Good Health Information and Where to Find It

The sheer size of the Internet is both a blessing and a curse. If you are looking for good health information, it can be hard to know what is reliable and what is snake oil. There are thousands of health-related sites out there, and just one on-line bookstore lists over half a million books on health. Given that different sources often contradict each other, how do you find dependable answers to questions like these:

  • does exercise help weight loss?
  • is there a link between vaccines and autism?
  • who should cut out gluten?

You could take advice from a supermodel, a chef, or a celebrity doctor. But where do they get their information, and is it solid? Skeptics often tell people to “do your own research”, but how do you know you are doing a good job at that?

This article outlines sources and strategies you can use to find information to inform your lifestyle choices. Even so, if you are planning to make major changes, we recommend you run them past your doctor first. A professional familiar with your medical history can help you avoid making decisions that could risk your health.




A Note on Evidence

Before going into specifics about searches, a few words about evidence. Very few of us explicitly believe in knowledge handed down from the heavens. (If you do, I won’t argue with you, as it is unlikely to shift either party’s views.) If we rule out divine revelation, then for any commonly accepted fact, we can ask how and when it came to be part of human knowledge. We can also ask if the evidence is still considered to be strong, how it fits into the rest of our knowledge, or whether more recent discoveries have superseded it. 

The knowledge base of mainstream medicine has been acquired gradually, largely in the past two or three centuries. While there have been missteps, the fundamental contribution of evidence-based medicine has been to ask: what works? Setting aside long-held beliefs, do we know if blood-letting makes patients better? If we compare suitable groups, the answer turns out to be no, so we stop using it. Does fresh fruit cure scurvy? Testing shows it does, so we incorporate that into our practice. For any treatment, a similar question should be asked. What is the evidence that it works, and is there any risk of harm? 



google scholar

Discovering Google Scholar

The quality of evidence becomes an immediate issue if you use Internet search engines. When you run a search on Google, Bing or Yahoo, the top hits are determined by a constantly evolving algorithm that ranks pages with your search terms. Commercial websites invest heavily in search engine optimization to rank highly for a particular term, e.g. coconut oil. But there is no guarantee your top hits will provide balanced information on that topic. In fact, it is more likely you will find sites that want to sell you coconut oil, regardless of the evidence base. 

A better approach is to search a database of journal articles, written by and for specialists. One such site is Google Scholar. Another is Pubmed, maintained by the US National Institutes of Health. Academic researchers use subscription services like Web of Science, but for lay purposes, the free databases will do. Search results link to a summary (abstract) of a journal article giving the main results in a standardized format. In some cases, the full-text article may be available, but you will rarely need or want to read this.



coconut oil

Searching Journal Articles

Let’s try a sample search on Google Scholar, to see what we can find on coconut oil. Typing “coconut oil” (quotation marks search it as a phrase) gives you >100,000 results, many not relevant to human health. You could try “coconut oil” human, but you will still have >50,000 hits. Another option is “coconut oil” “health benefits”, but this could bias the results to documents from coconut industry representatives. You can use “coconut oil” “clinical trial” to find studies in humans, but you are still really asking an open-ended question, so your results may not be very focused.

Perhaps you’ve seen reports coconut oil can help weight loss and you want to know if this is true. In that case, you can search for “coconut oil” “weight loss” or “coconut oil” “weight loss” “clinical trial”. If there are too many hits, you can limit results to the last few years by clicking the button at the top left. You shouldn’t lose much information this way, as recent articles tend to summarize the earlier literature. Alternately, you can replace “clinical trial” with “meta-analysis” or “systematic review” (more on this below). But in this case our search has already identified one useful article. Let’s cast an eye over the results.



­coconut oil weight loss

Making Sense of the Abstract

We now have the abstract of a paper by Harris and others from the Journal of Medicinal Food. Journal abstracts are typically limited to about 300 words. They have one or two sentences each of an introduction, methods, results and discussion or conclusion. (Full-length articles follow a similar layout, but in much more detail.) The language is formal, but with a little effort you can get the gist. In this study, 12 women followed their usual diets for 4 weeks while taking two tablespoons per day of either virgin coconut oil or safflower oil. After 4 weeks without a supplement, they switched to the other oil and repeated the test. The researchers measured body fat, blood cholesterol and some markers of inflammation.

A look at the abstract suggests there was no weight loss from using coconut oil. We conclude this simply because the authors don’t mention any effects on weight (*). There were changes in blood cholesterol and inflammation, but nothing worth reporting for weight loss. Of course, this is a small group, a very defined population, and a short study period. These factors weaken the study and make it hard to generalize. At the same time there is nothing that suggests you will get weight loss at home with a sample size of one. The authors acknowledge the study limitations, indicating that while coconut oil might be safe in terms of heart disease, more work is needed.

(*) The full-text article confirms no weight loss as a result of using coconut oil.




kills cancer cells in a petri dish

A Hierarchy of Evidence

In looking over this or any other paper, it is worth keeping in mind that a single research study has limited evidence value. This is particularly true if you are trying to extrapolate from molecules or cells or animals to humans in general. Biology is always more complicated than you first thought. Our bodies have multiple levels of regulation affecting functions like inflammation or blood sugar or appetite. Partly due to this complexity, partly due to different study designs, different clinical trials often give different results. What if there are two studies showing less cancer in people who eat more fruit, but a new trial reports no benefit? The media sometimes announce the most recent finding as the definitive truth, but professionals have to take a longer-term view. 

Collating the results of different clinical trials into a single report is the function of a systematic review or meta-analysis. In some cases it is possible to combine results from different studies into a single numerical outcome. For example, relative to people eating 0-1 pieces of fruit per day, those who eat 3-4 pieces might have a relative risk of cancer of 0.94, meaning a 6% lower risk. But the statistical spread collated from multiple trials might be 0.86-1.02, meaning the answer probably lies somewhere between a 14% decrease in risk and a 2% increase. This invented example would suggest a slight cancer prevention benefit from fruit, but more studies would probably help. Conversely, if there haven’t been enough trials for a meta-analysis on a topic of interest, we probably don’t know enough to make an informed decision.



coconut oil meta analysis

Medium-Chain Triglycerides and Coconut Oil

If we want more specific information on coconut oil and weight loss, we can look a bit further. A meta-analysis has been published, though it is limited by a low number of participants (749) and a short time period (>3 weeks). The review found a minor effect, around 0.5 kg more weight loss, in the treatment group. But the quality of some trials was low, and a statistical analysis suggests commercial bias, e.g. failure to publish studies that showed no benefit.

In any case, there may be a small benefit in people randomized to take medium-chain triglycerides (MCTs) rather than long-chain triglycerides. Wait, what is that? Is this even about coconut oil? Well, not specifically. A scan of the abstract (and a quick check with Wikipedia) shows the study relates to purified oils with 8 or 10 carbon atoms in a chain. While there are some of those in coconut oil, it has a much higher content of longer-chain fats. And similar MCT levels are present in palm kernel oil, which no-one seems to be promoting as a path to a lean body.

In summary, while there is some buzz around coconut oil and weight loss, our search hasn’t found studies that strongly support this. Well, you might say, but absence of evidence isn’t evidence of absence. That’s very true. There could well be circumstances where coconut oil, with its medium content of medium-chain triglycerides, significantly assists weight loss. But as things stand right now, we don’t have a good reason to believe it.



A Note on Looking Further

You might notice the “cited by 27” link below the abstract of the meta-analysis. For articles more than a year or two old, the number of citations (other papers that refer to this paper) is a measure of whether other scientists consider the work to be significant. If you click on the “cited by” link, you will get a list of those papers, which can lead you to interesting recent information. If you sort “by relevance” rather than “by date”, Google Scholar tends to show you more highly-cited papers first, but it is still worth keeping an eye on the number. Papers with citation numbers in the thousands can be considered a fair representation of consensus in the field.



Cochrane Library

All this sifting through information might seem like a lot of work. No doubt it is (though still less wasteful than buying expensive oils without good reason). For something more directed, you might try the Cochrane Library. This a medical site that collates information on clinical trials and best-practice care. Searches on coconut oil won’t get you a lot of hits at the Cochrane Library, however. Where it comes into its own is providing good health information on topics that have already been thoroughly studied. In those cases, summaries from medical experts can help you get to the heart of an issue without needing to wade through many individual papers.

The Cochrane Library contains several databases, including Clinical Answers, Cochrane Reviews and a registry of clinical trials. While the site is aimed at doctors in the first instance, the reviews include a plain language summary for a broader audience. In many cases, the full-text article is also available. To get a sense of what is available at Cochrane, you could have a look at this article on whole grains and heart disease. Or try reading this article on restricted diets for autism, or this one on exercise for weight loss. With a good GP, you should be able to use this kind of information as a basis for discussion on your health choices. 



mayo clinic

Mayo Clinic

Still, maybe an academic search process is getting on your nerves. Fair enough: not everyone enjoys all that formal language. A different option for good health information is the Mayo Clinic. This is a non-profit network of several thousand physicians based in Minnesota. The Mayo Clinic does clinical and research work, but also has a useful website with plain-language information on medical topics. 

Searching for coconut oil on this site takes you straight to an article on weight loss, which begins:

The few small studies that have looked at coconut oil for weight loss suggest that coconut oil may help reduce waist size, but it doesn’t lead to significant weight loss or improved body mass index (BMI).

and finishes:

Although eating coconut oil in moderation isn’t going to result in great harm to your health, it’s not likely to help you lose weight either. For successful, long-term weight loss, stick to the basics — an overall healthy-eating plan and exercise.

It’s a pretty straightforward answer. The author is named and you can check her credentials. The articles she refers to aren’t specifically cited, but you now know where to look for them.



There is one last site to mention, but not really to recommend. In theory, Wikipedia should be somewhere you can find good health information. In practice the quality of information is highly variable. Anonymous authors and delayed editing processes make the site open to partisan views on hot topics—and there are few topics hotter than health. You can certainly use Wikipedia for factual background (what are medium-chain triglycerides?). However, getting medical information there can be hit-and-miss. At a minimum, if working from a Wikipedia entry, you should check the footnote references to see if they say what is attributed to them. Even then, if you really want to do your own research, you’ll probably have to head back to Google Scholar.




Further Reading

There is good health information with a solid evidence base, and then there are philosophies that do not accept a need for proof. It is good to be wary of the latter.

Killing Us Softly : The Sense and Nonsense of Alternative Medicine, Paul Offit, HarperCollins Publishers, London 2013

The Angry Chef: Bad Science and the Truth About Healthy Eating, Anthony Warner, Oneworld Publications, London 2017

Trick or Treatment: Alternative Medicine on Trial, Simon Singh and Edzard Ernst, Corgi Books, London 2009