Towards identifying emergency patients with unconcerning brain - PowerPoint PPT Presentation

Towards identifying emergency patients with unconcerning brain computed tomography scans Presented by Dr Aldo Saavedra Dr Madhura Killedar on behalf of Prof Jonathan Morris Dr Joel Nothman Seven Guney Dr Catherine Naidoo Dr Lucy Blumer Peter Thiem Dr Felicity Gallimore The University of Sydney Page 1

Introduction • Overall aims: • Harness the information captured by the eMR • Identify problems that can be solved with the data • Multi-disciplinary team from Sydney University and Royal North Shore Hospital (RNS). • Increasing number of presentations to the emergency department (ED) are placing • RNS is a tertiary teaching stress on an already busy health care hospital of the University of system (4% between 2017 and 2018) Sydney • It services 17% of the population in NSW The University of Sydney Page 2

Computed Tomography Scan – CT Scans - Focus on a resource intensive investigation: CT scans - At RNS there are 3 scanners with a daytime team 4-6 consultant radiologist and 6-10 registrars - One registrar after 8:pm - The actual scan can take 5-15 minute. - It can take 3-4 staff to transfer an elderly patient on/off the bed – 10 minutes. - A few minutes for write a report if radiologist is not interrupted. University of Virginia The University of Sydney Page 3

Motivation The focus • 69 different of our study types of scans performed • CT Brain scans by ED in correspond to ≈ 50% three years of all ED Scans The number of An increase of scans in Q1 of ≈ 13 per quarter 2013 was 387 The University of Sydney Page 4

Project goals using data Data subset: EMRs for 5600 encounters from RNS emergency between 2013 & 2016 in which CT brain scans were taken for patients >16 yrs old – are CT scan requests increasing with time? – do these CT scans have positive/negative findings? – can we classify findings based on text from the CT scan report? – is the information captured by the electronic medical record (eMR) during an emergency presentation enough to predict whether the scan will have a negative finding? The University of Sydney Page 5

Determining negative findings from CT scan report text – Two radiology registrars labelled 843 CT Brain performed on 29-MAY-2015 at 11:53 AM Dictated By: Dr X. Xxxxx reports manually Typed By: Dr X. Xxxxxx Approved By: Dr Yyyyy Yyyyyyy – patient demographics made available but Clinical History: not needed Unwitnessed fall at NH; on anticoagulants - clopidogrel. Findings: – conclusion text was sufficient No acute intracranial haemorrhage. A lacunar infarct is noted in the left basal ganglia (likely chronic) unaltered – high level of agreement since 23/05/2015. Bilateral periventricular white matter hypodensity is unaltered since the previous – Two classes of findings study compatible with chronic microvascular ischaemic change. Grey-white matter differentiation is otherwise – Negative if both registrars agree negative maintained. The ventricles and CSF spaces are enlarged in or expected for a given age keeping with volume loss. Patchy mucosal thickening involving the visualised paranasal – Positive : otherwise sinuses. No fracture of the skull vault or visualised facial bones. – Remain conservative Conclusion: No acute intracranial haemorrhage. The University of Sydney Page 6

Classifying CT scan findings 1. Extract conclusion text 2a. Rule-based system assesses sentences, No acute intracranial eliminates if describing normality or absence of pathology identified. No finding, looks for any remaining space-occupying lesion • very high sensitivity identified. 2b. Machine learning system used bag-of- No acute intracranial words features pathology identified. No • more balanced sensitivity and specificity space-occupying lesion identified. 3 . Hybrid machine learning classifier (used rule-based prediction as ML feature) – trained against manual classification – very high sensitivity (93%) , but improved specificity (96%) The University of Sydney Page 7

Increasing Negative Findings? 14% 5% 14% 18% 3% 5% 81% 81% 77% 82% 80% 80% 81% 77% 83% 80% 81% 76% The University of Sydney Page 8

Information from EMR Demographics – age, gender, indigenous status, preferred language Vital Signs ( on arrival) – blood pressure, SpO2, respiratory rate Pathology test results – Haemoglobin, Platelet Count, White Cell Count, C-Reactive Protein, Creatinine, AST, PT, APTT, INR, Glucose (Rand) Emergency Triage Form (Presenting Problem) – headache, falls, injury, head, dizziness, pain, limb, altered, syncope, sensation, weakness, faint, seizure, eye, mh, abnormal, vision, loc, care, speech, difficulties, chest, facial Emergency Triage Form (Presenting Information for Tracking Board) – pain, headache, head, fall, loc, alert, biba, well, neck, onset, nausea, limb, dizziness, chest, facial, side, perfused, hit, equal, weakness, limbs, vision, sided, back, arm, speech, droop • Clinical consultation to determine relevant fields • Raw data categorized relative to medical standard reference values • Ensure data is recorded before CT scan • Mutual information points at redundancies The University of Sydney Page 9

Cohorts of patient encounters Cohort Cohort Cohort Cohort Distinguishing feature: presenting problem The University of Sydney Page 10

Most common presenting problems Presenting Number of # encounters with % encounters with problem encounters positive findings positive findings Headache 999 129 13% Falls 866 86 10% Injury 719 74 10% Dizziness 450 46 10% 283 36 13% Limb Syncope 244 30 12% Weakness 221 35 16% Pain 290 30 10% 257 26 10% Altered Seizure 172 20 12% Eye 154 17 11% Mental Health 133 8 6% The University of Sydney Page 11

Learning and Predicting – Predict +ve/-ve CT scan findings via a percentage probability of +ve – Bayesian Generalized Linear Model – performs logistic regression – each of EMR fields are features – trained against CT scan report classification – Bayesian framework accounts for uncertainties, and model removal of features – Calibrate separately for each presenting problem: headache, falls, injury – Learns from 80% of data (2400 encounters) – Predicts for each of remaining 20% (600 encounters) The University of Sydney Page 12

Evaluating the machine Combined 600 results for encounters: headaches, falls, injury Set conservative threshold: 2% of positive findings fall below 28 out of 482 (6%) negative findings are identified Corresponds to 300 encounters from total dataset of 5600 Likely negative Likely positive The University of Sydney Page 13

Summary – Increase in brain CT scans at RNS – Large value in electronic medical records – machine learning on report text to classify CT scan findings – characteristics of encounters – identifying cohorts – predict findings for patient encounters – Potential to increase sensitivity: – other models – incorporate previous studies – varied diagnoses – additional fields in EMR The University of Sydney Page 14