Dataset Information

Building interpretable models for polypharmacy prediction in older chronic patients based on drug prescription records.

ABSTRACT:

Background

Multimorbidity presents an increasingly common problem in older population, and is tightly related to polypharmacy, i.e., concurrent use of multiple medications by one individual. Detecting polypharmacy from drug prescription records is not only related to multimorbidity, but can also point at incorrect use of medicines. In this work, we build models for predicting polypharmacy from drug prescription records for newly diagnosed chronic patients. We evaluate the models' performance with a strong focus on interpretability of the results.

Methods

A centrally collected nationwide dataset of prescription records was used to perform electronic phenotyping of patients for the following two chronic conditions: type 2 diabetes mellitus (T2D) and cardiovascular disease (CVD). In addition, a hospital discharge dataset was linked to the prescription records. A regularized regression model was built for 11 different experimental scenarios on two datasets, and complexity of the model was controlled with a maximum number of dimensions (MND) parameter. Performance and interpretability of the model were evaluated with AUC, AUPRC, calibration plots, and interpretation by a medical doctor.

Results

For the CVD model, AUC and AUPRC values of 0.900 (95% [0.898-0.901]) and 0.640 (0.635-0.645) were reached, respectively, while for the T2D model the values were 0.808 (0.803-0.812) and 0.732 (0.725-0.739). Reducing complexity of the model by 65% and 48% for CVD and T2D, resulted in 3% and 4% lower AUC, and 4% and 5% lower AUPRC values, respectively. Calibration plots for our models showed that we can achieve moderate calibration with reducing the models' complexity without significant loss of predictive performance.

Discussion

In this study, we found that it is possible to use drug prescription data to build a model for polypharmacy prediction in older population. In addition, the study showed that it is possible to find a balance between good performance and interpretability of the model, and achieve acceptable calibration at the same time.

SUBMITTER: Kocbek S

PROVIDER: S-EPMC6187991 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:The objectives of this study were to investigate older adults' knowledge of prescription drug safety and interactions with alcohol, and to identify pharmacists' willingness to disseminate prescription drug safety information to older adults.The convenience sample consisted of 48 older adults aged 54-89 years who were recruited from a local pharmacy and who completed surveys addressing their alcohol consumption, understanding of alcohol and prescription drug interactions, and willingness to change habits regarding alcohol consumption and prescription drugs. To address pharmacist willingness, 90 pharmacists from local pharmacies volunteered and answered questions regarding their willingness to convey prescription drug safety information to older adults.Older adults reported low knowledge of alcohol and prescription drug safety, with women tending to be slightly more knowledgeable. More importantly, those who drank in the previous few months were less willing to talk to family and friends about how alcohol can have harmful interactions with prescription drugs, or to be an advocate for safe alcohol and prescription drug use than those who had not had a drink recently. Pharmacists reported that they were willing to convey prescription drug safety information to older adults via a variety of formats, including displaying or distributing a flyer, and directly administering a brief intervention.In this study, older adults were found to have inadequate knowledge of prescription drug safety and interactions with alcohol, but pharmacists who regularly come in contact with older adults indicated that they were ready and willing to talk to older adults about prescription drug safety. Future research should focus on interventions whereby pharmacists disseminate prescription drug safety information to older adults in order to improve healthy prescription drug and alcohol behavior and reduce medical and health costs associated with interactions between alcohol and prescription drugs.

Project description:BackgroundAs survivors of childhood cancer age, development of cancer treatment-related chronic health conditions often occur. This study aimed to describe the pattern of chronic prescription medication use and identify factors associated with polypharmacy among survivors of childhood cancer.MethodsThis was a retrospective study conducted at the pediatric oncology long-term follow-up clinic in Hong Kong. Eligible subjects included survivors who were (1) diagnosed with cancer before 18 years old, (2) were at least 3 years post-cancer diagnosis and had completed treatment for at least 30 days, and (3) receiving long-term follow-up care at the study site between 2015 and 2018. Dispensing records of eligible survivors were reviewed to identify medications taken daily for ≥30 days or used on an "as needed" basis for ≥6 months cumulatively within the past 12-month period. Polypharmacy was defined as the concurrent use of ≥5 chronic medications. Multivariable log-binomial modeling was conducted to identify treatment and clinical factors associated with medication use pattern and polypharmacy.ResultsThis study included 625 survivors (mean current age = 17.9 years, standard deviation [SD] = 7.2 years) who were 9.2 [5.2] years post-treatment. Approximately one-third (n = 219, 35.0%) of survivors were prescribed at least one chronic medication. Frequently prescribed medication classes include systemic antihistamines (26.5%), sex hormones (19.2%), and thyroid replacement therapy (16.0%). Overall prevalence of polypharmacy was 5.3% (n = 33). A higher rate of polypharmacy was found in survivors of CNS tumors (13.6%) than in survivors of hematological malignancies (4.3%) and other solid tumors (5.3%) (P = .0051). Higher medication burden was also observed in survivors who had undergone cranial radiation (RR = 6.31; 95% CI = 2.75-14.49) or hematopoietic stem-cell transplantation (HSCT) (RR = 3.53; 95% CI = 1.59-7.83).ConclusionAlthough polypharmacy was observed in a minority of included survivors of childhood cancer, chronic medication use was common. Special attention should be paid to survivors of CNS tumors and survivors who have undergone HSCT or cranial radiation. These individuals should be monitored closely for drug-drug interactions and adverse health outcomes that may result from multiple chronic medications, particularly during hospitalization in an acute care setting.

Project description:BackgroundPolypharmacy is common in older people and associated with potential harms. The aim of this study was to analyse the characteristics of an older multimorbid population with polypharmacy and to identify factors contributing to excessive polypharmacy in these patients.MethodsThis cross-sectional analysis is based on the PRIMA-eDS trial, a large randomised controlled multicentre study of polypharmacy in primary care. Patients' baseline data were used for analysis. A number of socioeconomic and medical data as well as SF-12-scores were entered into a generalized linear mixed model to identify variables associated with excessive polypharmacy (taking ?10 substances daily).ResultsThree thousand nine hundred four participants were recruited. Risk factors significantly associated with excessive polypharmacy were frailty (OR 1.45; 95% CI 1.22-1.71), >?8 diagnoses (OR 2.64; 95% CI 2.24-3.11), BMI ?30 (OR 1.18; 95% CI 1.02-1.38), a lower SF-12 physical health composite score (OR 1.47; 95% CI 1.26-1.72), and a lower SF-12 mental health composite score (OR 1.33; 95% CI 1.17-1.59) than the median of the study population (?36.6 and???48.7, respectively). Age???85 years (OR 0.83; 95% CI 0.70-0.99) led to a significantly lower risk for excessive polypharmacy. No association with excessive polypharmacy could be found for female sex, low educational level, and smoking. Regarding the study centres, being recruited in the UK led to a significantly higher risk for excessive polypharmacy compared to being recruited in Germany 1/Rostock (OR 1.71; 95% CI 1.27-2.30). Being recruited in Germany 2/Witten led to a slightly significant lower risk for excessive polypharmacy compared to Germany 1/Rostock (OR 0.74; 95% CI 0.56-0.97).ConclusionsFrailty, multimorbidity, obesity, and decreased physical as well as mental health status are risk factors for excessive polypharmacy. Sex, educational level, and smoking apparently do not seem to be related to excessive polypharmacy. Physicians should especially pay attention to their frail, obese patients who have multiple diagnoses and a decreased health-related quality of life, to check carefully whether all the drugs prescribed are evidence-based, safe, and do not interact in an unfavourable way.Trial registrationThis trial has been registered with Current Controlled Trials Ltd. on 31 July 2014 (ISRCTN10137559).