Clinical Response Model Pilot Evaluation Report February 2019 · Doc: Clinical Response Model Pilot...

Clinical Response Model Pilot Evaluation Report

February 2019

Authors: Dr Jim Ward, Medical Director Dr Patricia O’Connor, Director of Care Quality and Strategic Development Neil Sinclair, Consultant Paramedic – Clinical Standards and Innovation

DRAFT

Doc: Clinical Response Model Pilot Evaluation Report v1.00 Author: As detailed on cover sheet

Date: February 2019 Version 1.00 Review Date:

New Clinical Response Model Executive Summary at a Glance

New Response Model Aims The „New Clinical Response Model‟ (NCRM) aims to;

Save more lives

Provide a safer, more efficient and more effective response model by better matching responses to identified patient need based on clinical evidence.

Improve patient and staff experience Background This New Clinical Response Model has been developed following the most extensive clinically-evidenced review of its type ever undertaken in the UK.

This review of around half a million emergency calls highlighted that in certain cases, the previous time-based system – introduced in 1974 as a one-size-fits-all solution for the whole of the UK and not updated in over 40 years – did not identify some patients‟ conditions as immediately life threatening. Other patients were getting an eight minute response, even though their condition had no time critical response requirements and this was potentially diverting life-saving resources away from those cardiac arrest cases were time was a critical factor for a patient‟s survival. Under the new system, patients with immediately life threatening conditions, such as cardiac arrest, or who have been involved in serious road traffic incidents, are prioritised and receive the fastest response.

In less urgent cases, call handlers may spend more time with patients (or their representatives) to better understand their health needs and ensure they send the most appropriate resource for their condition, first time – not necessarily the nearest or quickest response. This approach takes into account the advances we have seen in clinical expertise in recent years and focuses on patient outcomes. NCRM Implementation Phases

NCRM Phase

NCRM Development Implementation Date

Phase 1 NCRM Code Acuity Level Changes 23 November 2016

Pre-Determined Attendance assigned to codes in C3 System

Updates to Standard Operating Procedures

Phase 2 Enhanced Defibrillation Module/PAD Mapping

29 March 2017

First Responder Module 15 May 2017

DRAFT



ProQA MPDS Triage System Upgrade Version 13

14 June 2017

C3 Nexus ELAN

Pre-Entry Questions Module (PEQ 1) 28 August 2017

PEQ 2 and Dispatch on Disposition (DoD)

10 October 2017

Phase 3 Key Phrases Module 9 April 2018

Overview of Response Categories and Results Purple Response Category Key Points The intention was to more accurately identify patients with an immediate need for resuscitation. This meant an expectation of a significant increase in volume in the category. This has been observed as predicted. Inclusion criteria: include cardiac arrest rate over 10%. Response: Two closet emergency resources with a minimum of three clinicians attending Results:

43% increase in 30-day survival (23.9% - 34.1%) for all patients in the Purple Category (1182 people)

The cardiac arrest rate in the purple basket was 52% in the first year of the pilot, indicating that we are more accurately identifying patients likely to go into cardiac arrest and ensuring they get our fastest response

21% increase in patients in Cardiac Arrest achieving Return of Spontaneous Circulation (i.e patients who have a pulse when they arrive at hospital after suffering a cardiac arrest)

100% increase in double responding to these codes in first year of NCRM (i.e attempting to ensure two separate crews are on scene in cardiac arrest cases as evidence shows this increases the chances of survival)

Median and 90th percentile response time maintained following NCRM introduction – this is despite a significant increase in call volume in this category

Red Response Category Key Points Modelling demonstrated that patients had a defined need for specific clinical assessment and care on scene. Based on clinical need some codes were re-triaged to other response categories after more detailed questioning by call handlers about the patient‟s condition. The expectation therefore was that there would be a significant reduction in red call volume compared with the previous Cat A cohort. This has been observed as predicted. Inclusion criteria: include cardiac arrest rate between 1% and 10%.(Actual figure is 1.5%).

DRAFT



Response: Immediate dispatch of nearest SAS resource and if this is not a conveying or paramedic resource, appropriate automatic back up is allocated. Results:

No observed change in 24 hour or 30 day survival.

Median and 90thpercentile response times maintained following NCRM introduction.

Clinical interventions identified were as predicted by the pre NCRM modelling.

Amber Response Category

Key Points The modelling demonstrated a cohort of patients who had a high requirement for diagnosis and conveyance to definitive care but a low requirement for immediate resuscitation. This included many patients with chest pain and suspected stroke. The expectation was that there would be a significant increase in call volume in this category. This has been observed as predicted. Inclusion criteria: cardiac arrest rate under 1% (actual figure 0.3%) but high likelihood of need for conveyance to definitive care. Response: identification and dispatch of nearest conveying resource with the preferred clinical skill set available. Results;

No increase in cardiac arrest rate for patients moved from Cat A to Amber response category.

Dispatch of best and most appropriate resource increased from 77.3% to 86.4% following NCRM introduction.

No observed changes in 24 hour or 30 day survival data

Increase in number of STEMI cases identified and taken directly to a specialist facility for treatment (i.e we were able to better identify heart attack patients and then take them to a specialist unit for further care, rather than just the nearest hospital).

Yellow Response Category Key Points This cohort of patients presented with a low need for immediate resuscitation and a variety of clinical responses that required a face to face consultation and assessment with onward care options including on scene definitive care or referral to the nearest Emergency Department. A range of responses had been defined and ongoing testing of appropriate response is progressing. There was an expectation that volumes of calls in this category would remain broadly similar. This has been observed as predicted. Inclusion criteria: cardiac arrest rate under 1% (actual figure 0.1%) and no clinical need for definitive care pathway (e.g a specialist facility like a PCI or stroke unit). Response: identification and dispatch of the nearest appropriate resource. The type of resource and skill set are determined by the individual code and resource availability.

DRAFT



Results

Low rates of deterioration and stable mortality rates.

Optimal dispatch of most appropriate resource for the patient‟s condition increased from 85.8% to 93.1% post NCRM launch.

Significant reduction in „stand down‟ (where a crew is already on their way to the patients address, and subsequently instructed to stand down) post launch of Dispatch on Disposition element of NCRM.

Median response times stable at 13 minutes in year 1. In year 2 we have seen an increase in 90thpercentile response times. Efforts to address this trend robustly are ongoing and require a range of interventions out with the NCRM which are being implemented.

Conclusions

Changing the triage system for emergency care within the Scottish Ambulance Service has been a complex and complicated process. The new system of care required redesign across the service from patient call to definitive care to achieve the aim of responding based on clinical need and in doing so improve outcomes.

The systematic review of the data provided a clear evidence base for change. Implementing the changes required staff support and the new system was co-designed with frontline staff with the recognition that improvement was required to be more efficient and effective as a service. Several highlights from the data demonstrate the following improvements

The model is working well and as planned, particularly in relation to the identification response and improved outcomes for our most seriously unwell patients.

The modelling predictions carried out pre-launch and post new clinical response have been realised. This has been validated by the data captured since NCRM launch, supporting the validity of the response category descriptors.

The purple response category shows an improvement in 30 day survival rates from 23.9% pre NCRM to 34.1% post NCRM, an increase of 43%. The survival data shows that this equates to 1182 people.

Response times are stable across the system for our most acutely unwell patients (purple and red categories).

During the pilot year, the Amber median and 90th percentile response times have shown improvement. Median response has reduced from 11.02 minutes to 10.43. 90th percentile has reduced from 23.87 minutes to 20.95. (Distribution Graph 5).

In the Amber category, there has been a welcome increase in the number of patients with chest pain overall and the sub set of patients presenting with STEMI conveyed to a specialist facility for treatment – Percutaneous Coronary Intervention (PCI) centres, rather than the nearest hospital. The model has demonstrated it has effectively dealt with this increase and this volume will be continually reviewed.

DRAFT



There has been an increase in response times to patients in the yellow codes as expected. However, linked mortality data appears to be stable for this category of patients who are at lower clinical risk. We recognise, however, that there is more improvement work to do for this category of patients and are implementing new approaches to improve our response. (Distribution Graph 8 and Table 16).

With the use of median and 90th percentile response data, this has improved SAS ability to understand and manage demand for all of patients and presents a much improved measurement framework than the previous 75% in 8 minute target for only 32% of our patients.

There have been no Serious Adverse Events reported related to the new response re-triage and clinical hierarchy.

There is evidence of improved efficiency and more accurate allocation of resources since NCRM implementation. There has been an increase in allocation of the right resource first time, with a reduction in inappropriate double resourcing to amber and yellow calls (Run chart 5).

11 Recommendations

Implement the system and process changes developed within the pilot period of NCRM to business as usual within SAS.

As with services in England and Wales, replace the 8 minute response target with agreed median and 90th percentile measures for patients in the defined categories to allow a more accurate and meaningful analysis of performance.

Create additional quality indicator measures that focus on patient and staff outcomes, continually developing how this data for improvement is actively shared with clinicians.

Continue to work to improve the experience of patients within the yellow category who are waiting longer.

Further focus on how SAS can deliver effective care to patients in the yellow response category, developing access to alternative care pathways from face to face or additional telephone triage in line with the Service‟s 2020 Strategy.

Continue to review demand and capacity and implement models that ensure a better match of resourcing to demand.

Continue to invest in the additional workforce requirements determined by the independent demand and capacity modelling review (May 2018).

Continue to showcase our work in SAS as a key part of Scotland Health and Care system, nationally and internationally.

DRAFT



Contents Pages 1 Introduction ………………………………………………………………... 8

2 Background ……………………………………………………………….. 8

3 Prioritisation of Calls ……………………………………………………… 9

4 The New Model …………………………………………………………… 10

5 NCRM Approach and Methodology…………………............................ 11

6 New Clinical Response Model Implementation ……………………….. 12

6.1 Enhanced Defibrillator Module …….…………………………... 13

6.2 First Responder Module ………………………………………... 13

6.3 C3 Nexus ELAN …………………………………………………. 14

6.4 Dispatch on Deposition …………………………………………. 14

6.5 Pre-Entry Question Module ……………………………………. 14

6.6 Key Phrases Module ……………………………………………. 15

7 Measuring Improvement …………………………………………………. 15

8 Results from the NCRM Evaluation …………………………………….. 16

8.1 Purple Response Category …………………………………….. 16

8.2 Red Response Category ……………………………………….. 23

8.3 Amber Response Category ……………………………………. 29

8.4 Yellow Response Category ……………………………………. 35

9 Staff Engagement and Experience ……………………………………... 40

10 Conclusions ……………………………………………………………….. 41

11 Recommendations ……………………………………………………….. 42

Glossary …………………………………………………………………… 44

DRAFT



1 Introduction

The „New Clinical Response Model‟ (NCRM) aims to;

Save more lives

Provide a safer, more efficient and more effective response model by better matching responses to identified patient need based on clinical evidence

Improve patient and staff experience

This evaluation report describes the clinical and organisational impact of the NCRM that was introduced within the Scottish Ambulance Service (SAS) in November 2016, makes recommendations for further enhancement of the model and associated performance reporting. An external evaluation by Stirling University has also been undertaken and accompanies this report.

The Stirling University evaluation report focuses on a „snapshot‟ of data taken in the month of January 2016 before the new model was introduced as well as January 2017 and January 2018. Stirling University make clear that although their report „allows some relevant comparisons between years, the findings cannot be generalised to the whole year or whole time period, also that further analysis of the data from each month of the pilot would allow a much more robust/relevant evidence of change and impact on the service and patients‟.

This Scottish Ambulance Service Report detailing data for the whole two year time period and the Stirling University Report, taken together therefore provide a comprehensive and robust analysis of the impact of the New Clinical Response Model on the service and patients.

2 Background

This New Clinical Response Model has been developed following the most extensive clinically-evidenced review of its type ever undertaken in the UK.

This review of around half a million emergency calls highlighted that in certain cases, the previous time-based system – introduced in 1974 as a one-size-fits-all solution for the whole of the UK and not updated in over 40 years – did not identify some patients‟ conditions as immediately life-threatening. Other patients were getting an eight-minute response, even though their condition had no time critical response requirements and this was potentially diverting life-saving resources away from those cardiac arrest cases where time was a critical factor for a patient‟s survival. In the past, any clinical stratification of risk was based on opinion and consensus rather than data and evidence. To date, there is little consensus informing best practice in ambulance pre-hospital response models, within the UK or internationally.

In contrast, there have been significant developments in the clinical capability and professional growth of ambulance service clinicians and a wide range of innovations and improvements in practice and equipment within the Scottish Ambulance Service 1,2. In November 2016, the Scottish Ambulance Service was supported by the Scottish Government to develop and pilot a new model of emergency response: A

DRAFT



New Clinical Response Model (NCRM). The pilot focused on a range of improvements to better manage response to all 999 patients.

The new model therefore focuses on improving patient survival and treatment, rather than simply measuring the time it takes to respond. Under the new system, our most critically ill patients receive the fastest response. Call handlers may spend more time with non-immediately life threatening patients to better understand their health needs and ensure that they send the most appropriate response for their condition first time, taking into account the advances we have seen in clinical expertise in recent years.

Other UK Ambulance Services have developed new response models to improve the triage and response to emergency calls. This evaluation report relates to the developments in Scotland. NHS Wales and NHS England have fully adopted their new models; Wales in 16/17, England in 20183, 4. As a result, revised performance measures are now endorsed at policy level. In Scotland, the National Review of Targets and Indicators by the Scottish Government in 20175 underpins our approach.

3 Prioritisation of Calls

SAS Ambulance Control Centres (ACCs) recruit and train non clinical call handlers who use the Medical Priority Dispatch System (MPDS – an internationally adopted ambulance triage system) to manage Emergency calls. Callers are asked a series of questions to categorise the call to determine the level of illness of the patient. At the end of the triage process an MPDS dispatch code is generated to send the most appropriate resources. SAS is able to determine the response to any MPDS code supported by MPDS governance in a process known as local development and response.

Historically, SAS had three time targeted response categories associated with MPDS determinant codes; Category A within 8 minutes, Category B within 19 minutes and Category C within 60 minutes. The clinical hierarchy for these was linked to the alphabetical order denoting the most acutely ill patients to Category A. A range of resources (ambulances, cars and motorbikes) were tasked to scene at the discretion of the ACC dispatch staff to balance demand, whilst meeting the aim of achieving target response times. In the past, dispatch decisions were based on the quickest response and took little cognisance of the needs of the patient and how these would be addressed by the nature of the response. The whole system supporting the old response model was developed to optimise efforts to achieve the 8 minute target. The good intention of achieving the target could result in a mismatch of resource to need, such as sending a response car to a patient who then needed to be taken to hospital. This could result in a quick initial response followed by a delay in getting the patient to hospital or further treatment whilst they waited on a suitable vehicle to arrive to transport them (ambulance). Such events often led to a poor experience for the patients, their families and their carers. It also potentially diverted a valuable resource away from attending a patient in more urgent need.

Furthermore, where resources were available, the response was often initiated too early in the call (as soon as the address was verified) before an understanding of the

DRAFT



patient‟s clinical need could be determined. This system, despite the best intention to get to the sickest people as quickly as possible, often resulted in a mismatch of response to the patients‟ needs as well as inefficient use of ambulance resources. Resources were often sent to patients and subsequently „stood down‟ whilst they were still on their way to the address, once the patient‟s condition was known. This led to inefficiency, inappropriate use of resources, significant disruption to work patterns and staff frustration.

4 The New Model

In response to these problems, a NCRM whole system of improvement was designed with the aim to;

Save more lives

Provide a safer, more efficient and more effective response model by better matching responses to identified patient need based on clinical evidence

Improve patient and staff experience

High quality clinical data to underpin provision of evidence based clinical care is essential to achieve high quality outcomes for patients. Obtaining such data can be challenging, however, in this respect SAS is unique amongst the UK ambulance services as it has had Electronic Patient Records (EPR) for more than ten years. This substantial clinical data set has enabled a deep understanding of pre-hospital patient interventions and outcomes which provided the necessary evidence to inform and develop the NCRM. A systematic review of the data was the basis for the proposed NCRM change. The Model (Table 1) was agreed and endorsed by the Chief Medical Officer and Scottish Government prior to introduction. To achieve the NCRM aims, an understanding of all elements of the patient pathway was necessary. As part of that process, the system was designed to move away from a singular focus on initial response times and instead where clinically safe, for the call handler to spend a little more time with the patient to better understand their clinical needs in order to ensure the best and most appropriate response is sent to assist them, first time. Delivering this model required redesign of the dispatch systems with ACC staff able to identify the clinical skill and resource type required to deliver care that most closely matched the patient‟s needs. All aspects of the NCRM have been continuously reviewed since the pilot introduction through an agreed governance framework. This framework includes daily, weekly and monthly reviews of the data within SAS and reporting to Scottish Government to ensure the clinical safety of patients and to assess the reliability of the process changes implemented in practice.

DRAFT



Table 1 - NCRM Categories

Purple Response Category (ILT) Identified by; Patient codes with a likelihood of Cardiac Arrest Rate of over 10% Response; Two closest emergency resources with a minimum of three clinicians attending Clinical Skill; Paramedic skill set defined as essential

Red Response Category (ILT) Identified by; Patient codes with a likelihood of Cardiac Arrest Rate of over 1% Response; Closest emergency resource + an emergency ambulance if initial response is Paramedic Response Unit (PRU) Clinical Skill; Paramedic skill set defined as essential

Amber Response Category Identified by; Acute pathway need I.E FAST+/STEMI Response; Right resource first time with conveying emergency ambulance due to high need for onward acute pathway care Clinical Skill; Paramedic skill set preferred

Yellow Response Category Identified by; Exclusion of ILT and Amber categories Response; Right resource first time to match patients need PRU for patients with potential for alternative pathway care, emergency ambulance for patients with a defined need for hospital transport Clinical Skill; Paramedic skill set preferred

5 NCRM Approach and Methodology

Early in 2016, SAS began to investigate and develop a plan for a new clinical response model. Proposed changes to the response model were based on clinical data. A rationale was developed and agreed by the SAS Clinical Advisory Group (CAG), which is a sub-group of SAS Clinical Governance Committee and includes both internal and external clinical experts. External critical appraisal was also undertaken prior to submission to the Scottish Government.

A NCRM project group was established to analyse and review the data set along with the development of a proposal for the new response model. The project group carried out the review using historical data from the MPDS call handling system and internal clinical outcome data to analyse over 500,000 calls to identify actual patient acuity levels, clinical interventions and outcomes associated with MPDS codes. This allowed the team to develop new procedures, the proposed acuity level/priority associated with each MPDS code and the required resources by clinical skill and resource type.

To enable a greater understanding of the clinical interventions and actual outcomes associated with MPDS codes, the NCRM project group in partnership with the CAG developed a 10 point data analysis capture tool (Table 2). This analysis provided an insight into patient pre-hospital clinical acuity level by MPDS code and supported assignment of acuity levels to codes for the new model, taking account of the cardiac arrest rates and interventions carried out on particular patient groups.

DRAFT



Table 2 – 10 Clinical Measures for each Code

Measure 1 Transported to Hospital

Measure 2 Airway Intervention

Measure 3 Breathing Intervention

Measure 4 Circulation Intervention

Measure 5 Drugs Administered

Measure 6 NEWS Score +4

Measure 7 Pre-Alert

Measure 8 STEMI

Measure 9 FAST+ Stroke

Measure 10 Cardiac Arrest

There was clear evidence from the data to highlight that the SAS response model was over prioritising patients as immediately life threatening for some MPDS codes and under categorising others. The associated risks with over and under categorising of patients in the old model became the primary driver for developing a better model to deliver improved patient outcomes.

The review focused on gaining an understanding of the patient‟s clinical acuity level and consideration of their complete care pathway/journey. This process allowed the NCRM project group to propose a new response system and hierarchy for the 1254 dispatch codes linked to the response of public 999 calls, developing the previous three tier hierarchy to a wider five tier response hierarchy (Table 1). This further stratification of clinical acuity delivers a much more differentiated and appropriate response hierarchy. A systematic process was developed to match clinical skill and resource type, with the patient need.

In practice the review highlighted that previously, patients at highest risk of cardiac arrest were identified across a wider range of codes than was expected. The NCRM process created the ability to group all of these codes into a single category (Purple) based on clinical need. This means that all of these codes with higher cardiac arrest rates now receive the highest priority response whereas pre NCRM they did not.

6 NCRM Implementation

The NCRM was implemented in three phases (Table 3) using an improvement approach. A measurement, audit and monitoring framework was established to enable continuous evaluation throughout all phases.

DRAFT



Table 3 – NCRM Implementation Phases

NCRM Phase NCRM Development Implementation Date

Phase 1 NCRM Code Acuity Level Changes 23 November 2016

Pre-Determined Attendance assigned to codes in C3 System

Updates to Standard Operating Procedures

Phase 2 Enhanced Defibrillation Module/PAD Mapping 29 March 2017

First Responder Module 15 May 2017

ProQA MPDS Triage System Upgrade Version 13

14 June 2017

C3 Nexus ELAN

Pre-Entry Questions Module (PEQ 1) 28 August 2017

PEQ 2 and Dispatch on Disposition (DoD) 10 October 2017

Phase 3 Key Phrases Module 9 April 2018

Phase 1 of NCRM went live on 23 November 2016. Phase 1 consisted of changes to acuity levels associated with codes, also the configuration of pre-determined resources on the command and control system to guide dispatchers on the type of resource(s) required to meet the clinical needs of the patient.1 This was supported by a change to any standard operating procedure that impacted on dispatch decision making to ensure that staff had clear guidance on dealing with all eventualities (for example; how to respond to a high priority call when a crew were already on their way to another call, on a rest period, or approaching the end of their shift). Phase 2 of the NCRM consisted of enhancements in technology to aid and support dispatch decision making along with a number of process improvements aimed at ensuring the earliest possible recognition of critically ill patients. Phase 3 saw the implementation of the key phrases module in the command and control system. „Key Phrases‟ are an additional mechanism by which patients with high acuity symptoms or in dangerous circumstances are identified in order to enable dispatch of a response as early as possible in the call cycle.

6.1 Enhanced Defibrillator Module

The enhanced defibrillator module is an enhancement within the Ambulance Control command and control dispatch system (C3) that allows public access defibrillators to be entered and mapped onto the system. They were also configured to be highlighted only if the dispatch code is appropriate and the public access defibrillator is at a suitable distance from the incident. This development allows the call handler to signpost bystanders to the location of the defibrillator.

6.2 First Responder Module

The First Responder module is an enhancement within the C3 Command and Control dispatch system that aids and supports better dispatch decision making. This module allowed the system to be configured with set criteria for dispatching first responders, including dispatch codes and age criteria suitable for first responders as well as distance of first responder from location of incident. Prior to

DRAFT



the implementation of the first responder module, the dispatcher had to manually assess the call suitability for first responders which was time consuming and inconsistent.

6.3 C3 Nexus ELAN

The C3 Nexus ELAN upgrade to the system was a development that enabled auto population of address locations of incidents on the system taking account of the location of the callers by using the landline number used to call 999, to save vital time in address location verification.

6.4 Dispatch on Disposition (DoD)

The essence of DoD is that, for patients who are not presenting with an immediately life threatening condition, taking a short amount of additional time to clearly understand the clinical needs of the patient will result in a better match of resource to patient need. This optimises on scene care and clinical management and optimises the onward progress to either definitive emergency care or a community based care pathway. In January 2015, approval was reached from the Department of Health in NHS England (NHSE) to pilot DoD; this was the first phase of the new Ambulance Response Programme (ARP) approved for implementation in NHSE. In early February 2015, the first phase of the ARP went live in the London Ambulance Service (LAS) and the South Western Ambulance Service NHS Foundation Trust (SWAST) with the North East, South Central, West Midlands and Yorkshire Ambulance Trusts following suit in late October 2015. SAS went live with DoD on 10 October 2017. Prior to the implementation of DoD, practice was broadly centred around allocating the nearest resource on confirmation of address, in pursuit of delivering against an 8 minute response time target. The old dispatch on address model meant that dispatchers were sending resources to incidents were the patient‟s condition and acuity level had not yet been established. There was recognition that moving away from dispatching on address and taking the time to understand the condition of the patient before sending a resource would mean a potential delay in allocation of resources to all patients. As such there was a requirement to have a robust system in place that would allow call handlers to identify cardiac arrest and patients with immediately life threatening symptoms at the earliest possible stage in the call cycle. This was achieved through the development and implementation of the Pre-Entry Questions and Key Phrases as described below.

6.5 Pre-Entry Question Module

Response times to patients in the purple response category, including those in cardiac arrest, have remained very fast following the implementation of DoD owing to the ability to identify a high proportion of these patients at a very early stage in the call, enabled by the introduction of two new Pre Entry Questions (PEQs). These questions come before the formal triage process and allow for the identification of critically ill patients at the earliest point in the call cycle. Since DoD, the dispatcher

DRAFT



allocates a resource on confirmation of the condition of the patient unless final coding hasn‟t been reached at 180 seconds, in which case the call presents to the dispatcher for allocation of a resource. The PEQ Module within the command and control system was introduced to ensure early identification of patients in Cardiac Arrest and patients with immediately life threatening symptoms to allow rapid dispatch of resources. Pre Entry Question 1 “Is the patient breathing” was implemented in August 2017 followed by Pre entry Question 2 “Is the patient awake” on 10 October 2017 alongside DoD. A key enabler for delivering the out of hospital cardiac arrest strategic aim to save more lives is by quickly and accurately identifying patients in cardiac or peri-arrest. This allows the ACC call taker to provide CPR telephone advice instructions and signposting bystanders to public access defibrillator were available. This ACC call taker input happens at the same time as dispatchers are notified of a purple code for immediate dispatch if there is a „no‟ answer to PEQ1.

6.6 Key Phrases Module The Key Phrases Module is an enhancement within the C3 command and control dispatch system that enables earlier identification of patients with immediately life threatening symptoms by associating key words used to describe the patient‟s problem with key words that historically have resulted in an ILT outcome code.

7 Measuring Improvement

Developing and measuring the NCRM required a new approach to capturing, monitoring and understanding the data. Measuring the 8-minute target in isolation creates a binary system with a pass or fail result. This single way of measuring does not enable understanding of the degrees of success when passing the target or the units of failure when the target is not achieved. Given the detailed changes to the codes within the NCRM, monitoring and understanding variation in practice was essential. Developing and displaying data overtime helps to more clearly identify the opportunities for learning and improvement across the entire range of activity. A whole new approach to data reporting was created using run charts, line charts, control charts and frequency distribution charts to understand how the changes implemented were impacting on the service delivered. Data presented in this way gives the opportunity to dig deeper to explore reasons behind changes and understand the conditions that created that result, either positive or negative. Understanding variation, particularly unintended variation, is crucial to making changes that result in improvement. When measuring a range of responses within the NCRM system the mean is not always the best 'average' to use as some events are rare and affect the whole system average. The median is the middle value, 50% of incidents are above it, and 50% below it. When the data is not symmetrical, this form of 'average' gives a better representation of the data and is also used to minimise the impact of small numbers and outliers. The NCRM measures on time based targets were designed to record response across the entire system as opposed to only 75% of incidents in the previous system:

DRAFT



Purple; Median under 6 minutes. 90th percentile under 15 minutes

Red; Median under 7 minutes. 90th percentile under 15 minutes

The median measure applies to 50% of the patients in the category and the 90th

percentile applies to 90% of patients in the Category.

During the NCRM pilot SAS have continued to measure and report on the Scottish Government 75% target for 8-minute response as well as report on these new measures and monitor internally the response times for patients in the other response categories.

8 Results from the NCRM evaluation 8.1 Purple Response Category

Key Points The intention was to more accurately identify patients with an immediate need for resuscitation. This meant an expectation of a significant increase in volume in the category. This has been observed as predicted. Inclusion criteria; include cardiac arrest rate over 10%. Response; Two closest emergency resources with a minimum of three clinicians attending Results;

43% increase in 30-day survival (23.9% - 34.1%) for all patients in the Purple Category (1182 people)

The cardiac arrest rate in the purple basket was 52% in the first year of the pilot

21% increase in patients in Cardiac Arrest achieving Return of Spontaneous Circulation

100% increase in double responding to these codes in first year of NCRM

Median and 90th percentile response time maintained following NCRM introduction despite significant increase in call volume in this category

The purple response category is the highest response priority. The hierarchy of response is based on clinical measures as detailed in Table 1 above. The observed activity and interventions in the purple response category since the NCRM launch in November 2016 is described in Tables 4 and 5 below.

Table 4 presents the initial pilot year data November 2016-November 2017. Table 5 details results from December 2017- June 2018 (extended period of the pilot). It is important to note that Table 5 included data over the winter period of 2017/2018 where service demand was at the highest level ever recorded (27% higher than previous year) and there was extreme and prolonged adverse weather (for example; Beast from the East).

DRAFT



The Tables (4 and 5) below describe that the clinical parameters were as predicted by the pre-pilot NCRM modelling, and specifically for purple codes, the cardiac arrest rate was well above the 10% threshold.

The NCRM modelling is based on the risk or likelihood of a patient requiring a clinical intervention. Within the Purple category, these interventions are the most time critical of all the categories. The most acute example of this is the cardiac arrest rate and the requirement for resuscitation, including basic and advanced life support.

Table 4 - Purple Response Category (Pilot period November 2016 – November 2017)

Volume Transported Airway Breathing Circulation Medicines NEWS+4 Pre-Alert

STEMI FAST+ Cardiac Arrest

7781 40.7% 33.1% 23.2% 27.1% 38.4% 51.9% 8.9% 0.1% 0.6% 52.1%

95% CI 39.6- 41.8%

32.1- 34.2%

22.3- 24.1%

26.1- 28.1%

37.3- 39.5%

50.8- 53.0%

8.3- 9.6%

0.0- 0.2%

0.6- 0.4%

51- 53.2%

Table 5 - Purple Response Category (Beyond Pilot December 2017 – June 2018)

Volume Transported Airway Breathing Circulation Medicines NEWS+4 Pre-

Alert STEMI FAST+ Cardiac

Arrest

4640

42.2%

29.9%

13.8%

24.9%

43.6%

64.3%

26%

0.3%

0.6%

48.7%

The pre-NCRM data analysis revealed that there were patients in a number of MPDS codes with unexpectedly high cardiac arrest rates. Within NCRM, all of these codes with high cardiac arrest rates (over 10%) have been brought together into a single purple category requiring the fastest and most skilled response. This, along with an improved recognition of patients in cardiac arrest, explains why the volume of patients in the purple category has increased over time (Control Chart 2). In terms of improving survival, this is a new way of considering how the system works to save lives. Historically, we have focussed on patients who were confirmed to be in cardiac arrest. It remains to be seen how this way of working can further develop in order to prevent critically ill patients deteriorating to cardiac arrest, however the initial results from linked survival data are very positive. As an example of using the methodology described above, consider the changes in response to three groups of patient codes, namely 31, 29 and 12 (as detailed below). Prior to the NCRM system, patients within these groups would not have been responded to as our highest priority unless they were actually identified as being in established cardiac arrest. However, the pre-NCRM data analysis identified higher rates of deterioration to cardiac arrest (over 10% cardiac arrest rates in each of these codes) and consequently these patients are all triaged into the purple category and responded to as our highest response priority.

DRAFT



MPDS Dispatch code 31 “Unconscious”, descriptor “Unconscious Agonal/Ineffective Breathing”

Re-triaged from Category A to purple response category, 1483 patients responded to over the pilot period with a reported average 20% cardiac arrest rate (approximately 296 patients).

MPDS Dispatch code 29 “Road traffic collision”, descriptor “Possible death at scene” Re-triaged from Category B to purple response category, 64 patients responded to over the pilot period with a reported 19% cardiac arrest rate (approximately 12 patients). MPDS Dispatch code 12 “Seizure”, descriptor “Seizure ineffective breathing”

Re-triaged from Category A to purple response category, 173 patients responded to over the pilot period with a 37.5% cardiac arrest rate (approximately 65 patients).

The patients identified in these three sub set groups, were identified and re-triaged more accurately in NCRM and received a higher priority response and a multi resource response which results in a greater chance of survival4, 5, 6.

Cardiac Arrest and Return of Spontaneous Circulation data In addition to understanding the overall risk of deterioration to cardiac arrest as described in the previous section, NCRM has also focussed on improving outcomes for those patients in established cardiac arrest. One of the key internationally accepted markers for successful pre hospital resuscitation is the number of patients who achieve Return of Spontaneous Circulation (ROSC) in the pre-hospital environment. One important aim of the NCRM was to develop a specific focus for this patient group and target them for the highest priority response. Multiple responders are tasked to optimise resuscitation at scene including a clinical skillset as essential. As previously stated, the purple response category has a cardiac arrest rate of over 50% since the launch of NCRM. Over the initial 12 month pilot period of NCRM there has been an increase in the volume of patients achieving ROSC. Compared to the same 12 month period pre NCRM all rhythms ROSC has an increased average number of 55.3 to 71.1 patients achieving ROSCs per month (28.6% increase) and VF/VT ROSC has an increased average number from 22.0 to 33.4 patients achieving ROSCs per month (51.8% increase). Control Chart 1 shows the percentage of patients per month with all rhythms and VF/VT ROSC and Control Chart 2 shows the number of VF/VT patients per month.

DRAFT



Control Chart 1 - All rhythms ROSC Percentage of patients per month

(November 2015 - June 2018)

Control Chart 2- VF/VT ROSC Number of patients (November 2015 - November 2018)

The NCRM is part of SAS Clinical Services Transformation strategic programmes of improvement. There have been several concurrent service developments to support the response and resuscitation of critically ill patients. For example, there has been continuous development and investment in our ACC‟s to optimise early identification of patients in cardiac arrest, along with guidance and support from our bystander CPR programme and Public Access Defibrillator (PAD) access and mapping. The coordination of a whole system approach to the quality of

NCR

M In

trod

uced

Mean 22.0

Mean 33.4

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

0

10

20

30

40

50

60

Nov

-15

Dec

-15

Jan-

16

Feb-

16

Mar

-16

Apr

-16

May

-16

Jun-

16

Jul-1

6

Aug

-16

Sep-

16

Oct

-16

Nov

-16

Dec

-16

Jan-

17

Feb-

17

Mar

-17

Apr

-17

May

-17

Jun-

17

Jul-1

7

Aug

-17

Sep-

17

Oct

-17

Nov

-17

Dec

-17

Jan-

18

Feb-

18

Mar

-18

Apr

-18

May

-18

Jun-

18

VF/VT ROSC

VF VT Actuals Mean UCL LCL

DRAFT



resuscitation on scene has been supported by a continuous programme of clinical education and national Learning in Practice programme. Examples of recent initiatives implemented to improve our responses to our most seriously ill patients include the national roll out of the 3RU (Rapid Resuscitation Response) unit model, a specialist team of responders who are experts in cardiac care and are supported by doctors/nurses who are sent out to patients in cardiac arrest, also the Advanced Critical Care Paramedics pilot in the West of Scotland who support the work of Scotland‟s Out of Hospital Cardiac Arrest (OHCA) strategy and the Global Resuscitation Alliance in implementing the „It Takes a System to Save a Life‟ approach to patient care. Linked Mortality and Survival Data The aim of responding effectively to critically ill patients is to enable people to survive to get home to carry on with their lives. This means that it is important to understand markers such as 30 day survival. In Scotland, we have a system of data linkage that allows SAS data to be linked via the Unscheduled Care DataMart (UCD) to provide data on 24 hour and 30 day mortality. In order to understand survival data more clearly, the NCRM pilot team collaborated with Information Services Division (ISD) and NHS National Services Scotland (NSS) colleagues to link outcome mortality/survival data from the NHS Scotland Unscheduled Care DataMart to explore the survival rate at 24 hours and 30 days‟ post incident for all NCRM response categories. Significant work has been undertaken to match data and improvements have been made in matching data over time. Where data cannot be matched, it means that SAS has not been able to identify the identity of patient to a sufficient level of certainty for the SAS data to be linked to the UCD. Table 6, shows the linked survival data pre NCRM and one-year post NCRM. At its most simple the percentage of patients surviving 30 days in the purple category post NCRM is 34.1% compared with 23.9% pre NCRM. This equates to 1182 people. Interpretation of the data is complex and there are three significant areas of variation between the data pre and post NCRM. Three variables that require to be considered in interpreting this data are;

Percentages of data linked

Volumes of patients pre and post NCRM

24 hours and 30 day survival

Variability of percentages of data linkage In the year pre NCRM the data linkage between SAS and ISD was 66.8%, and in the year post NCRM launch data linkage had increased to 77%. In understanding the changes in survival rates, there was a concern that survival within the non-

DRAFT



linked cases pre and post NCRM were unknown and that this could affect the ability to understand the survival/mortality percentages. Therefore, an audit of unmatched patient reports from the pre and post NCRM time periods was undertaken by SAS. This data relates to survival status for patients in the purple category at the point where SAS care stopped. We are unable to say with certainty that all unmatched patients who were alive at the point where they left SAS care would survive a further 30 days so if anything the non-linked 30 day survival figures could be an over-estimate as the assumption made where data cannot be linked is usually that the patient has not survived. Audit of outcomes for non-linked purple codes

Deceased Alive Duplicate Record Other

Pre NCRM 59.5% n=1100 13% n=240 25.5% n=471 2% n=37

Post NCRM 44.7% n=810 16.8% n=304 32.7% n=592 5.8% n=105

In mapping this data to the number of unmatched cases and assuming that patients survived 30 days it can be estimated that approximately an additional 240 patients may have been alive in the pre sample after 30 days, taking the pre sample survival figures to 1127. This represents an estimated overall survival rate as a percentage of the total number of incidents in the pre NCRM sample of 20.2%. In relation to year 1 post NCRM it can be estimated that an additional 304 patients may have been alive at 30 days taking the survival numbers to 2373. This represents an overall survival rate as a percentage of the total number of incidents in the 1 year post NCRM sample of 30.1%. This represents a percentage improvement in 30 day survival of 49%, which is higher than the percentage improvement for linked data alone. The audit data provides a high degree of reassurance that non linked data does not negatively impact the findings in the linked data section and indicate that 30 day survival may be even higher. Volumes of patients pre and post NCRM The data presented in table 6 shows an increase in purple coded calls post NCRM introduction. There are a number of possible reasons for this including;

Increase in actual volume of high acuity patients due to demographic and disease prevalence

Improved recognition of critically ill patients by Ambulance Control Centre (ACC) staff due to improvements in training, audit and feedback within ACC

Impact if the trauma desk in identifying patients with traumatic critical injuries

Improved ACC practice in terms of more accurate coding of critically ill patients into the purple category

DRAFT



In relation to the increased numbers of purple incidents observed since the NCRM launch, the University of Stirling independent statistical analysis states; ‘that the increase in volume represents patients correctly identified with the highest requirement for immediate response’. Volumes of patients in the purple category have now stabilised including taking into account seasonal factors, indicating the NCRM has embedded an overall response model that creates the foundations on which to build further improvements in outcomes. Improvements in 24 hour and 30 day survival Taking into account the two variables discussed above, there are large improvements in both 24 hour and 30 day survival. Understanding the outcomes for patients in the unlinked data sets shows an improvement in 30 day survival from 43% in the data linked section to an estimated 49% when all incidents are considered.

In summary, the most accurate data reflecting 30 day survival is the data that has been linked with the ISD UCD. The percentage increase in survival for patients whose records could be linked has increased from 23.9% to 34.1%, an increase of 43%. Table 6 - Linked 24 hour and 30-day survival data (1 year pre NCRM and 1-year post NCRM introduction)

24 Hour 30 Day Total

Total SAS Incidents recorded by SAS

% Matched Survival Survival

SAS Incidents linked by ISD

Number of

people %

Number of people

% Number of incidents

Number of incidents

%

All Purple Codes Pre Sample

1010 27.2% 887 23.9% 3,719 5,568 66.8%

All Purple Codes Post Sample

2325 38.3% 2069 34.1% 6,070 7,881 77.0%

DRAFT



8.2 Red Response Category

Key Points

Modelling demonstrated that patients had a defined need for specific clinical assessment and care on scene. Based on clinical need some codes were re-triaged to purple with some being re-triaged to amber. The expectation therefore was that there would be a significant reduction in red call volume compared with the previous Cat A cohort. This has been observed as predicted. Inclusion criteria; include cardiac arrest rate between 1% and 10% (Actual figure is 1.5%) Response; immediate dispatch of nearest SAS resource and if this is not a conveying or paramedic resource, appropriate automatic back up is allocated. Results;

No observed change in 24 hour or 30 day survival.

Median and 90thpercentile response times maintained following NCRM introduction.

Clinical interventions identified were as predicted by the pre NCRM modelling.

The red category is the second highest acuity response. The aim of the NCRM work in this category is to identify and target a focused response to patients who require intervention to prevent further clinical deterioration or cardiac arrest. From the evidence base patients are identified by having a cardiac arrest rate of over 1% and under 10% (over 10% would be triaged to the purple response category). This high clinical acuity within this group of patients can be demonstrated by the data in Tables 7 and 8. The observed data was as predicted pre NCRM launch for example;

patients requiring airway and breathing (ventilation) interventions

Medication intervention rates, discussed in more detail later in this report

Deranged physiology in the response category defined by National Early Warning Score +4 rates

Emergency Department pre-alert rates demonstrating high acuity patients The NCRM review methodology used clinical data in combination with wider MPDS descriptors and application of the NHS Scotland Risk Matrix review process which allowed for the review of specific clinical need related to patient presentation (for example; a conscious patient suffering an acute anaphylactic reaction has a defined need for medication resuscitation at scene, with the aim of reversing life threatening symptoms and hopefully minimal resuscitation markers would follow after drug resuscitation). This detail and clinical need would be difficult to extract from the clinical data alone, but this level of clinical consideration was given to each of the 1254 MPDS codes and conditions reviewed.

DRAFT



A further review was undertaken to understand the related practice and response category accuracy to highlight this element of the NCRM methodology and to allow for a further analysis of the 26% of medications administrated to the red response category patients (initial 12-month pilot). This initial review has demonstrated that 94% of the 26% of medication interventions (presented in Table 7 below) are linked to medications that could be considered resuscitative and in a life threatening situation, the timely response and administration of these medications would prevent further clinical deterioration.

The data demonstrates that our response to the patients with anticipated high acuity factors within the purple and red categories worked as predicted, operating reliably over the 19 months since introduction. As the model has matured, it is clear that the levels of clinical acuity and clinical need of patients in the purple and red categories is materially different, as evidenced by the significantly different cardiac arrest rates in the purple and red categories (Purple 52% and Red 1.5%). Note the increase in pre-alert percentage in red and amber categories. This relates to the ease in which data can be captured in the new Electronic Patient Record (EPR).

Table 7 - Initial 12-month pilot period data

Volume Transported Airway Breathing Circulation Medicines NEWS+4 Pre-Alert


45520

68%

4.6%

1.3%

9.4%

26%

24.1%

6.6%

0.1%

1.0%

1.5%

95% CI

67.6-

68.4%

4.4 – 4.8%

1.2-

1.4%

9.1- 9.6%

25.6-26.4%

23.7- 24.5%

6.4 – 6.8%

0.1%-0.1%

0.9-

1.1%

1.4%-1.6%

Table 8 - Extended 7-month pilot period data.

Volume Transported Airway Breathing Circulation Medicines NEWS+4

Pre-Alert


32079

72.6%

3.6%

0.7%

11.8%

32.3%

31.9%

22.4%

0.4%

0.3%

1.7%

Linked Mortality Data The NCRM project has linked SAS data to outcome mortality data from NHS Scotland datasets. Using like for like codes in a pre and post sample analysis the red response category has demonstrated stable mortality rates across both pre and post time periods. Most of these patients would have been within the Cat A category previously, therefore, there was no expectation of any significant change in clinical outcome or mortality data as the model response to this cohort was not changed significantly.

DRAFT



Table 9 - LinkedSurvivalData Red Category


Total SAS Incidents

recorded by SAS


SAS Incidents linked by

ISD

Number % Number % Number Number %

Red Codes Pre Sample 27,823 96.0% 26,795 92.4% 28,986 39,860 72.7%

Red Codes Post Sample 34,266 96.3% 32,922 92.5% 35,574 46,201 77.0%

Response Times for Purple and Red Calls Distribution Graphs 1 and 2 (below) demonstrate that response time to the purple response category over the 12month pilot and the extended pilot period (19 months) have remained stable despite the increase in volume of incidents. Post NCRM median response time for 8449 purple patients was 5.73 minutes compared to a median response time of 5.88 for 4528 patients on the old response model. Both graphs demonstrate a reliable response to this patient group over the 19 months of the NCRM pilot and the response times aims described on page 8 have been achieved.

Distribution Graph 1 - Purple Response Times (NCRM year 1)

0

200

400

600

800

1,000

1,200

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25+

Dem

and

Response Time Minutes

Purple Incidents Distribution

Incident Count - Pre NCRM Incident Count - Post NCRM

DRAFT



Distribution Graph 2 - Purple Response Times (November 2017- June 2018)

Distribution Graphs 3 and 4 (below) demonstrate the response to the second tier red response category over the 19-month pilot period. The data in the initial 12month pilot period graph, demonstrates a 58.08% reduction in the incident volume in this response category in NCRM as predicted by the modelling. Both graphs demonstrate a consistent and stable response to this patient group over the 19 months of the NCRM pilot. There is a marginal increase in response, this is related to the introduction of DoD and there are developments including the key phrases module that continue to reduce the response time for red calls. Distribution Graph 3 - Red Call Response Times (NCRM Year 1)

0

100

200

300

400

500

600

700

800

900

1,000

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25+

Dem

and


Purple Incidents Distribution 27/11/2017 - 30/06/2018

Incident Count

DRAFT



Distribution Graph 4 - Red Response Times (November 2017 - June 2018)

Resources Dispatched to Scene

Dispatch guidance was developed to ensure that the most appropriate resource was sent to all patients with the aim “right resource first time”.

Multi Response to Purple Codes

For patients in established cardiac arrest, a multi resource response (two ambulances or one plus a Paramedic Response Unit) is best practice. In order to improve survival for those critically ill patients, NCRM extended this level of multi resource response to all Purple category calls, meaning that both cardiac arrest and many peri-arrest patients have a more robust response. This change in practice has demonstrated (Control Chart 3 an almost 100% increase in patients receiving a multi resource/clinician response over the NCRM pilot. With the roll out of the 3RU model to Glasgow and beyond, this results in the provision of a resuscitation team to scene as quickly as possible. Further analysis is needed to understand the contribution of this intervention to the observed improvement in survival within the purple category.

0

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

4,500

5,000

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40+

Dem

and


Red Incidents Distribution 27/11/2017 - 30/06/2018

Incident Count

DRAFT



Control Chart 3 - Purple Response with 3+ crew on scene

Improved back up times to Paramedic Response Units (PRU) Delays to back up PRU‟s sent to scene were identified as a clinical risk in 2015/2016. In considering appropriate dispatch arrangements, dispatch arrangements were re-configured to mandate that any PRU sent to a purple or red call would be automatically backed up by the nearest ambulance. This was in appreciation of both the need for early response and also the requirement for onward conveyance to hospital in most cases. Control Chart 4 (below) demonstrates there has been a significant improvement in back up times for PRU‟s at ILT incidents with response times improved by over 100% over the 12 month pilot period.

NC

RM

Intr

od

uce

d

Dat

a C

on

tin

ued

fro

m 2

7/11

/201

8

Mean 71

Mean 126

Mean 145

0

1

2

3

4

5

6

7

8

9

10

0

50

100

150

200

250

Purple Incidents Where Crew on Scene 3+

Incident Count Where COS >=3 (absolute numbers) Mean UCL LCL

DRAFT



Control Chart 4 - Back up times to PRUs.

8.3 Amber Response Category

Key Points The modelling demonstrated a cohort of patients who had a high requirement for diagnosis and conveyance to definitive care but a low requirement for immediate resuscitation. This included many patients with chest pain and suspected stroke. The expectation was that there would be a significant increase in call volume in this category. This has been observed as predicted. Inclusion criteria; cardiac arrest rate <1% (actual figure 0.3%) but high likelihood of need for conveyance to definitive care. Response; identification and dispatch of nearest conveying resource with the preferred clinical skill set available Results;

No increase in cardiac arrest rate for patients moved from Cat A to Amber response category.

Dispatch of best resource increased from 77.3% to 86.4% following NCRM introduction.

No observed changes in 24 hour or 30 day survival data

Increase in number of STEMI cases identified.

The NCRM data for the amber response category has demonstrated over the 19 month pilot period the consistent identification of patients with a low identified need for resuscitation and a defined need for specific acute pathway care. This confirms that for this category of call, the predicted outcomes have been observed.

NCR

M In

trod

uced

Data

Con

tinue

d fr

om 2

7/11

/201

8

Mean 14.51

Mean 6.46 Mean 6.68

0

1

2

3

4

5

6

7

8

9

10

0.00

2.00

4.00

6.00

8.00

10.00

12.00

14.00

16.00

18.00

20.00

Back up time for PRUs (minutes)

Avg Backup Time Mean UCL LCL

DRAFT



The amber response category is the third tier in the NCRM response hierarchy. This category was defined by a combination of the exclusion of a high chance of cardiac arrest or need for resuscitation to prevent deterioration to cardiac arrest, the identified clinical requirement for a timely response to assess a patient and convey without delay to definitive care. This clinical focus is applied to the patient cohort with an acute pathway need (for example: ST Elevation Myocardial Infarction (STEMI), patients with Hyper-Acute Stroke symptoms and breathing problem patients with a high volume ED resuscitation room pathway). Identification Following the NCRM clinical governance arrangements patient clinical need was identified and triaged using the agreed 10 clinical measures; these are displayed below for the amber response category.

Table 10 - 10 Clinical measures 12-month initial pilot period

Volume Transported Airway Breathing Circulation Drugs NEWS+4 Pre-


Arrest

82103

77.9%

0.4%

0.2%

7.2%

35%

19.6%

6.0%

2.1%

6.2%

0.3%

95% CI

77.6-

78.2%

0.3- 0.4%

0.2-

0.3%

7.1-

7.4%

34.7-35.4%

19.3-

19.9%

5.9- 6.2%

2.0-2.2%

6.0%-6.3%

0.2%-0.3%

Table 11 - Extended pilot period data



Arrest

56860

83%

0.3%

0.1%

13.6%

33.3%

26.5%

30.6%

2.2%

14.9%

0.3%

It is important to state that these patients are still responded to as emergencies, although they present with lower immediately life threatening symptoms than patients in the purple and red response categories. In addition to confirming the accurate identification of these main sub set of patient groups in the amber response category, the data demonstrated no observed increase in clinical deterioration patterns identified in the clinical data nor were any concerns raised through SAS Datix or patient experience reporting systems.

Linked Mortality Data Using like for like codes in a pre and post sample analysis the amber response category has demonstrated stable mortality rates across both samples. This is reassuring in respect of SAS taking many of these patients out of the 8-minute category (Cat A) and prioritising the right response (A paramedic led conveying ambulance rather than the nearest resource regardless of type which was pre NCRM model).

DRAFT



Table 12 - Amber Category, Linked Survival Data


Total SAS Incidents recorded

by SAS


SAS Incidents

linked by ISD


Amber Codes Pre Sample

58,259 99.0% 65,187 95.1% 68,513 81,718 83.8%

Amber Codes Post Sample

76,329 99.1% 73,356 95.2% 77,051 90,341 85.3%

Response Times to Amber Codes As demonstrated by the data in the 10 clinical measures the amber response category identified patients who have a low need for resuscitation but the requirement for an acute onward clinical pathway. This category historically had an optimal response aim of 75% of patients attended within 19 minutes. The distribution graphs (5 and 6 below) demonstrate a consistent response to this patient group over the pilot period with improved median and 90th percentile response times post NCRM introduction. We have also observed a significant reduction in numbers of patients experiencing delays beyond 60 minutes in the first year of NCRM. In the subsequent months into year two, an increase in response times has been noted but no associated evidence of clinical deterioration. This reflects the significant and sustained increase in demand experienced from November 2017 until April 2018. Distribution Graph 5 - Response Times to Amber Codes (November 2016 – November 2017)

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59

Dem

and


Amber Incidents Distribution


DRAFT



Distribution Graph 6 – Response Times to Amber Codes (November 2017 – June 2018)

Resource Dispatched to Scene Dispatching the “right resource first time” to amber response category patients is important. This is due to the NCRM principles that consider facilitating the onward acute transport need as one of the key interventions. There has been a considerable increase in this practice since the launch of NCRM, although as outlined in run chart 1 below, this has fallen off somewhat in the last 6 months for amber calls. There are a number of factors that have influenced this; including increase in emergency demand and a change to dispatch were we will send a single PRU to an amber call if there is no conveying resource (ambulance) available within 20 minutes.

This change in dispatch arrangements was made to the system after engagement with and feedback from staff.

0

1,000

2,000

3,000

4,000

5,000

0 10 20 30 40 50 60 70 80 90 100 110 122 142 159

De

ma

nd


Amber Incidents Distribution 27/11/2017 - 30/06/2018

Incident Count

DRAFT



Run Chart 1 - % Incidents with 1 resource on Scene

Identified Benefits

Amber response category patients have a defined onward acute pathway need with minimal identified need for resuscitation at scene. Following NCRM principles of “responding with the right resource first time” a key response to these patients is to respond with an emergency ambulance with the ability to transport the patient to hospital/acute pathway as a key treatment intervention has been delivered over the pilot period.

The NCRM concurrent development with the SAS national stroke lead, integrated a change in practice in relation to stroke patients for whom the onset of symptoms is unknown. The previous response was dealt with conservatively due to the unknown onset element of the stroke symptoms; the change in practice is now to treat such patients as hyper-acute status until a diagnosis can be made by a CT scan in hospital. Over the 12 month initial pilot this change has improved the response to 804 unknown time frame stroke patients with 25.3% (203) of patients being recorded as having a FAST positive hyper-acute stroke from clinical assessment.

Acute Pathway Care - Patient Journeys One of the main focuses of the NCRM was to consider the patients‟ clinical needs on scene and the onward transport requirements to an in-hospital pathway care. The amber response category contains two main patient groups who were re-triaged from the old Cat A model. These were certain codes of patients with chest pain symptoms and certain codes of patients with breathing problems.

The focus on patients with a presenting symptom of chest pain is important as a high volume patient group that SAS clinicians can diagnose with a 12 lead ECG in the pre-hospital environment, identifying those with ST elevation Myocardial Infarction (STEMI) suitable for direct conveyance to definitive care in a specialist Percutaneous Coronary Intervention (PCI) centre.

NCRM

Intr

oduc

ed

Data

cont

inue

d fro

m 2

7/11

/201

777.5% Median

90.8% Median

86.4% Median

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

80.0%

90.0%

100.0%

60%

65%

70%

75%

80%

85%

90%

95%

100%

% Incidents with 1 Resource on Scene (Amber)

DRAFT



The following data in Distribution Graph 7 (12 month pilot period) demonstrates that post NCRM, there has been an observed increase of 22% in the volume of patients with STEMI diagnosed, treated, triaged and transported, where geographically possible to PCI centres. Information around the entire journey time for this specific sub set of STEMI patients to definitive care outlined below (Table 13) demonstrates there is a slight increase in the median and 90th percentile of journey times to definitive care post NCRM. However, considering a more detailed journey breakdown the small increase in time to PCI relates to an increase in all elements of the response cycle, with initial response only being a small part of this. We are continuously monitoring and reviewing on scene times with a view to optimising this pathway.

Distribution Graph 7 - Amber STEMI Incident times to Definitive Care (November 2016 to November 2017)

Table 13 - Journey break down STEMI patients Year 1 NCRM

Allocation to mobilisation

Mobilised to at scene

Arrived scene to left scene

Left scene to arrive hospital

Median

Pre NCRM 0.62 6.32 32.10 14.38

Post NCRM 0.63 7.67 35.53 15.85

90th

Percentile

Pre NCRM 1.68 15.55 50.47 39.52

Post NCRM 1.93 17.37 52.84 41.96

0

10

20

30

40

50

60

18 23 27 31 35 39 43 47 51 55 59 63 67 71 75 79 83 87 91 95 99 103 107 111 115 119 123 127 131 135 139 143 147 152 156 160 167 172 178

Inci

dent

s

Call to Hospital Time

Amber MI Incidents

Incident Count Pre NCRM Incident Count Post NCRM

DRAFT



8.4 Yellow Response Category

Key Points This cohort of patients presented with a low need for immediate resuscitation and a variety of clinical responses that required a face to face consultation and assessment with onward care options including on scene definitive care or referral to the nearest Emergency Department. A range of responses had been defined and ongoing testing of appropriate response is progressing. There was an expectation that volumes of calls in this category would remain broadly similar. This has been observed as predicted. Inclusion criteria; cardiac arrest rate < 1% actual figure 0.1%) and no clinical need for definitive care pathway (eg PCI or stroke unit) Response; identification and dispatch of the nearest appropriate resource. The type of resource and skill set are determined by the individual code and resource availability. Results

Low rates of deterioration and stable mortality rates.

Optimal dispatch of resourcing increased from 85.8% to 93.1% post NCRM launch.

Significant reduction in „stand down‟ (where a crew is already on their way to the patients address, and subsequently instructed to stand down) post launch of Dispatch on Disposition element of NCRM.

Median response times stable at 13 minutes in year 1. In year 2 we have seen an increase in 90th percentile response times. Efforts to address this trend robustly are ongoing and require a range of factors out with NCRM to be addressed.

The yellow response category is the fourth tier of emergency response categories. The category is defined as an emergency response category, but without the high likelihood of resuscitation efforts or the requirement for a defined care pathway out with Emergency Department conveyance. Identification Following the NCRM governance arrangements patient clinical need was identified and triaged using the agreed 10 clinical measures; these are displayed below for the Yellow response category.

Tables 14 and 15 below indicate a stable system in terms of clinical acuity and

interventions that are broadly as predicted pre-NCRM. The increase in patient

interventions noted post year 1 NCRM reflects improvements in data collection

following the introduction of the new EPR.

Table 14 - Yellow Response Category Initial 12 Months. Volume Transported Airway Breathing Circulation Drugs NEWS+4 Pre-


Arrest

187569 68.7% 0.4% 0.1% 5.1% 15.4% 10.3% 2.3% 0.1% 0.7% 0.1%

DRAFT



95% CI 68.5- 68.9%

0.4- 0.4%

0.1- 0.1%

5.0- 5.2%

15.2- 15.6%

10.1-10.4%

2.3-2.4%

0.1-0.1%

0.6-0.7%

0.1- 0.1%

Table 15 - Yellow Response Category Extended 7-month pilot period.



Arrest

107600

71.4% 1% 0.59% 10.6% 27% 18% 15.7% 0.4% 1.6% 0.2%

Linked Mortality Data NCRM project has linked outcome mortality data from linked NHS Scotland data. Using like for like codes in a pre and post sample analysis the yellow response

category has shown no significant change in 24 hour or 30 day survival rates. Table 16 - Yellow Category, Linked Survival Data


Total SAS Incidents recorded by SAS


SAS Incidents linked by

ISD


Yellow Codes Pre Sample 148,702 98.4% 144,964 95.9% 151,138 200,106 75.5%

Yellow Codes Post Sample 148,459 97.5% 145,347 95.5% 152,234 196,394 77.5%

Table 16 above outlines a stable system in terms of 24 hour and 30 day morality,

call volumes and data linkage.

DRAFT



Distribution Graph 8 - Yellow Incident Response Times (November 2016 to November 2017)

Distribution Graph 9 - Yellow Incident Response Times (November 2017 – June 2018)

Our work in the yellow response category highlights a 4.2% incident volume increase in the NCRM. The distribution graphs 8 and 9 above, demonstrates a consistent volume of yellow response category incidents. The response time to these patients has extended across the response model pilot, with a small proportion of patients experiencing longer delays. The causes of the extended response times is under investigation and is linked to multiple operational interdependences, including increase in higher category demand (particularly

0

5,000

10,000

15,000

20,0000 2 4 6 8

10

12

14

16

18

20

22

24

26

28

30

32

34

36

38

40

42

44

46

48

50

52

54

56

58

60

62

64

66

68

70

72

74

76

78

80

82

84

86

88

90

92

94

96

98

10

0

De

ma

nd


Yellow Incidents Distribution


0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

9,000

10,000

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180+

Dem

an

d


Yellow Incidents Distribution 27/11/2017 - 30/06/2018

Incident Count

DRAFT



since November 2017), increasing job cycle times and delays in hospital turnaround times. The clinical data supporting the yellow response category demonstrates clinical stability throughout the NCRM project, with a small reduction in mortality across this group. The Service has now received an independent report which explains the additional resources required to better match supply of resources to patient demand in order to fundamentally resolve the issue of delays in response which are experienced by a small proportion of these patients. This independent report is broadly in accordance with the Service‟s 2015-2020 Workforce Plan which underpins the 2020 strategy delivery and will see 1000 additional paramedics trained over these five years.

Resource Dispatched to Scene One of the aims of the project was to use resources as effectively as possible and dispatch “the right resource first time”. This aim was applied across the yellow response category to analyse each individual patient need, and dispatch the most appropriate resource. A conveying resource was often aligned as the optimal response to these patients.

Run Chart 4 - % Incidents with 1 Resource on Scene (Yellow)

One sub set of yellow response category patient were identified as being suitable for response by a single crewed Paramedic Response Unit (PRU) resource. This sub category consists of patients with an identified low onward transport need to hospital and within the scope of practice of a PRU clinician. As demonstrated in Run Chart 5 there is an increase in PRU resources directed to these appropriate patients.

NCR

M In

trod

uced

Dat

a co

ntin

ued

from

27/

11/2

017

Median - 85.8%

Median 93.1%Median - 92.3%

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

80.0%

90.0%

100.0%

70%

75%

80%

85%

90%

95%

100%

% Incidents with 1 Resource on Scene (Yellow)

DRAFT



Run chart 5 - % Yellow B3 Incidents where PRU was first on scene

Demand Profile The line charts below outline the profile of demand that SAS has experienced over time. Despite sustained increases in demand, response to and outcomes for patients have been well maintained, however where we have seen delays these are to patients in the lower acuity categories.

Line Chart 1 – Emergency Incidents attended per day

NCRM

Intro

duce

d

Data

cont

inued

from

27/1

1/20

17

15.32%

30.85%

27.08%

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0.00%

10.00%

20.00%

30.00%

40.00%

50.00%

60.00%

30/1

1/20

15

31/1

2/20

15

31/0

1/20

16

29/0

2/20

16

31/0

3/20

16

30/0

4/20

16

31/0

5/20

16

30/0

6/20

16

31/0

7/20

16

31/0

8/20

16

30/0

9/20

16

31/1

0/20

16

30/1

1/20

16

31/1

2/20

16

31/0

1/20

17

28/0

2/20

17

31/0

3/20

17

30/0

4/20

17

31/0

5/20

17

30/0

6/20

17

31/0

7/20

17

31/0

8/20

17

30/0

9/20

17

31/1

0/20

17

30/1

1/20

17

31/1

2/20

17

31/0

1/20

18

28/0

2/20

18

31/0

3/20

18

30/0

4/20

18

31/0

5/20

18

% Yellow B3 Incidents where PRU First on Scene

% Incidents where PRU First on Scene Median % Incidents where PRU First on Scene

1,200

1,250

1,300

1,350

1,400

1,450

1,500

1,550

1,600

Dec-16 Jan-17 Feb-17 Mar-17 Apr-17 May-17 Jun-17 Jul-17 Aug-17 Sep-17 Oct-17 Nov-17 Dec-17 Jan-18 Feb-18 Mar-18 Apr-18 May-18 Jun-18

Average Emergency Incidents Attended per Day

DRAFT



Line Chart 2 – Total Purple incidents attended

Line Chart 3 - Total Red incidents attended

9 Staff Engagement and Experience

Throughout the planning and delivery of NCRM, staff partnership colleagues have been closely involved in shaping the processes and have been a key element of SAS staff engagement internal communication planning. The following elements, among others, have been in place;

Partnership representation on the NCRM project team

0

50

100

150

200

250

300

Purple Incidents Attended

Purple Median

0

200

400

600

800

1,000

1,200

1,400

1,600

1,800

Red Incidents Attended

Red Median

DRAFT



Partnership representation on SAS Clinical Services Transformation Steering Group

Engagement with Partnership around agreed changes to Standard Operating Procedures

Internal communication plan delivered as planned

Webinar Q&As with CEO and Medical Director

Dedicated email for staff to ask questions, provide feedback or raise concerns

NCRM discussed at multiple station visits by Board Members and Exec Team

Changes to model based on staff feedback around reducing stand downs and mobilising PRU resources.

Among the biggest concerns of SAS staff that adversely affect experience are the areas of non-compliance with rest periods, end of shift overruns and delays in hospital turnaround. Agreement was reached to revise the rest period and end of shift SOPs early in the model development which has enabled rest period & end of shift compliance to remain at stable levels broadly since spring 2017. Significant efforts to minimise hospital turnaround times are progressed through SAS engagement with the National Unscheduled Care work streams and through local discussion with Health Board partners.

10 Conclusions

Changing the triage system for emergency care within the Scottish Ambulance Service has been a complex and complicated process. The new system of care required redesign across the service from patient call to definitive care to achieve the aim of responding based on clinical need and in doing so improve outcomes.

The systematic review of the data provided a clear evidence base for change. Implementing the changes required staff support and the new system was co-designed with frontline staff with the recognition that improvement was required to be more efficient and effective as a service. Several highlights from the data demonstrate the following improvements.

The model is working well and as planned, particularly in relation to the identification response and improved outcomes for our most seriously unwell patients.

The modelling predictions carried out pre-launch and post new clinical response have been realised. This has been validated by the data captured since NCRM launch, supporting the validity of the response category descriptors.

The purple response category shows an improvement in 30 day survival rates from 23.9% pre NCRM to 34.1% post NCRM, an increase of 43%. The survival data shows that this equates to 1182 people.

Response times are stable across the system for our most acutely unwell patients (purple and red categories).

DRAFT



During the pilot year, the Amber median and 90th percentile response times have shown improvement. Median response has reduced from 11.02 minutes to 10.43. 90th percentile has reduced from 23.87 minutes to 20.95. (Distribution Graph 5).

In the Amber category, there has been an increase in the number of patients with chest pain overall and the sub set of patients presenting with STEMI conveyed to PCI centres. The model has demonstrated it has effectively dealt with this increase and this volume will be continually reviewed.

There has been an increase in response times to patients in the yellow codes as expected. However, linked mortality data appears to be stable for this category of patients who are at lower clinical risk. We recognise, however, that there is more improvement work to do for this category of patients and new approaches to managing these patients are being implemented. (Distribution Graph 8 and Table 16).

With the use of median and 90th percentile response data, this has improved SAS ability to understand and manage demand for all of patients and presents a much improved measurement framework than the previous 75% in 8 minute target for only 32% of our patients.

There have been no Serious Adverse Events reported related to the new response re-triage and clinical hierarchy.

There is evidence of improved efficiency and more accurate allocation of resources since NCRM implementation. There has been an increase in allocation of the right resource first time, with a reduction in inappropriate double resourcing to amber and yellow calls (Run chart 5).

11 Recommendations

Implement the system and process changes developed within the pilot period of NCRM to business as usual within SAS.

As with services in England and Wales, replace the 8 minute response target with agreed median and 90th percentile measures for patients in the defined categories to allow a more accurate and meaningful analysis of performance.

Create additional quality indicator measures that focus on patient and staff outcomes, continually developing how this data for improvement is actively shared with clinicians.

Further improve the experience of patients within the yellow category who are waiting longer by developing access to alternative care pathways from face to face or additional telephone triage in line with the Service‟s 2020 Strategy.

Continue to review demand and capacity and implement models that ensure a better match of resourcing to demand.

DRAFT



Continue to invest in the additional workforce requirements determined by the independent demand and capacity modelling review (May 2018).

Continue to showcase our work in SAS as a key part of Scotland‟s Health and Care system, nationally and internationally.

DRAFT



Glossary 3RU Rapid Resuscitation Response (Unit) ACC Ambulance Control Centre CAG Clinical Advisory Group CA Clinical Advisor CI Confidence Interval CQI Clinical Quality Indicator ED Emergency Department EPR Electronic Patient Report HCP Health Care Professional ISD Information Services Division ILT Immediately Life Threatening NCRM New Clinical Response Model NEWS National Early Warning Score NHS National Health Service NSS National Services Scotland OHCA Out of Hospital Cardiac Arrest PAD Public Access Defibrillator PCI Percutaneous Coronary Intervention PRU Paramedic Reponses Unit ROSC Return of Spontaneous Circulation SAER Significant Adverse Event Review SAS Scottish Ambulance Service SOP‟s Standard Operating Procedure‟s STEMI ST Elevation Myocardial Infarction VF Ventricular Fibrillation VT Ventricular Tachycardia References 1) Bigham BL; Kennedy SM; Drennan I; Morrison LJ (2013) Expanding paramedic

scope of practice in the community: a systematic review of the literature Prehospital Emergency Care 17 (36) 361 - 372

2) Eaton G. and Broughton W. The Four Pillars: Outlining and aligning our future Journal of Paramedic Practice Vol. 10:, 22-23

3) PACEC with Turner J (2017) Clinical model pilot evaluation (final report) NHS Wales Ambulance Service

4) Turner J; Jacques R; Crum A; Coster J ; Stone T; Nicholl J (2017) Ambulance Response programme. Evaluation of Phase 1 and Phase 2 NHS England

5) Review of targets and indicators 6) Association of Ambulance Chief Executive National Ambulance Service Medical

Director Statements of Best Clinical Practice Adult Cardiac Arrest Care September 2016

7) Scottish Government Out of Hospital Cardiac Arrest Strategy for Scotland 2015

Clinical Response Model Pilot Evaluation Report February 2019 · Doc: Clinical Response Model Pilot...

Documents

Transcript of Clinical Response Model Pilot Evaluation Report February 2019 · Doc: Clinical Response Model Pilot...