Effectiveness of the BBC's 999 training roadshows on cardiopulmonary resuscitation: video performance of cohort of unforewarned participants at home six months afterwards

Introduction

Bystander cardiopulmonary resuscitation improves survival in people who have a cardiac arrest outside hospital.1 Consequently, increased emphasis has been placed on training the public in these techniques as a means of reducing mortality from ischaemic heart disease.2 3 Since 1994 the BBC has organised annual training roadshows on cardiopulmonary resuscitation throughout the United Kingdom to coincide with the broadcasting of its 999 television programme.

Although the effectiveness of bystander cardiopulmonary resuscitation has been shown, the effectiveness of mass training courses such as the BBC's 999 roadshow is less clear. Much evaluation of cardiopulmonary resuscitation training has been concerned with the performances of medical and allied professional staff rather than the lay public.4 5 6 With few exceptions,7 studies that have tried to evaluate training of the lay public have pre-warned subjects of testing either explicitly or immediately before retraining.8 Despite this, most results have been disappointing.9 We describe an innovative method of evaluating how an unforewarned lay person trained on a roadshow would perform cardiopulmonary resuscitation should a cardiac arrest occur in their home.

Subjects and methods

At each of the 10 roadshows held between April and June 1994, participants were asked to complete an optional card consenting to take part in further unspecified research. A total of 7584 course attenders completed an anonymised demographic questionnaire, and of these 6123 (81%) completed a consent card. The research entailed “cold calling” on a sample of trainees and video recording a simulated attempt at cardiopulmonary resuscitation using a Laerdal Recording Anne Manikin (Norway). Trainees from Londonderry, Northern Ireland, were excluded owing to the sensitivity at that time of undertaking unsolicited house calls there. Those who lived in a postal district outside a 32 km radius of their training centre were also excluded on logistical grounds. This left 4651 cards, from which a 6% (n = 280) stratified random sample was chosen by ordering the cards by adjacent postal district and selecting the 16th and 17th card alternately from a starting point generated by computer.

All tests were conducted between 27 and 35 weeks after the initial training. Each test was videoed for later evaluation in conjunction with the manikin recording strip. Two researchers, a man (CLlM) and a woman (see acknowledgements) visited subjects at home, without warning, and, after showing the completed consent card, asked their permission to conduct the test. The manikin and video equipment were prepared while the subject completed two questionnaires on sociodemographic data and attitudes to emergency life support. The subject then read the following passage:

“On entering a room you see a person lying still upon the floor. Please describe and demonstrate how you would approach and assess the casualty and what action you would take. I will not offer any advice or opinion other than to report the status of the casualty during your assessment.”

The subject was not informed that the patient was in cardiac arrest but was expected to deduce this by assessment. The only interruption from the researchers was to confirm the status of the casualty after the appropriate assessment—that is, to report that the casualty was unconscious, not breathing, and pulseless. If subjects stated that they would phone for help they were told that help was on its way. Subjects were encouraged to continue cardiopulmonary resuscitation for a maximum of three cycles, after which they were asked when they would have stopped had this been a real life situation.

All videos and recording strips were analysed by one of us (CLlM) using guidelines developed at Cardiff. The test criteria were based on the European Resuscitation Council's guidelines,10 as these formed the core of the 999 teaching syllabus, with the categories of effective and safe representing what we considered to be reasonable deviations from these guidelines (table 1). A 20% sample was analysed by another of us (CAL) as a check on interobserver variation. All analysis was performed using SPSS for Windows.

Results

Of the original 280 subjects, 186 (66%) took part in the study. To complete the remaining 94 tests, 136 replacements were made for the reasons listed in table 2. The final sample was comparable with the total trained cohort in terms of age, sex, and social class. Table 3 shows the kappa score of inter-rater agreement between the two assessors for the 10% sample.

Table 4 shows the number of subjects performing each procedure and, when appropriate, whether that procedure was effective, ineffective, or potentially injurious according to the European Resuscitation Council guidelines. Of the 210 (75%) who effectively checked for breathing, 173 did not first open the airway.

Twenty seven (10%) subjects said that they would call an ambulance before beginning cardiopulmonary resuscitation, but when subjects were asked when they would stop cardiopulmonary resuscitation in a real situation 138 (49%) stated they would continue until an ambulance arrived. Of the remaining 142 (51%), 51 (36%) said that they would continue for a defined period and 55 (39%) that they would continue until the patient showed signs of revival. Sixty of these 106 (57%) subjects were asked a follow up question about what they would do when either the defined period had elapsed or the patient had failed to respond. Nineteen of them (32%) said that they would wait for the ambulance to arrive while 29 (48%) said that they would leave the patient and telephone for an ambulance after having started cardiopulmonary resuscitation.

The overall performance of cardiopulmonary resuscitation was assessed for effectiveness and safety (table 5). None performed cardiopulmonary resuscitation to the European Resuscitation Council guidelines, but 19 (7%) performed it in an effective and safe manner. A further 14 subjects (5%) were effective but performed at least one procedure that was potentially injurious. Ineffective performances were recorded for 243 subjects (87%); of these, 110 (45%) were also classed as potentially injurious. Four subjects (1.4%) did not attempt ventilation or compression.

Sex and social class had no influence on the effectiveness of cardiopulmonary resuscitation (table 6), but age was a significant factor, with 31 (14%) of those under 45 performing effectively compared with only two (4%) of those aged 45 or over (P<0.05). Those who felt confident before testing performed more effectively (26 out of 133, 20%) than those who were not confident or unsure (seven out of 127, 6%) (P<0.005). Subsequent training also improved performance, with eight of the 20 (40%) who had attended a subsequent course giving an effective performance against 25 of the 256 (10%) who had not. The numbers were, however, too small to make this result conclusive.

There was no difference by age group or social class in those performing safe cardiopulmonary resuscitation, but only 34 (42%) men performed resuscitation safely compared with 117 (62%) women (P<0.005).

Discussion