WS MoreOrLess: Alcohol risk

Summary

Podcast Summary: More or Less – "Alcohol Risk"

Host: Tim Harford
Date: January 27, 2014
Episode Theme:
This episode delves into statistical claims about alcohol and cancer risk as presented by public health authorities, specifically examining how risk is communicated and understood. The show also includes a second segment analyzing an unusual statistical anomaly in tennis, focusing on Roger Federer's record.

---

1. Alcohol Consumption and Mouth Cancer: How Risk Is Framed

Listener Question & NHS Claim

• (00:19) Michael Green, a listener, noticed an alarming statistic in an NHS leaflet:

  “Two large glasses of wine a day or more could make you three times more likely to get mouth cancer.”
  Michael wonders, “Gosh, could it be that bad? And what evidence was there?” (00:44)

Interpreting the Claim

• Tim Harford (01:07):
  • The leaflet states a “three times more likely” risk but doesn't specify the absolute risk of getting mouth cancer.
  • Quote:

    “3 times 0 risk would still be 0 risk. And unless we know how common mouth cancer is, it's hard to know how worried we should be about the risk of it tripling...”

  • Experts advise against communicating risks only in relative terms.

Establishing Context: How Common Is Mouth Cancer?

• David Conway, Lecturer in Dental Public Health, University of Glasgow (01:43):
  • Around 3,000 cases of mouth cancer per year in the UK.
  • Not as common as breast or prostate cancer but not extremely rare.

What Counts as 'Two Large Glasses'?

• David Conway (02:12):
  • “Two glasses of wine... 250 millilitres times two...about two-thirds of a 750ml bottle of wine -- quite a lot of alcohol.”
  • The "high risk" threshold is quite a substantial intake.

Where Does The ‘Three Times’ Figure Come From?

• Tim Harford (02:27):
  • Cites a 2004 meta-analysis by Professor Giovanni Corrao et al., reviewing international studies.
  • Challenges:
    • Difficult to standardize between populations (e.g., wine in France vs. vodka in Russia).
    • The bigger issue is confounding factors, especially tobacco.

Separating Alcohol, Tobacco, and Combined Risk

• David Conway (03:08):
  • Tobacco is “the elephant in the room” in alcohol risk studies.
  • Reference to the EURO-wide Archeage study:
    • Tobacco alone: 22% contribution to head and neck cancer risk.
    • Alcohol alone: 0% in their data.
    • Tobacco and alcohol together: 61% risk.
  • Quote:

    “The combination of their effect is actually, I think, where the heart of a lot of the risk is.” (03:44)

Key Point:

• The major risk comes from the combination of heavy drinking and smoking rather than heavy drinking alone.

Public Health Messaging

• Tim Harford (04:19):
  • NHS campaign intentions: Promoting overall healthier lifestyles, not just focusing on mouth cancer.
  • “The general claim that drinking less will reduce the risk of illness is indeed well supported by the evidence.” (04:49)

---

2. Statistical Surprises in Tennis: The Case of Roger Federer

The Irregularity in Wimbledon 2010

• Tim Harford (04:47):
  • Refers to the marathon Wimbledon match between John Isner and Nicolas Mahut:
    • Mahut, despite losing, won more points than Isner.
  • Introduces Ryan Rodenberg, sports writer/statistician.

How Often Does the "Losing More Points but Winning the Match" Happen?

• Ryan Rodenberg (06:38):
  • Analysed 61,000+ men’s matches since 1990.

  • “A little over 4.5% exhibited this characteristic. So it’s rare but certainly not unheard of.”

Roger Federer the Outlier

• Tim Harford (06:50):
  • Federer is the most frequent player involved in these anomalies:
    • 28 matches where the winner won fewer points, but Federer lost more often in these situations (4:24 lifetime record).
  • Ryan Rodenberg (07:18):
    • “When Roger Federer is involved in a match where the person who wins most points loses the match, that guy’s usually Roger Federer.”
    • “Exactly.” (07:30)

Why Might This Happen?

• Rodenberg (07:32):
  • Federer's opponents may employ a high-risk strategy (“go for broke”) because playing conservatively is unlikely to succeed against Federer.

  • “If I adopt a very high risk, high reward strategy, perhaps my slim chances of winning go up.” (08:00)

  • Occasionally, this strategy allows them to win just enough crucial points, despite losing more overall.

Tactical “Tanking”

• Rodenberg (08:18):
  • Opponents may deliberately slack off in sets they are likely to lose (“tanking”) to conserve energy for key sets.
  • Quote:

    “The fourth set is just beyond me. I’m not going to win it...Hopefully then I will have my energy back and I’ll be able to win the fifth set. And if I win the fifth set, I win the match.” (08:44)

Federer’s Approach

• Rodenberg (08:51):
  • No evidence that Federer ever “tanks”; he consistently fights for every point.

---

3. Notable Quotes & Memorable Moments

• Tim Harford on Risk Communication (01:07):

  “3 times 0 risk would still be 0 risk. And unless we know how common mouth cancer is, it’s hard to know how worried we should be about the risk of it tripling...”

• David Conway on Tobacco and Alcohol (03:44):

  “The combination of their effect is actually, I think, where the heart of a lot of the risk is.”

• Ryan Rodenberg on Federer’s Statistical Oddity (07:18):

  “When Roger Federer is involved in a match where the person who wins most points loses the match, that guy’s usually Roger Federer.”

• Ryan Rodenberg on High-Risk Tennis Strategy (08:00):

  “If I adopt a very high risk, high reward strategy, perhaps my slim chances of winning go up.”

---

4. Key Timestamps

• 00:44 – Listener question: NHS claim about wine and cancer
• 01:07 – 01:54 – Risk communication and mouth cancer statistics
• 03:08 – 04:19 – Role of tobacco and alcohol together in cancer risk
• 04:47 – 05:21 – Wimbledon case: More points but losing the match
• 06:38 – 07:30 – Roger Federer and anomaly of ‘losing winners’
• 07:32 – 08:55 – Possible explanations: opponent tactics & “tanking”
• 08:56 – Reference to Ryan Rodenberg’s analysis in The Atlantic

---

5. Conclusion

This episode unpacks both the flaws and intentions in how health risks are communicated to the public, with a critical eye on absolute vs. relative risk. It also ventures into a curious statistical occurrence in tennis, highlighting the nuance behind raw numbers and targeting Federer’s unique involvement.

For further information or to submit your own questions:

• Email: [e-mail]
• Find more content at bbcworldservice.com/moreorless

Transcript

[BBC Announcer]

This is the short edition of More or Less, first broadcast on the BBC World Service.

[BBC Podcast Host]

Thank you for downloading from the BBC. For details of our complete range of podcasts and our terms of use, go to bbcworldservice.com podcasts

[Tim Harford]

hello and welcome to More or Less on the BBC World Service. I'm Tim Harford. First, to your local doctor's surgery, or more specifically, to loyal listener Michael Green's doctor's surgery. He got in touch with us via more or lessbc.co.uk after spotting a statistic in the waiting room published by Britain's National Health Service, the nhs.

[Michael Green]

I came across this brightly coloured Department of Health leaflet and was going through the warnings about drinking too much and. And came across what to me was quite a scary warning, which was that two large glasses of wine a day or more could make you three times more likely to get mouth cancer. I wondered, gosh, could it be that bad? And what evidence was there?

[Tim Harford]

Before we get to the question of evidence, let's think about what the NHS leaflet is actually saying and what it's not saying, because it doesn't actually tell us what the risk of mouth cancer is. 3 times 0 risk would still be 0 risk. And unless we know how common mouth cancer is, it's hard to know how worried we should be about the risk of it tripling, which is why risk communication experts often advise against reporting risks in this way. So what are the risks of getting cancers of the mouth? This is David Conway, Lecturer in Dental Public Health at the University of Glasgow in the uk.

[David Conway]

They are relatively low volume cancers. So I think across the UK we have around 3,000 cases per year of mouth cancer.

[Tim Harford]

So 3,000 cases of mouth cancer, that's not by any means an extremely rare condition, but it is a lot rarer than breast cancer or prostate cancer. So the second thing to think about is the amount of alcohol required to produce this higher risk. Here's David Conway again.

[David Conway]

Two glasses of wine, and we're talking large glasses here, I presume, which is what, 250 millilitres high strength wine? I mean, 250 milliliters times two is about two thirds of a 750 milliliters bottle of win quite a lot of alcohol.

[Tim Harford]

Well, now we've put the claim into context. Now, is it true? The key claim, two glasses equals three times the risk, seems to come from a study published in 2004 by Professor Giovanni Correio and colleagues from the University of Milan Bikotka. They reviewed all the studies they could find and brought them together to try to produce a clearer picture. But do we fully believe it? It's a robust, decent meta analysis, but even so, it's difficult to standardize the data. The French are drinking wine while the Russians are drinking vodka. How do we make these alcohol consumption studies comparable? However, this is not the main difficulty that David can spot in the study.

[David Conway]

I think tobacco is often the elephant in the room when you are looking at the risks associated with alcohol. It's difficult to totally disentangle that risk to just focus on the alcohol consumption.

[Tim Harford]

Which brings us to a new and important problem. When you see that drinkers tend to suffer from cancer, is that correlation caused by the drink or is the pattern in the data explained by something else? For instance, that drinkers are more likely to smoke? Of course, the Correo paper tries to make the necessary adjustments and the authors defend their methodology as robust. But David Conway isn't sure how effective that is.

[David Conway]

We've looked at that in a European wide case control study called the Archeage study. We had two and a half thousand head and neck cancer cases. From our data we see that tobacco alone seems to contribute 22%. Interestingly, alcohol alone, when you just try and tease out the contribution of alcohol alone, it contributed 0%. But tobacco and alcohol in combination had a 61% risk or population attributable risk. So the combination of their effect is actually, I think, where the heart of a lot of the risk is.

[Tim Harford]

In short, perhaps the risk factor for mouth cancer isn't alcohol or even tobacco, but alcohol and tobacco together. Although the NHS has based its claim on some well established studies, it's hard to be as definitive as this leaflet sounds. But one thing we must also consider, this fact shouldn't be seen in isolation. It was part of a bigger campaign run by the NHS trying to encourage people to have a healthier lifestyle to avoid all sorts of cancers. And the general claim that drinking less

[Ryan Rodenberg]

will reduce the risk of illness is

[Tim Harford]

indeed well supported by the evidence. Back in 2010, one of the most memorable tennis matches in Wimbledon's history took place between John Isner from the United States and France's Nicolas Maheu. The match was the longest in history and broke many other records. But meanwhile, over in the United States, a sports writer was firing up his spreadsheet because he'd noticed something else was remarkable about the match. Nicola Me who lost, but he actually won more points.

[Ryan Rodenberg]

It jumped off the page to me that Isner won the match, but his opponent had won a full set's worth of Points more than John Isner had. Something had to be going on there to have such an irregular result in terms of the number of points won.

[Tim Harford]

This is Ryan Rodenberg, sports writer and sports law specialist at Florida State University. Mahoud, by the way, won 502 points. Isna Amir, 478. One of the things that makes a tennis match exciting is that it's made up of lots of smaller dramas. A match is made up of sets, and to win a men's Grand Slam match, you need to win three sets. Each set is made up of games. You win a set by winning at least six games and then either winning a tie break game or getting two games ahead of your opponent. Even the games themselves are made up of points. You need to win at least four points and to be two points ahead of your opponent to win a game. What all that means is that if you win some games and some sets by large margins, you can end up winning more points than your opponent, yet losing the match. Now, the Mahuisna match was unique, but these topsy turvy matches, where the loser accumulates more points than the winner, aren't unique at all. As Ryan Rodenberg discovered, when we looked

[Ryan Rodenberg]

at every men's match since 1990, a sample of over 61,000 matches, a little over 4.5% exhibited this characteristic. So it's rare, but certainly not unheard of.

[Tim Harford]

And then Ryan stumbled across another surprise. There was one player in particular who stood out in this data, and you may just have heard of him. He's the winningest player in the history of men's tennis. A young chap called Roger Federer.

[Ryan Rodenberg]

Roger Federer was a complete outlier. During the course of his long career, he's participated in 28 of these type of matches, and, and he has a 4:24 lifetime record in them.

So what we're saying here is when Roger Federer is involved in a match where the person who wins most points loses the match, that guy's usually Roger Federer. He's usually the one who wins most points and yet loses the match.

Exactly.

[Tim Harford]

So why is that?

[Ryan Rodenberg]

We came up with two explanations. The first one primarily is on the other side of the net. It's not specific to Federer directly, but his opponents may have consciously decided, well, if I were just to play my normal tennis game, I'm going to lose 626 2. This guy's just too good. But if I adopt a very high risk, high reward strategy, perhaps my slim chances of winning go up.

So they're going for broke on the big points and they'll usually lose. That usually won't work. But every now and then that comes off. So this is a player who loses more points than Federer because, hey, they're playing Roger Federer, of course they're going to lose more points and yet somehow they get a few lucky points just when it counts and they manage to win the game.

Exactly. So that would be going on. Plus they strategically, throughout the course of the match may take certain stretches of points off to rest, what people would call tanking.

And so that's just saying, look, the fourth set is just beyond me. I'm not going to win it.

[Tim Harford]

I'll lose heavily in the fourth set. I'm not even going to try.

[Ryan Rodenberg]

Hopefully then I will have my energy back and I'll be able to win the fifth set. And if I win the fifth set, I win the match.

Exactly. There is little to no evidence that on an intra match basis Federer partakes in this strategy.

There is no point at which Roger Federer takes a rest. He's always trying to win the next point.

Exactly.

[Tim Harford]

Ryan Rodenberg and you can read more about his analysis online in the Atlantic magazine. If there are any numbers you'd like us to check or explain, do please email us at more or lessbc.co.uk and you can download this program and many others for free free at our website, bbcworldservice.com More or less there are dozens

[BBC Podcast Host]

of different podcasts now available from the BBC, including news, documentaries, science, business, arts and sports. The details of them all go to bbcworldservice.com podcasts.