Dataset Information

Predicting incident delirium diagnoses using data from primary-care electronic health records.

ABSTRACT:

Importance

risk factors for delirium in hospital inpatients are well established, but less is known about whether delirium occurring in the community or during an emergency admission to hospital care might be predicted from routine primary-care records.

Objectives

identify risk factors in primary-care electronic health records (PC-EHR) predictive of delirium occurring in the community or recorded in the initial episode in emergency hospitalisation. Test predictive performance against the cumulative frailty index.

Design

Stage 1: case-control; Stages 2 and 3: retrospective cohort.

Setting

clinical practice research datalink: PC-EHR linked to hospital discharge data from England.

Subjects

Stage 1: 17,286 patients with delirium aged ?60 years plus 85,607 controls. Stages 2 and 3: patients???60 years (n =?429,548 in 2015), split into calibration and validation groups.

Methods

Stage 1: logistic regression to identify associations of 110 candidate risk measures with delirium. Stage 2: calibrating risk factor weights. Stage 3: validation in independent sample using area under the curve (AUC) receiver operating characteristic.

Results

fifty-five risk factors were predictive, in domains including: cognitive impairment or mental illness, psychoactive drugs, frailty, infection, hyponatraemia and anticholinergic drugs. The derived model predicted 1-year incident delirium (AUC?=?0.867, 0.852:0.881) and mortality (AUC?=?0.846, 0.842:0.853), outperforming the frailty index (AUC?=?0.761, 0.740:0.782). Individuals with the highest 10% of predicted delirium risk accounted for 55% of incident delirium over 1 year.

Conclusions

a risk factor model for delirium using data in PC-EHR performed well, identifying individuals at risk of new onsets of delirium. This model has potential for supporting preventive interventions.

SUBMITTER: Bowman K

PROVIDER: S-EPMC7297278 | biostudies-literature | 2020 Apr

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Predicting incident delirium diagnoses using data from primary-care electronic health records.

Bowman Kirsty K Jones Lindsay L Masoli Jane J Mujica-Mota Ruben R Strain David D Butchart Joe J Valderas José M JM Fortinsky Richard H RH Melzer David D Delgado João J

Age and ageing 20200401 3

<h4>Importance</h4>risk factors for delirium in hospital inpatients are well established, but less is known about whether delirium occurring in the community or during an emergency admission to hospital care might be predicted from routine primary-care records.<h4>Objectives</h4>identify risk factors in primary-care electronic health records (PC-EHR) predictive of delirium occurring in the community or recorded in the initial episode in emergency hospitalisation. Test predictive performance agai ...[more]

PMID: 32239180

Similar Datasets

Project description:ContextEarly childhood caries (ECC) is a persistent public health challenge, affecting more than 56% of US toddlers and preschool-aged children. Despite this, ECC is largely preventable with routine oral hygiene practices, diet, and application of topical fluoride.ObjectiveThis study assessed the utilization of preventive oral health care in primary care practices and evaluated the variation in patient characteristic and geographic disparities.DesignWe conducted a retrospective study using electronic health records (EHRs) over a 2-year period. Patients' home addresses were geocoded and linked to census-based neighborhood statistics and fluoridated water accessibility. Multiple logistic regression modeling was used to assess the risk of ECC in patients with fluoride preventive care, controlled for demographics, comorbid conditions, and neighborhood risk factors.ParticipantsPatients aged 6 to 71 months who had primary care providers at family medicine and general pediatric clinics in a large academic medical center.Main outcome measureThe presence of dental caries based on diagnoses in EHRs.ResultsThe study consisted of 10 836 patients: 17% treated with topical fluoride varnish (TFV), 12% prescribed oral fluoride supplement, 6.1% with both TFV and supplement, and 64% without fluoride treatment. Patients with fluoride treatment were 24% to 53% less likely to have ECC. Children living in rural and nonfluoridated water communities had 1.7 to 1.8 times greater risk of developing ECC. Minority, under/uninsured, and low-income patients also were at an increased risk of ECC.ConclusionDespite continuing efforts to improve access to dental care for vulnerable populations, substantial disparities remain among socioeconomically disadvantaged children. To address dental care shortage, primary care clinicians should serve as the safety net to care for vulnerable and underserved children who have no or limited access to oral health services. Future research into the collaboration between primary care and dental providers at the level of both practice and professional education should be considered.

Project description:Timely outreach to individuals in an advanced stage of illness offers opportunities to exercise decision control over health care. Predictive models built using Electronic health record (EHR) data are being explored as a way to anticipate such need with enough lead time for patient engagement. Prior studies have focused on hospitalized patients, who typically have more data available for predicting care needs. It is unclear if prediction driven outreach is feasible in the primary care setting. In this study, we apply predictive modeling to the primary care population of a large, regional health system and systematically examine the impact of technical choices, such as requiring a minimum number of health care encounters (data density requirements) and aggregating diagnosis codes using Clinical Classifications Software (CCS) groupings to reduce dimensionality, on model performance in terms of discrimination and positive predictive value. We assembled a cohort of 349,667 primary care patients between 65 and 90?years of age who sought care from Sutter Health between July 1, 2011 and June 30, 2014, of whom 2.1% died during the study period. EHR data comprising demographics, encounters, orders, and diagnoses for each patient from a 12?month observation window prior to the point when a prediction is made were extracted. L1 regularized logistic regression and gradient boosted tree models were fit to training data and tuned by cross validation. Model performance in predicting one year mortality was assessed using held-out test patients. Our experiments systematically varied three factors: model type, diagnosis coding, and data density requirements. We found substantial, consistent benefit from using gradient boosting vs logistic regression (mean AUROC over all other technical choices of 84.8% vs 80.7% respectively). There was no benefit from aggregation of ICD codes into CCS code groups (mean AUROC over all other technical choices of 82.9% vs 82.6% respectively). Likewise increasing data density requirements did not affect discrimination (mean AUROC over other technical choices ranged from 82.5% to 83%). We also examine model performance as a function of lead time, which is the interval between death and when a prediction was made. In subgroup analysis by lead time, mean AUROC over all other choices ranged from 87.9% for patients who died within 0 to 3?months to 83.6% for those who died 9 to 12?months after prediction time.

Project description:BackgroundThe mission of medical schools is a sustainable commitment to orient education, research, and services based on the priorities and expectations of society. The most common complaints of patients from comprehensive health service centers (CHSCs) based on the data from electronic health records were assessed in order to determine primary health care (PHC) priorities for the educational planning of medical students in Iran.MethodsA population-based national study was designed to assess clinical complaints of patients in all age groups who were referred to CHSCs at least once to be visited by physicians. All the data in the census were extracted from electronic health records in PHC system during 2015-2020, classified by the International Classification of Primary Care 2nd edition (ICPC-2e-English), and statistically analyzed. The total number of complaints that were recorded in the system was 17,430,139.Results59% of the referring patients were women. The highest number of referrals was related to the age group of 18-59 years (56.9%), while the lowest belonged to the elderly people (13.3%). In all age and sex groups, the first ten complaints of patients with three top priorities in each category included process (follow-up, consultation, and results exam), digestive (toothache and gum complaint, abdominal pain, and diarrhea), respiratory (cough, sore throat, and runny nose), general (fever, pain, and weakness and fatigue), musculoskeletal (back pain, leg complaint, and knee injuries), endocrine and nutritional (weight gain, Feeding problem, and weight loss), cardiovascular (hypertension, palpitations, and Postural hypotension), neurological (headache, dizziness, and paralysis), sexual dysfunction (vaginal complaint, discharge, and irregular menstruation), and dermatological (pruritus, rash, and inflammation) problems.ConclusionHigh priorities in referring to PHC had a key role in assessing the country's health needs. Since this study was in line with the national pattern of complaints and patients' profile, the present findings can be helpful to amend policy-making, educational planning and curricula development in medical schools.

Project description:IntroductionThe extent of preventable medication-related hospital admissions and medication-related issues in primary care is significant enough to justify developing decision support systems for medication safety surveillance. The prerequisite for such systems is defining a relevant set of medication safety-related indicators and understanding the influence of both patient and general practice characteristics on medication prescribing and monitoring.ObjectiveThe aim of the study was to investigate the feasibility of linked primary and secondary care electronic health record data for surveillance of medication safety, examining not only prescribing but also monitoring, and associations with patient- and general practice-level characteristics.MethodsA cross-sectional study was conducted using linked records of patients served by one hospital and over 50 general practices in Salford, UK. Statistical analysis consisted of mixed-effects logistic models, relating prescribing safety indicators to potential determinants.ResultsThe overall prevalence (proportion of patients with at least one medication safety hazard) was 5.45 % for prescribing indicators and 7.65 % for monitoring indicators. Older patients and those on multiple medications were at higher risk of prescribing hazards, but at lower risk of missed monitoring. The odds of missed monitoring among all patients were 25 % less for males, 50 % less for patients in practices that provide general practitioner training, and threefold higher in practices serving the most deprived compared with the least deprived areas. Practices with more prescribing hazards did not tend to show more monitoring issues.ConclusionsSystematic collection, collation, and analysis of linked primary and secondary care records produce plausible and useful information about medication safety for a health system. Medication safety surveillance systems should pay close attention to patient age and polypharmacy with respect to both prescribing and monitoring failures; treat prescribing and monitoring as different statistical processes, rather than a combined measure of prescribing safety; and audit the socio-economic equity of missed monitoring.

Dataset Information

Predicting incident delirium diagnoses using data from primary-care electronic health records.

Importance

Objectives

Design

Setting

Subjects

Methods

Results

Conclusions

Publications

Predicting incident delirium diagnoses using data from primary-care electronic health records.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets