Now I’m courting comtroversy
Here’s a view arguing against the current zeitgeist around screening for diabetes.
I’m aware it’s more fiddly than I set out here.
I’m often asked what to do re diabetes screening and prevention
My (rather hardline) view is this one
NSC had said no
1 it IS screening
identification of a cohrt of patient WITH disease but whom are as yet asymptomatic and offering them a test to definitively determine a diagnosis and then ensuring they get appropriate treatment. this IS screening.
NSC have recommended against screening for diabetes – this is on the basis of whole population screening.
(Off hand can’t recall the timing of ADDITION and the NSC review. Was pretty close)
Waugh and NSC
The paper by Prof Waugh sets it all out beautifully.
The proposal foul of the optimised treatment of current patients – why add more when current lot sub optimally managed.
No RCT evidence of outcome.
Waugh, and NSC, and the ADDITION study was considering a high risk cohort (as did NICE PH 38 – prevention in high risk) not a whole population approach.
Waugh said a case for screening doesn’t meet the NSC tests
The NSC recommendation, based on Waugh and the committee considerations, was against pop screening in those considered at high risk –
The current patient cohort is not optimised This is pretty clear, locally as well as nationally.
The argument goes
1) why add more cases in when you sub optimally manage current cohort. There is probaly greater clinical gain to be had out of improving the management of current cohort; and
2) adding more may take the eye off the main game, and thus be to the detriment of the care of the current cohort ? Some, but not yet convincing, evidence of this in places that have gone down the path of PH38. Mixed signals. Needs to be watched carefully.
RCT evidence re morbidity and mortality
The ADDITION study
No difference in CV events or outcomes out to 10 years.
Some signals (not yet convincing) re cv risks
Arguably 10 years is too soon a time horizon to be measuring CVD endpoints. I agree with this. However this in itself is not a reason to implement a population screening programme out with a NSC recommendation.
The evidence now report out to 10 years
Griffin et al 2011. 5 yr data – http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(11)60698-3/abstract
And – http://www.ncbi.nlm.nih.gov/pubmed/23959571 – QOL study out to 5yrs
10yr data – Lancet 2012 Simmons et al- http://www.ncbi.nlm.nih.gov/pubmed/23040422 – ADDITION at 10yrs out
Obviously there’s a valid argument that 10 years too soon to see complications. This is valid.
But in 15 years time we might find that concentrated rhubarb juice is a cure for cancer….doesn’t mean we’d recommend it now
This 2012 letter from Yudkin and others is worth a read
But……PH38 said yes.
NICE had considered the cost effectiveness of screening, but not considered the whole range of factor that the NSC consider.
But NSC advises on screening policy.
Screening vs case finding
After having thought about the issue very deeply, I (and many many others) can’t find any distinguishing features between “screening” and “case finding”.
They are essentially the same – as such a proposal to screen, or case find, or whatever for diabetes should be subject to the same process
Back to the question of the difference between “screening” and “case finding in a high risk pop” (hint – theres no difference)
I was accused of simply ignoring what NICE had said.
NICE had recommended risk assessment in high risk groups.
Did NICE overstep their brief in recommending screening here. The recommendation in ph38 has been enthusiastically adopted by those that want to believe screening saves lives
No QA criteria established
Population based screening is fundamentally based on the premise that there is robust QA in place I have heard some mixed (and potentially worrying) signals on this. Particularly wrt to DCA for diagnostics and it’s use in this particular population There is no, to my knowledge, QA in place for this
As a side point – Some residual concerns re point of care DCA machines.
Don’t know extent to which these were resolved.
5) will it save money?
There is a widespread assumption that this programme will save money
The evidence base is contrary to this.
NICE base case was (from memory) £12k / QALY (Im happy to be corrected there)
It’s therefore arithmetically impossible to try to make a case that this will save money.
This is on assumption that screening programme implemented perfectly, as per NICE base case.
If in implementing ph38, an area finds 1000 cases of diabetes it is relatively easy to calculate the cost of this
Category 4a of the NHSE programme budget data is diabetes. By combining this with the number of cases it is relatively easy to calculate the crude cost per case (yes I’m aware it’s more complex). From memory it’s c £4-550 per case looking at a spread of CCGs (someone may improve on this estimate)
So finding 1000 cases will add an annual cost of £400k to £550k + the cost of the programme implementation
Difficult to see the outcome improvement if you believe ADDITION so far.
We are giving people more time with a diagnostic label
More time and cost of treatment – xxxxx person years of Glycaemic drugs, atorva 40, retinal screen, foot care etc etc)
Maybe to the detriment of those that had previously been diagnosed
without (on the basis of the evidence we have at the moment) changing CVD endpoints
And risk of over treatment is very obvious here…… – we will find no end of elderly people with diabetes that we didn’t know about that we may harm through over treatment, and spend resource that could be better spent elsewhere.
But no improvement in outcomes comparing screen found vs normal clinical care
So by doing so, we’ve made the NHS less efficient.
6) Yes I know it’s all a bit complex and fiddly
And the separation of the “screening for diabetes bit” with the “prevention” bit is difficult to impossible – but by implementing the DPP we are de facto screening for diabetes.
And that the DPP is going to happen as it seems to be a political imperative, so we’d might as well make the best of it (gulp – rare pragmatism from me)
We are medicalising the prevention bit – at the expense of policy solutions?
The prevention bit of this is really important.
The original Diabetes Prevention studies (USA, Europe and China) are pretty compelling
NNT to prevent new incident diabetes is c7-8. Consistent efficacy evidence on this. Ie you have to “treat”or effect change to active living and diet in 8 people to prevent one case of diagnosed diabetes
But…….actually delivering this in real life devilishly difficult ….so everyone, including I fear this CCG, have adopted the “easiest” line….screening
Hopefully the evaluation will shed light on this. I guess the conversion between impaired glucose tolerance through to invited to the sweat club…through to successfully lost weight and kept it off
Some essential things to read
a) John Yudkin Kenote : Overdiagnosis and the Epidemic of Prediabetes
b) The epidemic of pre-diabetes: the medicine and the politics
c) Time to question the NHS diabetes prevention programme
The BMJ editorial
d) go back to prof Waugh report
It was masterful
The recent BMJ meta analysis. Barry et al
“Results 2874 titles were scanned and 148 papers (covering 138 studies) reviewed in full. The final analysis included 49 studies of screening tests (five of which were prevalence studies) and 50 intervention trials. HbA1c had a mean sensitivity of 0.49 (95% confidence interval 0.40 to 0.58) and specificity of 0.79 (0.73 to 0.84), for identification of pre-diabetes, though different studies used different cut-off values. Fasting plasma glucose had a mean sensitivity of 0.25 (0.19 to 0.32) and specificity of 0.94 (0.92 to 0.96). Different measures of glycaemic abnormality identified different subpopulations (for example, 47%of people with abnormal HbA1c had no other glycaemic abnormality). Lifestyle interventions were associated with a 36% (28% to 43%) reduction in relative risk of type 2 diabetes over six months to six years, attenuating to 20% (8% to 31%) at follow-up in the period after the trails.
Conclusions HbA1c is neither sensitive nor specific for detecting pre-diabetes; fasting glucose is specific but not sensitive. Interventions in people classified through screening as having pre-diabetes have some efficacy in preventing or delaying onset of type 2 diabetes in trial populations. As screening is inaccurate, many people will receives an incorrect diagnosis and be referred on for interventions while others will be falsely reassured and not offered the intervention. These findings suggest that “screen and treat” policies alone are unlikely to have substantial impact on the worsening epidemic of type 2 diabetes.”
And the editorial
- most of the studies in the Public Health England review lasted only 12 months.
- improvements might not be sustained once the intervention has finished.
- balance between the medical model of screening and treating of individuals and the public health model of changing behaviour in the whole population at risk.
- The main problem with both models is adherence.
- Preventing or delaying type 2 diabetes requires effective measures to motivate the general population to protect their own health.
As is always the case the really detailed arguments can be found in the rapid responses section…….
Two years later I’m still struggling to see an evidential case