Hi David. I’m with James on this. Arbitrary thresholds, and qualitatively labelling risk as low, medium, high etc are constructs I have fought against for several decades in guideline groups, but without success. The response has always been
that busy doctors and their unprepared patients need clear decision points.
In our most recent (New Zealand) guidelines, we have two thresholds, but at least they have some clinical justification.
We have a 5% 5-year CVD risk threshold, which was identified as a level, below which, we could not confidently state that the benefits of drug treatment would consistently outweigh possible small harms, given the very small absolute benefits. So our recommendation
was that we would not routinely recommend drug treatment below this level.
Our other threshold was a 15% 5-year CVD risk, which was a risk we have demonstrated similar to most people in NZ with prior CVD. As we routinely recommend lipid lowering and blood pressure lowering drugs for most people with prior CVD, for consistency,
we decided to recommend drug treatment to most people without prior CVD who have a 5-year CVD risk over 15% in 5 years.
Between 5% and 15% we don’t recommended that drug treatment should be routinely offered, but that clinicians should begin by discussing the absolute benefits of treatment, in terms of events prevented, and also describe possible harms. If the patient decides
they want to take medications, we recommend that this decision should be supported.
We do also encourage a discussion about the benefits and risks for patients with a risk less than 5% and over 15% in 5 years, but the supporting advice about drug treatment is different.
Recently we published risk prediction equations for patients with prior ACS, most of whom have a 5-year risk over 15% in 5 years (Poppe KK, et al. Heart 2019;0:1–6. doi:10.1136/heartjnl-2019-315809), and which we believe should inform the intensity of
treatment. This supports the argument not to use thresholds, given almost all these patients have risks above typical thresholds, yet the information can still inform important clinical decisions.
With regard to your second question about meaningful differences between scores, it depends on your question. A CVD risk score is designed to estimate risk for a specific patient, so the context will be key to answering your question.
Cheers Rod
Rod Jackson
CVD epidemiologist
University of Auckland, New Zealand
David:
I’ve tried to get guidelines to stop doing this for years with no luck. They seem to desperately need treatment thresholds.
In our simplified lipid guideline we used thresholds but they were thresholds for discussion NOT treatment.
The way it should be done IMHO is your risk is roughly x%, your benefit would be roughy Y% - here are the harms (side effects/cost etc). If you want treatment I’ll support you, if you don’t I’ll support you.
In general CPGs for HTN, diabetes and cholesterol appear to have no desire to have shared decision making be front and centre of the decision making process.
I gave a talk on that here
More than happy to answer any questions you may have.
James
Disappointed as many of you may be, I have a query unrelated to coronavirus/COVID-19.
I’m trying to determine two things in relation to CVD risk scores (obtained by any of the various methods):
-
Why do we lump everyone over 20-30% risk as the same = high or very high risk? Why is risk not assessed on a continuous scale from 0% to 100%
-
If comparing mean/median absolute risk scores in two groups, what would folk consider as a meaningful (minimal) difference in risk score between groups?
Many thanks, David.
To unsubscribe from the EVIDENCE-BASED-HEALTH list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=EVIDENCE-BASED-HEALTH&A=1
To unsubscribe from the EVIDENCE-BASED-HEALTH list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=EVIDENCE-BASED-HEALTH&A=1