Article Text

Download PDFPDF

Low level of knowledge of heart attack symptoms and inappropriate anticipated treatment-seeking behaviour among older Chinese: a cross-sectional survey
  1. Pui Hing Chau1,
  2. Gordon Moe2,
  3. Siu Yin Lee3,
  4. Jean Woo4,
  5. Angela Y M Leung5,
  6. Chi-Ming Chow2,
  7. Cecilia Kong1,
  8. Wing Tung Lo1,
  9. Ming Hay Yuen1,
  10. Julie Zerwic6
  1. 1 School of Nursing, The University of Hong Kong, Hong Kong, China
  2. 2 Department of Medicine, University of Toronto, Toronto, Canada
  3. 3 Department of Health, Elderly Health Service, Hong Kong, China
  4. 4 Department of Medicine and Therapeutics, The Chinese University of Hong Kong, Hong Kong, China
  5. 5 Centre for Gerontological Nursing, School of Nursing, The Hong Kong Polytechnic University, Hong Kong, China
  6. 6 College of Nursing, University of Iowa, Iowa City, Iowa, USA
  1. Correspondence to Dr Pui Hing Chau, School of Nursing, The University of Hong Kong, Hong Kong, China; phchau{at}


Background Prehospital delay of acute myocardial infarction (AMI) is common globally, and Hong Kong—home of a rapidly ageing Chinese population—is not an exception. Seeking emergency medical care promptly is important for patients. Treatment-seeking behaviours have been shown to be associated with knowledge of AMI symptoms and specific cultural beliefs. This study aimed to assess the level of knowledge of AMI symptoms and expected treatment-seeking behaviour among older Chinese in Hong Kong.

Methods A cross-sectional population-based survey was conducted at the Elderly Health Centres in Hong Kong from March to September 2016. Face-to-face interviews were conducted with a structured questionnaire based on previous studies and local adaptations.

Results Among 1804 people aged 65 years and above who completed the questionnaire, chest pain (80.2%), palpitations (75.8%) and fainting (71.9%) were the major symptoms recognised as AMI related. Meanwhile, stomach ache (46.9%), coughing (45.4%) and neck pain (40.8%) were recognised as not AMI related. The mean expected discomfort intensity during AMI onset was 7.7 out of 10 (SD=2.1). Regarding the expected treatment-seeking behaviour, seeking non-emergent medical care was the most popular action when AMI symptoms emerged during the day, without chest pain or with lower discomfort intensity, whereas calling an ambulance was the most common option when AMI symptoms emerged at night or with high discomfort intensity.

Conclusions To minimise delays in seeking treatment, future health education should focus on increasing the public knowledge of AMI symptoms and the need to call an ambulance during an emergency.

  • elderly
  • health behaviour
  • ischaemic heart disease

Statistics from

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.


Worldwide, ischaemic heart disease, including acute myocardial infarction (AMI), account for 164 million disability-adjusted life-years (DALYs) that accounted for 6.7% of DALYs for all causes.1 Patients with heart attack need a timely medical treatment, and early presentation in a hospital is essential to reduce morbidity and mortality. However, prehospital delay of heart attack is common globally.2 A patient’s decision to seek medical treatment for AMI is a complex process that involves correctly recognising AMI symptoms, determining the need for professional medical care and finally taking appropriate actions to seek emergency medical evaluation.3 Failure in any of these steps will result in prehospital delay, which increases the time to treatment.

Treatment-seeking behaviours have been shown to be associated with knowledge of heart attack symptoms.2 4–7 However, populations did not have good knowledge towards AMI symptoms. While many people recognise typical symptoms like chest pain, many fail to recognise atypical symptoms like sweating, neck pain or jaw pain. According to population-based surveys both in Western and Asian populations, the majority (ranging from 50.0% to 79.1%) of respondents could name chest pain as an AMI symptom.6 8–12 Among Hong Kong Chinese population, such proportion was as high as 90%.13 Meanwhile, the proportions of respondents who could name jaw pain as AMI symptom were only 6% in Birmingham6 and around 20% in Kuwait and Hong Kong.9 13 Only 20% of Chinese people in Beijing could name four or more AMI symptoms,12 and 11% of South Korean people could name five symptoms.11 An additional area of confusion is the discomfort intensity. Nearly 90% of individuals surveyed expected the discomfort to be severe, a ‘9’ on a 0–10 scale (with 10 indicating maximum discomfort).14 However, there was a lack of studies among Asian populations in terms of the perception towards discomfort severity for comparison.

Apart from knowledge level, there are other factors associated with prehospital delay. Those with older age and low education level have been consistently reported in Western and Asian populations as having a longer prehospital delay.15 16 It was because older patients with AMI were more likely to attribute their symptoms to normal ageing, experience more atypical symptoms (absence of chest pain/discomfort) and tend to have a longer patient delay.16–18 In China and Hong Kong, some traditional Chinese beliefs, which are more prevalent among older people, such as internal coping strategies, maintaining a sense of normality and avoiding bothering others, might further delay medical treatment.4 16 Interestingly, the belief of avoiding bothering others also persisted in non-Chinese populations.18 Meanwhile, previous AMI experience did not necessarily shorten or lengthen prehospital delay.15 16 In China, onset at night, accessibility of transportation and absence of medical insurance coverage were also factors that lengthened prehospital delay.16 Whereas in Western populations, seeking primary healthcare consultation was reported as a risk factor for prehospital delay.15

Hong Kong is a rapidly ageing Chinese population. In 2016, 1.16 million people were aged 65 years or above. In 2012–2013, 14% of Hong Kong Chinese patients with AMI below 75 years old had atypical symptoms, while this proportion increased to 40% among those aged 75 years or older.19 These percentages were much larger than those reported in those patients (<10% of the younger group and >20% for the older group) during 1991–1994.20 This implied a large proportion of older Chinese patients with AMI might experience longer delay since they might not realise atypical symptoms as related to AMI. Before developing public education programmes to reduce prehospital delay, the expected decision-making process for recognising AMI symptoms and seeking treatment must be evaluated in the local population. This highlighted the needs to study the knowledge level of AMI among the older Hong Kong Chinese population and explored their expected treatment-seeking behaviours. There was one local study addressing knowledge of AMI symptoms and another addressing AMI patients’ treatment-seeking behaviours,4 13 a population-based study including both AMI symptom knowledge and anticipated treatment-seeking behaviour would address the research question from a different perspective and enable investigation of the association between the two domains. Therefore, this study aimed to assess the level of knowledge of AMI symptoms and expected treatment-seeking behaviour among older Chinese in Hong Kong.


Study design and participant recruitment

A cross-sectional population-based survey was conducted with older Chinese people recruited from Elderly Health Centres (EHC). The Department of Health has established 18 EHCs in Hong Kong to promote health among Hong Kong residents ≥65 years old through a multidisciplinary approach in a primary care setting. Services provided by EHCs include health assessment, physical check-up, individual health counselling, health education and general outpatient services. Elders receive a comprehensive health assessment on their enrolment. Enrolment is voluntary. In 2017, EHCs have performed 45 000 health assessments for older people, roughly covering 4% of the Hong Kong older population. Since those receiving services at EHCs have access to primary healthcare, they might not be representing those who have poor access to healthcare. Details of the EHCs have been reported elsewhere.21 22

Six EHCs were randomly selected within each geographical region, with one in the Hong Kong Island, two in Kowloon and three in the New Territories. Participants were referred by the medical staff at EHCs. Sample size calculation was based on the precision of the estimated proportion of older people with poor knowledge of AMI symptoms. Assuming such proportion was 50% (ie, the most conservative scenario), a sample of 385 older people would be required to estimate a 95% CI of the proportion with a margin of error of 5%. As age (65–74 and ≥75 years) and gender specific rates would be estimated, 385 subjects from each of the four age-specific and sex-specific groups were required. Taking into account 15% incompletion or item-wise missing, at least 1812 older people would be recruited. Hence, a quota sample of 450 participants was targeted for each of the four subgroups defined by age group (65–74 or ≥75 years) and sex. The inclusion criteria were Chinese ethnicity, age ≥65 years, resident in Hong Kong for ≥3 months per year and no communication problems or cognitive impairment as identified in the health assessment by the medical staff at EHC. Trained research assistants explained the study to eligible participants and obtained their informed consent. Face-to-face interviews were conducted in a room at the EHC to avoid influencing their responses or distraction from other respondents. To prevent respondents from learning AMI symptoms from other parts of the questionnaire, they were not allowed to change their answers once the questions were answered.


A structured questionnaire was developed in Chinese, modified from similar reported surveys.10 14 An expert panel (one cardiologist, one geriatrician and two nurses) rated the validity of the questions. Cognitive debriefing was conducted with 30 participants to ensure that the respondents could understand the questions with no misinterpretations. The average interview time was 15 min.

The first part of the questionnaire listed 20 symptoms.10 14 Participants rated the likelihood of each being an AMI symptom as ‘likely’, ‘unlikely’ or ‘unsure’. Participants were asked to identify a key sign of heart attack and rate their expected overall intensity of discomfort on a scale of 0 (no discomfort) to 10 (maximal discomfort).

In the second part of the questionnaire, participants were presented with six heart attack scenarios based on local AMI survivors’ experiences (box 1).4 23 These scenarios were chosen to investigate the influence of time of onset, chest pain and chest pain intensity on their anticipated treatment-seeking behaviour. In scenarios A and B, ‘heart’ was explicitly mentioned as a clue to respondents that the discomfort is of cardiac origin but with different times of onset, to determine if onset at night would prolong the delay, as reported in other studies.16 18 24 In both scenarios C and D, pain in the shoulder, arm and abdomen and feeling of indigestion were included. The only difference was that chest pain was explicitly stated in scenario D. Different responses between scenarios C and D would indicate that the presence of chest pain was critical to decision making. In both scenarios E and F, chest pain was included. The only difference was that the pain intensity in scenario E (scored 5 out of 10) was lower than that in scenario F (scored 8 out of 10). Different responses between scenarios E and F would reflect whether pain intensity was critical to decision making. For each scenario, the anticipated treatment-seeking behaviour of the participants and their reasons were collected through open-ended questions, ‘What would you do first? And why?’. If the participants replied answers like ‘wait and see’, they would be asked ‘How long would you expect to wait until you take other actions?’.

Box 1

Heart attack scenarios for investigating participants’ anticipated treatment-seeking behaviour

  1. You experience some ‘heart discomfort’ continuously for half an hour in the daytime.

  2. You experience some ‘heart discomfort’ continuously for half an hour but in the night-time.

  3. You experience pain in the shoulder, arm and abdomen, along with a feeling of indigestion continuously for half an hour in the daytime.

  4. You experience pain in the shoulder, arm and abdomen, along with a feeling of indigestion continuously for half an hour in the daytime, but also experience chest pain.

  5. You experience pain in the middle of your chest continuously for half an hour in the daytime, with a pain score of 5 (out of a scale of 0–10, with 10 being most painful).

  6. You experience pain in the middle of your chest continuously for half an hour in the daytime, with a pain score of 8 (out of a scale of 0–10, with 10 being most painful).

In the last part of the questionnaire, demographic and medical variables and a question about the source of respondents’ health information were also included.

Outcome measures

‘Poor knowledge about AMI symptoms’ was defined as failure to identify at least half of the common symptoms experienced by the local patients with AMI (ie, five out of nine symptoms).4 The lower the proportion correctly identifying the symptoms, the higher the probability that the population would miss the AMI symptoms when they actually occur, resulting in a delay to seek treatment. ‘Inappropriate expected treatment-seeking behaviour’ was defined as failure to call an ambulance in at least three of the six hypothesised scenarios of AMI onset. The lower the proportion correctly assessing the scenarios, the higher the probability that people would make inappropriate treatment-seeking decisions when an AMI actually occurs, resulting in delayed treatment. Only completed questionnaires were used in the analysis, which included completed all questions on knowledge about AMI symptoms and completed at least four questions on expected treatment-seeking behaviour for the six scenarios.

Data analysis

The responses to the open-ended questions on the expected immediate treatment-seeking behaviour and its reasons were coded. The proportions of respondents who satisfied the outcome measure criteria were tabulated by dividing the frequency by the number of respondents. Proportions were weighted by the age and gender composition of the 2016 older Hong Kong population. Matched-pair odds ratios (ORs) were used to investigate the associations between the expected treatment-seeking behaviour of calling ambulance (as against not calling ambulance) and factors like onset time (scenarios B vs A), chest pain (scenarios D vs C) and chest pain intensity (scenario F vs E). A matched-pair OR greater than 1 implied the increased intention of calling ambulance in the scenario with the corresponding factor and less than 1 implied decreased intention. Multiple logistic regression was used to identify demographic (eg, sex, education level) and medical characteristics (eg, history of AMI and history of diabetes) that predicted poor knowledge of AMI symptoms. Multiple logistic regression was used to identify demographic and medical characteristics, as well as the knowledge level of AMI symptoms, that predicted an inadequate expected treatment-seeking behaviour. Missing data were excluded from analysis. The OR and 95% CI of each predictor were estimated. SPSS V.23 was used for statistical analyses, with a significance level of 5%.


A total of 1839 participants were recruited at EHCs from March to September 2016. Thirty-five participants did not complete the interview, leaving 1804 complete interviews for analysis. Among the respondents, 52.1% were 75 or older, and 51.1% were female. Approximately half of the respondents had an education level of primary school or below. About 4% of respondents reported a previous history of AMI (table 1).

Table 1

Characteristics of the respondents (n=1804)

Most respondents regarded chest pain (80.2%), palpitations (75.8%) and fainting (71.9%) as AMI symptoms but considered stomach ache (46.9%), coughing (45.4%) and neck pain (40.8%) to be unlikely AMI symptoms. Quite some respondents were unsure whether sweating (49.4%), headache (44.9%), chills (41.6%), coughing (41.3%) and neck pain (40.6%) were symptoms of AMI. One-third of respondents recognised at least five of the common symptoms reported among patients with AMI in Hong Kong (table 2). About 33.0% of participants perceived chest pain as the key sign of AMI onset. The respondents’ mean expected discomfort intensity during AMI onset was 7.7 (SD=2.1). For 24.4% of respondents, the expected discomfort intensity score was 8 or 9, and 20.4% expected the highest intensity pain, with a score of 10.

Table 2

Acute myocardial infarction symptoms identified by the respondents (n=1804)

Calling an ambulance was the most common treatment-seeking behaviour for scenarios F (chest pain with a score of 8/10) (53.4%) and B (night-time ‘heart discomfort’) (38.9%). Only 25.6% of respondents opted for calling an ambulance in at least three scenarios (table 3). The proportion of respondents who opted for calling an ambulance was greater for scenario B (night-time ‘heart discomfort’) than scenario A (daytime ‘heart discomfort’), for scenario D (indigestion and pain sites including chest) than scenario C (indigestion and pain sites excluding chest) and for Scenario F (chest pain score 8/10) than scenario E (chest pain score 5/10) (matched-pair OR=24.7 (95% CI 14.7 to 41.3), 54.2 (22.4 to 131.3) and 118.3 (53.0 to 264.3), respectively).

Table 3

Percentage* of responses in different scenarios (n=1804)

Seeking non-emergent medical care was the most popular expected treatment-seeking behaviour in scenarios A (daytime ‘heart discomfort’; 26.2%), C (indigestion and pain sites excluding chest; 40.0%), D (indigestion and pain sites including chest; 38.6%) and E (chest pain score 5/10; 35.2%). The most common reason for participants to opt for this action was the perceived shorter time to receiving treatment.

For participants who expected not to seek emergency care immediately and were able to quantify a specific observation time, most of them expected that they would observe for less than 60 min, except for scenarios C (indigestion and pain sites excluding chest) and D (indigestion and pain sites including chest), in which 65.0% and 48.2% of respondents, respectively, expected to observe symptoms for at least 60 min (Table 3).

Multiple logistic regression analysis indicated that age ≥75 years, male sex, lower education level and absence of heart disease were associated with greater likelihood of poor knowledge of AMI symptoms (table 4a). Furthermore, non-married status (ie, single, widowed and divorced), living with others, without hypertension and poor knowledge of chest pain and neck pain as AMI symptoms were associated with greater likelihood of inappropriate expected treatment-seeking behaviour (table 4b).

Table 4a

Predictors of poor knowledge of AMI symptoms by multiple logistic regression (n=1804)

Table 4b

Predictors of inappropriate expected treatment-seeking behaviour at AMI onset scenarios by multiple logistic regression (n=1804)

The respondents received health information from multiple sources. Thirty per cent relied on two sources, and 18.5% had three sources or more. The majority of respondents stated that their sources of health information were from mass media, namely television (44.6%) and newspapers (29.4%). Since they were members of EHCs, 23.0% of them received health information from their EHC (table 5).

Table 5

Sources of health information (n=1804)


This cross-sectional population-based survey explored older Chinese people’s knowledge of AMI symptoms and is unique in its inclusion of scenarios to further investigate their anticipated treatment-seeking behaviour. We found generally poor knowledge of AMI symptom, particularly among those ≥75 years old, male, with a lower education level and free from heart disease. Moreover, inappropriate expected treatment-seeking behaviours were revealed, particularly among those who were not married, lived with others, without hypertension and demonstrated poor knowledge of chest pain and neck pain as AMI symptoms.

Previous studies have demonstrated patients postpone treatment because they perceived their symptoms to be non-cardiac in origin, such as indigestion.25–27 This issue is particularly prominent in cases where the affected person only expected chest pain, but not atypical symptoms.5 Unfortunately, older patients, females and patients with diabetes were more likely to experience atypical symptoms than younger ones.19 28 Furthermore, when atypical symptoms occurred, older people tended to observe them for a more prolonged period before taking further action, as compared with the onset of typical symptoms. Consistent with a previous study conducted in the USA,29 about 40% of our participants did not recognise neck pain as AMI symptoms. If an AMI occurred, these patients would likely arrive at the hospital more than 90 min after onset, which does not meet the American Heart Association recommendation that the patients with AMI should be admitted to the hospital within 90 min of onset in order to start treatment within 120 min.30

In our survey, most participants did not have any experience of AMI, and their perception of AMI onset was based on mass media exposure. Older people may falsely expect the discomfort of AMI onset to be dramatic, as depicted in movies or television dramas. Participants in our study also incorrectly expected a very high intensity of discomfort. In fact, the chest pain intensity was reported to range from 2 to 10 (on a 0–10-point scale) among local patients with AMI,23 indicating the possibility of mild pain. A false expectation of high-intensity discomfort (scoring at least 8/10) might lead to a longer delay in seeking treatment.

We found that men had poorer AMI knowledge than women after controlling for age, education level and history of heart disease. Inconsistent findings have been reported concerning sex differences in AMI knowledge, with some studies proposing that women31 or men32 33 had poorer knowledge, while others found no difference.6 34 35 The reasons for these discrepant findings are unclear, although it is possible that men less actively seek health information compared with women.36 37

Our study found that seeking non-emergency care was a popular treatment-seeking behaviour, which is not consistent with the results from local AMI patient surveys indicating that general practitioner consultation before seeking emergency care was uncommon.4 20 Our survey participants indicated that the waiting time for emergency care was too long; therefore, they would prefer to see their private general practitioners first. This response reflects a failure to understand that under the triage system, cardiac cases are given priority. Also, rather than taking a wait-and-see approach, older people trust their general practitioners for advice. This response could be due to the fact that our participants were members of primary care centres, and thus were more resourceful in seeking medical care. It is also possible that general practitioners are becoming more accessible. Among both Western and Chinese patients with AMI, seeking general practitioner consultation was more common than calling an ambulance (Quan X, under review).25 38

It is understandable that people with only atypical symptoms may be less likely to seek emergency care immediately. However, for those individuals who have significant risk factors for cardiovascular disease, particularly those at high risk of experiencing atypical symptoms,19 should remain alert to the possibility of atypical symptoms and seek the medical care accessible to them at the shortest time. Furthermore, general practitioners should be alerted that individuals with AMI symptoms are likely to contact them instead of activating the emergency care system. General practitioners or primary care centres should also be equipped with an automated external defibrillator in case patients with AMI deteriorate rapidly while waiting at the clinics. Our findings also indicated that when elderly people perceived the need for emergency care, they would prefer to use transportation other than an ambulance. Their reasoning likely reflected a belief that other transportation options were better choices in an emergency. Only six respondents mentioned avoiding an ambulance due to possible abuse of the ambulance system.

Our survey participants revealed lower tendency for calling ambulances as expected immediate action particularly in the absence of chest pain and high pain intensity. These findings are consistent with surveys of patients with AMI conducted worldwide.16 38 39 A Canadian study reported that patients with better knowledge of cardiovascular diseases tended to call an ambulance as the immediate action.7 This finding indicates the need to educate the population before AMI occurrence. Prior studies of Chinese patients with AMI indicated that night-time onset correlated with shorter prehospital delay.18 This is consistent with our findings that people would be more likely to call an ambulance during the night-time, with a reason that the emergency department and ambulance were the only options at night-time.

While this study recruited a large sample of 1804 respondents, it was subject to limitations. First, response rates were not available since it is a quota sample. Also, the older people attending EHCs might be healthier than the general age-matched population and are expected to be more knowledgeable. Nevertheless, we recruited participants from EHCs, because it was cost-effective, and the findings would be suitable for informing public education programmes to be held at EHCs. Thus, the results from this study represent an optimistic scenario for the situation in Hong Kong. Moreover, public education programmes targeted to the knowledge gap revealed by this survey could be administered through EHCs. The expected treatment-seeking behaviours might only capture intent that might be different from the actual actions. Nevertheless, our results were consistent with the actual actions as revealed by surveys with patients with AMI. While the study sample was better educated than the general older population in Hong Kong (ie, the proportion of older people with at least secondary education level was 47.8% in the sample and 39.6% as reported in the 2016 By-Census), the results of our study served as the optimistic scenario. This implied even if the better educated older people were having misconception and knowledge gap about AMI, the general older people in Hong Kong would have even lower knowledge level. Therefore, the results of our study can be applied to the general Hong Kong population particularly in terms of health education purpose.

Our findings have several practical implications. Future public education programmes should emphasise the atypical symptoms of AMI, possible lower intensity level of discomfort at onset and need to call an ambulance during an emergency. Individuals aged 75 years or older and men should be particularly targeted. Additionally, those who may be perceived to have a lower risk of AMI should not be overlooked. It would also be useful to disseminate relevant health information via television or newspaper. Dissemination through EHCs could also be considered. In the long term, increasing public awareness of AMI symptoms and appropriate treatment-seeking behaviour would increase the survival rate of AMI and improve outcomes.


To minimise delay in seeking treatment, public education programmes should target specific demographic groups within elderly populations that are more likely to have poor knowledge of AMI symptoms and expected treatment-seeking behaviour. The finding that television and newspapers are the main sources of health information among older people can also inform the planning of the dissemination mode of health information.

What is already known on this subject

  • Prehospital delay of acute myocardial infarction (AMI) is common globally.

  • Treatment-seeking behaviours have been shown to be associated with knowledge of AMI symptoms and specific cultural beliefs.

What this study adds

  • Older Chinese people have poor knowledge of AMI symptoms, particularly the atypical symptoms.

  • There is inappropriate expected treatment-seeking behaviour among the older Chinese people, particularly in scenarios with no chest pain or low pain intensity.

  • To minimise delays in seeking treatment, future health education should focus on increasing the public knowledge of AMI symptoms and the need to call an ambulance during an emergency. 


We would like to thank the nursing officers and colleagues at the Elderly Health Centres for their kind assistance in data collection. We would also like to thank the research team and the older people who participated in this study. Without all these supports, this study would not be completed.



  • Contributors All authors were involved in the planning of the study, implementation of study, writing and approving the manuscript. PHC, WTL and MHY were involved in data analyses as well. PHC, GM, SYL, JW, AYML, CMC and JZ were involved in interpretation of results as well.

  • Funding This study was supported by the Early Career Scheme (no. 27105415) of the University Grants Committee, Hong Kong.

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval Ethical approvals were obtained from Institutional Review Boards of the University of Hong Kong/Hospital Authority Hong Kong West Cluster and the Department of Health of Hong Kong Special Administrative Region.

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