Article Text

Download PDFPDF

Inappropriate 999 calls: an online pilot survey
  1. Helen M Kirkby1,
  2. Lesley M Roberts2
  1. 1MRC Midlands Hub for Trials Methodology Research, Public Health and Epidemiology, The University of Birmingham, Edgbaston, Brimingham, UK
  2. 2Primary Care Clinical Science, The University of Birmingham, Edgbaston, Birmingham, UK
  1. Correspondence to Dr Lesley Roberts, Primary Care Clinical Science, The University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK; robertslz{at}


Background Abuse of ambulance services is high, and there is concern among healthcare professionals that misuse of ambulances places stress on services, which may jeopardise patient care. This study aims to determine the proportion of people who correctly identify appropriate situations to call for an ambulance, and determine the characteristics of those most likely to call inappropriately.

Methods An online questionnaire presented 12 common scenarios that may require medical attention and required participants to identify when they would request an ambulance. Proportions correctly responding to each scenario were calculated and each respondent was given a total score. t-Tests compared mean scores between groups (with and without first aid (FA) training), and χ2 tests compared between-group proportions of correct answers for scenarios. Backwards stepwise logistic regression analyses determined the characteristics of those most likely to call inappropriately.

Results 150 respondents completed the questionnaire. 5.2–47.8% responded with an inappropriate answer, depending on the scenario. Almost all participants identified the need for an ambulance in 3/5 scenarios when it was required; however, fewer (74.8%) respondents identified the need for an ambulance to a suspected stroke. The majority correctly identified an ambulance was not required in only 2/7 scenarios. Those with FA training were less likely to call inappropriately in all scenarios (significant in three situations). However, no participant characteristics were predictive of calling an ambulance inappropriately once confounders were taken into account.

Conclusions The majority would call for an ambulance appropriately when a real emergency occurred, and most inappropriate classification occurs when an ambulance is not required.

  • Ambulance
  • emergency ambulance systems
  • emergency medical service
  • first aid
  • inappropriate
  • public understanding

Statistics from

Despite clear guidelines available that relate to situations requiring ambulance attendance,1 abuse of ambulance services is high,2–5 with research suggesting between 16% and 51.7% of calls are inappropriate.2–4 6 7 This misuse may be because those guidelines, while available for the public to view, have not been subject to a mass media campaign and therefore the public are not aware they exist. There is concern among healthcare professionals that misuse of ambulance and emergency departments places stress on services that may jeopardise patient care.2 3 There is also evidence that patients who have the ability to get to hospital by other means call an ambulance as a convenient alternative.4

Evidence suggests those most likely to use emergency medical services and use them inappropriately are white, under 50 years old, with a high educational level.4 5 The initiator of the call can be important in determining inappropriate calls, with shop staff and employers least likely to make inappropriate calls, which may be due to essential basic first aid (FA) training.6 Teachers are most likely to order unnecessary ambulances to cover their liabilities,6 and friends and passersby are also not good judges, potentially due to panic or lack of basic medical or FA knowledge.6

Qualitative research concerning patient's reasons for calling an ambulance found the most common reason for calling when no emergency was present was lack of awareness of other sources of help.7 Others believe ambulances are the safest and most secure form of transport, and emphasise the need for care so they may expedite treatment in the emergency department.7 Pallazzo et al8 provided evidence of patients' reasons for calling 999 inappropriately, with only 60% calling because they believed they had a serious or life-threatening condition. Of those inappropriate calls, 16% were unaware an emergency general practitioner service was available, 16% had no means of reaching the hospital, and 8% wished to avoid the long wait to see an emergency department doctor.8 The sample size was, however, very small, given that only 25 patients were willing to state motives for their call.

Abuse of the ambulance service has always been a salient topic in the media, and recent times are no exception.1 9–12 Reports of obviously inappropriate calls made to the ambulance service include asking if a Chinese take-away is safe to eat.13 Ambulance attendance is now not required to each call received, and while criteria-based dispatch facilitates appropriate service use, much of the situational context cannot be gathered by these systems alone. For example, a patient dials 999 to complain of back pain, which could potentially be a serious symptom of an aneurysm. A paramedic is sent, and discovers the patient has had it for 6 months and called for an ambulance because they wanted stronger painkillers.9 Appropriate use of services is therefore dependent on public understanding, because an ambulance has to be sent to these potentially serious situations even though the caller may have additional information and know it is not an emergency. There is a deficit of evidence relating to the situations in which public awareness of appropriate ambulance use is lacking. This evidence is needed to target future campaigns, and detailed information on public understanding is required to inform such an educational strategy.

This pilot study aimed to determine the proportion of people who correctly identify appropriate situations in which to call for an ambulance, and determine what alternatives would be taken if they deemed calling an ambulance inappropriate. The secondary aim of the study was to determine the characteristics of those most likely to call for an ambulance inappropriately, to enable targeting of future strategies to improve knowledge and understanding.


An e-mail invitation was distributed to 133 friends, family and colleagues of the researcher, requesting they complete the online questionnaire, and contained a link to the domain at which the survey was hosted. They were also asked to forward the e-mail onto everyone in their mailing list, which would set up a chain e-mail for participation. This method of recruitment is likely to create bias in the results. However, as this is a pilot study, the use of an extended group of friends/family as a study population was deemed reasonable. Anyone over the age of 18 years (the study was not testing the knowledge of minors), living in the UK (the study related to the provision of UK services) with a valid e-mail address was eligible to take part. The questionnaire comprised three parts:

  1. Demographic questions used to demonstrate generalisability of results and to identify characteristics of those not using services appropriately.

  2. Vignettes of common medical scenarios, which asked participants to respond if they would or would not call for an ambulance. (Table 1 presents a summary of vignette scenarios.) Scenarios were devised using information regarding situations in which to call for an ambulance from the West Midlands Ambulance Service. Once developed, an agreement was made on the appropriateness of ambulance attendance in each scenario in conjunction with the West Midlands Ambulance Service via e-mail and a paramedic. Respondents were asked if they would call an ambulance or not in each scenario. If they responded ‘no’, they were asked what alternative action they would take.

  3. Questions to determine knowledge of other forms of medical help.

Table 1

Vignettes and numbers of scenarios

Several approaches have been used to achieve community-based samples, and increasingly the use of the internet has meant forums and social networking sites are an easy way to approach large numbers of people.10 11 While e-mails are used to cascade petitions for political causes, humour or warnings, we were unable to find any evidence that they have been used to achieve a community-based research sample. This study piloted a novel sampling technique of using chain e-mails as a way to generate research participants, therefore embedding a methodological question within the design (which will be reported elsewhere).

Ethics approval was obtained from the BMedSc (Medicine in Society) Internal Ethics Committee at the University of Birmingham before commencement of the study.

Data were collected from 10 March to 11 April 2009. Analyses were performed using SPSS version 16. Proportions correctly responding to each scenario, and an overall score (0–12) were calculated for each participant. Respondents who did not answer all questions were disregarded from analyses involving total scores. Individual scenario-based analyses included all available answers. Medical personnel were identified as indicating they had ever studied as a medical student, dentist, nurse, or ambulance crew, and those with FA training as having had St John's, Red Cross or other formal FA training. t-Tests were used to compare mean scores between those with and without FA training after excluding those with medical training. χ2 tests were used to compare proportions of correct answers for each scenario between these groups. Backwards stepwise logistic regression analyses were undertaken; with covariates sex, age (<50 or >50 years), marital status (single or other), country of birth (UK or other), ethnicity (white British or other), level of education (GCSE or below, or A-level and above), FA training, medical training, and having children or not; to determine the demographic characteristics of those most likely to ring for an ambulance inappropriately.


The e-mail was distributed to 133 people initially, 161 attempted the questionnaire by the analysis cut-off date, and 150 completed all vignettes. Sixty-six per cent (66.1) were women, 51% were aged 18–29 years, 23.2% were aged 30–44 years, 18.5% were aged 45–59 years and 7.3% were aged 60–74 years. Ninety-six per cent (95.7) were white, 2.5% were black, 1.2% were Asian and 0.6% self-classified as other ethnicity.

Of the 150 respondents, 37 had undertaken medical training, 68 had undertaken some FA training, and 45 had undertaken no medical or FA training. Table 2 reports mean scores and individual scenario scores. Most inappropriate decisions occur when an ambulance is not required, with between 5.2% and 47.8% indicating they would inappropriately call for an ambulance in five scenarios in which it was not deemed necessary.

Table 2

Mean scores and proportions of incorrect responses for each scenario between groups (with 95% CI)

Appropriate call scores ranged from 5 to 12, with a mean score of 10.15. Those with FA training were less likely to make an inappropriate decision in all scenarios, and the between-group difference for mean score obtained was 0.63 (FA training (10.41, SD 1.149) vs no FA training (9.78, SD 1.277); p=0.007). Those with FA training were significantly less likely to ring for an ambulance inappropriately in three scenarios; for a man with chronic back pain (2.9% vs 15.4%; p=0.013), a drunk but conscious friend (7.2% vs 21.6%; p=0.023) or if they had an episode of haematuria (0% vs 6.2%; p=0.035).

Logistic regression analyses

The mean score was used to determine good and poor scores, with poor score used as the independent variable in backward stepwise logistic regression analysis with covariates sex, age (<50 or >50 years), marital status (single or other), country of birth (UK or other), ethnicity (white British or other), level of education (GCSE or below, or A-level and above), FA training, medical training and having children or not. No variables were predictive of participants calling an ambulance inappropriately.

The same regression analysis was subsequently undertaken for each scenario, using correct response as the independent variable. For scenario 1, an ambulance was more likely to be called inappropriately when the caller was female (OR 5.964; p=0.015), over 50 years of age (OR 4.573; p=0.032) and single (OR 4.512; p=0.034). There were no significant characteristics for any other scenario. Once confounding factors were taken into account, those with FA training were no less likely to ring for an ambulance inappropriately than those without FA training.

Alternative actions to calling for an ambulance

Summaries of key themes identified as alternatives to calling for an ambulance, and the proportions of respondents giving each alternative response, can be seen in table 3.

Table 3

Key themes identified by participants as alternatives to calling for an ambulance, and proportions of respondents giving each alternative response


Across the range of scenarios, between 46.7% and 100% of respondents correctly identified appropriate situations in which to call/not to call for an ambulance. Situations least likely to be responded to appropriately included meningitis, stroke and labour. Most participants indicating they would not call for an ambulance, could identify an appropriate alternative action, with these actions being split into three main categories: seek medical advice elsewhere; seek lay advice (from friends and family); or self-medicate/monitor.

A large number of responses from those with FA training allowed the exploration of differences of this subgroup. Those with FA training were less likely to ring for an ambulance inappropriately in all scenarios. Initial analyses suggested those with FA training were significantly less likely to ring for an ambulance in three situations (back pain, drunk friend and haematuria), although these results appear to be due to confounding factors. FA training should allow a superior level of knowledge required to evaluate situations and enable appropriate ambulance use more often than the lay public, so it is possible that FA training does have a moderating effect and this should be explored further.

In three out of five scenarios in which an ambulance was required the majority correctly identified the need. This is encouraging and suggests that most can identify and respond appropriately to genuine emergencies. One of these exceptions is the case of a stroke, in which only 74.8% (95% CI 68.01 to 81.67) of respondents correctly identified the need for an ambulance to an elderly lady with signs of a stroke, and only 80% of people would either call for an ambulance or take them to the emergency department. This statistic is concerning, given the current government FAST campaign12 to raise awareness of the signs of stroke. This launched on the 9 February 2009,12 and uses the acronym FAST: face, arm, speech, time to call 999; to encourage people to call for an ambulance straight away if they see someone with any one of these signs. Although this is a 3-year initiative to raise awareness of stroke and is in its initial stages, there has already been substantial media attention and this finding is therefore surprising.

There are high proportions of respondents who would call an ambulance for five scenarios in which ambulance attendance was not appropriate. All of these scenarios may require medical advice or help ranging from FA at home to an urgent emergency department visit, but none require ambulance attendance. It is highly likely there is confusion between the need for medical treatment and the need for an ambulance. However, some confusion may be introduced by the use of vignettes. An example is scenario 5, which only describes a patient who has had too much alcohol and is being sick. In the situation described, an ambulance would not be required. However, if the patient was also unconscious, this would warrant an ambulance. It is possible that readers of vignettes may infer additional factors that would alter their decision. Another issue making ambulance need uncertain is recent changes to the ambulance service, such as the extended role of ambulance crews who now provide some primary care and discharge patients at scene. Such issues mean that providing clear-cut guidelines as to when an ambulance should/should not be called may be difficult. Media campaigns target specific situations and advise on the course of action that should be taken. In the case of previous meningitis media campaigns, ‘seeking immediate medical help’ was encouraged if signs/symptoms of the illness were experienced. This campaign did not stipulate how immediate medical help was to be sought, and is likely to have led to the high numbers indicating they would not call for an ambulance for signs of meningitis. The FAST campaign explicitly told the public that an ambulance should be called if any sign of stroke is seen, but there was still a large proportion of people who indicated they would not call an ambulance. If the public are not taking the correct course of action when told explicitly what to do (as in the FAST campaign), they are even less likely to make the right choice when they are not explicitly told (as in the meningitis campaign). Further research should be conducted to determine how advertising campaigns impact on the population, and how the finer points can be differentiated.

Future campaigns to tackle the inappropriate use of the ambulance service can be tailored to those situations in which an ambulance is not required. Campaigns are needed to target and reduce numbers of inappropriate calls that place a financial burden upon the ambulance service. If the public do not know the difference between needing an ambulance and needing to go to hospital, this needs to be addressed, possibly by media campaigns. The most notable problem is the case of the woman in early labour, in which the use of an ambulance as a ‘maternitaxi’ is already well documented on the internet.14 15 Most people attend antenatal classes, so this problem could be addressed here by assuring women and their partners that calling for an ambulance in the early stages of labour is not appropriate.

When respondents deemed an ambulance inappropriate, the majority knew of an appropriate alternative action, and there is a high level of knowledge of alternative sources of medical help.


The sample size required to estimate the proportion of those who would make inappropriate decisions to an error margin of 5% was 377. Only 162 respondents completed the questionnaire meaning an error margin of approximately 7.5% applies.

The recruitment strategy used in this pilot study may have introduced bias because friends, family and colleagues are likely to be of similar educational background and to a large extent the same sex and ethnicity. This may lead to the population sample not being generalisable to the UK population in many aspects. These include educational level (77.43% of the study population was educated to A-level or higher compared with 30.13% of the UK population), white ethnicity (95.7% of the study sample was white compared with 90.9% of the UK population), and the age group 18–24 years (31.93% of the study population was in this age group compared with 10.85% of the UK population). Although these results may not be generalisable, the high level of the inappropriate use of ambulances shown in this study is still a concern. Further to this, given that the study population could be perceived to have had an interest in the research (by the nature of participant selection), it is likely this study has underestimated inappropriate calls in the wider population.

Unmeasured possible confounders mean logistic regression results may have failed to account fully for other explanatory factors. One such possible confounder is previous experience of a situation. The number of participants who had experienced each of the situations presented is likely to be low and the impact of previous experience therefore potentially negligible at the population level; further work in this area should consider this issue.


Most people would call for an ambulance appropriately when a real emergency occurred, but there are high levels of inappropriate calls when emergencies are not present. Those with FA training may be less likely than the lay public to call for an ambulance inappropriately in all situations, but these associations disappear once confounding factors are taken into consideration.

There may be value in nationwide FA and education courses in order to reduce the numbers of inappropriate calls to the ambulance service, but this needs to be explored further. If successful, it could reduce unnecessary costs and improve response times. Campaigns should initially be targeted at the misuse of the ambulance service by those in the early stages of labour, as this returned the highest proportion calling for an ambulance inappropriately. There should be education of the public that needing to go to hospital urgently and requiring an ambulance are different things. Recommendations for future research are for these findings to be conducted on a larger sample representative of the UK population.


The authors would like to thank the participants who completed the questionnaire, the University of Birmingham for funding the research, the Internal Ethics Committee at the University of Birmingham for support and feedback during conduct of the research, Dr Roger Holder for providing statistical support, the Royal College of Physicians and the Primary Care Research Trust for awarding HMK personal bursaries to fund the intercalated degree for which this research was conducted in part fulfilment of.



  • Funding This research was undertaken as part fulfilment of a bachelor of medical science, public health and epidemiology degree, and funding was provided by the University of Birmingham.

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the BMedSc (Medicine in Society) Internal Ethics Committee, University of Birmingham.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Primary survey
    Darren Walter