Using machine learning to model older adult inpatient trajectories from electronic health records data.

Electronic Health Records (EHR) data can provide novel insights into inpatient trajectories. Blood tests and vital signs from de-identified patients' hospital admission episodes (AE) were represented as multivariate time-series (MVTS) to train unsupervised Hidden Markov Models (HMM) and represent each AE day as one of 17 states. All HMM states were clinically interpreted based on their patterns of MVTS variables and relationships with clinical information. Visualization differentiated patients progressing toward stable 'discharge-like' states versus those remaining at risk of inpatient mortality (IM). Chi-square tests confirmed these relationships (two states associated with IM; 12 states with ≥1 diagnosis). Logistic Regression and Random Forest (RF) models trained with MVTS data rather than states had higher prediction performances of IM, but results were comparable (best RF model AUC-ROC: MVTS data = 0.85; HMM states = 0.79). ML models extracted clinically interpretable signals from hospital data. The potential of ML to develop decision-support tools for EHR systems warrants investigation.

Anticholinergic burden in older adult inpatients: patterns from admission to discharge and associations with hospital outcomes.

BACKGROUND: Anticholinergic medications are associated with adverse outcomes in older adults and should be prescribed cautiously. We describe the Anticholinergic Risk Scale (ARS) scores of older inpatients and associations with outcomes. METHODS: We included all emergency, first admissions of adults ⩾65 years old admitted to one hospital over 4 years. Demographics, discharge specialty, dementia/history of cognitive concern, illness acuity and medications were retrieved from electronic records. ARS scores were calculated as the sum of anticholinergic potential for each medication (0 = limited/none; 1 = moderate; 2 = strong and 3 = very strong). We categorised patients based on admission ARS score [ARS = 0 (reference); ARS = 1; ARS = 2; ARS ⩾ 3] and change in ARS score from admission to discharge [admission and discharge ARS = 0 (reference); same; decreased; increased]. We described anticholinergic prescribing patterns by discharge specialty and explored multivariable associations between ARS score categories and mortality using logistic regression [odds ratios (ORs), 95% confidence intervals (CIs)]. RESULTS: From 33,360 patients, 10,183 (31%) were prescribed an anticholinergic medication on admission. Mean admission ARS scores were: Cardiology and Stroke = 0.56; General Medicine = 0.78; Geriatric Medicine = 0.83; Other medicine = 0.81; Trauma and Orthopaedics = 0.66; Other Surgery = 0.65. Mean ARS did not increase from admission to discharge in any specialty but reductions varied significantly, from 4.6% (Other Surgery) to 27.7% (Geriatric Medicine) (p < 0.001). The odds of both 30-day inpatient and 30-day post-discharge mortality increased with admission ARS = 1 (OR = 1.21, 95% CI 1.01-1.44 and OR = 1.44, 1.18-1.74) but not with ARS = 2 or ARS ⩾ 3. The odds of 30-day post-discharge mortality were higher in all ARS change categories, relative to no anticholinergic exposure (same: OR = 1.45, 1.21-1.74, decreased: OR = 1.27, 1.01-1.57, increased: OR = 2.48, 1.98-3.08). CONCLUSION: The inconsistent dose-response associations with mortality may be due to confounding and measurement error which may be addressed by a prospective trial. Definitive evidence for this prevalent modifiable risk factor is required to support clinician behaviour-change, thus reducing variation in anticholinergic deprescribing by inpatient speciality. PLAIN LANGUAGE SUMMARY: We describe how commonly medicines which block the chemical acetylcholine are prescribed to older adults admitted to hospital as an emergency and explore links between these medicines and death during or soon after hospital admission Backgroud: Medicines which block the chemical acetylcholine are commonly prescribed to treat symptoms such as itch and difficulty sleeping or to treat medical conditions such as depression. However, some studies in older adults have found potential links between these medicines and confusion and falls. Therefore, doctors are recommended to prescribe these drugs cautiously in adults aged 65 years and over.Methods: In our paper we use data collected as part of routine medical care at one university hospital to describe how often these medicines are prescribed in a large sample of older adults admitted to hospital as an emergency. We look at the medicines patients are prescribed on admission to the hospital and also when they are later discharged.Results: We find that these medicines are frequently prescribed. We also find that, in general, patients are prescribed fewer of these potentially harmful medicines on hospital discharge compared with hospital admission. This suggests that clinicians are aware of advice to prescribe acetylcholine blocking medicines cautiously and they are more often stopped in hospital than started. However, we find a lot of variation in practice depending on which hospital specialty was caring for the patient during their inpatient stay. We also find potential links with these medicines and death during the admission or soon after hospital discharge, but these potential links are not always consistent.Conclusion: Further study is needed to fully understand links between medicines that block acetylcholine and late life health. This will be important to reduce variation in prescribing practices.

A Minimal Information Model for Potential Drug-Drug Interactions.

Despite the significant health impacts of adverse events associated with drug-drug interactions, no standard models exist for managing and sharing evidence describing potential interactions between medications. Minimal information models have been used in other communities to establish community consensus around simple models capable of communicating useful information. This paper reports on a new minimal information model for describing potential drug-drug interactions. A task force of the Semantic Web in Health Care and Life Sciences Community Group of the World-Wide Web consortium engaged informaticians and drug-drug interaction experts in in-depth examination of recent literature and specific potential interactions. A consensus set of information items was identified, along with example descriptions of selected potential drug-drug interactions (PDDIs). User profiles and use cases were developed to demonstrate the applicability of the model. Ten core information items were identified: drugs involved, clinical consequences, seriousness, operational classification statement, recommended action, mechanism of interaction, contextual information/modifying factors, evidence about a suspected drug-drug interaction, frequency of exposure, and frequency of harm to exposed persons. Eight best practice recommendations suggest how PDDI knowledge artifact creators can best use the 10 information items when synthesizing drug interaction evidence into artifacts intended to aid clinicians. This model has been included in a proposed implementation guide developed by the HL7 Clinical Decision Support Workgroup and in PDDIs published in the CDS Connect repository. The complete description of the model can be found at https://w3id.org/hclscg/pddi.

Minocycline and the risk of acute psychiatric events in adolescence: A self-controlled case series.

BACKGROUND: Minocycline has neurological anti-inflammatory properties and has been hypothesised to have antipsychotic effects. AIM: The aim of this study was to investigate, using routinely collected United Kingdom primary health care data, whether adolescent men and women are more or less likely to receive an urgent psychiatric referral during treatment for acne with minocycline compared with periods of non-treatment. METHOD: A self-controlled case series using United Kingdom Clinical Practice Research Datalink to calculate the incidence rate ratio of urgent psychiatric referrals for individuals, comparing periods during which minocycline was prescribed with unexposed periods, adjusted for age. RESULTS: We found 167 individuals who were at the time exposed to minocycline for a mean of 99 days and who received an urgent psychiatric referral. There was no difference in psychiatric referral risk during periods of exposure compared with periods of non-exposure: incidence rate ratio first 6 weeks of exposure 1.96, 95% confidence interval 0.82-4.71, p=0.132; incidence rate ratio remaining exposure period=1.97, 95% confidence interval 0.86-4.47, p=0.107. CONCLUSIONS: We found no evidence in support of a protective effect of minocycline against severe psychiatric symptoms in adolescence.

Examining the potential preventative effects of minocycline prescribed for acne on the incidence of severe mental illnesses: A historical cohort study.

BACKGROUND: Animal studies suggest that the antibiotic and microglial activation inhibitor, minocycline, is likely to have a protective effect against the emergence of psychosis but evidence from human studies is lacking. The aim of this study is to examine the effects of exposure to minocycline during adolescence on the later incidence of severe mental illness (SMI). METHODS: A historical cohort study using electronic primary care data was conducted to assess the association between exposure to minocycline during adolescence and incidence of SMI. The Incidence Rate Ratio (IRR) was measured using Poisson regression adjusted for age, gender, time of exposure, socioeconomic deprivation status, calendar year and co-medications. RESULTS: Early minocycline prescription ( n=13,248) did not affect the incidence of SMI compared with non-prescription of minocycline ( n=14,393), regardless of gender or whether or not the data were filtered according to a minimum exposure period (minimum period: IRR 0.96; 95% CI 0.68-1.36; p=0.821; no minimum period: IRR 1.08; 95% CI 0.83-1.42; p=0.566). CONCLUSIONS: Exposure to minocycline for acne treatment during adolescence appears to have no effect on the incidence of SMI.

DINTO: Using OWL Ontologies and SWRL Rules to Infer Drug-Drug Interactions and Their Mechanisms.

The early detection of drug-drug interactions (DDIs) is limited by the diffuse spread of DDI information in heterogeneous sources. Computational methods promise to play a key role in the identification and explanation of DDIs on a large scale. However, such methods rely on the availability of computable representations describing the relevant domain knowledge. Current modeling efforts have focused on partial and shallow representations of the DDI domain, failing to adequately support computational inference and discovery applications. In this paper, we describe a comprehensive ontology for DDI knowledge (DINTO), which is the first formal representation of different types of DDIs and their mechanisms and its application in the prediction of DDIs. This project has been developed using currently available semantic web technologies, standards, and tools, and we have demonstrated that the combination of drug-related facts in DINTO and Semantic Web Rule Language (SWRL) rules can be used to infer DDIs and their different mechanisms on a large scale. The ontology is available from https://code.google.com/p/dinto/.

Lessons learnt from the DDIExtraction-2013 Shared Task.

The DDIExtraction Shared Task 2013 is the second edition of the DDIExtraction Shared Task series, a community-wide effort to promote the implementation and comparative assessment of natural language processing (NLP) techniques in the field of the pharmacovigilance domain, in particular, to address the extraction of drug-drug interactions (DDI) from biomedical texts. This edition has been the first attempt to compare the performance of Information Extraction (IE) techniques specific for each of the basic steps of the DDI extraction pipeline. To attain this aim, two main tasks were proposed: the recognition and classification of pharmacological substances and the detection and classification of drug-drug interactions. DDIExtraction 2013 was held from January to June 2013 and attracted wide attention with a total of 14 teams (6 of the teams participated in the drug name recognition task, while 8 participated in the DDI extraction task) from 7 different countries. For the task of the recognition and classification of pharmacological names, the best system achieved an F1 of 71.5%, while, for the detection and classification of DDIs, the best result was an F1 of 65.1%. The results show advances in the state of the art and demonstrate that significant challenges remain to be resolved. This paper focuses on the second task (extraction of DDIs) and examines its main challenges, which have yet to be resolved.

The DDI corpus: an annotated corpus with pharmacological substances and drug-drug interactions.

The management of drug-drug interactions (DDIs) is a critical issue resulting from the overwhelming amount of information available on them. Natural Language Processing (NLP) techniques can provide an interesting way to reduce the time spent by healthcare professionals on reviewing biomedical literature. However, NLP techniques rely mostly on the availability of the annotated corpora. While there are several annotated corpora with biological entities and their relationships, there is a lack of corpora annotated with pharmacological substances and DDIs. Moreover, other works in this field have focused in pharmacokinetic (PK) DDIs only, but not in pharmacodynamic (PD) DDIs. To address this problem, we have created a manually annotated corpus consisting of 792 texts selected from the DrugBank database and other 233 Medline abstracts. This fined-grained corpus has been annotated with a total of 18,502 pharmacological substances and 5028 DDIs, including both PK as well as PD interactions. The quality and consistency of the annotation process has been ensured through the creation of annotation guidelines and has been evaluated by the measurement of the inter-annotator agreement between two annotators. The agreement was almost perfect (Kappa up to 0.96 and generally over 0.80), except for the DDIs in the MedLine database (0.55-0.72). The DDI corpus has been used in the SemEval 2013 DDIExtraction challenge as a gold standard for the evaluation of information extraction techniques applied to the recognition of pharmacological substances and the detection of DDIs from biomedical texts. DDIExtraction 2013 has attracted wide attention with a total of 14 teams from 7 different countries. For the task of recognition and classification of pharmacological names, the best system achieved an F1 of 71.5%, while, for the detection and classification of DDIs, the best result was F1 of 65.1%. These results show that the corpus has enough quality to be used for training and testing NLP techniques applied to the field of Pharmacovigilance. The DDI corpus and the annotation guidelines are free for use for academic research and are available at http://labda.inf.uc3m.es/ddicorpus.