Feb 18 2013
Contradictory studies can help bring researchers as close as they can ever hope to achieve the ‘absolute’ answers
First things first, let us delve into the oat bran saga. The whole drama began in the early 1980s, when a handful of researchers, taking a ‘cue’ from past research and expanding on them, suggested that the ‘water-soluble’ fibre in oat bran could lower cholesterol levels. Things reached the crescendo when Robert E Kowalski’s book, 8-Week Cholesterol Cure, eulogised some of the ‘wonderful’ medicinal qualities of oat bran.
This bid fair to yet another scientific review of high-fibre diet studies, with oat bran and beans. Oat bran was soon bestowed with an ‘effective certificate’ from The Journal of the American Medical Association (JAMA) — this was a major scientific avowal, emerging as it did from a well-respected, peer-reviewed publication. It highlighted why oat bran was special — because it was much more cost-effective than any other treatment ‘form’ available for high cholesterol levels. The spin-off was imminent. By the late 1980s, sales of oat brain reached astounding levels, so much so there was trouble keeping in step with its ever-increasing demand.
You’d sure call it ‘soda water bottle scientific enthusiasm’ — certainly, it wasn’t going to last. The magic oracle of the ‘one-piece’ oat bran orchestra slid and tumbled down the scientific alley when a 1990 study in the United Kingdom demonstrated that refined wheat had just as much healing effects as oat bran on high cholesterol readings — even without its water-soluble properties. Soon enough, conflicting opinions and views became commonplace. What’s more, while a handful of studies showed that oat bran had tangible healing effects, one new study reported an alarming finding — that oat bran could trigger intestinal distress. The inference was obvious. The oat bran impasse emerged as a conventional case of a scientific inquiry gone askew — one that the public had to contend with in stoical, albeit frustrating silence.
To cull another paradigm — the great coffee dilemma. A study in the 1980s, again, concluded that coffee, the cuppa that cheers, could be a likely trigger for more than half of all pancreatic cancers. Follow-up studies showed the whole idea was flawed. What do we have now? New studies have revealed the goodness of coffee in improving our reflexes, memory and reasoning power, along with the brew’s protective ‘cascade’ on healthy cells, following radiation therapy. This is cheering news for coffee lovers — but, again, the jury is guarded. A body of research now suggests that the better part of valour is discretion. The best thing to do, they argue, is to sip not more than three-four cups of coffee every day. Well, the deduction is obvious. All we need to do, perhaps, is wait for another day, another time, to know whether the coffee-driven roller-coaster ride is really going to take us places. For the time being though, the pleasure in the coffee cup is most welcoming, albeit just one revolutionary finding, in the opposite direction, could make us go bonkers, while discounting all the favourable scientific claims on the refreshing brew.
Let us now examine the billion-rupee question in detail. How can science, or medicine, in the wake of conflicting evidence, get to the root of scientific reality and persuade people not to throw up their arms in disgust? One plausible answer is meta-analysis. What’s more, modern science demands it. Because, the tools for making sense of reams — nay oceans — of data collected through painstaking scientific studies, are already available in the practice of meta-analysis.
As all of us know clinicians and others are up against a huge challenge — the overwhelming wealth of information, even in taut, small areas of speciality interest. Reviews designed to summarise the outsized volumes of information are, as a result, frequently published. When a review is done systematically, following certain criteria and the results are pooled and analysed quantitatively, it becomes meta-analysis. You get the point. A well-designed meta-analysis not only provides valuable information for researchers, policy-makers and clinicians, but also a host of decisive caveats in performing and interpreting them to the best extent possible.
So far, so good. Yet, a major limitation continues to haunt the scientific community. Meta-analysis is not a panacea. It is, at best, a useful remedy. This is, all the same, more than half of the battle won. Meta-analysis is well-established in medicine — although the whole idea got its first major boost in the social and behavioural sciences. This was primarily because of the presence of a large number of confounding factors capable of disproving a tentative claim. To paraphrase Turner E H, a researcher, who recently analysed the publication status of studies of anti-depressants. “Based on studies registered with the US Food and Drug Administration (FDA), it was found that 97 per cent of the positive studies were published versus only 12 per cent of negative studies. Furthermore, when the non-published studies were not included in the analysis, the positive effects of individual drugs increased between 11 per cent and 69 per cent,” said Turner. One reason for the publication bias, Turner further explained, “is that drug manufacturers are not generally interested in publishing negative studies... Yet another (factor) may be that (medical) editors favour positive studies, because they are the ones that make headlines and give the publication visibility.”
Yet, there is a huge upside. Take for instance, the early meta-analysis on aspirin and other anti-coagulant drugs, which provided a prop for the efficacy of the former, when taken together with the latter. It took just two years for the US Food and Drug Administration (FDA), thereafter, to approve the use of aspirin for survivors of myocardial infarction. You guessed it right. If only the use of meta-analysis for aspirin was established a good 15 years before, it could have saved over 15,000 lives every year in the US alone.
The use of meta-analysis in medical research during the past few years has brought about a phenomenal transformation. It has also, in fact, brought an end to many medicinal- or drug-propelled ‘slam-bang’ effects. This is because the future of medicine is enormously keyed to seeking new agents with incremental benefits and therapeutic effects. If this is not a signal proof, or the whole essence of meta-analysis, what is? To pick another typical example — the meta-analysis on the therapeutic effectiveness of streptokinase in heart attack. This is how it all happened and emerged — nearly eight previous studies did not significantly confirm the drug’s beneficial effects. Conversely, a meta-analysis conducted in the US, showed that streptokinase lowered death rates for patients treated immediately after symptoms of a heart attack occurred. This was a shot in the arm for medicine. The streptokinase perestroika exemplified — a la the promise the potent antioxidant, coenzyme Q10 (CoQ10), holds as a supplement in the treatment of hypertension — the solid ground for meta-analysis, including the big leap to explore new theories, decipher and analyse possibilities and variations between subject populations and also augment therapies in ways that would be next to impossible in just one isolated study — howsoever immaculate or meticulous.
All said and done, meta-analysis has its share of critics. Detractors call it philosophical, or prejudiced, with its hard and soft end-points — or, publication bias, because studies that show a positive results — usually in favour of a new treatment or against a well-established protocol — are more likely to be published than those that do not, as we discussed earlier. On the realistic side too, there is a lurking fall-out of meta-analysis — the imprudent use of sub-group analysis, most often with computer assistance. Yet, despite its pitfalls, most scientific observers reckon that meta-analysis is the best there possibly is — and, that it is a definitive advance for better protocols to emerge in the future. This is also precisely how science expands and has expanded all along.
The big point, on the brighter side, is meta-analysis can help bring researchers and scientists as close as they can ever hope to arrive at the ‘absolute’ answers they relentlessly seek to their niggling questions. Well, there is a big ask too — if only scientists and researchers avoid the temptation to regard meta-analysis as a completely objective question-emblem with solutions, they will, doubtless, find the methodology a prudent tool to detach the chaff from the grain and vice versa. The reason being, science does not begin with a nice question, nor does it end with a nice answer. zz
(The writer is a wellness
physician and author)