Saturday, May 23, 2009

Is red meat bad for you?

Below is a research summary circulated on a mailing list for medical practitioners. It arrived under the heading: "Red meat is bad for you —and bad for everyone else". Further below is the journal abstract (summary) concerned -- from a very respectable medical journal. The whole thing is, however, one big confidence trick and will achieve nothing other than frightening people off perfectly harmless food that they would otherwise enjoy. The entire report is a scientific, statistical and ethical nothing. Let me tell you very quickly why.

For a start, the "hazard ratios" (relative risks) reported are negligible -- at 1.2, 1.3 etc. The Federal Reference Manual on Scientific Evidence, Second Edition says (p. 384): "the threshold for concluding that an agent was more likely than not the cause of an individual's disease is a relative risk greater than 2.0."

OK. So who cares about a silly old Federal Reference Manual on Scientific Evidence? But it gets worse. The findings are reported in terms of upper and lower quintiles. In other words they threw away three fifths of the information that they had in order to arrive at their reported conclusions. That is quite simply dishonest and unethical. NO categorization of such data for analytical purposes is now ethically defensible. In pre-computer days, when all calculations had to be performed by hand, doing so could in some cases be justified but with the advent of computers there is NO reason why regression techniques that include ALL the data cannot be used. I note that before I had access to computers, I analysed the data from my first ever piece of research (in 1966) using a regressional technique. Even at that early stage I did not contemplate throwing away any of my data in the course of analysing it.

Had the whole of the data been analysed using a regressional technique, there is no doubt the the resultant correlation between meat consumption and disease would have been derisorily small and maybe even of negative sign, indicating that red meat eating is NOT a cause of cancer, heart disease etc. It is certainly not "bad for you —and bad for everyone else". The authors would of course be aware of that but have nonetheless chosen to present their data in a way that makes mountains out of pimples, which seems to me quite simply unethical.

So how did such a piece of utter crap get published in a medical journal? More particularly, why is such crap ROUTINELY published in medical journals? I am afraid that it is a sad outcome of the "publish or perish" regime that prevails in academe. Researchers need to get papers published in order to be promoted. So a well-meaning consensus has emerged among journal editors that they will accept extreme quintile reports out of solidarity with their colleagues. Otherwise they would have to reject more than half of what they currently publish. That the practice routinely results in the public being deceived is of no account. It is an utter disgrace but I doubt if I will live to see it stopped. An ethical vacuum prevails where the public would normally expect the highest ethical standards.

The emailed circular from DocAlert Messages below:
Further evidence of a link between red meat and poor health has emerged from a large cohort of older US adults. Men and women in the top fifth of red meat intake had a significantly higher risk of death over 10 years than men and women in the bottom fifth (hazard ratio for men 1.31, 95% CI 1.27 to 1.35; for women 1.36, 1.30 to 1.43). The authors also found a link between death and a high intake of processed meat such as bacon, ham, and sausage.

The 545 653 adults were between 50 and 71 when they filled in a detailed food frequency questionnaire in 1995. By 2005, more than 71 000 had died. These large numbers mean the authors were able to estimate with some precision the risks associated with eating red and processed meats for both men and women. The analyses were fully adjusted for other lifestyle factors likely to influence lifespan, especially smoking.

These data add to other observational studies that suggest we should all eat less red and processed meats. Not least because the increasing consumption of meat in many countries is putting a strain on global supplies of water, energy, and food in general, says a linked comment (p 543). It is costlier in all these precious resources to grow meat to eat than to grow vegetables and grains instead.

Journal abstract below:
Meat Intake and Mortality: A Prospective Study of Over Half a Million People

By Rashmi Sinha et al.

Background: High intakes of red or processed meat may increase the risk of mortality. Our objective was to determine the relations of red, white, and processed meat intakes to risk for total and cause-specific mortality.

Methods: The study population included the National Institutes of Health–AARP (formerly known as the American Association of Retired Persons) Diet and Health Study cohort of half a million people aged 50 to 71 years at baseline. Meat intake was estimated from a food frequency questionnaire administered at baseline. Cox proportional hazards regression models estimated hazard ratios (HRs) and 95% confidence intervals (CIs) within quintiles of meat intake. The covariates included in the models were age, education, marital status, family history of cancer (yes/no) (cancer mortality only), race, body mass index, 31-level smoking history, physical activity, energy intake, alcohol intake, vitamin supplement use, fruit consumption, vegetable consumption, and menopausal hormone therapy among women. Main outcome measures included total mortality and deaths due to cancer, cardiovascular disease, injuries and sudden deaths, and all other causes.

Results: There were 47 976 male deaths and 23 276 female deaths during 10 years of follow-up. Men and women in the highest vs lowest quintile of red (HR, 1.31 [95% CI, 1.27-1.35], and HR, 1.36 [95% CI, 1.30-1.43], respectively) and processed meat (HR, 1.16 [95% CI, 1.12-1.20], and HR, 1.25 [95% CI, 1.20-1.31], respectively) intakes had elevated risks for overall mortality. Regarding cause-specific mortality, men and women had elevated risks for cancer mortality for red (HR, 1.22 [95% CI, 1.16-1.29], and HR, 1.20 [95% CI, 1.12-1.30], respectively) and processed meat (HR, 1.12 [95% CI, 1.06-1.19], and HR, 1.11 [95% CI 1.04-1.19], respectively) intakes. Furthermore, cardiovascular disease risk was elevated for men and women in the highest quintile of red (HR, 1.27 [95% CI, 1.20-1.35], and HR, 1.50 [95% CI, 1.37-1.65], respectively) and processed meat (HR, 1.09 [95% CI, 1.03-1.15], and HR, 1.38 [95% CI, 1.26-1.51], respectively) intakes. When comparing the highest with the lowest quintile of white meat intake, there was an inverse association for total mortality and cancer mortality, as well as all other deaths for both men and women.

Conclusion: Red and processed meat intakes were associated with modest increases in total mortality, cancer mortality, and cardiovascular disease mortality.

Arch Intern Med (2009) Vol. 169 No. 6. 562-571

In addition to the statistical and ethical failures that I have detailed above, there are of course other large problems with the interpretation of the study. The first or second thing you learn in Statistics 101 is that "correlation is not causation". The authors above were cautious NOT to make causative inferences from their data but that message got lost downstream. Even well-informed people reading the report DID assume a causative relationship. They assumed that red meat eating CAUSED heart disease etc. But NO epidemiological study enables causative inferences. There could easily be third or fourth factors producing the observed association.

Just to give a top-of-the head example of how that could have played out: Given the weak associations reported, maybe a substantial proportion of those who ate little or no meat were Seventh Day Adventists. Adventists are an exceptionally healthy group who encourage vegetarianism. So WHY are they exceptionally healthy? Nobody really knows but it seems likely that the strong social and psychological support that they get from their heavy church involvement reduces stress and thus also reduces stress-related disease. And heart disease is partly a stress-related disease. So even if we accept as proper the statistical jiggery pokery reported above we may be basing our conclusions entirely on the doings of Seventh Day Adventists -- which is not of much relevance to the rest of society.

Research hints kids can defeat autism

I believe this. Austistic people are often highly intelligent and high intelligence can help people learn more adaptive behaviour. The famous John Nash even overcame schizophrenia by himself

Leo Lytel was diagnosed with autism as a toddler. But by age 9, he had overcome the disorder. His progress is part of a growing body of research that suggests at least 10 percent of children with autism can "recover" from it - most after undergoing years of intensive behavioral therapy.

Skeptics question the phenomenon, but University of Connecticut psychology Professor Deborah Fein is among those convinced it's real. She presented research last week at an autism conference in Chicago that included 20 children who, according to rigorous analysis, got a correct diagnosis but years later were no longer considered autistic.

Among them was Leo, a boy in Washington, D.C., who once made no eye contact, who echoed words said to him and often spun around in circles - all classic autism symptoms. Now he is an articulate, social third-grader. His mother, Jayne Lytel, says his teachers call Leo a leader.

The study, funded by the National Institute of Mental Health, involves youths ages 9 to 18. Autism researcher Geraldine Dawson, chief science officer of the advocacy group Autism Speaks, called Fein's research a breakthrough. "Even though a number of us out in the clinical field have seen kids who appear to recover," it has never been documented as thoroughly as Fein's work, Dawson said. "We're at a very early stage in terms of understanding" the phenomenon, Dawson said.

Previous studies have suggested between 3 percent and 25 percent of autistic kids recover. Fein says her studies have shown the range is 10 percent to 20 percent.

But even after lots of therapy - often carefully designed educational and social activities with rewards - most autistic children remain autistic. Recovery is "not a realistic expectation for the majority of kids," but parents should know it can happen, Fein said. Doubters say "either they really weren't autistic to begin with . . . or they're still socially odd and obsessive, but they don't exactly meet criteria" for autism, she said. Fein said the children in her study "really were" autistic and now they're "really not."

University of Michigan autism expert Catherine Lord said she also has seen autistic patients who recover. Most had parents who spent long hours working with them on behavior improvement. But, Lord added, "I don't think we can predict who this will happen for." She does not think it's possible to make it happen.

The children in Fein's study, which is still ongoing, were diagnosed by an autism specialist before age 5 but no longer meet diagnostic criteria for autism. The initial diagnoses were verified through early medical records. Because the phenomenon is so rare, Fein is still seeking children to help bolster evidence on what traits formerly autistic kids may have in common. Her team is also comparing these children with autistic and non-autistic kids. So far, the "recovered" kids "are turning out very normal" on neuropsychological exams and verbal and non-verbal tests, she said.

The researchers are doing imaging tests to see if the recovered kids' brains look more like those of autistic or non-autistic children. Autistic children's brains tend to be slightly larger than normal. Imaging scans also are being done to examine brain function in formerly autistic kids. Researchers want to know if their "normal" behavior is a result of "normal" brain activity, or if their brains process information in a non-typical way to compensate for any deficits. Results from those tests are still being analyzed.


No comments: