The impact of health literacy-sensitive design and heart age in a cardiovascular disease prevention decision aid: randomised controlled trial and end user testing
Summary (4 min read)
METHODS
- The literacy-sensitive design included simple language, supporting images and white space to improve readability and understandability 28 , and a novel action plan format their team developed that has been shown to reduce unhealthy lifestyle behaviours amongst people with low health literacy29.
- The authors added options for physical activity and smoking to the existing tool to reduce unhealthy snacking, drawing on previous literature on effective if-then plans in these areas39,40. .
- CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint.
- The copyright holder for thisthis version posted September 22, 2021.
Design
- The randomised trial was based on a 3×2 factorial design to test the effect of literacy-sensitive design (literacy-sensitive DA, standard DA, or control: Heart Foundation patient information) and risk format (explaining CVD risk only (as a percentage risk), or CVD risk percentage + heart age) on psychological and behavioural outcomes.
- The snacking action plan was previously developed by their team 29, and the exercise and smoking plans were in the same style using research in those areas 39,40 .
Recruitment
- A national sample was recruited through Qualtrics, an online social research agency, with stratified sampling based on gender/age groups (5 year age groups from 45-74 years).
- Participants completed a CVD risk assessment based on Australian guidelines and the New Zealand approach to calculating heart age1,38.
- If blood pressure or cholesterol were not known, the average by age/gender based on non-diabetic participants in the AusDiab cohort was used (accessed via JD), and all participants were advised to see a GP for a more accurate risk assessment.
- Those with established CVD or taking CVD prevention medication were excluded.
- Duplicate IP addresses were replaced, and stratified sampling was relaxed with additional quality checks added if hard to reach groups did not reach quota after 2 weeks.
Measures
- Established measures were used for the primary outcome of behavioural intentions (validated Theory of Planned Behaviour scale applied to smoking, diet, exercise and GP visit) 45–47 .
- CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint.
- The copyright holder for thisthis version posted September 22, 2021.
A priori sample size calculations determined that 85 participants per randomised group (total n=510)
- Would yield 90% power to detect a moderate effect size of d=0.5 (standardised difference) in the primary outcome of intention to change lifestyle or any of the secondary outcomes, assuming a twosided alpha of 0.05.
- The authors aimed to recruit 20% more cases to account for potential missing values, totalling 600 participants (100 per group) at follow-up.
- This sample was inflated for recruitment to.
850 to account for potential attrition of up to 30% between the intervention and follow-up.
- Continuous outcome variables were modelled using linear regression.
- Dichotomous outcomes were analysed using modified Poisson regression (using a log-link function with robust error variances).
- All regression models included decision aid group (literacy-sensitive decision aid; standard decision aid; or basic Heart Foundation patient information) and risk format (CVD risk only, or CVD risk + heart age) as categorical variables and controlled for health literacy adequacy (categorical based on the Newest Vital Sign (NVS) measure51,52: low; moderate; adequate) and absolute risk .
- CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint.
- The copyright holder for thisthis version posted September 22, 2021.
3. Adding heart age to absolute risk will be more effective than absolute risk alone.
- Iterative end user testing with varying health literacy levels Participants in the trial were invited to opt-in to a “think aloud” interview to provide further end user testing and feedback for the literacy-sensitive version of the intervention, also known as Stage 3.
- Participants went through the risk calculator in full, while saying out loud everything they were thinking, for example any areas of confusion.
- Transcripts were thematically coded and discussed after each set of 4-5 interviews, and improvements were made to the intervention before the next set of interviews.
- The copyright holder for thisthis version posted September 22, 2021.
RESULTS
- The authors used the question format and style of the current national heart age calculator as the basis for the risk factor questions in all groups, and also based the heart age presentation on that tool.
- The CVD risk results and decision aid were presented based on: 1) their existing GP decision aid tool 37 (standard decision aid group); 2) a simplified version of the standard decision aid with supporting images ; and 3) the current risk calculator from the National Vascular Disease Prevention Alliance41.
- The copyright holder for thisthis version posted September 22, 2021.
- The authors recruited 859 participants for the intervention (including the 100 in the soft launch), with a target of 600 at 4 week follow-up, for which they recruited 596.
- CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint.
Demographics
- . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint.
- The copyright holder for thisthis version posted September 22, 2021.
- For Hypothesis 2, there was no difference between the standard and literacy-sensitive decision aids for any outcome (Table 6).
- There were significant interactions between decision aid and health literacy for intention to talk to a doctor about medication (p=0.019) and emotional response (positive p=0.010; negative p=0.006).
At follow-up after 4 weeks, there were no significant differences between the control and decision
- There were significant differences between decision aid groups by health literacy levels for self-reported calls to the Heart Foundation helpline (p<0.001) and verbatim knowledge of CVD percentage risk at follow up (p<0.001).
- CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint.
4 week differences between heart age groups
- At 4 week follow-up there were no significant differences between the heart age groups in terms of lifestyle behaviour change, seeing a doctor for a Heart Health Check or gist knowledge of risk level (Table 7).
- The copyright holder for thisthis version posted September 22, 2021.
- From this pool, 20 were selected to represent a range of ages, genders, risk levels, and health literacy.
DISCUSSION
- The authors used both a mixed methods development and evaluation process to produce a CVD decision aid that is effective for improving verbatim and gist knowledge of CVD risk and fruit consumption after 4 weeks.
- The results show literacy-sensitive decision aids can support people with lower health literacy to make informed decisions, while still being suitable for the general population.
- CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint.
- The copyright holder for thisthis version posted September 22, 2021.
- Analyses of people who received an older heart age result suggests it may be useful as a marketing tool to get attention and initiate behaviour change, but knowledge of heart age did not translate to knowledge of risk.
Strengths and limitations
- The authors had sufficient follow-up to run the study per protocol despite COVID-19 disruptions, and observed no difference in dropouts for key variables.
- A limitation is that the online panel sample may not be representative of the general population, and may better reflect users of online heart age tools than patients presenting to primary care for CVD risk assessment.
- Different countries use also different CVD risk models/heart age algorithms, which may affect the results given the differences the authors observed in the older heart age sample.
- Finally, the authors used validated outcomes where possible, but behaviour change was self-reported.
Conclusion
- This study shows the value of combining health literacy-sensitive design with best practice risk communication and behaviour change tools.
- This approach improved knowledge of CVD risk and heart age, and behaviour, in a sample with varying health literacy levels.
- Further research should investigate implementation pathways for integrating such consumer tools with clinical practice, and distinguish between older and younger heart age results.
- CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.(which was not certified by peer review)preprint.
- The copyright holder for thisthis version posted September 22, 2021.
Did you find this useful? Give us your feedback
References
8,889 citations
5,042 citations
3,748 citations
3,457 citations
1,941 citations
Related Papers (5)
Frequently Asked Questions (10)
Q2. What are the future works in "Title: the impact of health literacy-sensitive design and heart age in a cardiovascular disease prevention decision aid: randomised controlled trial and end user testing" ?
Further research should investigate implementation pathways for integrating such consumer tools with clinical practice, and distinguish between older and younger heart age results.
Q3. How many people are taking medication to prevent a heart attack or stroke?
up to 75% of high risk patients are not receiving recommended medication to prevent death and disability from CVD, while 25% of low risk patients are taking medication they are very unlikely to benefit from7.
Q4. How many people in Australia, Europe and the US have low health literacy?
Low health literacy is common in many countries, with estimates ranging from 36 – 60% of the population in Australia, Europe and the US15–17.
Q5. What are the main strategies to address the issue of communicating CVD risk to people with lower health?
There are several evidence-based strategies to address the issue of communicating CVD risk to people with lower health literacy:1) Use literacy-sensitive design to improve the readability of health information and reduce the cognitive load of action plans for behaviour change27–29.2)
Q6. What is the definition of prevention of cardiovascular disease?
Prevention of cardiovascular disease (CVD) includes lifestyle interventions and medication for those at highest risk who are most likely to benefit.
Q7. What are the main reasons for the lack of a systematic review of online CVD risk assessment?
their systematic review of 73 online CVD risk assessment tools available to consumers found they were not suitable for people with lower health literacy: their readability level was too high; they frequently used unexplained medical terms; few used best practice risk communication formats such as frequencies in icon arrays; and they rated poorly on actionability (i.e. clarity in instructions of.
Q8. What was the main strength of this study?
; https://doi.org/10.1101/2021.09.20.21263868doi: medRxiv preprint31A major strength of this study is that the authors were able to recruit a large, diverse sample in terms of health literacy and risk results.
Q9. What is the current definition of a shared decision making approach?
Recent guideline changes have led to calls for a shared decision making approach, to ensure medication prescribing for blood pressure and cholesterol is more in line with patient values12–14.
Q10. What is the effect of a decision aid on the health of people with low health literacy?
As this sample was predominantly low risk, the authors would not want a decision aid to lead to greater actual medication uptake in this group, but speaking with a doctor about risk and how to reduce it may be a positive outcome in line with guidelines to assess risk in this age group1.