Electronic health records in ambulances: the ERA multiple-methods study

Research aim

To understand how EHRs can be most effectively implemented in a pre-hospital context in order to support a safe and effective shift from acute care to community-based care, and how their potential benefits can be maximised.

Objectives

1. To describe the current usage of EHRs and associated IT in ambulance services in the UK:

   1. – To describe processes of implementation, uptake and usage.

   2. – To investigate what use is currently being made of EHRs in terms of identifying and managing repeat callers, information transfer to other providers, linking with other electronic resources (e.g. for decision support and referral), and research and audit.

   3. – To investigate the use and development in ambulance services of other handheld technologies (including apps) to support decision-making and referral to community-based care.

2. To understand how the ambulance workforce responds to the introduction of EHRs and associated infrastructure, and what impact they are perceived to have on the role of ambulance clinicians.

3. To investigate risks, benefits and unintended consequences of implementation of EHRs in terms of changes to patient care, working practice of ambulance clinicians, management and organisational practice within ambulance services, and planning and commissioning processes in the wider health economy.

4. To understand the factors that lead to successful implementation of EHRs and adoption by the workforce, and how risks can be minimised and benefits can be maximised.

5. To assess the potential to further develop and implement EHRs, computerised clinical decision-support and referral tools to support the shift to out-of-hospital care.

Work package 2

All free-standing UK ambulance services (n = 13) were used: the three national services for Wales, Scotland and Northern Ireland, and the 10 independent ambulance service trusts in England. The only service not included was the small integrated health-care trust that provides an emergency ambulance service on the Isle of Wight. In each site, we invited, by e-mail, up to two senior or middle managers with knowledge of EHRs to take part in an interview. Potential participants were identified through our professional networks.

Work package 3

Four selected ambulance services, along with their associated health economies, were used. We selected the ambulance services to give a range of experiences. At the time of selection (2015), one service was an established user of EHRs, one had partially adopted EHRs, one was anticipated to be in the early stages of implementation during the time of the fieldwork and one was still discussing introduction of EHRs. Representatives of each ambulance service were co-applicants on this proposal.

Work package 2: survey

We conducted semistructured telephone interviews with senior and middle managers in each service to gather descriptive date regarding the implementation, usage and development of IT in pre-hospital care settings in the UK. Building on the rapid review, we developed an interview schedule that covered the current stage of implementation of EHRs in the service, reasons for introduction or non-introduction of EHRs perceived benefits and uses of EHRs, impediments or barriers to their use and other electronic tools used by ambulance clinicians and how they interface with EHRs. If EHRs had not yet been introduced in the ambulance service, we asked about future plans, perceived benefits of EHRs and perceived impediments to their introduction. Verbal consent to participate was taken at the beginning of the telephone call and participants could decline to answer questions or cease to participate simply by indicating that they wished to do so. Interviews were recorded and transcribed in full; respondent identifiers were removed from the transcripts. Interviews were conducted by Victoria Williams during the period February to August 2017.

Work package 3a: case studies

We undertook case studies in four ambulance services and their associated health economies to address objectives 2, 3 and 4. Each case study examined the story of the implementation and use of EHRs in the site, incorporating descriptive analysis of any available routine data on uptake, use and impact of EHRs. We adapted some principles of ethnographic research to incorporate observation of practice on ambulances. To facilitate comparison across sites, we included a particular focus on three tracer conditions known to have potential for increased non-conveyance rates: falls in older people, diabetic hypoglycaemia and mental health crises. We examined how, in each study site, out-of-hospital care pathways are being provided for these patient groups, and how ambulance service data are (or could be) informing both care provision and planning/commissioning decisions.

In each site, we had a part-time site researcher; the role was split between two members of staff in site 2, although only one of them carried out observations. Two researchers (in sites 1 and 4) were registered paramedics. In three sites, the site researchers were employed by the ambulance trust; in the fourth (site 3), the researcher was employed by one of the collaborating universities but maintained close links with the trust. The site researchers’ reflections on their role are summarised in Reflexivity and the site researchers. Fieldwork took place during the period April to October 2017, and involved collecting the following data (also summarised in Table 2):

• Documentation (business case, minutes, etc.) to track the story and timeline of implementation, including costs (planned and actual) and collection of descriptive data on the EHR systems themselves in terms of hardware, software, data categories and flexibility.

• Interviews with relevant senior and middle managers and training staff (identified by the local ambulance service lead) within ambulance services to gather their views on the rationale for introducing EHRs in their ambulance service, the way in which the records were introduced, the challenges to introduction and the opportunities presented by EHRs. Interviews were carried out face to face and recorded and transcribed in full. We asked about reasons for non-introduction or delay in introduction, and the choices that may have been made, including decisions relating to the opportunity costs associated with introduction of records. We also asked about care pathways (actual or in development) for the three tracer conditions, and the perceived relevance of EHRs to these.

• Focus groups with paramedics and ambulance technicians – recorded and transcribed in full, involving participants from a range of ambulance stations within the service. In collaboration with each ambulance service, the site researchers selected a random sample of ambulance clinicians to invite to the focus groups, and thus aimed to ensure that we talked to people with a range of levels of enthusiasm for EHRs. Site researchers had aimed to carry out a focus group in each of the stations from which they carried out their observation, giving a target of three focus groups per case study area. However, it proved difficult to recruit sufficient numbers of staff in several of the areas, largely due to shift patterns that made it difficult to get enough people to attend at any one time. Site researchers tackled this in several ways. In one case study area, the site researcher held a large focus group with clinicians attending a training session, thus capturing views from practitioners from a range of localities in the service. Another researcher held several smaller groups to try to gather a range of views.

  Focus groups were conducted within ambulance stations, and we paid participants an honorarium for taking part. The discussion covered participants’ views on the rationale for the introduction of EHRs in their ambulance service, the way in which the records were introduced, the challenges to the introduction, their experience of using EHRs in practice (where relevant) and the opportunities presented by EHRs. We also asked about care pathways (actual or in development) for the three tracer conditions, and the perceived relevance of EHRs to these.

• Observation of EHRs in use. Site researchers carried out a total of 12 full-shift observations with different crews from different ambulance stations. Although our original plan had been to carry out three observations in each site, owing to local issues in site 3 the site researcher was able to carry out only two observations. In site 2, where a new system was being rolled out, the site researcher decided to observe a fourth shift in an attempt to capture some of the changes to the way patient data were gathered and transferred.

  Observations were carried out by the site researcher over the course of a full shift and covered both day and night shifts, modelled on an approach previously used in a study by O’Hara et al. 12 of decision-making in pre-hospital care. The site researcher made recorded observations and reflections using a pre-formatted site researcher notebook (the edited version is in Appendix 3; the full version is available at www.journalslibrary.nihr.ac.uk/programmes/hsdr/144722#/) and a digital recorder, and discussed the use of records with the crews between episodes of patient contact. Shifts were selected to ensure that different clinicians were shadowed, and that observations took place at different times and in different localities. The site researchers were trained in participant observation. They observed the use of electronic and other records (when completed, by whom, processes undertaken) and the use of other data sources to support decision-making. At times when there was no patient in the ambulance, the researchers asked ambulance clinicians for clarification or explanation of processes and decision-making, and then recorded this information.

• Interviews with other stakeholders, including representatives of the commissioning or funding organisation(s), EDs and community health-care providers. Interviews covered participants’ views on the rationale for the introduction of EHRs in the ambulance service, and their perception of the opportunities and challenges presented by EHRs. We also asked about care pathways (actual or in development) for the three tracer conditions, and the perceived relevance of EHRs to these.

Work package 3b: routine data

• Extraction and collation of a sample of routine data to explore the use of electronic patient records. Following the signing of data-sharing agreements with each case study ambulance service, we collected computer-aided dispatch (CAD) data relating to incidents (calls) and resources (vehicles and crews) alongside EHR data for a 2-month ‘snapshot’ period of January and February 2017 from all four ambulance services. Table 3 indicates the variables requested from each service, and the data source from which they were obtained. Patient identifiers were removed before data transfer. We also requested from each service a cover sheet describing the parameters of the data (i.e. date extracted, data sources, data coverage and contact details of those extracting data) and a data dictionary describing and defining the variables (e.g. data source – CAD, EHRs, other; and date ranges).

  Following data cleaning, we analysed and profiled the data to show, at each site, the total number of calls attended with an EHR, and identified associations between completion of an EHR and the patient demographic (age and sex), clinical factors (condition type and severity from triage category) and operational factors [crew member identifier (ID), ambulance station, time of day, disposition of patient, hospital to which a patient was conveyed], as recorded in the call centre.

• We investigated what routinely available data there were (time period, nature of data source) on ambulance attendances and consequent care for the three tracer conditions, in order to understand patterns of disposition and care pathways. This work was primarily exploratory. We sought to establish a minimum data set across all sites based on dispatch codes in order to describe patterns of activity, and looked at how any additional data were collected and used to inform decision-making.

Work package 4: knowledge exchange workshop

We presented our initial findings at a knowledge exchange workshop on 3 July 2018 in central London. The workshop functioned both as a validity check of the findings and as a source of new data. We supplemented the workshop with online and telephone contact with stakeholders both before the workshop (to help to inform the plan for the workshop and to gain initial responses to study findings) and after it (to seek endorsement/comment in relation to conclusions and recommendations). This contact allowed us to obtain input from people who were not able to attend the workshop in person, to gather reflections formed over time rather than immediate responses and to obtain comments in confidence that people might be reluctant to express in a group setting.

The workshop was audio-recorded and notes were made from the recording. WP4 brought together UK ambulance services, secondary care clinicians, patient and public involvement (PPI) representatives, relevant IT suppliers and other relevant policy-makers. We shared learning on successful implementation and discussed potential for, and possible challenges to, future development of EHRs in relation to identifying and managing repeat callers, information transfer, linking with other electronic resources, and research and analysis.

Framework approach

Analysis of the qualitative data drew on the framework approach to analysis for applied policy research. 45 Framework provides a method for researchers to systematically sort and analyse data in relation to key issues and themes within a finite time frame. Framework involves five stages of analysis. The first is familiarisation, whereby a researcher reads the data and begins to form ideas about emerging issues and themes. Next is identifying a thematic framework, based on these issues, and informed by the original research aims and questions; this framework may be revised as analysis progresses. In the indexing stage, the framework is applied to the data, and the data are linked to the themes. Charting the data then entails the researcher reviewing the data theme by theme, and relating text to themes and subthemes. Finally, the mapping stage involves the researcher synthesising and interpreting the data in order to define concepts, map the range and nature of phenomena and provide explanations in relation to the research questions and objectives.

Work package 2: survey

Analysis of the transcripts of the survey interviews was carried out by two members of the research team (VW and AP).

Work package 3: case studies

Analysis of the large number of data (transcripts, observation notes and documentation) generated by the case studies was supported by the use of NVivo 11 software (QSR International, Warrington, UK). As above, our analysis followed the framework approach. The analysis was carried out on a team basis by up to nine members of the research group with experience in qualitative analysis (AP, VW, ZM, HP, KM, DS, RS, HM, YL-W). Analysis began with a workshop event bringing together members of the research team and PPI representatives. Prior to the event, each participant read a sample of texts (observations, interview transcripts and focus groups transcripts). Each text was analysed by a minimum of two researchers for validation.

Routine data about the uptake and use of EHRs were analysed descriptively, and multivariate analysis was used to identify associations and predictors of use.

Analysis of all case study data was first carried out within sites, and was followed by cross-case analysis across sites. 46 Our analysis drew on SST with a technological dimension in order to understand how new processes are, or are not, adopted in practice. 20,44 In each site, we examined the role and practice of the key agent (in this case, the ambulance clinician) using data from observation and focus groups; the political and organisational context, using interview data and documentation; the technology itself, drawing on documentation and observation; and the recursive relationship between these elements, using routine data on usage, interview and focus group information relating to feedback and performance management, and data on response to and provision for the three tracer conditions. We then compared across sites in order to understand consistency or variation in experience, the significance of time in the implementation process and the role of local context.

Work package 4: knowledge exchange workshop and online forum

Consensus methods were used during the workshop to develop shared understanding of future opportunities around use of EHRs in ambulance services. 47

Synthesis

Table 4 summarises how each WP addressed the study’s objectives. The research team members were in regular contact with each other, allowing the different WPs to inform each other’s development. A final overarching synthesis is presented in the discussion and conclusions chapter (see Chapter 9), bringing together the findings of each WP, presenting a summary, interpreting findings, drawing out common themes and presenting any tensions and conflicts between the WPs’ conclusions as higher-order data.

Eligibility criteria

We included articles that described experimental or observational studies, including case studies, carried out in the pre-hospital setting and that made reference to EHRs. Articles that did not describe any form of investigation – opinion pieces, advertorial articles, editorials and theoretical design articles – were not included.

Information sources and search strategy

We searched the following databases: Cumulative Index to Nursing and Allied Health Literature (CINAHL), The Cochrane Library, Health Management Information Consortium, MEDLINE, Scopus^® (Elsevier, Amsterdam, the Netherlands) and Web of Science™ (Clarivate Analytics, Philadelphia, PA, USA). We did not limit for date, nor did we use any study design restrictions. Keywords and medical subject heading (MeSH) terms used to conduct the search are presented in Box 2.

Study selection

Studies were initially screened by title and abstract content following database searches. Full papers that were deemed suitable based on title and abstract content were then assessed to confirm suitability. The search and selection process is detailed in Figure 2. Our search of the literature in seven databases returned 1464 results. A total of 229 duplicates were removed and 1235 articles were screened. Of the 1235 articles, 1217 articles did not meet the eligibility criteria and were excluded. Data were extracted from the remaining 18 articles, using a standard record sheet, by Matthew Jones and Anisha Badshah.

FIGURE 2.

Study selection flow diagram.

Data reporting

We present findings within seven themes that we developed inductively: (1) implementation and adoption of EHRs, (2) EHRs in use, (3) determining clinical status using EHRs, (4) quality of EHR documentation, (5) EHRs’ impact on call duration, (6) EHRs and real-time data-sharing and (7) EHRs and retrospective data linkage.

Overview

The characteristics of the 18 included studies are summarised in Table 5. All but one study were set in a single country, with one study sampling from the USA and Canada. 49 The majority of the studies (n = 10) were conducted in the USA,50–59 two were conducted in Japan,60,61 two in Australia65,66 and one was conducted in each of Finland,62 Italy63 and the UK. 64 The identified literature is of limited quality. The majority of the empirical research was conducted with small, private emergency services providers with small sample sizes. Three of the studies consider one EHR system within scenario testing over the course of its development. 55,56,58 Publication dates range from 2004 to 2016; the age of the literature is an important consideration as earlier studies illustrate software deficiencies that would not be anticipated today. 59

Objectives of system implementation

Three studies detailed the objectives of EHR system implementation; all of them were conducted in North America. Baird and Boak50 carried out a case study of one small EMS in the USA 20 months post implementation of an EHR system. They gathered both quantitative and qualitative data from users of the system through interviews with representatives of leadership, mid-management, trainers and junior paramedics, and 29 follow-up surveys completed by paramedics. They found improved billing and quality assurance to be primary motivating factors behind the implementation of an EHR system. Similarly, Landman et al. ,49 through interviews with 23 representatives of 20 EMS agencies in the USA and Canada (medical directors, emergency physicians and EMS agency administrative staff), found improved quality assurance to be a primary objective of the implementation of an EHR system through the increased availability of patient records and improved reporting functionality. Other motivating factors were improved billing, improved legibility of patient records, reduced data loss and state mandates requiring services to adopt EHRs. Similarly, Broussard57 identified quality improvement to be a motivating factor in the development and implementation of an in-house EHR system by a large private ambulance service in the USA. The service was using EHRs for billing purposes, such as generating invoices and for reimbursements, clinical audits, workforce training needs analysis and education.

Barriers to implementation

Three studies identified barriers to the implementation of EHR systems. Landman et al. 49 discussed the challenges posed by high start-up costs, lack of financial resources, lack of leadership, complex organisational structures, poor usability of software, unreliable vendors, concerns about privacy and security, concerns about increased run times and other service disruption, and concerns about difficulty integrating systems within existing ED or hospital systems. In Baird and Boak’s50 study, integration with ED systems was found to be challenging, as 67% of participants reported frustration with a tablet stylus interface. Similarly, Broussard57 found that paramedics faced initial difficulty with the interface for reporting, as long hold times of 15–20 minutes were problematic when EMTs were calling in their reports to live operators who would follow the EHR software template.

Strategies to support implementation

Having identified challenges to implementation, Broussard57 reported on the introduction of two options for alternative interfaces through which information could be recorded onto EHRs: (1) leaving a voicemail that was then transcribed, or (2) entering reports on a laptop or computer during downtime or after shifts and e-mailing them to dispatch centres. The latter was the preferred method, used by 70% of staff; 20% used dictation and 10% called live operators. Baird and Boak50 found that the introduction of optional additional hardware and software tools (e.g. keyboard and private laptop) aided adoption, with the majority of users (86%) finding the system easier to use with the addition of extra hardware tools. All participants in Baird and Boak’s50 research suggested that additional formal training may have aided earlier adoption, and the majority (79%) found informal learning from peers to be very helpful or somewhat helpful. Both Landman et al. ’s49 and Baird and Boak’s50 findings highlight the importance of support and training in aiding adoption.

In Landman et al. ’s49 study, those EMSs that had already adopted EHR systems (14/20) were using strategies to overcome challenges, such as identifying alternative sources of funding (e.g. state road safety officials, billing companies or EHR system vendors) working with regional information organisations to exchange EHR systems data and funding dedicated IT personnel to support the process of implementation and adoption.

How electronic health records were used

In Baird and Boak’s50 study, participants were found to have various ways in which they used EHRs: filling in EHRs before arriving at the scene of a call, on the way back to the station, on returning or at the end of their shift after several calls. Buswell et al. 64 found varying preferences of ambulance personnel between the use of paper records and EHRs when attending to older patients. They conducted a retrospective chart review of 373 patients aged > 65 years who were attended by paramedics over a 1-month period. They found that ambulance clinicians were more likely to create a paper record than to use EHRs if discharging patients at their homes, leaving patients and their carers with a copy of the record; only 9 out of 187 (4.85%) patients for whom an EHR was created were left at the scene, whereas 67 out of 186 (36.0%) patients with paper clinical records were left at the scene. Bledsoe et al. 52 surveyed 228 emergency physicians to find out how they used EHRs generated by ambulance clinicians. Handwritten records were found to be more readily available at the time of physician decision-making than EHRs (22.4% vs. 20.4%), and EHRs were not used to support medical decision-making during the provision of care.

Perceived benefits of electronic health records

Five studies examined the perceived benefits of using EHRs, mainly through the comparison of EHRs with paper clinical records. Burley et al. 65 conducted 45 interviews with various stakeholders of an EMS organisation in Australia, including 38 paramedics and seven administrative, IT and management staff, before and after the implementation of an in-house EHR system, which replaced a paper-based system. They found perceived advantages to the new system to include improved legibility, standardisation of reports, reduced documentation, fast data acquisition, speedier record retrieval and scope for easier statistical analysis of patient data. Bledsoe et al. 52 found that the majority (81.6%) of emergency physicians in their survey had encountered EHRs in their practice at least once, and over half (52.2%) preferred EHRs to handwritten records (17.1%); 75% of respondents cited legibility as a benefit of EHRs, whereas only 1% reported legibility as a beneficial aspect of paper reports. Chittaro et al. 63 evaluated the usability of a personal digital assistant (PDA)-based EHR system. Following a short informal training session, six emergency responders in Italy filled in an EHR based on a described scenario of a rescue operation written by an emergency physician. Agreement was registered to a total of 21 statements describing the system using a Likert scale. The highest favourable scores were reported for screen legibility, ease of use, system information, system capabilities and overall reaction. In Baird and Boak’s50 study, the improved quality of charts was mentioned by all interviewees as a benefit to adopting an EHR system. Chittaro et al. 63 found that all six respondents were able to quickly learn how to use a PDA-based system without the need for specific training. Killeen et al. 58 also found this to be the case, with participants of a full-scale disaster drill being able to readily adapt to and use EHR devices (PDAs/tablet computers) with little or no training on the device beforehand. However, Baird and Boak50 found that more training could have aided adoption.

Perceived disbenefits of electronic health records

Four studies identified disbenefits of EHRs. The only disbenefit identified in Chittaro et al. ’s63 study of the PDA-based EHRs was the system’s handwriting recognition aspect, which was perceived as difficult to use, particularly on a moving ambulance and because screen space was insufficient, although two users suggested that this method of input could perform better with practice. Baird and Boak50 also noted the challenges of completing an EHR on a moving ambulance, with interviewees stating that many paramedics felt sick when trying to do so. Burley et al. 65 noted a loss of ‘narrative’ in records, suggesting that EHR systems did not facilitate the recording of qualitative data. Paramedics in Burley et al. ’s65 study also reported that during handover ED staff paid little attention to printed reports from EHRs, speculating that ED staff found it difficult to find what they were looking for. In Bledsoe et al. ’s52 study, 7.5% of respondents stated that there was no advantage to EHRs; however, a greater number (18.9%) believed that there was no advantage of paper records.

Improvements in documentation

The WIISARD EHR system was used by Chan et al. 55 and Lenert et al. 56 for evaluating EHR documentation during mass-casualty scenarios. Both studies compared the quality of documentation using EHRs with the quality of documentation using a paper documentation method. Chan et al. 55 compared 50 EHRs with 50 paper records during a simulated mass disaster. Lenert et al. ’s56 study similarly involved a simulated terrorist attack with 100 victims; 50 went through the WIISARD system and 50 through a paper-based system, and 39 EHRs and 32 paper records were included in analysis. Chan et al. 55 found that the greatest improvement in EHR quality of documentation was for perfusion rate (89.7%) and decontamination status (59%), which were not recorded at all in paper records. Lenert et al. 56 found that there were no data integrity problems in EHRs, whereas 47% of patients in the paper records group had some data integrity problems, including lost tags and duplicate identifiers. Field treatments received were recorded in the EHRs in 59% of cases, but not recorded at all in the paper records groups. Improvement of documentation of respiration rate was seen in Chan et al. ’s55 study: 92.3% of EHRs compared with 5.3% of paper records. Lenert et al. 56 found that age was recorded in 95% of cases and gender was recorded in 92% of cases in the EHRs, compared with 26% and 24% on paper, respectively. Chan et al. 55 similarly found that patient age was documented 94.9% of the time in the EHRs, compared with 26.3% in the paper group; gender was documented in 92.3% of cases in EHRs versus 23.7% in paper records, and mental status was documented in 64.1% of EHRs versus 2.6% in paper records. In Lenert et al. ’s56 EHRs, 73% of clinical information fields were completed, compared with 30% using a paper system.

Katzer et al. 53 carried out a retrospective chart review of a larger sample on the quality of documentation, at a small student-run EMS in the USA, comparing quality of physical examination documentation between paper records and EHRs. A total of 154 patient records were analysed (77 paper and 77 electronic). Records were grouped into three complaint categories (respiratory dysfunction, neurological dysfunction and those presenting problems from both). Each record was assessed for documentation completion of vital information depending on the category of complaint. For example, for the respiratory dysfunction category, completion rates for respiratory rate, visual appearance, quality of speech and auscultation were compared. Findings showed a mean increase of 36% in the completion of documentation of physical examination EHRs.

Loss of quality in documentation

Only one study, by Tollefsen et al. ,59 found a loss of quality in documentation using EHRs through the evaluation of an iRevive system, which comprised a PDA running EHR software, with decision-support functionality and wireless sensors that could monitor patients’ vital signs, with data relayed via a central server to receiving EDs. A simulated field trail was carried out alongside existing manual methods over 16 patient transports, involving 12 ambulance clinicians in the USA. The EHRs recorded and documented the same data as the existing manual approach. Problems were experienced by clinicians using the EHRs, including data fields in the ‘wrong’ order for system users (e.g. times and mileage pages placed too early in the process – requiring going back and forth); requirements to record data already recorded by dispatch centre such that EHRs could not be saved unless certain fields were completed; inability to return to previously saved EHRs until synched with the server; lack of appropriate values in some pull-down menus; and insufficient space to record certain clinical data, such as multiple allergies. Although the EHR system complied with local data set requirements, the need for more data-recording capabilities was identified along several different domains. The study did not find the EHRs superior to manual paper-based recording in terms of ease of use, data completion or data content. This may have been due to the technical capabilities of the EHRs given the age of the study, and such functional limitations may be less significant today.

Systems in use

Use of EHR systems at the time of the interviews (February to August 2017) is summarised in Table 7. Seven services had EHR systems in place. At the time of the interviews (February to August 2017), seven services had EHR systems in place. The length of time that systems had been in place ranged from over a decade to still being in the final stages of roll-out, with most services having had their systems in place for 2–3 years. Three of the services using EHRs were in the process of changing their systems completely, ceasing to use one provider and recommissioning from another. Others had opted for continuous changes and updates to both software and hardware in response to local need.

Six services did not use EHRs, of which two [ambulance service (AS) 11 and AS12] had never had an EHR system in place, although both were planning to introduce one. The other four (AS2, AS6, AS8 and AS9) had temporarily reverted to paper records while they were at various stages of designing, procuring and implementing new systems. These services reported having previously introduced EHRs under the Connecting for Health contract, and using the end of the contract as an opportunity to look for a new system better suited to their needs. Two services (AS4 and AS10) had not fully rolled out the old system to the whole service, largely due to the associated costs. Respondents from two services (AS8 and AS10) commented that they had received negative feedback from staff regarding the old system so were happy to have the opportunity to develop a new system.

AS4 took a different approach from all the other services, choosing, instead of a tablet, a digipen, a device used to write on a paper form that stored information through recording pen strokes.

AS1 and AS7 had installed printers on board ambulances, but most had chosen not to, citing them as unnecessary and a potential source of technical problems.

Time scale of implementation

Respondents described phased roll-out of new systems, generally area by area across a site. For most, this was a process that took around 6 months. As a representative of AS7 reported, this staged approach allowed attention to be given to the support the workforce needed to embed the new system into day-to-day practice:

Individual support through their clinical mentors . . . they’ve had presentations as a group and then we’ve also from the beginning run an [EHR] help desk which is a telephone support line, which we stopped running January of this year but we had the ability 24/7 . . . on the phone to take queries . . . That, I think, was one of the biggest drivers for getting people to use it.

AS7

An exception to this pace was AS13, which transitioned from one EHR system to another, and was keen to limit the return to paper in the interim. A respondent described their approach as:

We went big bang . . . it took about 2 months but we managed to move our full ambulance service across a 2-month period.

AS13

By contrast, AS1 had a much longer roll-out time, of around 19 months. This was determined by the fact that installing the new hardware (and removing the previous generation of EHR software) entailed taking vehicles off the road, so had to be done one at a time to minimise disruption to the service.

Support for the implementation process

Training was regarded as an important part of encouraging adoption. Although respondents did not think that EHR technology was particularly complicated (it was often likened to using a smartphone), they did acknowledge that ambulance clinicians were required to input information in a different way from how they would input information in paper forms, and that it took some time to learn their way around the system. Examples of training and support that were perceived as effective were a package of 1 day of training followed by support provided at each receiving hospital for the first few weeks after the system went live (AS1), and a telephone support line for front-line staff combined with the use of local clinical mentors to encourage a culture change (AS7).

Having a reliable system (devices and the EHR system) was also seen as an important part of staff adoption of EHRs. Battery failures, poor connectivity and an overcomplicated form that led to increased time on the scene all caused front-line staff to lose faith in the EHRs.

Completion rates

We asked respondents to estimate the proportion of patient calls attended that resulted in the creation of an EHR. Most of those currently using EHRs reported that rates of generation of EHRs were ≥ 90%. Services in the process of transition between systems reported that rates were much lower than this when their first generation of EHRs was taken offline, with use ranging from 34% down to 3%, these figures representing a decline in use over time. Some of the reasons presented for non-completion of EHRs are discussed in Problems and challenges associated with electronic health records.

Accessing existing patient records

Only a minority of respondents described ambulance clinicians being able to access previous medical records for patients they were attending. AS1, AS3 and AS7 appeared to be at the forefront of enabling this to happen. Although the EHR software did not link directly to previous patient records, ambulance clinicians were able to use the tablet to access the PDS and NHS Spine to look up information about their patients:

They can access the PDS from the tablet and they can do that at any point really. So they can put in a few details about the patient and then they can click the PDS lookup and it actually then connects to the Spine to look up that patient so that they can check things like GP surgery and all that sort of information.

AS7

If ambulance clinicians in AS7 wanted more detailed information, they could insert their NHS smartcard into the hardware to allow them to access the full SCR via NHS Spine. This would allow them to see information such as the reason for the patient’s last contact with their GP, what medication they are currently on and any existing or long-term conditions. For about 85% of patients, the PDS was looked up, either by the ambulance clinician or in the control room. Rates of access for the SCR were slightly lower, but it appeared to be highly valued as a source of information that could potentially support non-conveyance decisions:

The crews absolutely love it because it can give them an indication of, you know, some sort of preterm conditions, long-term conditions, that might be exacerbated in that particular incident and therefore they can look at what the management for that is and then make a decision as to whether they’re going to convey or not convey.

AS7

In AS7, information on previous ambulance service call-outs was available via the PDS. In AS1, as well as having access to information on NHS Spine, ambulance clinicians could use their EHR devices to search for information on previous call-outs if they had the patient’s NHS number. Clinicians in other services were not yet able to access this information.

Integration of assessment tools and prompts into the electronic health records

The AS7 system had prompts to carry out tests if the paramedic reached the end of the patient record without inputting certain data.

A respondent from AS1 suggested that the level of integration of the EHR system with other data sources meant that it functioned as ‘a clinical decision-support system’, with additional benefits in terms of supporting the ambulance clinician in carrying out their job:

We do an assessment of the patient, we capture other information in terms of social and medical detail, we will perform a series of diagnostics and those diagnostics have been built into the system so there’s not just an electronic patient record system, it’s actually a clinical decision-support system . . . So we have got a whole set of algorithms, things like NEWS score built in, we’ve got pain assessments and other assessments which . . . we find very useful and let’s say a typical paramedic should be performing as a matter of course on all patients but basically we’re guiding them through those things based on the presenting conditions. So fairly intelligent in that respect and . . . it is taking pressure off the paramedic in terms of them needing to remember all of the things that they need to remember when they go down a certain pathway.

AS1

Telemedicine and linking with other apps and devices

AS10 used the hardware to enable telemedicine, which then informed decisions of whether or not to convey the patient:

It then allows us to do a telemetry/telemedicine function live, whilst the crew are with the patient. So I’ll give you a really good example of that from a couple of days ago. I went to a patient that was burnt, took photos of the patient and was able to have a conversation with the burns consultant.

AS10

Respondents gave other examples of integration: AS1 ran its vital signs device through the hardware used for its EHRs and AS13 linked its defibrillator and monitor to the hardware used for the EHRs. AS2 planned to integrate Paramedic Pathfinder into its device.

Accessing information resources

Some devices could be used to access the internet, enabling paramedics to find information about other resources, such as the local directory of services, or drug information.

Others had access to national and local clinical guidelines, which could be used to aid decision-making but should not be considered decision-making tools.

Audit, research and analysis

Respondents from all ambulance services with EHRs described how data from the records were used to generate reports for audit, for both external reporting and internal use. Clinical audit teams generally had access to complete records, but, to protect patient confidentiality, other members of staff had only limited access, and queries could be run on specific fields and summaries produced.

The main external reporting function was in relation to the Ambulance Quality Indicators collected by NHS England and equivalent performance measures in the devolved nations. Services also generated audit reports for their own use; for example, AS10 reported running reports to see how many EHRs were being created and completed. There were also limited examples of audit reports triggering an investigation of specific issues: AS7 found that it was dealing with a spate of respiratory issues in one locality, looked at the records and found that they were clustered near a swimming pool where there had been a chlorine leak:

We just seemed to be doing a lot of respiratory jobs in one particular building and it was student accommodation for [name] University, yeah, and it was highlighted there was an issue, we reported it and public health looked into it and that’s what they found.

AS7

There was acknowledgement from most respondents that the scope for analysis and report making was not being fully realised because of capacity issues. Similarly, any engagement in research using the EHRs was inhibited by a lack of capacity within services, although several respondents hoped that this might change in the future:

No immediate plans at the moment – not from not wanting to do it, more from the point of view ‘cause we’re currently working through this process of getting the basic reports we need out we just haven’t really had the time to sort of sit down with anything on a research side.

AS5

Staff feedback and professional development

Three respondents described ways in which EHRs were used to provide feedback to staff and look at the quality of the records to monitor and improve paramedics’ record creation. In AS13, team managers were allowed access to records generated by their own teams, but not by others, to review activity and provide feedback. Ambulance clinicians in AS3 were able to access their own EHRs, giving them the potential to use the records for their own professional development.

Medicolegal uses

Respondents described the ways in which EHRs supported medicolegal requirements, particularly in providing information for the coroner after a patient death or responding to access requests, and were seen as having considerable advantages over paper records because of both the ease with which the record could be retrieved and the quality of information it contained. Respondents stated that there were far fewer missing or incomplete EHRs than missing or incomplete paper records, so providing reports when required was much more straightforward:

From a medicolegal and subject access requests, the problem with paper records historically is even with a really good system it can take up to, you know, it can take 2 weeks to get paper records in. We’ve got these records within minutes of them being completed.

AS13

An additional benefit was reported as the better quality of patient records, which had knock-on effects in terms of demands on clinicians’ time, with EHRs seen as providing:

More accurate and clear information about what’s happened to patients and [they] are calling less of the paramedics to give evidence in coroner’s courts. So it actually releases a significant amount of time back.

AS6

Management of frequent callers

Respondents from two services described ways in which EHRs, in addition to the more generally used CAD data, supported the management of people who make frequent calls to the emergency ambulance services. AS7 had dedicated call demand practitioners, who used records to build a picture of individual behaviour patterns and to see what the ambulance service has done for them:

What they basically do is work on a caseload of the top persistent callers in their area, so they get regular information about that. They can then review all the records that the crews are completing for that particular individual, and they tend to work across services, so they liaise with social services, police or whoever else has got interest in that particular individual. So they use the [EHRs] to really get a true picture of what’s actually been done for that patient when we’ve been there, so that’s quite a big use.

AS7

AS10 took a different approach, reporting that since it began sending EHRs to GPs it had seen a change in the way people were managed, as GPs were becoming aware of those patients who repeatedly called 999.

Data quality and completeness

Respondents identified the advantage of EHRs in producing data in a standard format, allowing records to be readily compared and for information transfer to other providers to be more effective. They also noted the value of records that were not handwritten:

It’s actually been very well received by the medical side at the hospitals as well. They love the clarity of the record compared to the carbon copies that ambulance services have always handed over in the past, we’re now handing over a very clear concise record laid out in the same order in every record basically.

AS5

Ease of handling the data

Respondents identified ways in which EHRs streamlined the handling and storing of data, because the data were automatically uploaded to a central hub. This contrasted with the labour involved in dealing with paper records, which had to be physically transported and then scanned. EHRs were seen both as easier to keep track of and as easier to search:

I’ve never ever been not able to find an electronic one, if there’s an electronic one it’s always been found, it’s been the paper ones that have not been there . . . easier to use the data, so what it’s telling you, it should be more easy to mine that data, to get better meaningful intelligence from it.

AS2

Data transfer and referrals

Respondents discussed the value of EHRs in allowing data to be shared with other health-care providers in real time, bringing benefits to patients and to other health-care providers:

And also I think that most of our benefits sit with the hospitals because if they use our system for pre-emptive care, they saw somebody coming in and they got a room ready or a surgeon ready for that patient, you know, they could assist a lot quicker in the pathway. They don’t have to wait for somebody really to make a phone call and turn up, they’ve got a lot more detail behind them.

AS1

Benefits were also identified for the ambulance service in terms of saving time for ambulance clinicians: the respondent from AS5 said that paramedics were able to do things on the road that they previously had to return to the station to do, such as send a report on a vulnerable adult.

Confidentiality and security

Respondents reported that EHRs offered greater security than paper records. As the respondent from AS7 pointed out, once the record was completed and signed off, it was closed and off the device, unlike paper records, which would still be in the ambulance. There was a higher level of confidentiality when records were being searched, as searches could be run without patient identifiers, and only the fields that needed to be accessed were included in the results:

It’s less risky if you’ve got it on a system than a piece of paper. You have the information governance risks, they’re not as high.

AS2

Costs

Respondents mentioned the cost savings associated with EHRs from not having to process, transport and store paper patient records. However, for both AS1 and AS7, EHRs were not primarily about saving money:

The thing about this system, it isn’t a cash-releasing benefit, it’s societal benefit and it’s quality benefit really.

AS1

I think it’s a very difficult one and the business case is based on something which – you know, there’s physical things like printing, running paper records around, scanning them in, you know, there’s all those sorts of cost savings, but how do you actually put a value on the fact that they can access a Summary Care Record for a patient whilst they’re in front of them? These sorts of things you can’t put financial money on, but what you can put is the value to the patient and the patient care.

AS7

Technical problems

Although EHRs seemed to be associated with fewer missing records than paper systems, some technical glitches were reported. AS3 had lost records from its first-generation devices, for example if the battery had run out before the device could be docked. Other services had taken proactive measures when designing their systems to protect against the loss of information. For example, AS10’s device continuously uploaded data to the server; it had a mobile signal so that even if the device lost power or was damaged, the record would still exist. There were some issues around ‘missing’ records due to lost server connections when making the record, but these had largely been resolved. AS1 had taken similar steps and although the clinician needed to log back in to access the record following a loss of connection, the record still existed on the server.

For some services, poor connectivity and the associated fears that records would be lost were reported as having a negative impact on adoption as crews lacked confidence in the system:

There’s been a big drive on the connectivity behind it, because historically in [county] we’ve got a really poor signal and the crews are finding it – that nothing is uploading, so that was one of the barriers for its use really.

AS2

We had issues with connectivity – serious issues with connectivity. Something to do with the [devices] themselves and the way they’d been installed which took us months and months and months to find out what the issue – and then something else was to do with [mobile network] and a switch that needed switching on and people didn’t recognise this. And it took a lot of fact finding and took a lot out on crews ‘cause they just . . . lost faith in the system so went back to paper very regularly.

AS6

AS4, which used the digipen system, reported that records were not always linked to CAD data as this had to be done manually. Although the EHR was not lost, it was separated from the corresponding dispatch record:

The staff themselves can write the incident number from the original call on the top of the form and then in their vehicles when they’ve finished the job they can put the clinical record number into the field in the job before they shut the job down. That’s how it should happen to make sure that there’s a link to every incident. But unfortunately again they’re forced to put eight characters in and some people don’t put the correct characters in, some of them by accident genuinely, but some it’s deliberate because you get eight zeros and that sort of thing.

AS4

Technical problems were cited by a minority of services as being one of the factors that kept completion rates below 100%. AS2’s first-generation system depended on mobile phone connectivity to work fully, so areas of poor coverage made it challenging. AS9 also reported problems with its first-generation system, which led to a decline in use:

A couple of significant IT, technical and then performance issues. We saw that drop from its high to probably a baseline I reckon of around . . . between probably 50 and 60 per cent for a fair period of time. And then we reached a period about just over a year ago where we had some significant technical issues with the old software it developed all sorts of issues with screens freezing, Wi-Fi links not working da de da de da. That all happened at the same time as the national programme sort of winding up.

AS9

For AS10, technical problems came in the following form:

Occasional failures of the tablet and the very, very occasional failures of our staff, like when they leave them on the roof of the car. So we’ve had a few tablets that have ended up on motorways or being driven over by the ambulance following the car out of the door. But we only see that on our RRVs, on our rapid-response vehicles.

AS10

In AS7, the respondent described difficulties associated with the design of the EHRs, with paramedics reluctant to use the drop-down menus and instead writing the bulk of information into free-text boxes, making it difficult to extract for audit. To try to correct the way clinicians used the records, AS7 had redesigned the EHRs, removing the bulk of the free-text boxes.

Interoperability with acute services

Four ambulance services (AS1, AS2, AS7 and AS10) identified problems with transfer of information to acute services, and suggested this as a reason for ambulance clinicians not completing EHRs. Each ambulance service was conveying patients to a number of different acute services, some of which were able to receive patient records electronically, but some of which were not able to, or perhaps were not willing to:

They like the paper ones at our acute, so the crews find that when they’re going in that the hospitals are saying, ‘oh we prefer the paper ones’ and they’re getting a bit of a hard time and then the administrators at the hospitals are saying, ‘we haven’t got smart cards’, they can’t access them. And then they were having to rewrite the patient information on paper so they just didn’t bother with the electronic one, they did it straight onto paper.

AS2

AS10 reported that there had been some regional variation due to one particular hospital not wanting to receive information on EHRs, so paramedics would use paper records; intervention from AS10’s medical director led to the hospital changing its system to accept EHRs.

Training and compliance issues

Three services (AS8, AS9 and AS2) described challenges with training staff to use the EHR system and then with maintaining compliance, sometimes in the face of technical challenges. AS8, where completion rates were around 3% immediately before it took its first-generation system offline, reported that this represented a substantial drop-off in use from an earlier peak:

In some areas it got to kind of 60–70 per cent but without sort of frequent – because there’s a lot of staff movement across the system, so you know, as soon as we scaled back on the support then it just fell away exponentially, in terms of its use because people were just not – they just weren’t using it. They just, you know, the people would move into an area where [EHRs] was enabled but they didn’t know how to use it. And equally, the people who did know how to use it, you know, would move on to another area . . . No, we never hit a critical mass.

AS8

Although AS8 did put some champions in place once it was clear that there were problems, the respondent felt that they were brought in too late:

They put some people down at ED so if crews would come in they would be there to just help them out and, you know, help them complete the record and things like that. But I think the horse had bolted by that stage and it was too late, there was the reluctance to even start using it . . . you’re in your ambulance for 12 hours with your crew mate and . . . you’ll just follow the path of least resistance and go back to pen and paper. So whilst they did put some champions in, there was nothing really for them to champion.

AS8

Learning from this, AS8 planned to make it compulsory to use the second-generation system once it was implemented:

Respondents from AS2 and AS9 reported problems with delivering training to staff. In AS9 they had found it difficult to find time to release staff for the EHR training. This impacted negatively on adoption as ‘Untrained staff means not using EHRs’.

AS9

Regional variations in completion rates, from 60% to 20%, were also reported by AS9, attributed by the respondent to local champions encouraging take-up in certain localities. They also reported resistance in certain areas, mostly in particularly busy areas where managers were reluctant to release staff for training, describing the reason as:

It’s a cultural thing, if I’m brutally honest. I think we just haven’t managed the usage very well in one area, whereas we will be moving forward. So I think we’ll see that change, so we’ll get all our divisions to a very similar level.

AS2

AS9 kept paper records on board ‘in case’ a backup was required for the EHR system; however, the respondent reported that ambulance clinicians found ‘inventive’ reasons for not using EHRs, so a more arduous system was being introduced to stop this from happening. Similarly, AS2 ascribed some of the challenges to uptake with their first-generation EHR system as being due to clinicians still being allowed the option to use paper – an option that would be removed with the transition to the second-generation system.

Plans for introduction in the future

The four services that had previously used EHRs intended to reintroduce them and all except one had clear timelines for doing so. As one respondent remarked:

I think they are completely essential.

AS6

Following the failure to successfully implement EHRs previously, AS6 was working to ensure that its second-generation system was straightforward to use as it anticipated that clinicians’ concerns about time on scene could be a potential reason for non-compliance. AS6 was involving staff in the design and procurement of the new system to encourage better buy-in:

We’ve actually involved a lot more staff in the – in the design of how this would work, we’ve involved a lot more clinical staff in how it looked and how the buttons work and then we’re also involving clinical staff in the actual purchase of the equipment so that hopefully there’s a lot more buy-in from the beginning. And then the other side to this if we add things to it that enable them to have a better experience of say something that they need to get done, something that they need then that will enhance the system for us.

AS6

The two services that had never previously used EHRs were keen to introduce them in the near future. AS12 had recently had its business case approved and was about to go out of procurement. AS11 stated that the introduction of EHRs was a long-term aim, but they had no fixed timeline for doing so owing to the cost of implementation and the difficulties surrounding the logistics of installing devices and training staff.

AS4 planned to continue using its current system but thought that its next steps would be to move to tablet-based EHRs rather than a digipen.

Plans for developing existing electronic health record systems

Respondents discussed their services’ plans for developing the use and value of EHRs in the future, and using them to their full potential. These plans covered a range of themes:

• improved transfer of patient information to primary care (AS1, AS3, AS5 and AS7) and to secondary providers (AS7)

• more analysis of patient data to improve service delivery (AS2 and AS5)

• using NHS numbers to link ambulance service records with other patient information, such as secondary care data, in order to track patient outcomes, described by one respondent as the ‘missing link in the patient care journey – you know, the hospital information’ (AS13)

• use feedback from patient records to support paramedics’ professional development (AS13).

Context

Case study site 1 covered a large area of some 20,000 square miles, spanning seven counties; the area is predominantly rural, but also includes some large urban centres as well as island communities. AS1 served a resident population of approximately 5.3 million people, and also responded to calls from the 17.5 million seasonal visitors per year. AS1 had nearly 100 ambulance stations. It was created in 2006, and merged with a neighbouring ambulance service trust in February 2013. AS1 covered 13 CCGs and served 18 acute trusts. It employed over 4000 mainly clinical and operational staff, including paramedics (n = 1788), emergency care practitioners, advanced technicians, ambulance care assistants and nurse practitioners, plus GPs and around 2785 volunteers. In the most recent Care Quality Commission (CQC) inspection, in 2016, AS1 received an overall rating (including its non-emergency services) of ‘Requires improvement’, although in terms of caring it was rated as ‘outstanding.’ The annual budget for AS1 was £240.5M in 2016–17.

Records system

AS1 used a Panasonic TOUGHBOOK running Ortivus software.

Implementation

The timeline of EHR implementation in site 1 is summarised in Table 10.

Design and procurement

AS1 did not identify a suitable ‘off-the-shelf’ EHR system to purchase, so the data management team at AS1 worked closely with the software company in developing the system and identifying changes or improvements that were needed, and its ‘tweaking’ of the system was ongoing (Site 1, Int 1). During the development stage, the team gathered feedback from front-line ambulance staff and ED staff, and had made ‘significant changes’ to the system (Site 1, Int 5). However, there still seemed to be a perception from ED staff that the design of the system and, in particular, its output, was not tailored to the hospital’s needs:

We hadn’t actually seen the system at all, so we were presented with a system which was more or less fully developed and with a start date for implementation.

Site 1, Int 3

Managers reported that the speed and extent of development was limited by the funding available:

The government refused to fund a system that was fully integrated, what they funded was an electronic system that captures data but the output is paper, so it was a bit disappointing really.

Site 1, Int 2

Training and roll-out

The service used ‘cascade training’ to train paramedics in how to use the EHRs. Formal training was carried out in each area and was quite a slow process, partly due to the logistical difficulties in getting clinicians off the road to be trained (Site 1, Int 5) and partly due to the logistics of installing equipment in vehicles. Although the staff member responsible for implementation had found that ‘conquering the workforce’ turned out to be ‘a lot easier’ than expected (Site 1, Int 5), training had ‘cost a huge amount of money’ (Site 1, Int 2), and the slow pace led to some problems for front-line staff:

The way that we delivered training made it very difficult for the user because we delivered training alongside the implementation plan, but unfortunately the devices and the fleet and the logistics weren’t ready . . . so the lag between having their training and then being out and using the device and going live in an area meant that actually it was all of a sudden unfamiliar to people again.

Site 1, Int 9

Implementation officers had been employed during the roll-out of the system to meet with hospital staff and deal with queries they might have, and to make sure that the workstations in the hospitals were installed and able to display the clinical data. There were differences in perspective about how this had gone: a manager from AS1 thought that the training process in hospitals had been ‘easy’ (Site 1, Int 2), but one ED clinician felt that it had been a ‘massive issue’ as staff needed to be trained in how to read the record (Site 1, Int 3).

The roll-out was phased over 2 years, following a trial period in a single area. One respondent described the lengthy roll-out, following a long procurement phase, as causing some problems:

So the tablets that we currently use are now obsolete because you purchase them, you take 2 years to roll out the product, we’re almost a year beyond the final roll-out and the devices become obsolete very quickly.

Site 1, Int 2

Electronic health records in use by ambulance clinicians

On standard ambulances operated by AS1 (converted Mercedes Sprinter, Stuttgart, Germany), the EHR device was stored in the back of the ambulance on the bulkhead adjacent to the patient trolley, as shown in Figure 3. In the single-crewed rapid-response estate cars, the device was stored in the boot of the car.

FIGURE 3.

Plan of typical AS1 ambulance, showing the position of the EHR device when stored. ECS, electronic care system.

The EHR was a complex form with approximately 1400 fields. The person completing the record would be taken to different parts of the form depending on the answers to previous questions. Clinicians entered some data manually, often using free-text boxes to enter supplementary details and negative findings, in addition to using the drop-down menus.

Paramedics could use the tablet to access NHS Spine and view patient demographic information and their SCR, which could provide information on, for example, prescribed medication. Except on rapid-response vehicles (RRVs), data were also entered directly to the device from other pieces of equipment, via a Bluetooth connection, and one senior manager described the benefits of this:

We have the tablet on the vehicle, the tablet is also linked up to the various vital signs equipment, linked up to a piece of equipment that allows a 12-lead ECG [electrocardiogram], blood pressure, SpO₂ levels, pulse oximetry all of that feeds into the electronic patient records.

Site 1, Int 11

The tablet had a camera function enabled, which would be used, for example, to photograph a wound before it was dressed. We also observed the camera being used as an ingenious workaround, photographing the electrocardiogram (ECG) display on a RRV not equipped to transfer information directly.

Information was not always entered directly into the tablet, with paramedics also using the traditional method of writing on their glove or in a notebook before transferring the information onto the EHRs:

Everything will be on the back of my glove . . . I don’t have time to be fiddling around with that.

Site 1, FG1, participant 1

In observations, we noted that the tablet was generally taken into the patient’s home (if that was where they were) along with the rest of the kit. If the patient was conveyed to hospital, the EHR might be completed during the journey by the clinician who was sitting in the back of the vehicle with the patient. It might be completed in the patient’s house if they were not being conveyed, or on the way to hospital if they were conveyed. One respondent, an ED consultant, commented that, in some cases, paramedics would complete the record after physically handing the patient over, meaning that hospital staff did not know if the record they could see was the completed EHR or one still in progress, something he felt was a potential patient safety risk (Site 1, Int 3).

In the vignette in Box 3, the site researcher describes how the attending paramedic used a mix of ways to collect and store data when attending a small child: exploiting the connectivity of the system to draw in background information, using paper and pen, and in this case using the researcher as an assistant to process information while she focused on direct patient care.

Use of data recorded in electronic health records

Attitudes to electronic health records

Context

Case study site 2 covered six counties and approximately 6425 square miles. The area contained several large cities, as well as rural areas, and had a population of around 4.8 million people. AS2 employed 3290 staff over 60 locations, and in 2014–15 provided a face-to-face response to 643,115 emergency calls. The service was commissioned by 22 CCGs, with one of these taking the role as co-ordinating commissioner. At its most recent CQC inspection in 2017, AS2 received an overall rating of ‘requires improvement’.

Records system

Site 2 originally introduced EHRs in 2010 in the form of Panasonic TOUGHBOOKs running Siren software from Medusa. During the period in which the site researcher was undertaking the case study interviews, focus groups and observations, the service withdrew its EHR system and rolled out a new one. This new system used GETAC tablets (shown in Figure 4) to run an updated and modified version of the EHR from Medusa. One of the features of the new system was high-specification hardware and software that could be easily updated and adapted.

FIGURE 4.

The EHR device used in AS2. (a) Front view; and (b) rear view.

Implementation

Table 11 summarises the timeline of implementation of EHRs in site 2.

History

A first-generation EHR system had been installed in 2010, funded by the Department of Health and Social Care through the NPfIT. However, it was never installed in every vehicle. Problems with connectivity (Site 2, Int 7) and faults with some machines affected clinicians’ confidence, and consequently compliance, and there was a tendency to use paper instead:

The biggest problem with the old system was that it didn’t seem to be robust enough and the hardware didn’t seem to be up to the job enough. It was too slow, too cranky and the paramedics seemed to have a general thing they’ll try something twice and if it fails the second time they won’t use it again.

Site 2, Int 2

Design and procurement

When the national contract for EHRs associated with the Department of Health and Social Care funding came to an end, the service took the opportunity to procure its own system:

Now we’re not on a national system and we have the ability to do as we please as opposed to, well it’s the offering, take it or leave it kind of thing.

Site 2, Int 5

Getting funding for the new system has been ‘the biggest challenge’ (Site 2, Int 1). The costs associated with the new device and software had to be balanced against the costs of paper and, although initial costs were high, over time it was expected to be more cost-effective (Site 2, Int 3).

Clinical commissioners had high expectations of the new system and were engaged with its introduction and integration with primary and secondary care. However, they were possibly overambitious in how fast they thought full roll-out could be achieved (Site 2, Int 9).

Training and roll-out

At the time of our fieldwork, AS2 had partially reverted to paper records while rolling out the second-generation EHR system. Although a full, temporary reversion to paper had been discussed, it was decided instead that each area would move to paper and then receive the new EHR system. Most clinicians had had only limited experience of using the new system, and some had not used it. Although we tried to schedule observations to follow roll-out of the new system, on some of the observations the EHRs were not fully operational.

Training was being provided online. Clinicians were asked to do it in their own time and then claim back the time it took, which they felt was a ‘good deal’ (Site 2, FG3, participant 2). However, focus group participants who had experience of the training for the original system did not feel that it compared favourably:

When we did our first TOUGHBOOK training . . . We had 2 days in one of the suites. There was about 10 of us in a room, we’d all got one and we were literally going through page by page by page, stage by stage to log onto it and then to do this and then went through, and you’ve got like, lists of like – there would be about 10/12 patients that you had to work through and you’d be just tapping away doing everything under the guidance of a trainer. So, if you’d got any questions about it, you could ask them. You could ask your mate next to you, say, ‘Oh, where’s that thing? How do you do this?’. You know, this one, it’s assumed that everybody’s completely versed in it, everybody understands IT and can learn off a series of 3-minute videos.

Site 2, FG3, participant 1

Despite this, the focus group participants did not generally feel that they required more training as the system was not very different from the one that they had been using and was thought to be fairly straightforward to use.

Electronic health records in use by ambulance clinicians

Ambulance clinicians described a range of ways in which they used the EHR system, choosing when, how and sometimes whether or not to complete the EHRs. They reported that the position of the tablet in the ambulance meant that they did not automatically pick it up with their other kit, and whether or not they took it in when arriving at the patient would depend on the nature of the call, with providing care taking priority over record-keeping in serious cases:

If I’m going to a life-threatening job, I don’t tend to grab it and take it in with me . . . if it’s a non-life-threatening job, I’ll tend to take it in with me and I open it up straight away . . . I’ll put the observations in it straight away and it’s time stamped straight away.

Site 2, FG4, participant 1

Our observations suggested that use of the EHRs was shaped by the type of call-out, individual practice, and the stage of the roll-out in the area. Clinicians on one shift were observed entering data onto the EHRs once the patient had been transferred to the ambulance, whereas on another shift the tablet was routinely taken into the patient and the record started straightaway. On another shift, clinicians did not routinely take the tablet with them on arrival, instead waiting until they had decided whether or not to convey the patient before fetching the tablet. During the observation, clinicians created EHRs for patients who were not being conveyed, copying the data from the paper form onto the tablet, but took a paper record to hospital with conveyed patients, because hospitals were not yet able to accept EHRs. All EHRs were added to on the ambulance and clinicians often carried on entering information on arrival at the hospital, for example while waiting to hand the patient over.

The EHR system had the potential to allow ambulance clinicians to access GP records and view pertinent patient information, but only where GPs had put the appropriate permissions in place:

At the moment, in theory, we can actually see do not resuscitate decisions, advance care planning, current plans for that patient, e.g. if this patient with chronic airways disease calls you and their saturations are 90 that’s normal, please leave them at home.

Site 2, Int 3

In addition to being entered manually, data were also drawn from the mobile data terminal (MDT), which could autopopulate some parts of the EHR. Information from the LIFEPAK (Physio-Control, Redmond, WA, USA) monitoring equipment was printed and entered retrospectively into the EHRs; the ability to transmit the information via Bluetooth had not yet been activated at the time of the observations (Box 4).

Completion of the EHRs was observed to be something of a ‘team’ approach: one clinician might start the form but have it signed off by their colleague; or one doing observations might tell the colleague what to enter on the record. One clinician commented that the EHRs acted as a prompt to ask or record things that she might not have done if using a paper form, leading to the creation of a more comprehensive record.

Some focus group participants expressed disappointment that they did not have more access to data on previous patient contacts, particularly in relation to people who make frequent use of the service:

It’s a shame that [recent visits] wasn’t included. It doesn’t need to give the information just that you’ve been to this one . . . that’d be useful.

Site 2, FG2, participant 4

Use of data recorded in electronic health records

Attitudes to electronic health records

Context

Case study site 3 covered a large geographical area of roughly 30,000 square miles, which included several major cities and urban conurbations, numerous smaller urban areas and some very remote rural areas. AS3 provided emergency care for approximately 5.4 million people. Its operational area included 14 NHS health boards.

Implementation

The timeline of implementation of EHRs in site 3 is summarised in Table 12.

History

AS3 was the first ambulance service in the UK to introduce EHRs, in 2007, and EHRs were firmly embedded in practice. A fully digitised system was envisaged for the service, enabling digital patient records to be shared across the health system – but this had not yet been achieved. The shift to digital records was seen as a way to free up paramedics to deliver more mobile health care in the community; improve the performance of front-line services; improve the integration of national, regional and local health services; provide decision-making support for ambulance clinicians; capture all clinical data in real time; and provide easier analysis to improve clinical performance.

The first-generation EHR system introduced in 2007 used Terrafix rugged tablets and bespoke software, but was not fully integrated with other EHRs within the health service. It was described as:

Basically just a data-gathering tool, so you put your record and all your observations for the patient . . . effectively that was all there was.

Site 3, Int 4

Design and procurement

By 2013, AS3 saw a need to replace the equipment, which had been in continuous use for 6 years, and which was no longer cutting-edge technology and could not support the planned software updates. The expiry of the contract to supply the first-generation EHR system was due in 2016, offering an opportunity to upgrade to a system better able to support data-sharing and new technology, such as video consultation.

AS3’s preferred option was to completely update software and hardware and establish the vehicle as a wireless hub. AS3 planned to continue using EHRs throughout the update process. A business case prepared in 2016 set out plans for the new system with four components: a new EHR system with a user-friendly, intuitive interface; access to a range of back-office systems and internet-based applications and services to support mobile working and improve patient care; a new application to provide key information to service staff about services, pathways and guidelines that will help them to do their job; and infrastructure to enable ambulance telehealth capability.

Training and roll-out

Roll-out of new hardware (tablets, communication hubs and printers) took place during 2015–16. New software was rolled out in the second half of 2017, so was in progress at the time of the interviews. Observations were carried out before the new software was in use. The researcher had originally arranged for some observations to take place once the new technology was in use but delays to the roll-out meant that this was not possible in the fieldwork period. This software roll-out was intended to give paramedics better access to patient data, clinical guidance and electronic access to patient records.

Electronic health records in use by ambulance clinicians

An EHR was generated automatically for every patient for whom an emergency response was sent:

A record is created automatically on allocation of a new incident to the vehicle, so the record is created with some of the detail already pre-populated from our command and control system, so for example the incident number, patient name, patient – any patient demographics that we’ve already picked up from the call-taking is already in there, as well as what we call the dispatch code.

Site 3, Int 1

It was intended that ambulance clinicians would take the device in with them when they saw a patient. In practice though, this did not always happen. During the observations, the site researcher saw notes being written on gloves or other temporary surfaces, and the device then used for the final ‘formal’ version of the patient record, as described in Box 5.

Focus group participants told a similar story, reporting that they did not routinely take the device out of the vehicle unless they decided to keep the patient at home. There were concerns that they would not be able to connect to the internet inside the house.

Records for patients who were conveyed to the ED were likely to be completed on the way to the hospital. Focus group respondents thought that this was because they already took ‘probably too much equipment’ (Site 3, FG1, participant 1) in when they first arrived.

The second-generation EHR software was designed to reduce the amount of information recorded as free text by using more drop-down pick lists:

What we’ve tried to do with the new version EHR then is reduce the amount of writing required, so that everything that we would expect to write down, there’s usually a button or something for them to pull down a drop-down list for them to pick from, to choose that information.

Site 3, Int 4

At the time of our fieldwork, ambulance clinicians were not yet able to look at a patient’s previous EHR from earlier call-outs, although there were plans to enable paramedics to access individual patient EHRs for the preceding 3 months. A paramedic would therefore be able to see if a patient had called 999 and been seen in the previous 3 months, and review what had happened. Managers and clinicians reported that patients expected that this already happened:

When we first raised the issue of accessing previous [EHRs] . . . the question is patient confidentiality, etc., etc. And then when we went and assessed it the patients expected that we actually knew that, so the patients do expect our staff to know there was an ambulance there last week or what happened to them . . . people expect us to have that record there and they’re surprised when we don’t have.

Site 3, Int 4

They’ll quite often think you’ve got access to all their records, because they’ll see you ‘Oh, that’ll be on my records’.

Site 3, FG1, participant 2

Use of data recorded in electronic health records

Attitudes to electronic health records

At the time of case study observations, clinicians were not yet using the new EHR software, so much of the discussion around the impact on practice, patient experience and data collection was speculative. As EHRs had been used in AS3 since 2007, it was difficult for staff to draw comparisons between EHRs and paper records, as a significant number of staff had little or no experience of using paper records. In focus groups, paramedics were generally talking about the first-generation software system on the new device (the interim arrangement), but, in interviews, managers were often talking about the new system on the new device.

Context

Case study site 4 was an area of approximately 8000 square miles, with a population of 3.1 million living in extensive rural areas, busy seaside resorts and large urban conurbations. AS4 attended more than 250,000 emergency calls per year, in addition to providing non-emergency transport and telephone advice. It employed 1310 people on emergency duties working from 90 ambulance stations.

AS4 developed a new clinical model and formally introduced it in 2015, to support innovative ways of responding to and measuring service performance. This model is based on patient clinical outcome rather than the blanket 8-minute response target previously used across the whole of the UK for all high-priority calls.

Records system

Since 2015, AS4 used a hybrid EHR system. Ambulance clinicians wrote on a paper form known as a PCR that was printed with a microdot pattern using a digipen, which is a digitised clinical record pen. As the clinician wrote, making conventional ink marks on the paper, the digital pen took up to 70 images per second, each stroke digitally recorded using a small digital camera. These images were stored as data with the precise position of the pen within the digital pen’s memory and labelled with a time stamp. Each digipen was issued to an individual member of staff and when records were created, the pen’s unique ID was also recorded. At the end of each shift, the digipen was docked and data were uploaded. The digipen and paper form are shown in Figure 5.

FIGURE 5.

Digipen used by AS4, with paper PCR.

Implementation

The timeline of implementation of EHRs in site 3 is summarised in Table 13.

History

Before the digipen was introduced, AS4 processed approximately 500,000 paper records per year. Paper records were physically transported and then scanned into a system that was able to capture only eight of the available data fields. It could take up to 4 months from the time of the incident to this small proportion of data being available for audit or analysis.

Design and procurement

The digipen was an off-the-shelf product that was used with a revised version of AS4’s existing design of paper form. A transition to EHRs for the service had been under discussion for some time when an external review recommended an upgrade to data systems as a priority. Procurement of the new system then proceeded rapidly, supported by slippage funding secured from an NHS capital programme. The digipen was seen by senior managers as a ‘stepping stone’ to a fully digital electronic PCR, offering a cheaper, simpler alternative to tablet- or laptop-based EHR systems. It also allowed (as the digipen also writes with ink) AS4 to continue to hand over a handwritten paper record at EDs. Stated aims of the 2014 business case included cost savings on the printing and processing of paper forms; improved data quality, security and timeliness; the ability to provide feedback to staff on clinical performance; and the development of algorithms from the clinical data, which would inform the clinical response model.

At the time of the fieldwork, there were plans under discussion to change to a tablet-based system, which would allow the clinicians to make referrals directly to other care providers or safeguarding services, to access an electronic version of the JRCALC guidelines and to use clinical support for decision-making, but there was no agreed plan or time scale for this transition at the time of the fieldwork.

Training and roll-out

Originally, the service had planned to roll the system out gradually. However, the introduction of a new clinical response model, combined with the approval of the cost of the project, put the service under pressure to roll the system out faster than intended – over a 3-week period across the entire service area.

The ease with which the digipen system could be implemented had been part of the reason the ambulance service had chosen this system:

Much easier to teach someone to carry on using a pen and then dock the pen rather than teach them how to use [an EHR with tablet interface] with drop-down menus and so on, if they’re not familiar with computers.

Site 4, Int 1

There was, though, a formal programme of digipen training delivered throughout the implementation period, consisting of a 25-slide PowerPoint^® (Microsoft Corporation, Redmond, WA, USA) presentation created by the digipen provider, and a practical demonstration by local digipen champions or clinical team leaders. Training focused largely on the practical features of the pen, such as the battery life and how to dock it. Accompanying the training was information provided on the AS4 intranet: a digital pen PCR user guide and completion instructions, and a clinical memo, in the form of 12 pages of written frequently asked questions and supporting information.

In one area, focus group participants reported that the new system had already been rolled out by the time they received their training. However, the respondents generally thought that training had been sufficient, as the simplicity of the system meant that it was easy to use, and it did not differ much from what they had been doing before:

How much training do you need to write, to use a pen?

Site 4, FG1, participant 2

Electronic health records in use by ambulance clinicians

Clinicians were observed getting patient information from a variety of sources, including information from the ambulance dispatch team, via the MDT, via the Corpuls cardiac monitoring equipment (GS GmbH Kaufering, Germany), from friends and relatives of the patient, and from the patient themselves. They could take the job number from the MDT to record on the PCR and would enter the unique reference on the PCR into the MDT when the job was cleared. This meant that the PCR could be linked to the dispatch data.

Because the PCR was a paper form, clinicians were not able to access information through the PCR and information could not be transferred directly into the form. They decided what data to record and wrote them on the PCR using a digipen as and when they could, as described in Boxes 6 and 7.

At the time of this study, AS4 had recently purchased Corpuls ECG monitors/defibrillators and placed them into ambulances to increase the potential for telemetry. Although the Corpuls equipment had the capacity to share observation data, it could not be linked directly to the patient record, as one clinician reflected:

There’s information being shared but not fully integrated within systems. It’s still very much separate systems talking to each other with a manual input in between. So there is a flow of information but it’s not an integrated flow.

Site 4, Int 4

In two out of the three observations, clinicians did not take the PCR in with them, instead making notes on their gloves or notepads, using this along with data from the Corpuls, and their memory of events, to complete the form once patients were in the ambulance or the decision had been made that the patient would be left at home. The researcher noted that the PCRs were not kept with the equipment that would routinely be taken to a patient for the initial assessment. On the third observation, the crew took the PCR in for the initial patient assessment, and one clinician completed it while with the patient, with the colleague relaying information to them to enter into the form. This was reported by the crew as being a common procedure except in time-critical cases.

Use of data recorded in electronic health records

Attitudes to electronic health records

Electronic health record generation by call characteristics

Time and day when call was received

In Figure 7, we indicate rates of generation of EHRs by day of the week, which showed little variation across services. We calculated odds ratios for EHR generation by day (presented in Appendix 8), which confirmed that there was no significant variation; that is, an EHR was equally likely to be generated on each day of the week.

FIGURE 7.

Percentage of attended calls resulting in generation of an EHR, by day, for each ambulance service.

Data were collected over a 9-week period. Because three out of four sites were in a process of transition in terms of record system, we were interested to see whether or not rates of completion changed over that time, as systems became more embedded. In Figure 8, we show rates of generation of EHRs by sample week, over the sample period, by ambulance service. Although we found some variation week to week, there was no clearly discernible or consistent trend across the sites. There was no indication of a consistent relationship across sites between EHR completion and the calendar within our snapshot of the data. Full data, including odds ratios, are presented in Appendix 7.

FIGURE 8.

Percentage of attended calls resulting in generation of an EHR, by sample week, for each ambulance service.

We examined how generation of EHRs varied according to whether the call was received during regular hours or out of hours (AS2 and AS3); for AS1 and AS4, where the data had been coded differently, we compared calls made during the day versus during the night. As shown in Table 20, we found that in AS1, AS2 and AS3, EHRs were more likely to be created during night-time/out-of-hours calls than during daytime/regular hours calls, and the difference was statistically significant. The effect was much more pronounced in AS3. In AS4, EHRs were slightly more likely to be generated in the day, but the effect was not statistically significant.

TABLE 20 - Electronic health record generation by the time when the call was received

CI, confidence interval.

Time (Number of calls)	EHRs, n (%)	Odds ratio	95% CI	p-value
AS1
Day (74,311)	64,326 (86.6)	0.933	0.904 to 0.963	< 0.001
Night (61,630)	53,834 (87.4)	1	–	–
AS2
In hours (42,975)	12,796 (29.8)	0.910	0.886 to 0.935	< 0.001
Out of hours (63,171)	20,073 (31.8)	1	–	–
AS3
In hours (43,823)	43,393 (99.0)	0.563	0.487 to 0.650	< 0.001
Out of hours (59,694)	59,363 (99.4)	1	–	–
AS4
Day (36,802)	26,232 (71.3)	1.028	0.994 to 1.063	0.113
Night (29,415)	20,801 (70.7)	1	–	–

Electronic health record generation by ambulance service processes

Electronic health record generation by patient characteristics

What is the role and practice of the ambulance clinician in bringing electronic health records into use?

Ambulance clinicians are the key agents tasked with bringing EHRs into everyday use. We examined what they had to do to make it work, what motivated them to use the system and how they might resist or modify the intended model of implementation.

At site 4, where the digipen had been introduced, it seemed that the hopes of having a relatively seamless transition to the new system had been realised. Ambulance clinicians reported that it brought no significant change in practice:

No difference to us at all, ‘cause we’ve still got to fill out the [EHR] whether it be digital or biro or pencil or crayons.

Site 4, FG1, participant 2

In the other sites, we found a generally positive attitude towards new systems, a recognition that they did not represent a massive change in practice, and certainly no outright resistance. However, there was some suggestion that the process of bringing them into use was not entirely painless. In site 3, an ambulance clinician anticipated some reluctance from staff to engage with the second-generation system, as they had concerns about its predecessor. In site 1, one ambulance clinician reported what they felt had been the response among peers to the new system:

It’s mixed . . . it doesn’t matter what grade you are or what age you are, some love it and some hate it. Some don’t like it ‘cause it’s different to what they used to do. Some don’t like it ‘cause it’s not one page. Some don’t like it because the battery runs out really quickly.

Site 1, Int 5

As part of the process of the implementation of EHRs, staff in all sites had to undertake formal training. Logistical challenges to this were identified in all sites, for example:

In our busier areas . . . crews are not on stations for any length of time, they are literally on a 12-hour shift and they are out on the road for that whole 12-hour shift, just going from one job to the next, so downtime to complete training is difficult.

Site 3, Int 1

E-learning, sometimes in clinicians’ own time, was one option for training when shift patterns did not allow for formal training sessions. In site 2, clinicians who were asked to complete e-learning from home or an ambulance station reported some resistance to doing it. In another site where e-learning was meant to be taking place, we found that only one of the crews we observed had undertaken the training, and reported that they had not fully engaged with it, choosing instead to run the computer program while washing vehicles and then returning to the computer to answer the questions at the end. The crew members were critical of the new EHR system but admitted that their lack of engagement with training may have made their experience more negative.

Once ambulance clinicians started using a new EHR system, there was (except in site 4) still work involved in adapting to the new system. Whereas a paper form makes it immediately obvious which sections have already been completed, an EHR takes a little more effort to navigate, as one front-line clinician reported:

I do end up repeating things because I’ve forgotten that 20 minutes earlier I ticked that box, it’s taken me that long to go through the drop-down menu, I tend to write my comments bit last, by the time I get to that I’ve forgotten, or it doesn’t flow very well if I miss those bits out.

Site 2, FG5, participant 3

In site 2, a number of front-line clinicians expressed some scepticism about the new EHR system, and there was concern that if it proved slower than paper records, people might want to revert. However, it is important to remember that we were talking to people in the midst of change, and any concerns were likely to have been shaped by the disruption of transition as well as memories of problems with the first-generation system, and were likely to subside over time as people got used to the new system.

In the course of bringing EHRs into use, it seemed that ambulance clinicians developed their own variations in ways of using it, and also learnt from their peers to develop local practice. In site 1, one clinician commented on the way in which colleagues based in different stations would establish their own ways of doing things:

People who all work together very frequently have somewhat attained a kind of a bit of a crossover of how they do things . . . if you went to different stations now I reckon you would see patterns in how data is entered, what is entered, and when it’s entered, and the time it takes to do it . . . I reckon it’s different everywhere.

Site 1, FG1, participant 4

Although all EHR systems were designed to be used live, during patient contact, we found that this was not necessarily the case in practice. We also found that it was not a simple question of EHRs being used or not – instead, they were used in different ways (timing of data entry, order of data entry, use of free-text boxes, sharing of tasks between clinician). EHRs may have all fields completed, or some data missing.

Across all four sites, we observed that data were not always entered directly into the EHR system. Instead, observations would often be written on a glove or notebook, or stored in the paramedic’s memory, before being entered in the EHRs, just as has been traditionally done with paper records:

That’s one of the reasons I wear gloves, it’s just a notepad.

Site 4, observation 3, paramedic

The entry of data onto the EHR sometimes took place much later during the course of the patient contact than the point when the information was obtained. From our observations, it seemed that this should not be read as resistance to using the EHRs, but instead a response to the practical constraints on delivering care in the field.

What is the role of the organisational and political context in bringing electronic health records into use?

In our interviews and focus groups, much of the discussion about the role of the organisation in bringing EHRs into use concentrated on the training of front-line staff. All case study areas had offered training, although the amount varied across the sites. In several areas, there was a time lag between delivering training to clinicians and them being able to use the new device and software, which had an impact on how well clinicians were able to put their skills into practice.

Two services used ‘local champions’ to help support the implementation of EHRs. In site 3, these ‘super users’ were given a full day of training on the tablet being introduced before spending time in the ambulance stations in their local area:

They’d spend the day sitting there ‘right, how are you getting on with your e-learning’ and they would encourage the staff to get the e-learning done. They actually had a tablet with them and they would sit and let them play with it and give them hands-on information.

Site 3, Int 4

Managers in all sites discussed that, to ensure transition to EHRs, staff needed to be given a simple, clear message that they were mandatory. In site 2, this was clearly borne out of experience of implementing the first-generation of EHRs:

I think what we’ve learnt is that we have to make the system mandatory and used, we can’t make it optional.

Site 2, Int 1

For ambulance clinicians, though, it seemed that what was needed was something more than a simple mandate, but an organisational system that was designed to provide feedback to front-line staff and reassure them that the EHR was worth completing:

If we knew that information was going to be used . . . we might all be encouraged to use it . . . if we had the confidence to know that by using this new system, and using it properly, that would possibly sort some of those problems out, it might encourage people to use it.

Site 2, FG2, participant 3

Even before the point when ambulance services were working to encourage change in practice among clinicians, they had tasks to accomplish in terms of designing and specifying the system. Work on these tasks was not limited to only long-serving staff of the ambulance service: external collaborators were involved, as were project staff who may not have had the embedded experience of ambulance service operational requirements to ensure that these were met.

We noticed that there were sometimes differences in perspective between people involved in designing and planning the implementation of EHRs and those tasked with putting it into practice on the ground. For example, in site 2, managers reported that they had engaged widely with internal and external stakeholders in the development of the system, and to plan roll-out. However, a front-line ambulance clinician told a different story:

We’re going at it half-baked, we’re kind of going ‘oh yeah, we need to do this, let’s just run at it headlong’ instead of actually sorting it out and looking at it and getting it all in place and getting everything set up properly and then rolling it out. We’ve jumped straight in.

Site 2, FG1, participant 2

What is the role of the technology itself in bringing electronic health records into use?

For a new technology to become readily embedded into use, it needs to be easy to use, sufficiently similar to previous systems to be easy to learn, and reliable enough that people do not give up on it. We found that, broadly, the EHR systems being introduced were seen as reliable and user friendly. However, even though managers reported that the EHRs had been designed in such a way that they would not change the operational process for completing clinical records (Site 1, Int 2), it seemed that a shift to record-keeping model using drop-down boxes was having an effect on the work practices around data recording for clinicians. One manager from site 1 reported:

I think one of the things that we failed to acknowledge is that we spent the best part of 10 years teaching our staff that in order to make themselves safe and to cover themselves legally, they had to write in what’s known as the medicolegal model . . . so recording all the pertinent negatives as well as all of the things that you do find. And when you then move kind of pretty much wholesale overnight to a system where you have drop-down boxes and some of those things are implied as opposed to explicitly recorded, then people really balked at that.

Site 1, Int 9

These concerns were confirmed by a front-line clinician from the same site:

I’m very much in the school of if I haven’t written it down it hasn’t happened, which I find is a big problem . . . I have to physically write they haven’t got a headache, but what you need is a box saying ‘do they have a headache, yes or no,’ and you say no . . . you’re still having to write things.

Site 1, FG1, participant 4

Also relevant to the implementation process was the way in which the EHRs related to technology being used by other health-care providers. In each case study site, we heard how there were multiple systems in use, making integration difficult:

Every hospital you go to, or every GP surgery you go to, use different systems. If they use different systems they might not be compatible.

Site 1, Int 5

All of our case study services were to a greater or lesser extent transitioning from one system to another and one supplier to another. During our fieldwork period, we observed roll-out of new hardware and software in AS1 and AS2, and new software in AS3. AS4 had started initial discussions about a successor to its system. The progress from one system to another in AS2 had included a partial return to paper records, area by area, as the system was switched over.

Even where there was no major change to new hardware or to a new software supplier, systems were subject to constant minor upgrades. In AS2, for example, managers reported that they had learnt from their previous experience with EHRs and had chosen a records system that would be easier to update than the previous one, and they acknowledged that tweaks would be needed to keep the system up to date (Site 2, Int 6). The device itself had also been selected to be easier to update and adapt(Site 2, Int 3).

Impact on clinician–patient interaction

We found that the use of EHRs appeared to have only a very limited impact on clinician–patient interaction. Much of the time, the EHR devices did not even feature in the interaction, as they were left in the ambulance while clinicians attended the patient in the home. In only one of the case study ambulance services were clinicians able to access relevant information from past health service encounters in order to inform decision-making: paramedics in this site found this very valuable:

The more access to information that we’ve had the better that we can decide on the best course of treatment for your patient.

Site 1, FG1, participant 3

We also found that the model of information gathering for which the EHR was designed – sequential and complete – did not necessarily tally with the situation of imperfect information that ambulance clinicians were dealing with on the ground. They did not necessarily have access to all of the information they needed to complete the form in full and to inform decision-making; if patients had poor recall, or were unco-operative, there was nothing an EHR system could do to help this. Information does not flow to the clinicians in a pre-set order. Although clinicians could record clinical observations and presenting complaints, there would always be uncertainties and ambiguities in making diagnoses.

Impact on the workforce

We were interested in whether or not the introduction of EHRs had led to any change in ambulance crews’ perceptions of their role, particularly in terms of decision-making about whether or not patients could safely be cared for in the community rather than conveyed to hospital. We found little evidence of this. In site 4, one clinician expressed this very directly. He had had to learn to remember to dock the pen at the end of the shift but:

From the point of view of actually doing the job, it’s made absolutely no difference to us.

Site 4, FG1, participant 1

In site 1, clinicians described using the prompts built into the EHRs, and were seen doing so on shift observations, but emphasised that they were not using them as a diagnostic tool:

You are the clinician diagnosing and this is a clinical record and not a diagnostic tool.

Site 1, FG2, participant 3

Ambulance clinicians talked at length about the role of the EHR in providing a formal record of what they had done, and about how it had the potential to ‘cover their back’ if there was any problem in the future. Clinicians were aware that the EHR was a formal record of what had happened and that it could not be changed, potentially leading them to put extra data into the free-text boxes:

People are inputting so much data into those jobs which are not relevant, and they don’t know when to stop. They’re afraid to stop because they think, alright, when I press the end key here, this is it forever, and I’ve got to have justified this. But for 99% of your jobs it’s irrelevant and the hospital are just not looking at it, they’re not interested.

Site 1, FG2, participant 5

Although ‘covering their back’ is a function also served by paper records, the EHR was slightly different because of some of its technical qualities: it was ‘locked’ on completion, it recorded the time when data were entered (even though this may not have been the time when observations were taken and activities were carried out) and it was seen as leaving some key information implicit, which meant clinicians might enter more free text in order to confirm negative findings.

Impact on the organisation

Ambulance service managers were generally enthusiastic about the benefits of EHRs for their organisations. They reported on efficiency savings from handling data electronically rather than on paper, although none quantified these. AS3, the longest-established user of EHRs, had as long ago as 2008/9 been reporting that EHRs had contributed to improvements in clinical performance and effectiveness, demonstrated by indicators such as an increase in the numbers of patients treated at home and not conveyed to hospital. The exact role played by EHRs in these shifts, among all the complexities of pre-hospital care provision, was hard to identify. In other sites, the impact of EHRs on non-conveyance rates was even harder to assess.

In site 4, a range of potential benefits – including the provision of timely feedback to operational staff on clinical performance to improve delivery of care, and the ability to develop algorithms from the clinical data that inform the management and implementation of the clinical response model – were presented as reasons for implementing the digipen. However, at the time of the fieldwork, these were still no more than possibilities.

Managers from all sites talked very positively about the advantages of sharing EHRs with other health-care providers, and with patients themselves, to improve care. Sharing of data across NHS organisations was seen as one of the biggest potential benefits of EHRs, and one of the biggest challenges:

Absolute top priority for us is to be sharing our [EHR] far and wide and wherever it can go.

Site 3, Int 1

It’s about being able to share our records more easily with the rest of the NHS and ultimately with patients, so that we can provide a more seamless service to patients.

Site 3, Int 6

The true advance in what we need to do . . . that sort of seamless integration between health system and health providers is absolutely key.

Site 2, Int 5

However, these views were generally expressed in terms of the potential of future developments, rather than reporting actual change.

In all sites, respondents reported that a lack of interoperability with hospital systems was a major stumbling block to achieving change through the implementation of EHRs. A clinical commissioner in site 2 commented positively that the involvement of the CCGs was key to the successful integration of a fully electronic system across the NHS because it involved all parts of the health service, not just the ambulance service, across the region. Contact with hospitals was about more than the ED, as a manager from site 1 pointed out:

It’s not just about [site 1] taking patients into the emergency department . . . it’s about taking patients into acute medical wards, into surgical wards, into paediatrics, into maternity. So it wasn’t just about producing kind of workstations in one location, it was about making sure the data was visible across the hospital.

Site 1, Int 2

At a more micro level, however, encouraging GPs to engage with the electronic transfer of EHRs was a particular challenge:

To get GPs to check their e-mail boxes and then disseminate that information as necessary. I think for commissioners, that will be the hardest thing for us to be doing – getting the GP practices to check their secure inboxes regularly enough to pick up an alert, a referral, a letter that crews have been out to see a patient.

Site 2, Int 9

In site 2, we also heard how barriers in hospitals to accessing the ambulance service EHRs delayed the full roll-out of the system and meant that ambulance clinicians were working in an inconsistent environment as some hospitals were able to access the records but others would not receive the new EHRs.

There may be hard-to-resolve differences in what people from different organisations want from an electronic data set. In all sites, there seemed to be challenges in reconciling the interests of clinicians with those of the audit department. This situation was described to us by an ambulance service manager in site 1:

Right from the outset . . . I had the clinical research team in one ear saying ‘I want nothing but structured data’ and I get that completely . . . because structured data enables you to run reports, it enables you to analyse data in a much more timely manner, and then I had ED consultants from the acute trusts who wanted nothing but free text because that’s what they were used to and they didn’t like all the boxes and structure.

Site 1, Int 2

Impact on care pathways and outcomes

Clinicians and managers recognised the potential for developing multiagency working to improve care pathways and outcomes. EHRs could potentially support this in two ways: (1) by providing large data sets, that are relatively easy to analyse, which would help service planners to assess demand and patient flow and to monitor the impact of changes, and (2) by facilitating data transfer to smooth patients’ transitions along care pathways. The first of these functions was highlighted by a manager from AS4, currently as an ambition, but potentially achievable even with the relatively simple technology of the digipen:

Essentially what we want to be able to do is map the patient, be able to link this data to map the patient’s journey from a 999 call, from our arrival right through the system.

Site 4, Int 3

Older people who fall – one of our tracer conditions – were highlighted across all sites as a patient group that could benefit from more information sharing between services to increase the scope for interagency working, but this was made difficult by the need for better technological interoperability.

In terms of transfer of patient information along care pathways, some of the challenges have already been noted. An additional issue was raised by a manager at site 2: although EHRs can be very helpful in terms of facilitating alternative care pathways (via decision-making support and ease of referral), and non-conveyance to ED, their effectiveness depends on the working culture of the clinicians:

If those clinicians are just simply in the traditional mindset of somebody rings 999 for an ambulance, they get an ambulance . . . take them to hospital, even though that facility is available to them on the electronic patient record they simply aren’t going to use it . . . if you don’t address the culture and you just plough on with the strategy, then essentially the culture will win every time.

Site 2, Int 5

To realise all the potential of EHRs requires engagement from other parts of the local health economy and dealing with the variations between systems operated by different providers.

However, integrating systems had proved problematic. Site 3 provides a case study here, but similar issues were experienced in all sites. Different secondary care providers were running different records system, and were not going to be changing them to suit the ambulance service:

The issues we’ve had is with that is the territorial health boards . . . using a mixture of different systems. And we just don’t seem to be able to get onto their priority list to engage with us and our uniform style.

Site 3, Int 1

Negotiating access into secure systems had been a challenge:

What held us back in the initial stages [of developing systems to share electronic records] is because all the different health boards had different firewalls and different systems and in effect networks, that was a huge issue for us. The problems we get now range from questions like, well who do you actually send it to? So does it go to the A&E [Accident & Emergency] receptionist? Does it go to the nurses’ station? Does it go to the doctors’ desk? . . . so there’s some resistance to some of it.

Site 3, Int 4

Although greater interoperability was an aspiration for the new system, it was acknowledged that this was still a significant challenge to achieve owing to the range of systems in use across the NHS and the logistical challenges of implementing a unified system:

It’s frustrating from our point of view because we’ve got it, and we want to share it, but we just can’t quite get the single one size fits all solution for people, which is annoying.

Site 3, Int 1

Work package 1: literature review

The literature review found that, although there is an extensive and theoretically developed literature examining the implementation of technology into health care more generally, there was a very limited range of published literature specifically on EHRs in ambulance services. We identified 18 peer-reviewed papers, mostly not of high quality. They were primarily descriptive and/or concerned with implementation of EHRs rather than their impact.

Work package 2: national survey

The national telephone survey established that only half of the UK emergency ambulance services had a current EHR system in use. There was a range of systems in use. All services seemed, to some extent, to be in a state of flux in relation to EHRs. Some had experience of start–stop implementations, with reversion to paper records not uncommon. External drivers could play a strong and sometimes disruptive role in implementation. Although there were some positive examples of EHR systems acting as hubs for the real-time transfer of data to improve patient care, uses beyond simple data storage were not yet fully developed. The main advantages of EHRs were seen to be around data handling and audit.

Work package 3: case studies

Our case studies identified that all four services were, to a greater or lesser extent, in a state of transition in relation to EHR systems. Although we had planned to observe four services at distinct stages of implementation, we discovered that the situation was more complex and fluid than this.

One service had adopted a technology that was relatively cheap and straightforward, but very limited in its functionality. The others were working with more sophisticated tablet-based systems. Although we saw some indication of data being transferred into and out of systems to support patient care, none of these seemed to yet be being used to their intended potential.

Despite the diversity of systems and contexts, we made some common observations across all sites. Positive aspects of what we observed included:

• vision and enthusiasm from senior managers

• front-line staff who were willing and able to readily adapt to new systems

• benefits being realised in terms of data quality, confidentiality and efficient storage and searching for audit and medicolegal use.

However, we also identified a number of challenges that were being encountered by all services:

• interoperability in terms of information transfer to and from other parts of the health-care system – the challenges related to information governance and differing information needs of different clinicians, as well as technical aspects

• technical issues with EHRs – most technical issues were relatively minor or had been resolved, but appeared to have had an influence on clinicians’ response to the EHR system

• the need to update systems (software and hardware) without undue disruption.

Our analysis of routine data sets found that different sites recorded information in different ways: there was variation in fields and priority categorisation systems. Levels of EHR creation at the time of the data sample (January to February 2017) ranged across sites from fewer than one-third of calls to over 99%. Analysis of factors influencing EHR completion showed broad consistency across sites, although striking variation in relation to some factors. Factors that seemed to be associated with a lower rate of EHR completion included:

• the call being received within routine working hours (Monday to Friday, 8.00 to 18.30)

• the call being categorised as low priority

• the patient not being conveyed (3/4 sites only)

• the patient being a child

• a mental health condition as the reason for the call (2/4 sites).

In one site, we found strong associations between completion rates and which hospital the patient was conveyed to, and which locality the crew was based in; it is unclear whether these associations were characteristic only of that time period or if they will have continued with roll-out of the service’s new EHR system.

These independent variables were confounded with each other. The biggest effects tended to be by geography and call priority, and these may explain other observed effects.

Work package 4: knowledge exchange workshop

Ambulance service representatives attending the knowledge exchange workshop felt that the findings resonated with their own experiences. They found it valuable to share knowledge with other ambulance services. Discussion groups identified the following issues:

• the need to present front-line staff with the optimum software and to ensure that staff see its value

• the value of empowering staff by providing feedback on patient outcomes

• the multifunctional potential of EHR devices, for example accessing web resources

• the need for information handover at the ED to be simple and streamlined

• the role of a single point of ownership in the organisation (rather than dispersed responsibility) to help knowledge sharing between organisations

• anxieties about data currency.

Key messages

• Digital diversity. There was no standard hardware or software in use. We found great variety in how (and if) other technology and record systems were linked to the EHRs.

• Constant change. Services were transitioning from one system to another, and from one supplier to another. Software and hardware underwent routine updates. Progress could be to and fro, including switching back to paper records as a temporary measure.

• Imperfect information. In real patient encounters, clinicians are likely to be dealing with partial or unclear information, which does not arrive to them in a pre-ordered sequence. It may be challenging to fit this into the structured, sequential format of an EHR.

• Indirect input. Some patient data can be fed straight into EHRs. However, data entered by clinical staff were still sometimes written on a glove or notebook, or stored in the paramedic’s memory, before being entered into the EHR.

• Data dump. The primary function of EHRs in all services seemed to be to act as a store for patient data. There was, as yet, limited evidence of EHRs’ full potential being realised to transfer information or support decision-making or changes to patient care.

• The system is bigger than the service. To realise all the benefits of EHRs requires engagement with other parts of the local health economy. This, in turn, requires dealing with variations between providers and the challenges of interoperability.

• Different data demands. Clinicians and data managers are likely to want very different things from a data set, and need to be presented with only the information that they need.

• ‘Ford Fiesta, not a Ferrari’. Sometimes simple is best.

Implications for health care

• Living with change. Our findings suggest that the implementation of EHRs in emergency ambulance services is not something that can be considered ‘task and finish’ but is a continuous work in progress. This requires a flexible and ongoing approach to managing change.

• Managing ambitions. Although EHRs are showing benefits, these benefits are not yet as wide as they could be. Realistic planning and an acceptance that change moves slowly will help to sustain what has been achieved so far and reduce the risk of disappointment or cynicism that might inhibit further progress.

• Flexibility in data collection process. Although EHRs may support standardised processes of clinical observation, data collection and clinical decision-making, there will remain uncertainties, fluidity and ambiguity in the information available to ambulance clinicians. Seeing mismatches between the requirements of the EHRs and the data entered into records as something to be anticipated and planned for is likely to be more helpful than seeing them as failure in compliance.

• Addressing the interoperability requirement. Ambulance services face challenges in working with their local primary, secondary and acute health-care providers, and with social care, in ensuring that patient data are transferred securely and appropriately to support patient care. This issue is clearly on the policy agenda locally and nationally, but skilled negotiation and priority setting will be required to achieve effective and productive real-time flow of patient information both into and out of emergency ambulance services.

• Maximising the potential of EHRs as a multifunctional resource. As well as being instruments of data storage, tablet-based EHRs have the potential to act as a portal to the world of information, both on the web and through secure local information databases. Maximising the ability of ambulance clinicians to readily access guidelines, clinical advice or local service directories will ensure that EHR devices get embedded in use and achieve maximum benefit.

• Maximising use of EHRs to support staff development. Feedback from EHRs has the potential to support staff reflective learning and appraisal, especially if linked to data on patient outcomes following ambulance service contact.

• Identifying streamlined data sets. Patient records transferred at the time of patient contact are most likely to be used and useful if they are focused and relevant to the receiving clinicians. A streamlined data summary for transfer at the ED, rather than the full EHR, might support this.

• Coherence of perspective throughout an ambulance service. We identified some differences in perspective between different staff groups within ambulance services about the value and potential of EHRs, with managers typically being more enthusiastic. There may be scope for more mutually beneficial communication to strengthen effective implementation of EHRs.

Search strategy

Interview questions for work package 2

1. Please can you tell me your job title and role within the ambulance service?

2. Does your ambulance service currently use electronic patient records?

• If yes

  • What system or systems?

  • When were they introduced? Go to question 3.

• If no

  • Are there any plans for introducing electronic patient records?

  • What system(s) and when? Go to question 9.

3. What proportion of patient encounters lead to an electronic record being created?

• Does this vary by area? Over time? By anything else?

• Why are some records made electronically and some on paper?

4. What was the purpose of introducing electronic patient records into your ambulance service?

• Stated rationale.

• Any other purposes?

• Any value which has come to light subsequently?

5. Please describe the process which the service went through to bring electronic patient records into use.

• Preparing business case.

• Identifying funding.

• Service specification/tendering.

• Roll-out and training – logistics, time period.

6. How do you think paramedics and EMTs have responded to the introduction of electronic records?

• Training required.

• Resistance/reluctance/enthusiasm.

• Different responses from different staff groups.

• Changing perception of role.

7. Once an electronic patient record has been created, what then happens to that information?

• Transferred to another care provider – how, when.

• Downloaded and stored – how, where.

• Used for analyses – what.

8. What use does your service make of data from electronic patient records to support performance management or service improvements?

• Identifying and managing repeat callers.

• Audit.

• Research – what?

• Medicolegal.

9. Do the electronic patient records link with other electronic resources or hand held technologies?

• For decision support.

• For referral.

• Apps.

10. What do you think are the benefits of implementing electronic patient records in the ambulance service?

11. What are the risks of implementing electronic patient records in the ambulance service?

• Unintended consequences?

12. What do you think will be the next steps for electronic patient records in your ambulance service?

13. How do you think electronic patient records have changed the experience of those using the ambulance service?

14. Are they any reports or documents on the process of adoption (e.g. business cases) or use that you are happy to share?

Research Advisory Group terms of reference

Electronic Records in Ambulances Research Advisory Group members

• Amanda Brewster, PPI representative.

• Tjeerd Van Staa, Professor in Health e-Research, University of Manchester.

• Martin McTigue, Ambulance Operations Manager, London Ambulance Service.

• Alison Clements, Head of Operations, University College London Hospitals NHS Foundation Trust.

• David Davis, Operational Staff Governor and Deputy Lead Governor, South East Coast Ambulance Service.

• Janette Turner, Reader in Emergency and Urgent Care Research, University of Sheffield.

• Catherine Pope, Professor of Medical Sociology in the Faculty of Health Sciences, University of Southampton.