May 30, 2016

How to Find and Read Scientific Studies

… so you can check the evidence and decide for yourself

Scientific studies are actually pretty easy to find on a wide variety of topics. I normally use pubmed.com, it’s a website where almost all medical research is posted on. There are millions upon millions of scientific articles on it today, and the number is increasing at an accelerated rate every year.

Research articles are usually published in scientific journals (which cost thousands of dollars), and then they appear on pubmed either as an abstract or full text. Scientists complete their studies, write articles about their studies in as much detail as possible, and submit them to these journals. The journals review the articles in a process called peer-review where a group of other scientists read the study article and judge if it’s sound or not. If there are no problems with the study or the article, the article is published, both online and in print.

The way you use pubmed is simply open the front page and enter the name of the condition or disfunction you are interested in and perhaps a remedy in the search field. For example, type depression lavender essential oil. Today, 24 articles were fetched. Most of the articles are not free, and you can only read their abstract (=summary). But on the left you can see a “free full text” under “text availability”, and clicking that will list all free full text articles. I found 4. I liked this one: “Effect of lavender scent inhalation on prevention of stress, anxiety and depression in the postpartum period.” It’s a randomized controlled trial (RCT), the gold standard of clinical research nowadays. From the abstract you can see that the study concluded that inhaling lavender oil helps with postpartum depression. You can then click on the “Free PMC Article” for full article and read it.

Scientific articles are all written in the same format. First goes the abstract, then an introduction, then methods, results, discussion, and conclusions. Sometimes the methods and results are put at the end. The abstract is a summary of the entire article, it introduces a problem, then briefly states what methods were used and what results were found, and conclusions.

In the body of text, introduction talks about the problem or disease, stating the hypothesis (=educated guess) on the causes or possible treatments of the disease. The Methods section talks about how the experiment was set up and run. Normally, it should be presented in a way that can be reproduced easily following the methods section as instructions. This is because in science, you never take one study as a proof of anything. It takes multiple studies to gather enough support for one hypothesis.

One fact to remember is that science nowadays is not used to prove things. You will never find the word “prove” or “proven” in a scientific study. The purpose of science is rather to disprove hypotheses. It takes just one contradictory data to disprove something, and it can flop a whole theory no matter how much evidence exists for it. Therefore, evidence cannot be called proof. So, when you hear someone say “prove” when mentioning scientific studies, that person is far from science.

A hypothesis is a guess that predicts an outcome. Our lavender essential oil study guessed that lavender helps with depression. However, it is not a groundless guess. The researchers posted a vast list of references, or studies that were used to set up the experiment and write the article, and many of those mention lavender oil being beneficial for depression. You can in fact click on the reference numbers right in the text, and that will take you to the articles that the scientists took information from for this article. Most studies aren’t free, but some are.

A scientific theory is a statement about some process that has gained a lot of evidence. For example, there is a theory that free radicals cause aging and disease. Hundreds if not thousands of studies present evidence for the theory, and no contradiction has been found for it yet. Although there are alternative interpretation or critique of the mechanism of action. For lay people, when they talk about a theory (“it’s just a theory”) they actually mean hypothesis, an educated guess. For scientists, a theory is a solid evidence-based scientific reality as close to truth as you can get. Science in principle can never reach the absolute truth about the universe, but is perpetually approaching it.

Without solid theories, it takes me about two dozen studies or a couple of reviews to make me convinced of some effect. Reviews (some are called meta-analyses but it’s the same thing for lay people) are scientific articles that put together all the studies about a specific problem, analyze them all together, and draw a conclusion. Different reviews might have different outcomes because they may exclude certain types of articles, or be more strict judging their quality. The study about lavender for postpartum depression is a pretty believable outcome because there have been hundreds of studies on lavender oil for depression in various populations such as kids, the elderly, adults, and people sick with various diseases. If it was just one isolated study, I and other scientists would be extremely skeptical.

The results section of a study talks about the outcomes. Most of the time scientists measure multiple things and some or all of the results they get sometimes are not “significant.” This is a term from statistics, it means that the outcome was not due to chance. A significant difference between groups (treatment and control) means that the treatment was effective. It typically depends on how many people participated in the study. The more people in the study, the better is the strength of statistics, and the more assured you are that the outcome is due to the treatment and not some other unrelated effect (chance). Normally, I look for a hundred people, but it’s best when there are several hundred people in each group. In this study there were 140 women total studied, and they were separated in two groups of 70 women each (although it doesn’t say that in the article, but it’s implied). This particular study used depression and anxiety questionnaire scores, which is so-so. I like to look at hormone levels, neurotransmitters, or other blood work “hard” data. Questionnaires are self-reported and they can be vague or imprecise.

The discussion section talks about the results of the study, their significance, the implications of the effects seen in the study, as well as any shortcomings or any confounding variables. Confounding variables are things that may have caused the outcomes and gotten in the way of (confound) the results, either amplifying or dampening them. For example, if the women in the treatment group in our lavender study were staying in hospital rooms with large windows facing a park, then their lower depression scores could have been due to the nice scenery they enjoyed every day. If the women in the control group were in rooms with no windows, that could have affected their depression scores. There is no problem like that in this study, though, that was just an example of a confounding variable. Sometime confounding variables aren’t apparent, and this is why it’s best to do the same research study multiple times by different researchers in different countries or hospitals. The multiple positive outcomes in these various studies confirms the effect of a therapy.

One cannot talk about scientific studies without mentioning anecdotal evidence. Anecdotal evidence is just one or very few people’s experience with a therapy. In science it’s called a case study. One or up to five people are reviewed in detail in a scientific case study, and any effect of disease or treatment is watched closely. Case studies allow for an in-depth review of the effects of a treatment, including improvement in quality of life, psychological effects, life satisfaction, improved mobility, etc. Case studies are excellent in that they give a lot of insight into the details. But they are not generalizable, meaning, they cannot be applied to the population. Whatever one or several people experience is their unique experience. A case study is anecdotal evidence and cannot be used as evidence to support a wide use of the treatment. Again, it takes a hundred and ideally several hundred people in a controlled environment in order to gain evidence for the use of a treatment.

Same can be applied to real life. If a treatment worked for your friend, it doesn’t necessarily mean that it will surely work for you, too, and vice versa. In holistic medicine sometimes several treatments for the same thing are applied in case one of them doesn’t work in your specific case.

The conclusion section briefly states the outcome of the study and suggestions for possible real life implementations or for further research of the same problem to gain a more solid support for the treatment.

Studies are typically very specific and narrow. It takes quite a few studies in order to form a broad understanding of a problem. Depression is such a multifaceted disorder, and many things can be beneficial for it, lavender is just one of them. There are also quite a few unfamiliar words in studies which can impede their understanding. I normally read wikipedia articles on the words I don’t understand along with the article, but if it’s words I sorta know what they are but they are not important to understand the study (“variance analysis” in the lavender study), I just ignore them and continue reading. Most of the sciency-sounding words aren’t even needed to garner what you need from a study for your own purposes. Remember, scientific articles are mostly written for other scientists, and they come with their own lingo.

I would still exercise extreme caution in cases dealing with dangerous conditions such as cancer or stroke. I would trust a treatment only after reading a couple of hundred studies on a specific cancer, including studies explaining pathophysiology (=what specifically goes wrong in that specific cancer), the causes, the risk factors, comorbid conditions (=conditions that typically go together), and why the selected treatments work on the cancer. Same thing with autism, multiple sclerosis, and other debilitating conditions.

Most health care professionals don’t read scientific studies. They either don’t know how to or where to find them. You can help them by researching your condition and identifying experimental treatments for them. You can then talk with your doctor about trying them out. This will help your doctor big time, as doctors are usually genuinely interested in helping people feel better, and your doctor may even get excited to learn about a new treatment that has been found, and especially since one of his or her patients is willing to try it. Some may be skeptical or too set in their ways, however.

Leave a Reply

Your email address will not be published. Required fields are marked *