Jump to comment:

• Considerations for future age-based risk reviews
  Carissa Bonner
  Published on: 24 March 2020

• Published on: 24 March 2020

  Considerations for future age-based risk reviews

  • Carissa Bonner, Behavioural Scientist The University of Sydney

  I read this rapid review with interest and agree this topic has reached a point at which a full systematic review could be useful. I have a few suggestions to the authors and/or future reviewers in this area, as there are several nuances in the studies presented that have not been acknowledged.

  Future reviews need to consider what primary outcome the study was powered for, and what the comparator is. For example, Bonner et al. looked at physical activity and diet changes as well as smoking in a combined measure and found no statistical difference between absolute risk and heart age when presented in the same format, but the smoking result is isolated and reported as “clinically significant” in the review. In contrast, Lopez-Gonzalez et al. compared an interactive online format for heart age to usual care which generally involves a verbal description of absolute risk by the doctor, so it is difficult to determine whether heart age or a more engaging presentation format produced the results.

  Psychological outcomes should also be considered. Modest gains in behaviour change may not be outweighed by reduced accuracy of risk perception, negative affect or reduced credibility of the assessment as per Bonner et al. While these age-based formats appear to elicit a stronger emotional response, this may not be sufficient to produce greater behaviour change or even intentions without additional support (e.g. Soureti et al. and Witteman et al.). It is unclear from the r...

  Show More

  I read this rapid review with interest and agree this topic has reached a point at which a full systematic review could be useful. I have a few suggestions to the authors and/or future reviewers in this area, as there are several nuances in the studies presented that have not been acknowledged.

  Future reviews need to consider what primary outcome the study was powered for, and what the comparator is. For example, Bonner et al. looked at physical activity and diet changes as well as smoking in a combined measure and found no statistical difference between absolute risk and heart age when presented in the same format, but the smoking result is isolated and reported as “clinically significant” in the review. In contrast, Lopez-Gonzalez et al. compared an interactive online format for heart age to usual care which generally involves a verbal description of absolute risk by the doctor, so it is difficult to determine whether heart age or a more engaging presentation format produced the results.

  Psychological outcomes should also be considered. Modest gains in behaviour change may not be outweighed by reduced accuracy of risk perception, negative affect or reduced credibility of the assessment as per Bonner et al. While these age-based formats appear to elicit a stronger emotional response, this may not be sufficient to produce greater behaviour change or even intentions without additional support (e.g. Soureti et al. and Witteman et al.). It is unclear from the review whether any behavioural effects can be attributed to age formats directly or whether the interventions included additional behaviour change techniques such as action plans or referrals to lifestyle programs.

  In addition, these age-based models have generally not been validated as predictive models for long-term health outcomes, which is not clear from the discussion of validity in this review. For example, heart age assessment tools are often based on converting the absolute risk from validated CVD risk models but this is not the same thing as being validated in itself.

  Finally, the variability between the age produced by different models is important to consider, as discussed in other review papers on vascular/heart age. The same person may receive a younger or older age result depending on the underlying model, which is likely to affect the behavioural outcome.

  I agree with the authors that additional high quality trials are needed to better understand this issue, as the growing use of age-based risk tools appears to be outpacing the limited evidence for their efficacy.

  Show Less

  Conflict of Interest:
  I have published several papers on heart age including one of the trials included in this review, and am currently involved in projects to develop and evaluate heart age assessment tools.

  • Back to top