Improving recording and reporting of dementia and frailty via electronic patient record by ambulance staff in a single service (IDEAS)

Background: Dementia is common in older adults assessed by ambulance services. However, inconsistent reporting via the patient record may result in this diagnosis being overlooked by healthcare staff further down the care pathway. This can have a deleterious effect on subsequent patient care, increasing morbidity and mortality. We sought to understand how and where ambulance staff would like to record this finding on the electronic patient record (ePR). Methods: We designed and implemented a survey of ambulance staff in a single service to understand how they identify patients with dementia, how they record dementia on the ePR and how the ePR could be improved to better capture dementia. Scoping questions on frailty were included. The survey was tested using cognitive interviewing. Analysis was conducted using descriptive statistics for closed questions and thematic analysis for open questions as appropriate. Results: 131 surveys were completed; 60% of participants were paramedics and 40% were other grades of front line staff. Participants reported consulting electronic/paper sources, and individuals such as carers involved in the patients’ care, to establish whether dementia had been diagnosed. Frailty assessments were prompted by social context, reduced mobility, a fall or diagnosis of dementia. Staff reported documenting dementia in 20 different areas on the ePR and 46% of participants stated a preference for a designated area to record the information. However, 15% indicated it was not necessary to record dementia or that no ePR changes were required. Conclusions: We have highlighted the variation in ambulance staff practice in recording of dementia. Alterations to the ePR are required to ensure that dementia is recorded consistently and is easily retrievable. Clearer guidance on when to assess frailty may also enhance information provision to care staff in other sectors, resulting in more appropriate clinical and social care.

customisation according to needs. It was originally implemented from 2016 across the Trust as part of a national project to move ambulance Trusts away from paperbased systems. The same system is used within the South West Ambulance Service Trust (SWAST) and Northern Ireland. The ePR consists of a hand-held, touchscreen tablet device that uses the Ortivus MMM software to interact with the user via the screen. The screen layout of the MMM software consists of a series of tabs where users can record information using free-text fields, dropdown boxes and pre-configured lists. Currently, the information about dementia can be recorded in a number of locations throughout the record, but there is currently no designated area or specific field on the SCAS ePR where dementia must be recorded.
The question this research sought to answer was 'How and when do ambulance staff identify dementia and frailty, plus where do they record it and why?' This information may facilitate improved design of electronic recording systems and associated training.

Study design
The study design was an electronic survey, delivered to a cross-sectional sample of ambulance staff.

Introduction
Ambulance services provide emergency urgent and unscheduled care in response to 999/111 and healthcare provider calls (South Central Ambulance Service, 2020). Many of these calls are to older people. Due to the complexity of older people's care, the ambulance service often refers patients to general practitioners and hospital settings, as well as other community healthcare services specific to older people, and adult social care. It is therefore important that information on key diagnoses such as dementia and the presence of frailty are available to all health and social care partners and can be effectively communicated.
Dementia is an increasingly common presentation in older people accessing ambulance services and hospital emergency departments, from pre-diagnosis through to end-of-life scenarios (Buswell et al., 2015;Voss et al., 2018). An audit of the South Central Ambulance Service (SCAS) electronic patient record (ePR) showed that, out of 314,786 ePRs of patients aged 65+ in a 1-year period, 13.5% had 'dementia' recorded somewhere in the ePR, increasing to 16.5% of patients aged ≥75 (Pocock et al., 2018). The audit also found that dementia was recorded in 16 different free-text fields, and 38.4% of records had dementia recorded in more than one field. Similar issues were found in another ambulance service, with dementia being recorded across a range of data fields including previous medical history, social or family history and treatment advice or notes (Buswell et al., 2016). This lack of systematic recording may impact on the retrieval of this information by healthcare professionals, reduce the quality of information passed between healthcare services and delay subsequent provision of specialised care. The SCAS ePR system is a commercial product (Ortivus.com) used by UK ambulance services with local were included. The survey was tested using cognitive interviewing. Analysis was conducted using descriptive statistics for closed questions and thematic analysis for open questions as appropriate.
Results: 131 surveys were completed; 60% of participants were paramedics and 40% were other grades of front line staff. Participants reported consulting electronic/paper sources, and individuals such as carers involved in the patients' care, to establish whether dementia had been diagnosed. Frailty assessments were prompted by social context, reduced mobility, a fall or diagnosis of dementia. Staff reported documenting dementia in 20 different areas on the ePR and 46% of participants stated a preference for a designated area to record the information. However, 15% indicated it was not necessary to record dementia or that no ePR changes were required.

Conclusions:
We have highlighted the variation in ambulance staff practice in recording of dementia. Alterations to the ePR are required to ensure that dementia is recorded consistently and is easily retrievable. Clearer guidance on when to assess frailty may also enhance information provision to care staff in other sectors, resulting in more appropriate clinical and social care.
Keywords dementia recording; emergency medical services; electronic patient record analysis Bias Selection bias was minimised by inviting all staff of all grades to take part within the study period. By ensuring responses were collected electronically and anonymised, response bias was reduced. The survey was piloted by SCAS paramedics working out of the area where the study was undertaken, in order to minimise instrument bias. The managers of all teams in the South East Hampshire division were approached with an invitation, to offer all teams an equal opportunity to take part in the study.

Sample size
All teams in the study region were invited to participate. Accounting for absence and leave, the study size was estimated at 100-150 participants.

Data analysis
Closed questions were analysed with descriptive statistics, using Microsoft Excel. Open-ended questions were analysed with the use of the NVivo Suite (version 12), using principles of thematic content analysis. Data were independently coded and themes identified by two researchers. No a priori themes were postulated, so themes emerged from the data. Differences were settled by a third member of the team.

Number and characteristics of participants
Thirteen teams were approached, of which nine responded and participated in the study. No team declined the invitation and the non-responding teams have offered no explanation for not inviting the research team to their session. 133 individuals were invited to participate and, from this population, 131 (98.5%) participants were recruited, with non-consenting meeting attendees being students who did not feel they had the knowledge to respond. More than half of the participants had worked in the service for more than five years (53.4%, n = 70), 34.4% (n = 45) for one to five years and 12.2% (n = 16) for less than a year. The roles of participants are shown in Figure 1.

Recognition or suspicion of dementia
All participants (100%, n = 131) answered an open-ended question aimed at establishing how the staff identify a patient with dementia, and 22 meaningful codes were developed. The dataset was coded accordingly, and two main themes with a further five sub-themes were developed to categorise the data into meaningful clusters (Table 1). The most dominant code for source of information when A Think-Aloud Cognitive Interview approach was taken, to understand the users' perception of the meaning of the questions asked (Beatty & Willis, 2007). Both think-aloud and probing questions were used to understand participants' interpretation of the questions, a hybrid technique commonly used in the practice of developing and testing questionnaires (Pocock, 2013).
This article adheres to the CROSS survey reporting framework (Sharma et al., 2021).

Study setting
The study took place in the South East Hampshire division of South Central Ambulance Service between November 2018 and March 2019. SCAS covers the counties of Hampshire, Berkshire, Oxfordshire and Buckinghamshire (approximately 3650 square miles). This represents a combined population of 4.2 million people. In the annual data capture for the year when this study took place, 488,526 calls were made to SCAS. Of these, 80,220 were for incidents located in the South East Division area (SE), with 33,873 of these calls for patients aged ≥65.

Participants
Emergency front line ambulance ePR users, including nurses, paramedics, student paramedics, ambulance technicians, associate ambulance practitioners (AAPs) and emergency care assistants (ECAs), were invited to participate. All participants recruited were from the South East Hampshire division of SCAS. Purposive sampling was employed to recruit participants during their quarterly team training sessions. These events are part of the staff rota and are compulsory to attend for all team members. Internal and external guest speakers are often invited to these meetings, and they are designed to provide periodic face-to-face updates and training to staff. All teams in the South East node were invited to take part. The invitation was sent by the lead researcher, via email to the Team Leader. One follow-up invitation was sent to teams that had not responded to the invitation, after which no more contact attempts were made. Information about the study was provided via email, two weeks prior to the date of the training session, with a face-to-face presentation at the team session prior to inviting attendees to provide written consent to take part.

Data collection
Data were collected during team sessions, allowing participants to take as long as required to complete the survey. Study team members were available to assist with technical difficulties or answer questions. Each participant was provided with an electronic tablet. A link to the survey was presented on the tablet which, when clicked, opened a Microsoft Office Form questionnaire (Supplementary 1). Responses were sent to a password-protected Cloud used by SCAS.  establishing a history of dementia was history acquired from the family, followed by patients' paper clinical notes and information from care staff.

Documenting the presence of dementia
The majority of staff (95.4%, n = 125) stated that they would record the presence of dementia on the ePR once this has been identified. Respondents who would not record dementia (3.1 %, n = 4) stated reasons such as not having an appropriate field on the ePR to record it in, feeling that there was no need for it when the patient was being discharged in a care home and, in one case, accidental omission of documentation. Two (1.5%) responses were unrelated to the question and were discarded.

Current location of recording dementia on the ePR
Participants were asked on which section of the ePR they currently recorded that a patient has dementia, with the option to record up to three answers in order of preference. Although the 'Past medical history' (PMH) (26.7%), 'Other PMH' (21.4%) and 'Neurological' and 'Mental health' (13.7%) tabs on the ePR appear the most commonly selected areas, there was a broad spread of other locations that were also recorded ( Table 2).

Where should dementia be recorded?
When asked to identify where dementia should be recorded, past medical history was identified as the most appropriate location and was preferred by 37.4% (n = 49) of participants (Table 3). This field was followed in order of preference by 14.5% (n = 19) for the medical history field and 11.5% (n = 15) the neurological field. It should be observed that all three answers link to a form of medical history. As regards reasons for the choice of location, 47% (n = 62) of participants felt it was the most appropriate location, 22.1% (n = 29) stated it was to support other healthcare professionals, 16.8% (n = 22) described there being nowhere else to capture this information, 7.6% (n = 10) believed this was part of their medical model, 3.8% (n = 5) referred to the reason for attendance and 2.3% (n = 3) provided an invalid answer that could not be categorised.

What would make it easier to record dementia on the ePR?
When asked what would make it easier to record dementia on the ePR, 45.8% (n = 60) of participants asked for a separate, designated tab, which was the dominant answer ( Figure 2). Although other suggestions were listed, an overwhelming majority was associated with the presence of a designated button, tab or tick box, allowing the operator to explicitly record that a patient has dementia.

Assessment of frailty
Seven themes were identified in response to the question 'In what circumstances would you assess for frailty?' ( Table 4). The most prevalent theme indicated by 47.3% (n = 62) of responses was associated with advanced age.

Discussion
We found that ambulance staff report recognising or suspecting that a patient has dementia while taking a history 'on scene', with family members and care staff being the most frequent sources of information. Hard copies of clinical notes and social care notes found on scene were the next most accessed source of information. Fewer respondents acknowledged the role of selfdisclosure by the person with dementia, and only 16/131 (12.2%) respondents indicated that they would use ePRs from previous ambulance service attendances as a source of information. This is consistent with the previous work of Voss et al. (2018), who recognised the importance of 'on scene' information sources to ambulance crews rather than reliance on wider healthcare records. Although electronic patient records are designed to streamline the process of recording and sharing data and to enhance patient care and safety, evidently they do not always produce the intended outcomes, as users' perception of the suitable location of recording dementia in this particular study shows broad disparity. Ambulance staff are required to be information analysts, having to make sense of each scene to which they are called. Whether they suspect dementia and look for evidence to support/refute hypotheses (top-down approach) or piece together the information they find to arrive at conclusions (bottom-up), our data showed that multiple information sources are opening tab of a patient record, and the designated section for frailty act as visible prompts for consideration of frailty. Recognition of frailty can provide useful information to clinicians when considering a patient's risks and resilience as part of a holistic assessment of their needs. If a designated area for dementia recording were placed next to the frailty section in the ePR, this could increase the likelihood of both sections being completed, where clinically indicated. Our study found that most respondents would record the presence or suspicion of dementia; yet, with the absence of a dedicated section for dementia on the ePR, it is recorded across 20 different ePR sections, which closely correlates with the findings of Pocock et al. (2018). This inconsistent location of recording suggests that the current system, the ePR, is suboptimal, as information regarding a patient's dementia may not be readily apparent. This was clearly recognised by staff in our survey, the majority of whom preferred a single place to record dementia diagnosis. Inconsistent recording represents a risk if receiving medical staff cannot reliably source this information following clinical handover, and may contribute to the significant problem of under-coding of dementia during hospitalisation among most Organisation for Economic Co-operation and Development (OECD) countries (OECD, 2018). A recent retrospective review of medical records over a 10-year period found that among patients known to have dementia, its recognition in subsequent hospitalisation was influenced by the reason for admission (Cappetta et al., 2020). Patients were more likely to have their dementia documented when they were admitted to hospital for falls and less likely for medical conditions including pneumonia and urinary tract infection (UTI). Furthermore, their study also reported that the over used to support their recording of dementia. This is intrinsic to the sense-making process developed by Pirolli and Card (2005). The adopted version of this model is illustrated in Supplementary 2. A user-friendly ePR should act as both a recording framework and a prompt tool. Certain findings will act as prompts to seek other related information that may or may not fit the schema of dementia. Therefore, using software that allows for multiple locations for recording dementia may lead to missed prompts for further probing which can contribute to vital information being missed by staff. This, in turn, could have a detrimental impact on patient safety since patients with dementia are likely to have negative outcomes when admitted to hospital and are at further risk of deterioration if their specific needs are not addressed . All individuals involved in the care of dementia patients must, therefore, be aware of the diagnosis of this condition in order to better address this vulnerable group's complex needs. Frailty assessments in emergency departments and hospital wards are becoming more common due to the increased risk of poor hospital outcomes of people with frailty, but there are challenges to completing assessments in a timely way (NIHR Dissemination Centre, 2017). It is possible that an assessment in pre-hospital care may provide at least a guide to emergency department or admitting ward staff to provide adequate care during the initial hours of admission.
Frailty assessment is an optional section on the ePR system used by participants, with advanced age being the factor most likely to prompt assessment and recording of frailty. A history of dementia prompted an assessment of frailty in few cases, suggesting ambulance staff may perceive age as a greater risk factor than dementia for frailty. It may be that the prominence of a patient's age, on the the individual responses of each group of participants, ensuring that every voice was heard and given equal weight.
A potential weakness of the study was that staff were recruited from one geographic area and their experiences may not be entirely the same as staff in other areas in the Trust. However, all staff across the Trust receive the same role-specific statutory and mandatory training and use the same ePR system regardless of their location, so the results are arguably transferable within this Trust, and might be applicable to other services which use the same system.
Closed-response options may have resulted in participants not being able to find an answer that reflected their true opinion. We balanced this by also including free-text response options so that we did not limit or influence 65s were more likely to be admitted to the emergency department by ambulance, and patients presenting with delirium were 20% more likely to have dementia actively managed. Ensuring that information on a dementia diagnosis can be found in a consistent location on the ePR for any patient admitted by ambulance, regardless of their presenting complaint, may prevent a delay in awareness, and subsequent appropriate management, of this complex progressive condition.

Strengths and limitations
The use of individual tablet devices to capture participants' responses provided a secure method of data capture and transfer. This enabled the researchers to capture participants' suggestions about how dementia and frailty should be captured.
Recommendations for improvements to the ePR were generated by users themselves in this study. The importance of adopting user-centred interactive design has previously been highlighted (Horsky et al., 2012), as has the need to understand how well the existing model works before making changes (Jafar et al., 2018). Our study is an early attempt to engage in this process and offer findings that could aid the development of how ambulance services record dementia in the future, using the ePRs. However, it should be acknowledged that this study represents regional findings that may not be generalisable to all ambulance services.

Conclusions
Based on the findings of our study, we recommend implementation of a designated area on the ePR to record dementia and frailty, as all care providers involved in the patient's journey could refer to and record the information in the same place, thus minimising the risk of vital information being missed. This may also prompt increases in frailty assessments. To inform the broader community, we recommend a larger scale study of this design to be carried out across multiple organisations, in order to validate our findings or offer novel contributions to the evidence base. Further evaluation of the ePR after implementation of designated areas for recording dementia and frailty, and follow-up studies with healthcare professionals, families and patients as to the impact of collecting and transferring the information, are essential.