A pilot study evaluating the use of ABCD2 score in pre-hospital assessment of patients with suspected transient ischaemic attack: experience and lessons learned

We performed a pilot, prospective, multi-centered observational study undertaken in two counties served by a large regional EMS in the UK. The EMS serves a total population of approximately 4.5 million, employs approximately 1140 paramedics and emergency medical technicians (EMTs) and received 862,446 emergency calls in 2013/2014, of which 8940 (1 %) were coded by the Emergency Operations Centre dispatch computer as suspected stroke/TIA. The study included five receiving hospital sites.

A standardized training package (Additional file 1) for the study was developed by the study team in collaboration with EMS educators. This included an update on stroke and TIAs and introduced the concept of risk scoring, using the ABCD2 tool, and study procedures and documentation. Training was delivered by EMS clinical supervisory staff, all experienced paramedics, to 146 EMS personnel (paramedics and emergency medical technicians) from one county, representing 58 % of staff eligible to be trained. A ‘cascade’ approach to training was used as this is the standard method employed by the participating EMS. Cascade or ‘train-the-trainer’ approaches involve training a small group who then pass on what they have learnt to the rest of the workforce. Participating staff completed a Learning Validation Form confirming they had read and understood the training material and would adhere to study procedures, and a central log of all staff trained was maintained by the study co-ordinator. Staff in a neighbouring county (part of the same EMS) were not trained in the study procedures, but were asked to provide suspected TIA patients with a study pack, to help assess whether additional training resulted in increased patient identification and recruitment.

We collected data on whether EMS staff were able to recruit patients according to study inclusion and exclusion criteria, whether the EMS diagnosis of TIA was confirmed subsequently by a hospital specialist, and whether patients would consent for follow up. We also collected data on the number of patients who had suspected TIA at the time of EMS assessment but were not recruited to help in planning the main study. We compared recruitment rates for the county where EMS staff received additional training compared with the county where no additional training was provided.

Inclusion/exclusion criteria were developed with the advice of a multidisciplinary steering group comprising stroke physicians, EMS staff and patient representatives. All patients aged 18 years or over with symptoms suggestive of TIA and could speak and understand English were eligible for the study. Patients were excluded from the study in the following circumstances:

Patients were assessed by EMS staff and if they had a suspected TIA and met the study inclusion criteria, were provided with a pack containing a Patient Information Sheet and Consent Form for follow-up, which they were requested to return in a stamped addressed envelope to the study co-ordinator. Study packs were also available in participating hospital Emergency Departments and specialist TIA clinics. EMS staff completed a standard study proforma (Additional file 1), which included baseline demographic and clinical data, and calculated the ABCD2 score. As agreed with the Steering Group and Ethics Committee, the ABCD2 score was not used to determine risk or guide patient disposition as the tool has not been validated in this setting. Hence all patients will have received standard care and been taken to hospital for further assessment, unless they declined transport. Patients in the county where EMS staff were not provided additional training followed the standard EMS pathway into hospital (Additional file 1).

Patients underwent standard evaluation in hospital according to local protocols. Patients who had the diagnosis of TIA confirmed by a hospital specialist and who had returned a signed consent form were followed up by a structured telephone interview at 7 and 90 days by a researcher trained in the use of the Questionnaire for Validating Stroke-Free Status (QVSFS) (Fig. 1) [11]. The interviewer was blinded to patient characteristics and ABCD2 scores.

The pilot study underwent a random, announced, research governance audit by the Sponsor over a 2 day period in February 2014 (recruitment had, as planned, ended on 1st September 2013). This was a useful exercise in confirming compliance with the requirements of the National Health Service (NHS) Research Governance Framework. However this audit was conducted according to a Sponsor proforma designed for clinical trials of investigational medical products (CTIMP). Subsequent feedback from the audit team to the study team recommended that a Delegation of Responsibilities Log for EMS staff was required in addition to the existing training log held by the study co-ordinator, and signed curricula vitae were required from all participating EMS staff. These may be considered as excessive requirements given that EMS staff were merely handing patients an envelope containing study information and not obtaining consent or initiating an intervention. The chief investigator and study co-ordinator had current certification in Good Clinical Practice which is a standard designed to meet regulatory requirements for CTIMP studies rather than pilot, non-interventional studies such as ours. The team were otherwise commended on the quality of the study site file.

Each participating hospital had a different approach to granting approval for the study. In this pilot study, each hospital Research Office reviewed the protocol and study documentation, duplicating efforts already undertaken by the study Sponsor, and leading to an overall delay of 31 months before patient recruitment could commence. Snooks et al. [12] and Thompson et al. [13] have previously highlighted the ‘stifling’ bureaucratic structures and processes for research governance that lead to barriers to undertaking medical research. It should be noted that when the substantial amendment was requested the study passed through the review process with relative ease, reducing any further delay once recruitment was underway. The new UK Health Research Authority (introduced in late 2011), which has responsibility for oversight of ethical and governance processes for research in the NHS, is consulting on a new regulatory framework, which will lead to a more proportionate approach to low risk studies such as ours.

Our pilot study demonstrates that it is possible to conduct a prospective observational study of EMS use of the ABCD2 score in suspected TIA patients. The research-trained paramedics and technicians were able to identify suitable patients, complete initial assessments and follow study procedures, and the study team was able to obtain consent and complete follow up.

The dissemination of training for paramedics and technicians using cascade training is standard for introduction of new procedures across a busy UK EMS, reflecting the challenges of reaching large numbers of staff, distributed across large geographical areas and working shift patterns in the mobile, high pressure and unpredictable prehospital environment. This is a very different environment from the outpatient clinic or hospital ward, where arguably communication is simpler by comparison.

We trained almost two-thirds of eligible EMS staff in one county, but since only 49 TIA patients were recorded across the two counties during a 4 month period, it is likely that exposure of an individual paramedic or EMT to a patient with TIA is infrequent, in common with other conditions such as cardiac arrest where on average an individual paramedic will encounter a single patient in a year, as reported by a recent UK trial [14]. In a pilot randomized trial involving EMS recruitment of stroke patients, 14 patients were recruited over 14 months, a rate similar to our experience [15]. Our study experience and recruitment figures suggest that for a future study to recruit enough patients for an adequately powered study, significant efforts are needed to train very large numbers of EMS staff. We did not survey EMS staff about their perceptions of the research process for our study. Other UK investigators surveyed paramedics engagement in a feasibility trial in an ultra-acute stroke setting, and identified lack of institutional support for research, a learning curve, and (as we highlight) rarity of eligible patients and lack of time for training as important barriers to study success in the EMS environment. EMS research is a relatively recent phenomenon, and paramedics have mixed views on their responsibilities: a survey in the United States indicated that paramedics were interested in taking part in research, but that only 38 % of them would want the right not to participate [16]. This requires further study and education for UK EMS personnel around their professional responsibilities to engage with research.