing medical data. Neural networks show a good capacity in variant classification, and in the future, they may be useful in predicting more complex variants.

Project description:ObjectiveAntipsychotics are licensed for psychosis and are also prescribed for behavior control. This study aims to examine characteristics and outcomes of children prescribed antipsychotics.MethodsA cohort study using general practice and hospital records linked with education records for 1,488,936 children living in Wales between 1999 and 2015. The characteristics of the children who were prescribed antipsychotics are presented using descriptive statistics and outcomes such as respiratory illness, diabetes, and injury were analyzed using multilevel logistic regression and the prior event rate ratio (PERR).ResultsChildren with intellectual difficulty/autism were more likely to be prescribed antipsychotics (2.8% have been prescribed an antipsychotic [75% with autism] compared with 0.15% of children without intellectual difficulty). Those with intellectual disabilities/autism were prescribed antipsychotics at a younger age and for a longer period. Antipsychotic use was associated with a higher rate of respiratory illness for all (PERR of hospital admission: 1.55 [95% CI: 1.51-1.598] or increase in rate of 2 per 100 per year in those treated), and for those with intellectual difficulty/autism, there was a higher rate of injury and hospitalized depression. However, among those without intellectual difficulty/autism, there were lower rates of depression (PERR: 0.55 [95% CI: 0.51-0.59]).ConclusionsThis work shows real-world use of antipsychotics and provides information on the rate of possible adverse events in children treated. Antipsychotics are predominantly used for those with intellectual difficulty/autism rather than those with a psychotic diagnosis. There is evidence that rates of respiratory disease, epilepsy, and diabetes are also higher postantipsychotic use for all. In those with intellectual difficulty/autism, hospital-admitted depression and injury are higher postantipsychotic use. The use of antipsychotics for behavioral management is likely to have increased cost implications to the healthcare system.

Dataset Information

Machine-Learning for Prescription Patterns: Random Forest in the Prediction of Dose and Number of Antipsychotics Prescribed to People with Schizophrenia.

Objective

Methods

Results

Conclusion

Publications

Machine-Learning for Prescription Patterns: Random Forest in the Prediction of Dose and Number of Antipsychotics Prescribed to People with Schizophrenia.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets