QRisk3 on line form

[LionChild]

Has anyone else filled in the online questionnaire called QRISK3 which is meant to be able to predict your chances of a cardiovascular event in the next 10 years, and tell you what your Heart age is in that sense?

I notice that when I unclick Type 1 diabetes, the figure becomes according to my actual age, but other wise it puts me at a very high risk as if I am an 83 year old in terms of my heart health. I am actually 72.

The form does not allow you to say what exercise you do; what your diet is; or even what your A1C tends to be? It doesn't even ask you if you are on STATINS. Surely if Statins are of any use at all, they should be figured into this assessment of your risk level?

All of these variables are surely relevant?

What do people think about this, and about the questionnaire assessment's accuracy?

 

[Rachox]

I agree with you, so many assumptions. I think your HbA1c should be included in the calculation and yes statin use (are they as good as they reckon if they are not included?). If I fill it in truthfully it makes me 12 years older than I am too. I don’t hold much store in it to be honest.

 

[LionChild]

I am glad you say that because it was worrying me!!

 

[MrsA2]

Dr Malcolm Kendrick critises it in depth in his book "the Clot Thickens". He's looking from a heart disease point of view.
Seems its a blunt tool

 

[HSSS]

When something uses your postcode to decide how healthy you are as an individual I’m highly sceptical. It may work as a population tool but it’s way too blunt for a single person imo. To ignore hba1c, overall control, method of control (relevant to type 2’s diet or medication etc) seems crazy too.

To some extent statin use is considered as presumably its lowering your ldl thus the scores you enter have been mediated by the drug already. Whether it’s doing anything else (like to your hba1c, your muscles or brain or actually effecting your CVD risk/outcomes beyond ldl number is another conversation entirely)

 

[LionChild]

When something uses your postcode to decide how healthy you are as an individual I’m highly sceptical. It may work as a population tool but it’s way too blunt for a single person imo. To ignore hba1c, overall control, method of control (relevant to type 2’s diet or medication etc) seems crazy too.

To some extent statin use is considered as presumably its lowering your ldl thus the scores you enter have been mediated by the drug already. Whether it’s doing anything else (like to your hba1c, your muscles or brain or actually effecting your CVD risk/outcomes beyond ldl number is another conversation entirely)

Thank you for your response. I have read, and have been told by a doctor (not my own official GP/consultant) that, in fact, STATINS which are meant to improve CV risk, only address LDL lipids which are of the 'light and fluffy' type whereas the type of lipids that are responsible for the heightened risk of CV events are the 'small dense' LDL. If this is the case, this would, of course, indicate that STATINS are totally missing the mark.....?

Also, another angle on this is that it is probable that even if statins were effective at all, the level of amelioration of risk as a result of taking STATINS is minimal ---and nothing like the 30% often quoted - the reality being that statins reduce risk by less than 2%....

Further...... that STATINS can have a bad impact on cognitive ability. However, they can also help relax muscles - which may be a helpful thing....

I would emphasize that of course I am no expert, just reporting what I have read, and have been told.

 

[HSSS]

Thank you for your response. I have read, and have been told by a doctor (not my own official GP/consultant) that, in fact, STATINS which are meant to improve CV risk, only address LDL lipids which are of the 'light and fluffy' type whereas the type of lipids that are responsible for the heightened risk of CV events are the 'small dense' LDL. If this is the case, this would, of course, indicate that STATINS are totally missing the mark.....?

Also, another angle on this is that it is probable that even if statins were effective at all, the level of amelioration of risk as a result of taking STATINS is minimal ---and nothing like the 30% often quoted - the reality being that statins reduce risk by less than 2%....

Further...... that STATINS can have a bad impact on cognitive ability. However, they can also help relax muscles - which may be a helpful thing....

I would emphasize that of course I am no expert, just reporting what I have read, and have been told.

I’m not disagreeing at all.

But, the discussion of the the value, or otherwise, of statins is a separate discussion. I phrased my reply the way I did because there are many threads discussing them rather than derail this one about qrisk.

 

[LionChild]

I’m not disagreeing at all.

But, the discussion of the the value, or otherwise, of statins is a separate discussion. I phrased my reply the way I did because there are many threads discussing them rather than derail this one about qrisk.

Sorry HSSS, I did not perceive a problem.

 

[HSSS]

Sorry HSSS, I did not perceive a problem.

No problem. Just aware how easily threads mentioning statins get off track of the topic being discussed.

 

[Lupf]

The QRisk calculator is using your age, height, weight, blood pressure, total cholesterol/HDL and Diabetes (none, T1, T2) plus a few other things, such as post code. Thus it is a very crude calculator, i.e. a diagnosis of T2 adds about +7 to your score, independent of your Hb1Ac and how you control your diabetes. I would not make any decisions based on my QRisk score alone.

 

[LionChild]

When something uses your postcode to decide how healthy you are as an individual I’m highly sceptical. It may work as a population tool but it’s way too blunt for a single person imo. To ignore hba1c, overall control, method of control (relevant to type 2’s diet or medication etc) seems crazy too.

To some extent statin use is considered as presumably its lowering your ldl thus the scores you enter have been mediated by the drug already. Whether it’s doing anything else (like to your hba1c, your muscles or brain or actually effecting your CVD risk/outcomes beyond ldl number is another conversation entirely)

Just wanted to clarify something here HSSS. What makes you think that already taking the statins is implicit in the QRisk3 score? I have not read that anywhere in their explanation---unless I have missed that? My assumption is, that since it is not included in the form questionnaire, there is no implication that statins are playing any part in the resulting score at all?

 

[Jasmin2000]

The QRisk scores are not just for diabetics, but the population as a whole, and hence you cannot include A1c as not everyone has this test done - and in any case, it only reports over 3 months.

Frankly I don't know why the QRIsk score is such an issue for diabetics being offered statins, as you get them offered simply by being diabetic anyways.

 

[LionChild]

The QRisk scores are not just for diabetics, but the population as a whole, and hence you cannot include A1c as not everyone has this test done - and in any case, it only reports over 3 months.

Frankly I don't know why the QRIsk score is such an issue for diabetics being offered statins, as you get them offered simply by being diabetic anyways.

Hi Jasmin. The point here is simply regarding the accuracy of the QRisk3 score. It is the variables that affect that score that are being discussed.

 

[Jasmin2000]

Hi Jasmin. The point here is simply regarding the accuracy of the QRisk3 score. It is the variables that affect that score that are being discussed.

Well in the context of predicting cardiovascular (CV) risks for the general population it is currently the most accurate calculator there is for 10-year risk of CV risk, based on historical CV incidence and EHR analysis (in UK). It is great for healthcare providers to plan budgets and care resources.

However, it is not useful in isolation for assessment of an individual's health needs, and can only be used together with other metrics to create a plan for reduction of CV risk for each patient.

There are other CV risk calculators around such as the UKPDS Risk Engine that do include A1c but the results are similar to QRisk - here's a link to the calculator.

 

[LionChild]

Well in the context of predicting cardiovascular (CV) risks for the general population it is currently the most accurate calculator there is for 10-year risk of CV risk, based on historical CV incidence and EHR analysis (in UK). It is great for healthcare providers to plan budgets and care resources.

However, it is not useful in isolation for assessment of an individual's health needs, and can only be used together with other metrics to create a plan for reduction of CV risk for each patient.

There are other CV risk calculators around such as the UKPDS Risk Engine that do include A1c but the results are similar to QRisk - here's a link to the calculator.

But that UKPDS seems to relate to Type 2 Diabetics but no mention of Type 1s?

 

[HSSS]

Just wanted to clarify something here HSSS. What makes you think that already taking the statins is implicit in the QRisk3 score? I have not read that anywhere in their explanation---unless I have missed that? My assumption is, that since it is not included in the form questionnaire, there is no implication that statins are playing any part in the resulting score at all?

I did not say taking anything was implicit. I said their use is already factored into the result in so much as they will (presumably as that’s their intention) have modified your cholesterol scores which are considered. So an indirect effect. The same as blood pressure medication would be. If the same person was not taking the statins the assumption is their ldl/total cholesterol would be higher therefore their risk higher. It so sure thats the case but that’s the main stream position currently.

None of the parameters specify if the results entered are before or after medications or actions to alter the data, including if I move a few streets away or have only lived in my current address for days.

 

[LionChild]

I did not say taking anything was implicit. I said their use is already factored into the result in so much as they will (presumably as that’s their intention) have modified your cholesterol scores which are considered. So an indirect effect. The same as blood pressure medication would be. If the same person was not taking the statins the assumption is their ldl/total cholesterol would be higher therefore their risk higher. It so sure thats the case but that’s the main stream position currently.

None of the parameters specify if the results entered are before or after medications or actions to alter the data, including if I move a few streets away or have only lived in my current address for days.

If that is the case---- that these assumptions are made---- then in many cases, the result of the test seems even less dependable than I at first thought!

 

[Jasmin2000]

But that UKPDS seems to relate to Type 2 Diabetics but no mention of Type 1s?

Yes, it's a T2 calculator - they are working on a T1 but it's a long process gathering all the data

EDIT: here are some T1 cardiovascular risk calculators
STENO Type 1 Risk Engine https://steno.shinyapps.io/T1RiskEngine/
Swedish 5-year Risk Calculator https://www.msdmanuals.com/en-gb/pr...-of-cardiovascular-disease-in-type-i-diabetes
Australian CVD Risk Calculator https://www.cvdcheck.org.au/calculator

 

[LionChild]

Yes, it's a T2 calculator - they are working on a T1 but it's a long process gathering all the data

EDIT: here are some T1 cardiovascular risk calculators
STENO Type 1 Risk Engine
Swedish 5-year Risk Calculator
Australian CVD Risk Calculator

Thank you Jasmine2000. I think these are very interesting and will follow them up. I think these risk calculators could be very useful if sufficient data is allowed for, and put in. At the present state of affairs regarding the QRisk3 however, in my opinion it is really very misleading. If the significance of being a Type 1 for example, means that one's risk shoots up astronomically, just on the basis of that variable (quite significant rise also I notice re. mental health issues - presumably relating to hormonal issues? ), then I would question the accuracy of any result? Surely the main point about being Type 1, that is important, is glucose control, and with the low carb diet now becoming much more prevalent, as well as more and more helpful technology being available to help people's control, then A1Cs seem to be being kept much lower than earlier times perhaps? If this is the case, then this must surely have a bearing on the level of CV risk?

I have already mentioned other variables not included so I won't repeat them. But it IS interesting to see the various calculators that have been and are being developed and I am sure the potential to guide people towards better health is very welcome.

Thank you for your information!

 

[Redshank]

For anyone intersted in the Background to the development to QRisk3, here is a link to a BMJ article covering this

Development and validation of QRISK3 risk prediction algorithms to estimate future risk of cardiovascular disease: prospective cohort study

Objectives To develop and validate updated QRISK3 prediction algorithms to estimate the 10 year risk of cardiovascular disease in women and men accounting for potential new risk factors. Design Prospective open cohort study. Setting General practices in England providing data for the...

www.bmj.com

It shows how the model was developed.
I found the "Hazard Ratios" interesting (Table 3 for women and table 4 for men)
All of the Risk estimates are established from a population and are not a "prediction " for you or I as an individual.
Some of the factors we can do something about (e.g Smoking), possibly Blood Pressure.
Some we cannont - eg ethnicity, age
It does emphasise the importance of the Total Cholesterol/HDL ratio rather than just Total Cholesterol
The Post Code is a proxy for deprivation. If you leave it blank it just uses the average for the country. It is valuable for planners but not much so for individuals.
With Diabetes, I am sure the model could be improved using more detailed information - but I don't know how easy that is to collect.

I think one of the major problems is not the data, it is the use of language to describe the data:

I think this part is good and can be useful
"In a crowd of 100 people with the same risk factors as you, 12 are likely to have a heart attack or stroke within the next 10 years."

​

​

I think this part is innacurate and over interprets the data being used. In addition, the use of a decimal point gives a false sense of precision (in my opinion)
"Your risk of having a heart attack or stroke within the next 10 years is:"

12.3%