Study design and setting

This was a cross sectional survey using mall intercept interviews [19]. This method was chosen as it allowed the researcher recruit community dwelling adults from varied socioeconomic backgrounds.

The study was carried out in a hypermarket located at an urban area, Cheras district of Kuala Lumpur, Malaysia. This area was surrounded by housing estates with private and public primary care clinics. This hypermarket was selected purposively for the wide range of population that attended the facility.

Study population

All Malaysian aged 30 years and above who attended the hypermarket during the study period were invited to participate. The age group selected is accordance with the Malaysian Ministry of Health’s recommendation for screening of CVD risk factors [11]. Individuals with known history of stroke or coronary heart disease and those who could not understand the Malay language (Malaysia national language) were excluded from the study.

Survey instrument

A questionnaire was developed in Malay language as this is a common language used by all ethnic groups in the country. (Refer S1 Appendix).

The questionnaire consisted of four sections: Section I was on participants’ socio-demographic information, history of CVD risk factors such as diabetes, hypertension, high cholesterol, overweight or obesity, smoking and family history of heart attack or stroke and their awareness of stroke and heart attack.

Section II was on participants’ previous experience of health checks.

Section III was on the determinants of CVD health check questionnaire [20]. This questionnaire was developed based on the findings from the earlier qualitative study and grounded to local and cultural context [16,20]. It was validated with good content and structural validity [20]. The item-content validity index ranged from 0.83 to 1.00. The factors loading of the items were more than 0.40 with good internal consistency (Cronbach’s alpha range: 0.66–0.85) [20]. The nine determinants measured are: Individuals’ belief that the course of CVD can be changed for the better, perceptions of self being at risk of CVD, perception of benefits of CVD health checks, perception of drawbacks of CVD health checks, preferred method for CVD prevention, individuals’ readiness to know the results of CVD health checks, individuals’ readiness to handle the outcomes following CVD health checks, external barriers and influence by significant others.

A Likert scale of 1 to 5 was used to indicate participant’s level of agreement with each item; score of 1 indicated “strongly disagree” and score 5 “strongly agree”.

Section IV was on participant’s intention to undergo CVD health checks. Participants were asked about their likeliness to undergo health checks in a specified timeline (within 3 months, 6 months or 1 year). A Likert scale (from score 1 “very unlikely” to 5 “very likely”) was used to denote the intention to undergo CVD health checks.

Data collection

Recruitment of participants was carried out at the exit area of the hypermarket, where there was no health store, to minimize possible bias of recruiting shoppers who were likely to be aware of CVD risks if they were to visit a health store.

During data collection, a bunting was set up to advertise and attract shoppers. Participants were recruited in a convenient manner by two possible methods: The shoppers were either approached and invited by the researcher or they self-volunteered to participate.

Participants were briefed about the objective of the study and screened for eligibility. Those who agreed to participate were given the participants information sheet and informed consent were obtained. To minimize social desirability bias, participants were informed that there were no right, or wrong answers and a truthful answer would be the most appropriate.

Participants were encouraged to self-administer the questionnaire. In circumstances where participants had problems reading the questionnaire, for example those who had visual problems or had literacy issues, the researcher would provide assistance by reading aloud the questionnaire to the participants.

Data analysis

All statistical procedures were performed using IBM SPSS Statistics software version 22.

In section III of the questionnaire, a Likert scale of score 1 to 5 was used to indicate participant’s level of agreement with each item. These were treated as continuous data and mean scores of each determinant was computed from its item’s score and used for analysis.

For the determinants of “individuals’ beliefs that the course of CVD can be changed for the better”, “perception of self being at risk of CVD” and “perception of benefits of CVD health checks”, a higher mean score denotes greater perception of the relevance of health checks. For the determinants of “perceived drawbacks of CVD health checks” and “preferred method for CVD prevention (healthy practice vs. medical measures)”, a higher mean score denotes lesser perception of the relevance of CVD health checks.

Readiness of participants to face the outcomes of CVD health checks included two determinants: individuals’ readiness to know the results of CVD health checks and individuals’ readiness to handle the outcomes following these checks. A higher mean score of these determinants denotes higher readiness in facing the outcomes of CVD health checks.

The external barriers towards CVD health checks were assessed in terms of time, cost, transportation and distance from the health check facilities. The higher the mean score of this determinant denotes a higher degree of barriers towards CVD health checks.

Influence from significant others included influence from doctors, family members, friends, employers and people around. The higher the mean score of this determinant denotes a higher degree of influence by others for the participants to undergo CVD health checks.

In section IV, participants were asked about their likeliness to undergo health checks in a specified timeline (within 3 months, 6 months or 1 year). Two outcome variables (dependent variables) were derived from this question: the degree of likeliness to undergo health checks and the likely timeline to undergo health checks. A higher intention of CVD health checks was reflected by a higher degree of likeliness to undergo CVD health checks. A higher intention of CVD health checks might also be reflected by the likely time they would attend the CVD health checks. The earlier time a participant decided to attend for health checks would indicate a higher degree of intention. These outcome data were ordinal data.

The degree of likeliness of a participant to undergo a health check was indicated by the highest score the individual answered, regardless of timeline. It was ranked from very unlikely (score of 1) to very likely (score of 5) (refer S1 Table). For ordinal regression analysis, the scores of 1, 2 or 3 were combined into one category (not sure, unlikely or very unlikely) due to the small numbers of these scores (refer S1 Table).

Similarly, the likely timeline for a participant to undergo health checks was indicated by the earliest time the participant would likely or very likely undergo a health check (score of 4 or 5 for the indicated time period) (refer S1 Table). For example, in section IV, if a participant chose a score of 3 within 3 months, a score of 4 within 6 months and a score of 5 within 1 year, the earliest time they would likely undergo a health check was within 6 months. With this, they would be classified into the category of likely to attend within 6 months. Four categories were derived: not sure or not likely to attend, likely to attend within one year, likely to attend within 6 months and likely to attend within 3 months.

Regression analysis was used to determine the relationship between the nine determinants and the two outcome variables (individuals’ degree of likeliness and the likely timeline to undergo health check). Two models (model 1 and 2), included all cases and had the two outcome variables analyzed using ordinal regression. Another two models (model 3 and 4), had 16 cases removed from the initial cases, to take into account of a possible Hawthorne effect. These 16 cases were participants who did not have any past health check experience but indicated an intention to undergo CVD health checks. Hawthorne effect meant the participant’s indication of an intention to undergo health checks might not be true, but a consequence of reactivity towards a socially desirable effect.

Two sets of regression analysis were performed, the first set without and the second set adjusted with the sociodemographic data (age, gender, ethnicity, working status, history of diabetes, high blood pressure, high cholesterol, overweight or obesity and past history of regular health checks).

The outcome variables were measured in an ordinal scale. Simultaneous ordinal regression was used in the analysis, which is useful in determining relative influence of each determinant on the outcome [21]. The complimentary log-log link function was used. Model fitness was assessed using the Pearson and Deviance goodness-of-fit measures [22]. The goodness-of-fit was acceptable. The test of Parallel lines was used to assess the proportional odds assumption. The assumption of test of Parallel lines was met for all the models (p-value >0.05) except model 4 with control of sociodemographic data (p = 0.013) (Refer S2 Table).

Multicollinearity of the determinant variables was examined using correlation coefficient of the determinant variables before regression analysis was performed. There were no multicollinearity detected and the correlation coefficient of the determinant variables was <0.85 [23] (Refer S3 Table).

Ethical issues

The Medical Ethics Committee, University of Malaya Medical Centre (20145–274) approved this study.

Factors associated with the intention for CVD health checks

Table 3 shows the determinants of individuals’ intention for CVD health checks without adjusted for the sociodemographic data. In the analysis of all 413 participants, people’s perception of the benefits of CVD health checks, and external barriers were the significant determinants of both individuals’ degree of likeliness to undergo CVD health checks (model 1) and individuals’ likely timeline to undergo CVD health checks (model 2). Perceptions regarding drawbacks of these checks was also a significant determinant of model 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Determinants of individuals’ intention to undergo health checks without adjusted for sociodemographic data.

https://doi.org/10.1371/journal.pone.0201931.t003

After removing the 16 cases for possible Hawthorne effect, four significant determinants were identified (model 3 and 4). The three determinants: individuals’ perception of the benefits of CVD health checks, individuals’ readiness to handle outcomes following health checks and external barriers were significant determinants of both individuals’ degree of likeliness to undergo CVD health checks (model 3) and individuals’ likely timeline to undergo CVD health checks (model 4). Perceptions regarding drawbacks of these checks remained a significant determinant of individuals’ degree of likeliness to undergo CVD health checks (model 3).

There were two possible relationships based on the positive or negative value of the β of the determinants. A positive value of the β indicates a positive relationship and a negative value indicates a negative relationship. Individuals’ perception of the benefits of CVD health checks had a positive relationship with their degree of likeliness and likely timeline to undergo CVD health checks. A higher degree of perceived health check benefits and readiness to handle outcomes were significantly associated with a higher degree of likeliness of the individuals to undergo CVD health checks and to undergo these checks within a shorter timeline.

The perceptions regarding drawbacks of these checks and external barriers had a negative relationship with the individuals’ degree of likeliness to undergo CVD health checks. The lower the degree of perceived health check drawbacks and external barriers was significantly associated with a higher degree of likeliness of the individuals to undergo CVD health checks and a higher likelihood of the individuals to undergo CVD health checks within a shorter timeline.

The 4 models seemed to explain 22.7%, 16.3%, 24.8% and 17.8% (corresponding to the pseudo-R² of 0.227, 0.163, 0.248 and 0.178, respectively) of the observed variance in the individuals’ intention to undergo CVD health checks (Refer S2 Table). According to Cohen et al., the R² value of 0.13 and 0.26 are proposed as medium and large effect sizes for the population [24]. The models in this study were useful in explaining the individuals’ intention to undergo CVD health checks.

Table 4 shows the determinants of individuals’ intention for CVD health checks adjusted for the sociodemographic data. After controlling for sociodemographic characteristic (age, gender, ethnicity, working status, history of diabetes, high blood pressure, high cholesterol, overweight or obesity and past history of regular health checks) and removing possible Hawthorn effect, individuals’ perceptions regarding benefits and drawbacks of health checks, individuals’ readiness to handle outcomes following health checks and external barriers were significant determinants of individuals’ degree of likeliness to undergo CVD health checks (model 3). For individuals’ likely timeline to undergo CVD health checks (model 4), the significant determinants were individuals’ readiness to handle the outcomes following health checks and external barriers. These two models (model 3 & model 4) explained 45.3% and 30.1% of individuals’ degree of likeliness to undergo CVD health checks and the individuals’ likely timeline to undergo CVD health checks respectively (Refer S2 Table).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Determinants of individuals’ intention to undergo health checks adjusted for sociodemographic data^{^}.

https://doi.org/10.1371/journal.pone.0201931.t004

Interpretation of findings and comparison to previous findings

A high proportion of participants indicated positive intention to undergo CVD health checks. This finding is consistent with another study, in which most people showed a positive intention towards CVD risk factor screening [25]. In our study, 45% of participants reported they were “likely” and 39% reported “very likely” to undergo CVD health checks. These “likely” and “very likely” responses could correspond to a moderate and high degree of intention, respectively. Locally, the National Health Morbidity Survey in 2011 reported about 38% of respondents had undergone health checks such as screening of blood pressure and blood sugar over the past 12 months [14]; this figure was similar to the high intention group in the present study. This suggests that people with high intention to undergo CVD health checks might realise their intention into action. In contrast, people with moderate intention might not realise their intention into action. This was supported by literature that showed attenders for screening were more likely to have definite intention than non-attenders [26].

A sensitivity analysis was conducted to test possible Hawthorne effect by removing 16 cases of participants who did not have any past health check experience but indicated intention to undergo CVD health checks. This analysis had resulted in an additional significant determinant (readiness to handle outcomes following health checks) to the initial analysis. As it is likely that the initial models were influenced by the Hawthorne effect, the added factor is likely to be a true determinant.

We found significant determinants for intention to undergo CVD health checks were positive perception of benefits and negative perception of drawbacks of CVD health checks, negative perception of external barriers and readiness to handle outcomes following CVD health checks. These findings were in line with some of the results found in previous research. The findings of positive perception of benefits of health checks were reported as significant predictors for intention to health checks in surveys based on the health belief model [27,28], and in a Dutch Health Care Consumer Panel survey, its result showed that the belief in the check that could contribute to a higher chance for aging healthily, was significantly associated with the willingness to participate in health checks [29]. Unpleasant screening procedures such as pain from finger stick tests have been reported as a barrier for health check participation [30], which was reflected by the perceived drawbacks of health checks in this study. It is important to emphasize the benefits of health checks and address the drawbacks of these checks when disseminating health check educational material to the public.

External barriers such as time constraints have been reported as significant factors that deterred public from attending health checks [27,29]. However, in the present study, participants were more concerned about cost rather than time, with only 1% of them indicating time was a barrier, compared to 20% for cost. Cost was raised as a concern in the utilization of private practice health care services in the Malaysia National Health Morbidity Survey [31]. A systematic review found that providing financial incentives for screening was an effective intervention to increase uptake of CVD health checks [32]. This could be a potentially useful measure to improve CVD health checks locally.

Previous research had shown that fear of health check results, and perception of health checks as being unnecessary worries, had reduced people’s intention to health check participation [26–29]. These factors are conceptually similar to readiness to know the result in the present study, where concerns could reflect a lack of readiness. We found individuals’ readiness to know the results was not a determinant of intention to undergo health checks, but the individuals’ readiness to handle the outcomes such as preparedness to take medication, preparedness to adjust lifestyle and preparedness to bear the cost of subsequent treatment following abnormal health checks was a significant determinant. Readiness to handle outcomes provides further understanding about the factors that people consider during the decision-making process, which has not been highlighted in the medical literature. Our findings show that there is a need to address individuals’ readiness to handle outcomes when developing interventions to improve CVD health checks.

There were inconsistent findings on the association of individuals’ perception of susceptibility and seriousness of disease with the intention or participation in health checks. We found the perception of self being at risk of CVD was not a significant determinant for intention to undergo CVD health checks. The mean score of this concept was 3.15, which suggested the perceived susceptibility of the disease was not high. However more than two-thirds of participants indicated they have moderate or high intention to undergo CVD health checks. It was uncertain whether this was a consequence of bias due to participants volunteering to take part in the study or differing cultural context. Future studies would be needed to verify this using probability sampling. Some studies reported health check attenders had a higher level of perceived susceptibility to disease [33], but many did not find this to be so [27,28,30,34,35]. Intervention studies that sent health risk appraisal questionnaire to participants showed that most people at risk who received the questionnaire did not turn up for CVD risk factor screening [36,37]. On the other hand, a Cochrane systematic review reported that there was a small effect with low quality evidence that suggested personalized risk communication increases uptake of screening tests, which was mainly for mammography and colorectal cancer screening [38].

Recommendations/Implications

To improve uptake of CVD health checks, the four significant determinants (i.e. perception of benefits and drawbacks of CVD health checks, perception of external barriers and readiness to handle outcomes following CVD health checks) are important issues to address when developing strategies. For effective health communication, health materials should contain relevant information for the targeted group [39]. Hence health messages to the public must emphasize benefits of CVD health checks and address the drawbacks of these checks. Information that are important to the public such as side effects of medication, health check procedures or lifestyle management need to be included to reduce misconceptions and enhance readiness to face outcomes of health checks.

A systematic review had identified effective interventions to decrease external barriers such as financial incentives to individuals [32]. In this study, external barriers were shown to affect participation in health checks. There has been efforts to improve the uptake of CVD screening such as the provision of free vouchers to attend health screening by the Social Security Organisation (SOCSO) to its active members aged 40 years and above, and community outreach health check programmes [40,41] that help address some external barriers such as cost and accessibility. Members of the public who do not have access to these supports should be made aware regarding the availability of CVD health checks in government clinics, which is provided free of charge. Flexible government clinic appointment times, such as having the checks after office hours or at weekends, would probably increase public participation. However, there remains challenges for implementation such as availability of resources and healthcare personnel.

Strength and limitations

This study was conducted in the community and it allowed the researcher to recruit the public from various backgrounds. However, due to the convenience sampling of survey population, there is a lack of representativeness of the study population to the general population. Future studies could use multi-stage sampling or random sampling for representative population and cover a wider population, including those in rural areas.

This study only included those who could speak or read Malay; this would probably exclude a number of people especially those who were older or of a poorer sociodemographic class. However, there was only one Indian participant who was excluded due to language issue. Therefore, the results of this study may be applicable to populations with similar sociodemographic characteristics.