Tag Archives: Risk

Hair trumps helmet

Christchurch city councilor and Canterbury Health Board member Aaron Keown is on another crusade.  This time he wants the bicycle helmet law repealed because “Putting a lid on your head messes up your hair, and for a lot of people that is an issue,”  (see here ).
“Vanity, vanity, everything is vanity”
and that should be the end of it, if he wasn’t serious.  He is.  He thinks the look of a helmet and the cost is stopping Christchurch become the Amsterdam of the South Pacific.  Having lived and worked in Amsterdam, I can think of many reasons, not least being driver behavior, why we won’t emulate that city. As for being put off cycling, nothing is morre disturbing than having fat bottomed cyclist zooming past you on the cycle path in skin tight lycra.  Ban lycra I say…thousands more will dust off the old treadly and all will be well.
Having said all that, Mark #18, commenting on the Press web said states “Aaron Keown has some good science to support his case: http://www.cyclinghealth.org.nz/ But never let science and logic get in the way of so-called “common sense” – especially when “common sense” is supported by anecdotes.”   OK then, let’s see what that “good science” is.  The web page quoted has No science whatsoever on it.  It links to one and one only paper from a science journal, namely the NZ Medical Journal of February 2012 . That sounds promising, so let us look at that article. I regularly get to referee articles for inclusion in medical journals, so I shall apply the same scrutiny to this one.
Evaluation of New Zealand’s bicycle helmet law
Colin F Clarke
NZMJ 10 February 2012, Vol 125 No 1349
The first thing that strikes me is that it is an “evaluation” rather than trial or systemic review.  That is, it is some word of (expert) opinion.  Fine, journals have those kinds of articles all the time.  They are worth reading if the person writing them really has expertise (see the end of the article and judge for yourself if Colin Clarke is really expert or not). They should not be talked about, though, in the same breath as a clinical trial or systemic review or metaanalysis.
The second thing that strikes me as I read through is that is is full of numbers used to support the author’s opinions, but there are no statistics at all to back this up.  My PhD students would get a flea in the ear if they tried to present a differrence in means to me as meaningful without backing it up by the appropriate statistical test which tells me how likely the difference is to be real rather than random variation.  I would not agree to publication of any article that looks at relative risk, as Colin Clarke does without presenting 95% confidence intervals so that I could see if a relative risk of, say, 2.4 was really differnt from 1.  ie was the risk of death by cycling really 2.4 times that of walking? If the 95% confidence interval straddles zero (eg 95%CI -0.6 to 5.4) then the answer is “probably not.” Without that information,
“Meaningless, meaningless, everything is meaningless.”
Mr Clarke’s conclusions are just that. Furthermore, he makes the mistake of assuming that changes in incidences of death or injury since the advent of the helmet are because of the helmet.  He does not account for other changes including the lightness and speed of bikes, the greater density of cars on NZ roads etc etc etc.
Sorry Mark #18, this is defininetly NOT good scientific evidence.  As a scientists in a university Department of Medicine in New Zealand, I am ashamed that the NZ Medical Journal should allow such poor science to be published.
I am aware there is some research on the issue in other jurisdictions.  It is not so overwhelming as to have resulted in comissioned studies in NZ, let alone a change of law. Certainly vanity is not a reason to trump safety.  In the meantime, Cr Keown, I expect better of someone on the CDHB health board.  We have many issues on this city that are much more important. Drop this one and get on with your job

To PSA or not to PSA

As a male, 40 mumble years old, do I do it?  Do I get a prostate exam and PSA test?  Do I plan to keep doing tests every few years?

PSA (prostate specific antigen) is a blood test where elevated levels may indicate the presence of prostate cancer.  A powerful group, the US Preventative Services Task Force has come out against screening with PSA giving the test its lowest (D) grade.  They conclude “that many men are harmed as a result of prostate cancer screening and few, if any, benefit.”  Strong words. TV3 (misleadingly, but that’s another story!) and other media reported on this last night. The response of the Urological Society (at least its president) is to reject the report and urges men “not to be deterred” and to “discuss the PSA blood test with their GP.

This is approximately how my conversation went a couple of years ago.

GP: We’ll do a PSA test while we are at it.

ME:  Isn’t that a waste of time? Doesn’t it have a lot of false positives?

GP: Yes, but we can monitor for changes.

Hmmm…so it is not just the value of the test, but how it changes in time that is important.  A quick check on the internet I find that this is called the PSA “velocity.”  Interestingly in the evidence provided by the US Task Force I can find no mention of PSA velocity.

In the meantime, a quick check on the Canterbury Health Labs web site (see here) tells me that the test has a reference range of 0 to 4.0 ug/L (this is a concentration in plasma).  If a test is above this range a GP is likely to want to discuss it with you and may recommend a biopsy.

This is where life gets interesting.  A couple of weeks ago I talked of “False Positives” and introduced the diagram below.  A “False Positive” for myself would have been a PSA above 4.0 ug/L which didn’t turn out to be cancer.  The main issue with PSA tests is the high number of False Positives.  The Task Force suggested that in a screening regime after 3 or 4 tests (over several years) 12 to 13% of participants have a positive test.  Most, though, are False Positives.  Approximately 80% of Positive tests are False Positives!  Consider this – if screening happened in NZ and 500,000 men had a test every 5 years then after 15 to 20 years 500,000 * 0.12 *0.8 = 4800 men will have had a False Positive test.  Another 1200 a True Positive test.

Ideally every test result will lie in the dark blue (true negative) or dark red (true positive). In reality, there is always a few false positives and false negatives [A good test would have few (the narrow ellipse), a poor test would have many (broader ellipse)].

Importantly, the Urological Society put it this way “The PSA blood test does not diagnose prostate cancer. But it raises a red flag and identifies those men who need to have prostate cancer excluded through further investigation via a prostate biopsy.”

PSA does not diagnose – this is a very important point that a GP must communicate BEFORE a test is done.  I would be surprised if even 10% of men realize that PSA does not a diagnose.  So what happens to all the False Positives and True Positives?  This is what the Task Force focused on.

First they asked “Does PSA-Based Screening Decrease Prostate Cancer–Specific or All-Cause Mortality? Does PSA-Based Screening Decrease Prostate Cancer–Specific or All-Cause Mortality?

There was no clear evidence it does (contradictory studies).  In their useful “stats at a glance” publication they state “1 man in 1,000 – at most – avoids death from prostate cancer because of screening.”

If this is so, then it could be worth it (by the way – at a cost of $11.92 + GST + cost of GP visit – say $60 (low), then I estimate screening of 100,000 men a year would cost a minimum of $7.2M annually in NZ).

It is the next questions of the Task Force that are revealing.  The looked at the harms of screening.  The harms of those with Positive test (True or False) and then the harm to those finally diagnosed with prostate cancer.  Again the summary is revealing:

Most prostate cancers found by PSA screening are slow growing, not life threatening, and will not cause a man any harm during his lifetime. However, there is currently no way to determine which cancers are likely to threaten a man’s health and which will not. As a result, almost all men with PSA-detected prostate cancer opt to receive treatment. In addition to the frequent complications of biopsy that lead to a cancer diagnosis, there can be serious harms from treatment of screen-detected prostate cancer.

For every 1,000 men who are screened with the PSA test:

  • 30 to 40 men will develop erectile dysfunction or urinary incontinence due to treatment
  • 2 men will experience a serious cardiovascular event, such as a heart attack, due to treatment
  • 1 man will develop a serious blood clot in his leg or lungs due to treatment 

For every 3,000 men who are screened with the PSA test:

  • 1 man will die due to complications from surgical treatment

And they did not attempt to assess social or psychological harm!  Imagine the conversation at home:

Man: Hi honey, I’m home.  I got a positive PSA test today.

Woman:  That’s nice dear.  Did you get an appointment for a biopsy.

Man:  Yes, in 3 months time.

Woman: Great.  Shall we go out for dinner?

Somehow, I don’t think it would be like that, except perhaps the waiting time for a next appointment.

So where does this leave us.  My opinion, for what it is worth, is that:

  1. A PSA screening program should not take place in New Zealand.
  2. GPs should use PSA tests only where there are other risk factors
  3. Prior to any other procedure, repeat tests of positives should be done under strict conditions. Particularly the diet of the person involved should be changed to minimize the risk of false positives (there is still debate about the role of diet in false positives – so some research should be done at the same time: “Does changing diet change PSA levels in the short term?”).  Men – you can ask for this!
  4. GPs should explain that:
  •            a positive test does not mean cancer (most probably already do explain this, but it worth emphasizing),
  •            there are risks with biopsies, and
  •            there are great risks with treatment (prostectomy or radiation normally).

I qualify this with what appears to me to be a lack of assessment of the benefit of “changes in PSA” levels.  The sort of question which comes to mind is “How accurate is the diagnosis of a 2 ug/L or 100% increase (say) in PSA over 5 years?”

I wonder, would you have a PSA test?

Hold that Rice, it may be dangerous or perhaps not…

“White rice could cause diabetes”   The Independent UK

“White rice link seen with type II diabetes”  Yahoo News

“White rice raises T2 diabetes risk”  The Telegraph

Those were the headlines of the last few days.  Of the three posted, one is really wrong, one is sort of OK, one is almost OK and all are meaningless.  The headlines come out of a “meta-analysis” of population studies of risks associated with diabetes.  Meta-analysis are important because they consider the quality of studies and combine their results which, if done well, reduces the chance of giving erroneous results.  In this case there were only four studies included in the meta-analysis.  The paper was published in the British Medical Journal and is available here if anyone wants to read it.

So – the real question, is “can we keep eating rice?”  The overwhelming answer is “yes we can!”

To answer why, you may need to learn something about (in a whisper) statistics.  If you are new to reading medical journals then one of the terms you will run across and need to understand is “Relative Risk.”  What the number means is the increase in risk (chance of having the disease) of one group compared to another.  A Relative Risk of 1.5, then, means that one group is 1.5 times as likely to get the disease as another.  Right away, you can see that it is easy to get a high relative risk by comparing groups with very low absolute risk to those with high absolute risk.  Eg, for Type 2 diabetes if we compared healthy normal weight individuals with obese individuals then the relative risk would be high for the obese individuals.  In this meta-analysis they compared the group with the lowest consumption of rice with those with the highest consumption of rice.  Before I give the numbers, remember that most people do not have either the lowest or the highest consumption, but somewhere in between.  Anyway, for Western populations the relative risk was 1.12 and for Asian populations it was 1.55.  This looks like those Westerners eating heaps of rice are about 12% more likely to get Type 2 diabetes than those eating the least.  However, the devil is in the details…

Stats lesson number 2.  The 1.12 is presented with some numbers after it in brackets: 1.12 (0.94 to 1.33).  These numbers are crucial.  They are even more important than the 1.12 itself!  They are what is called a confidence interval. In this case a 95% confidence interval.  They are an indication of just how good the estimate of 1.12 really is.  In this case, not very good.  The authors are saying that they are 95% confident that the true relative risk (i.e. the number we would get if we included everyone in the world in the study) is somewhere between 0.94 and 1.33.  Because this number straddles 1 there is a good chance (maybe ~25%) that the real relative risk is actually less than 1, in other words that eating the least amount of rice has a greater risk than eating the most!  In the Asian population the relative risk was 1.55 (1.20 to 2.01) which suggests that it is unlikely that the real relative risk is less than 1.20.

OK then, let us for a minute assume that if we measured everyone in the world and found a true relative risk of 1.2 for those eating the most rice compared with those in the group eating the least.  What does that mean to you and me.  Probably nothing, because what is important is Absolute risk.  As far as Type II diabetes goes the equation is simple – if you are obese then you are at high absolute risk of Type II diabetes.  Eating rice is going to make next to no difference.  If you eat heaps of rice it is probably  because you eat heaps of food.  Cut back on the rice, but for goodness sake do not replace it with pasta or spuds or anything else for that matter.  If you eat well and exercise a bit and happen to prefer rice to spuds….then there is no big deal…bon appetite.