|The web is jam-packed with nutrition information of all kinds, and much of it cannot be trusted. You may hear something on the news and want to research it online. Resources from medical institutions, universities and government health organizations can give you trusted, accurate information on any subject.
A reliable search engine, called Medline Plus , can help you search for any topic of interest. You can be sure that you are getting truthful information, backed by scientific research, on this comprehensive medical search engine.
If you are feeling daring, you can perform a search at PubMed, provided by the US National Library of Medicine, and view the study abstracts found in scholarly health and medical journals and many of the research studies themselves. To get a good picture of what the present research says about a particular topic, you can read a recent literature review, which will list the studies that have been done and their outcomes and will draw a general conclusion based upon the existing research. A good literature review will cover all sides of an argument objectively and clearly and will provide an in-depth evaluation of the findings.
For the daring type, familiarize yourself with the different designs a researcher uses to conduct a study. A study design is chosen based upon the research objectives. For nutrition and medicine, epidemiological studies, case studies and randomized-controlled studies are often used.
Epidemiological studies, such as prospective and retrospective cohorts, observe a large group of people over time, determining links between certain behaviors and the end result. For example, researchers may choose an epidemiological study to observe the link between smoking and diagnosis of lung cancer. It is important to remember that a link does not establish a cause and effect relationship between the behavior and the end result. Strong epidemiological studies have a large number of participants and are performed over long periods of time, giving them added validity.
Prospective cohorts are best, enabling researchers to start with their participant group and record all of the necessary information required to build their conclusion at the end of the study. These huge databases of information can then be used to build new studies, as long as all of the required information was recorded initially. For example, a study of the link between smoking and lung cancer could be further studied to see if obesity might also play a role, as long as all of the pertinent information was recorded from the beginning of the study to enable the researcher to draw that conclusion.
A retrospective cohort is a bit more difficult to validate because the researcher chooses a participant group with a certain characteristic that he or she would like to study and then builds up the medical history of each of the participants with the use of medical records or questionnaires. The difficulty comes with collecting historical information. Medical records often do not offer a lot of information about lifestyle behaviors and other factors of interest and questionnaires are subject to “recall bias,” meaning that the information that the participant gives may not be completely accurate. Additionally, subjective information about moods or opinions is hard to capture accurately, since it tends to change day by day.
Another commonly used epidemiological study design is the cross-sectional study. A cross-sectional study takes a snap-shot of a certain population at a certain time. For instance, a cross-sectional study would be used to find the incidence of breast cancer in African American women. Cross-sectional studies tend to be complicated, expensive and difficult to analyze.
Unfortunately, the validity of epidemiological studies can fall prey to confounding factors, which are uncontrolled in the study, and may also have a significant effect on the validity of the study. For instance, let’s pretend that researchers are looking at a population of children. They want to see if there is a link between childhood obesity and school cafeteria food. They take the participants’ body measurements, record a detailed food record, thoroughly examine each medical chart and observe socioeconomic status, family life, and any predisposing genetic factors. Through all of this data, the researchers determined that there is indeed a significant link between childhood obesity and a higher fat content in school cafeteria food. What they didn’t record was a cofounding factor- the activity level of their participants- which may have skewed the results of their statistical analysis.
Once in a while, a good case study will get published in a peer reviewed journal and draw some attention of the media. Case studies are observational in nature and usually have very small numbers of participants, sometimes only one participant. Although case studies are extremely interesting and may be cause for further research, one is not able to draw a conclusion from a case study. For instance, if a treatment is used successfully on a very rare type of cancerous tumor and it was documented in a case study, this evidence is not sufficient enough to create a gold-standard protocol used to treat this type of tumor. It only suggests that more research needs to be done.
A randomized-controlled study randomly divides the study participants into either a control group or one or more treatment/intervention groups. The control group receives a placebo, instead of the true treatment or intervention, which allows the researchers to establish a baseline used to compare the changes seen in the treatment groups. A well designed randomized-controlled study always has a control group. A double-blind design is the best, meaning that the participants do not know the group to which they were assigned, and neither do the researchers. This eliminates any chance for bias during the study, and is more likely to show a true link between the treatment and the outcome. The results of the study are analyzed by using statistical methods to determine statistical significance. Although a randomized-controlled study is a strong study design, too much control can eliminate the “real life” scenario, making it less practical, depending upon its application.
Some studies are performed on animals and on cells in vitro, and the media tends to misguidedly draw a parallel from these studies to its application in the human body. Although many animals have similarities to humans, it is impossible to draw a conclusion about the human body from an animal study. Likewise, animal studies are usually undertaken for shorter periods of time with higher doses of the treatment or intervention, which would not correlate to humans in their natural environment. Since the environment is controlled in vitro, results may not be able to be replicated in vivo. Significant findings in these types of studies are an indication that more research is needed in this area and no conclusions can be drawn.
There is always room for error in even the most well-designed studies. Therefore, a conclusion about a cause-effect relationship cannot logically be drawn from one study alone. The entire body of research needs to be reviewed before a conclusion can be drawn. Additionally, a topic in its beginning phases of research needs much more research before a conclusion can be drawn. Luckily, research reviews and articles produced by the experts are readily available at your computer screen for you to be able to draw your own conclusion about any health topic up for discussion.