CLINICAL GUIDANCE FRAMEWORKS Protracted investigation of patients - - PDF document

21/07/2017 1

IMPROVING CARE PROCESSES FOR PATIENTS WITH POSSIBLE SUSPECTED ACUTE CORONARY SYNDROME (ICARE-ACS): A NATIONAL IMPLEMENTATION TRIAL OF A CLINICAL GUIDANCE FRAMEWORK

S Aldous; on behalf of the IcareACS team

CLINICAL GUIDANCE FRAMEWORKS

Protracted investigation of patients presenting to emergency departments with symptoms suggesting acute coronary syndromes (ACS) is a major hospital burden. Clinical Guidance Frameworks, which assist medical decision making, can improve the accuracy of diagnosis and facilitate more timely definitive care, or discharge from hospital if significant pathology is excluded

21/07/2017 2

ACCELERATED DIAGNOSTIC PROTOCOLS

A component of the guidance framework, an Accelerated Diagnostic Protocol (ADP), can provide direction regarding clinical risk stratification and interpretation of early biochemical markers so that an earlier, safe discharge from hospital can be achieved. An ADP can combine variables usually including cardiac troponin (cTn) results, ECG results and objective scoring of clinical variables. ADPs allow blood sampling for cTn measurement at early time-points in patients who are low risk by the other variables. This allows clinicians to rapidly proceed to the same ‘next step’ in clinical management (such as cardiac stress test, imaging and/or discharge) as would have occurred using a more prolonged time course for serial troponin testing to exclude AMI.

ADP RESULTS FROM CHRISTCHURCH - 2011

The ASia Pacific Evaluation of Chest pain Trial (ASPECT) was an observational study of 3582 patients presenting with to the ED with acute chest pain. Christchurch Hospital contributed 1000 patients. 421 (11·8%) patients had 30 day MACE. 9.8% were classified low risk by the ASPECT ADP  Modified TIMI=0  No new ischaemic ECG changes  No elevation of any 0/2h biomarkers in a POC multimarker panel. MACE occurred in 0·9% of low risk patients  Sensitivity 99·3% (95% CI 97·9–99·8)  NPV 99·1% (97·3–99·8)

21/07/2017 3

ADP RESULTS FROM CHRISTCHURCH - 2012

2h Accelerated Diagnostic protocol to Assess Patients with chest pain symptoms using contemporary Troponins as the Only biomarker (ADAPT) was a sub-study of ASPECT including 1975 patients. 302 (15.3%) had a 30 day MACE. 19.8% were classified as low risk by the ADAPT ADP  Modified TIMI=0  No new ischaemic ECG changes  No elevation of 0/2h laboratory cTn MACE occurred in 1 (0.25%) of low risk patients  Sensitivity 99.7% (98.1% to 99.9%)  NPV 99.7% (98.6% to 100.0%)

ADP RESULTS FROM CHRISTCHURCH – 2014-15

Development of new risk stratification ADP (EDACS) ADAPT cohort (1974 patients, 30 day MACE rate 15.5%) 37 potential candidate variables including age, gender, symptoms, PMH/risk factors, medication, observations and using boot strapping to aim for sensitivity of 99% for 30 day MACE whilst maintaining the highest specificity possible.  Sensitivity 99.0% (96.9-99.7); Low risk 42.4% (ASPECT 9.8%, ADAPT 19.8%) . Validated: Combined TIMI RCT and Australian cohort (608 patients, 30 day MACE rate 13.0%)  Sensitivity 100% (94.2-100); Low risk population of 51.3% Vancouver cohort (763 patients, 10% AMI)  Sensitivity 100% (94.2-100); Low risk population of 41.6%

21/07/2017 4

MINISTRY OF HEALTH

This successful transition at Christchurch Hospital prompted the New Zealand Ministry

• f Health to mandate all hospitals to implement Clinical Guidance Frameworks for

possible ACS. The actual Clinical Guidance Framework or ADP to be used is not specified by the Ministry as minimal data exists of head-to-head comparisons (even observationally) between strategies.

AIM

To evaluate the safety and effectiveness of adopting Clinical Guidance Frameworks for possible ACS across the New Zealand National Health system by evaluating their adoption in 7 diverse hospitals. The hypothesis was that introducing the guidance frameworks would increase the proportion of patients safely discharged home within six hours of presentation to ED.

21/07/2017 5

METHODS - DESIGN

The study design was a pragmatic stepped-wedge implementation trial of guidance frameworks with 6 months control (or pre-implementation) and at least 4 months intervention (or post implementation). A stepped-wedge design involves sites beginning the study one after another. This means that later sites could benefit from lessons learnt from earlier sites on how to best manage the change of practice.

METHODS - SETTING

Sites (acute hospitals) were included if they were willing to participate and intended to implement a guidance framework incorporating an ADP by 1 September 2015. This allowed at least four months post-implementation data to be collected from all sites by the end of the study. Sites were not eligible if they had a pre-existing clinical pathway or ADP for suspected ACS. There were no restrictions regarding cTn assay/ADP or clinical guidance framework.

21/07/2017 6

METHODS - IMPLEMENTATION

Participating sites were presented with evidence regarding published ADPs including ADAPT, EDACS, HEART, NACPR, TRAPID-AMI and the Vancouver Chest pain rule, and then chose which ADP they would use. Blood sampling for cTn was to be done on arrival and then at either 1, 2 or 3 hours after presentation Pre-implementation, there were meetings involving primary investigator MT and representatives from each stakeholder group Stakeholders then planned and facilitated implementation at each site. Clinical and management leaders were identified within the local health system. These people remained in contact throughout the process. In one hospital there was a local independent change management process.

METHODS – DATA COLLECTION

The data collection process was pre-planned and developed before study commencement. Participants were identified from electronic diagnostic laboratory records of all patients who had blood drawn in the ED for a troponin test. Each laboratory record is linked to a patient’s NHI. Using the NHI, each hopitals Decision Support department could pull basic demographic data including age, gender and ethnicity as well as follow-up data on all readmission ICD10 codes. The national mortality dataset (Ministry of Health) was used to capture deaths within the 30 day follow-up time frame. Additional data was collected on those discharged within 6 hours coded as having an adverse event by contacting the local hospital clinicians, GPs and/or patients.

21/07/2017 7

METHODS - OUTCOMES

The primary outcome: Proportion “successfully” discharged home within 6h (no major adverse cardiac event (MACE) during the following 30 days). MACE included:  Death  Cardiac arrest, (ICD10 codes I46.0, 46.1, 46.9)  Emergency revascularization procedure  Cardio-genic shock (R57.0)  Ventricular arrhythmia (I47.2), ventricular fibrillation (I49.0)  High-degree atrioventricular block needing intervention (I44.2)  Acute myocardial infarction (I21.0, 21.1, 21.2, 21.3, 21.4, 21.9, 22.0, 22.1, 22.8, 22.9). Secondary outcomes included:  Proportion of patients discharged within 6 hours who had a MACE within 30 days  Lengths of stay in hospital in patients without ACS

RESULTS - SITES

21/07/2017 8

RESULTS – SITE DEMOGRAPHICS

Site ED attendance Total people in DHB Waikato 68,383 359,310 Hutt 44,470 138,378 North Shore 49,600 525,555 Wellington 52,146 283,704 Wairarapa 15,841 41,112 Waitakere 41,482 525,555 Middlemore 96,764 469,293

RESULTS – PATIENT DEMOGRAPHICS

Pre-Implementation Post-implementation Site Number % female % Maori Mean age Number % female % Maori Mean age Waikato 1495 45.3 15.1 66.8 5039 44.5 15.9 66.8 Hutt 844 44.8 12.2 67.2 1738 44.6 12.0 65.4 North Shore 2820 48.4 4.9 66.8 2514 47.9 4.7 66.2 Wellington 1266 43.8 7.4 67.6 2320 45.3 7.5 66.2 Wairarapa 284 42.3 3.2 70.2 395 41.8 1.5 69.6 Waitakere 1355 49.2 9.9 63.4 1409 47.7 11.8 63.8 Middlemore 3465 46.1 14.2 61.9 3135 43.6 13.4 63.1

21/07/2017 9

RESULTS – CHOSEN ADP

Site cTn Assay ECG criteria Risk stratification Waikato hs-cTnT at 0/2h Ischaemic (not known to be old) EDACS Hutt hs-cTnT at 0/2h Ischaemic (not known to be old) EDACS North Shore cTnI at 0/2h Ischaemic TIMI Wellington hs-cTnT at 0/2h Ischaemic (not known to be old) EDACS Wairarapa hs-cTnT at 0/2h Ischaemic (not known to be old) EDACS Waitakere cTnI at 0/2h Ischaemic TIMI Middlemore hs-cTnI at 0/3h Dynamic ST changes EDACS

RESULTS – PRIMARY OUTCOME

Pre-Implementation Post-implementation Site % MACE % Successful discharge within 6h % MACE % Successful discharge within 6h Waikato 11.8 5.9 15.2 15.0 Hutt 11.4 37.2 10.9 43.6 North Shore 14.8 2.8 12.1 11.8 Wellington 14.1 9.9 12.5 38.2 Wairarapa 13.7 36.3 13.7 38.7 Waitakere 10.6 5.4 9.2 12.7 Middlemore 14.7 5.1 14.3 6.8 The proportion of study patients discharged within 6 hours increased in every hospital with an overall change from 8.3% to 18.4%.

21/07/2017 10

RESULTS – ADVERSE EVENTS

In the control cohort, 5 out of 962 patients (0.52%) discharged within 6 hours had a MACE within 30 days (1 NSTEMI, 4 Deaths). In the intervention cohort, 16 out of 3632 patients (0.44%) discharged within 6 hours had a MACE within 30 days (8 NSTEMI, 1 STEMI, 1 Ventricular Tachycardia, 1 Cardiac Arrest and 5 Deaths). The sensitivity for 30-day MACE or mortality did not change with guidance framework implementation; p=0.28.

RESULTS – EVENTS PRE-IMPLEMENTATION

Hospital Age/ Gender Days after index Coding Notes Review Waikato 87F 16 Death During bronchoscopy Hutt 75F 25 Death Lung cancer Hutt 78F 15 Death Cardiac - non ACS Hutt 69F 26 Death Non-cardiac Hutt 79M 23 NSTEMI NSTEMI

21/07/2017 11

RESULTS – EVENTS POST-IMPLEMENTATION

Hospital Age/ Gender Days after index Coding Notes Review Wairarapa 51F 1 NSTEMI Non-cardiac* Waikato 66M 28 NSTEMI NSTEMI* Waikato 56M 3 STEMI STEMI* Waikato 91M 2 NSTEMI/VT NSTEMI/VT* Waikato 77F 27 Death Stroke* Waikato 48M NSTEMI NSTEMI* Wellington 83M 9 Cardiac Arrest Asystole, PPM inserted* Wellington 81M 10 NSTEMI NSTEMI, PCI* Hutt 94F 25 Death Cause unknown* Hutt 63F 25 NSTEMI NSTEMI* Hutt 84M 1 VT Non-ventricular arrhythmia** Hutt 79M 1 Death Cause unknown* North Shore 54F 2 NSTEMI Unstable angina* North Shore 79F 2 Death AAA* North Shore 73F 1 Death STEMI* North Shore 75F NSTEMI ESRF, CP , +ve cTn*

RESULTS – LENGTH OF STAY

Pre-implementation Post-implementation Site Hospital LoS (days) Hospital LoS (days) p Waikato 1.2 (0.6-3.1) 1.2 (0.3-3.4) 0.006 Hutt 0.4 (0.2-2.3) 0.3 (0.2-2) <0.001 North Shore 1.2 (0.7-3.9) 1.0 (0.5-3.3) <0.001 Wellington 0.9 (0.4-2.8) 0.7 (0.2-2.1) <0.001 Wairarapa 0.8 (0.2-2.3) 0.6 (0.2-2.3) 0.37 Waitakere 1.1 (0.7-2.2) 1.0 (0.4-2.5) 0.004 Middlemore 1.1 (0.7-3.1) 1.1(0.7-3.3) 0.26 Non-ACS patients’ hospital length of stay was shorter (p<0.001), with a median reduction in hospital length of stay of 2.9h

21/07/2017 12

CONCLUSION

This study demonstrated that a clinical framework incorporating accelerated diagnostic pathways for patients with suspected acute coronary syndromes can be safely implemented into multiple EDs across a country over a short time interval. This indicates that national adoption of similar guidance frameworks would be possible in many countries. After implementation at each site, a greater proportion of patients were discharged within 6 hours without compromising safety. This released thousands of hours of time for medical staff and resources which could be used for other sick patients. In addition, more rapid investigative process allows patients to be reassured earlier that they are not having a heart attack and to return at an earlier time point to their families and normal life activities including work.

LIMITATIONS

More ideal to prospectively recruit individual patients and collect electronic data on each individual patient, but the grant for the research was insufficient to allow for this type of study. Hospitals do not automatically record which patients are being investigated for ACS, so we had to rely on serial troponin measurements to capture the population of interest We had to rely on ICD10 coding to quantify clinical outcomes . We were unable to identify admissions and deaths outside New Zealand.

21/07/2017 13

ICAREACS TEAM

Allan S. M. D. Jaffe‎ ‎ ; Andrew Kerr‎ ‎ ; Bernhard Kuepper‎ ‎ ; Brad Peckler‎ ‎ ; Chris Florkowski‎ ; Chris Frampton‎ ‎ ; David Smyth‎ ; Denise Aitken‎ ‎ ; Frank Peacock; Gerard Devlin‎ ; Greg Hamilton‎ ; Gregor Larkin‎ ; Jaimi Greenslade‎ ‎ ‎ ; Jeremy Dryden‎ ‎ ‎ ; Jo Deely; John Bonning‎ ‎ ; John Pickering‎ ‎ ; Kate Allan‎ ‎ ‎ ; Kim Yates‎ ‎ ; Laura Chapman‎ ‎ ‎ ; Louise Cullen‎ ; Mark Simmonds‎ ‎ ; Mark Richards‎ ; Michael Ardagh‎ ; Niels van Pelt‎ ‎ ; Peter Freeman‎ ‎ ; Peter George‎ ; Pieter Van Wyk‎ ; Richard Troughton‎ ; Sally Lord‎ ‎ ; Steve Goodacre‎ ‎ ; Sue Crengle]‎ ‎ ; Tim Matthews‎ ‎ ; Tony Scott‎ ; Vince Lambourne‎ ‎ ; Yawane Jansz ‎