Evidence Made Easy
A student nurse’s roadmap to understanding clinical study jargon.
This is for you if you need to read and/or discuss research in your role – that should include everyone in medicine and nursing. When there is so much information and misinformation around, it’s often hard to see what’s sense and what’s nonsense.
It’s about a 12-minute read, so…help yourself to a beverage, put your feet up for a moment and take it slowly.
There’s a video and a 15-minute podcast which are worth your time to put the subject in context and help you understand and talk about it more with more authority.
Let’s get started.
Hopefully, your school of nursing has helped you unpick research papers so you have a basic understanding of what is valid and what’s not.
This article is designed to help you revisit and build on that. There’s a lot to this topic, so I’m going to cover the basics so you’ll have an initial checklist of what to look for. That potentially makes reading papers easier and quicker.
Just because it’s in black and white in a journal or on the internet, it’s not necessarily true. If you’ve spent more than five minutes watching world news and commentary, this won’t be a surprise.
Good science and research will strike a humble balance between answering questions plainly and making clear where there are more questions to answer.
Science is ‘this is what the evidence shows us now, in this study’. Not ‘this is it, the holy grail and the one-and-only answer to the problem/question’.
Here, we’re talking about research into new medicines, treatments, therapy, or getting information about how the body works. But research goes on every day outside medicine and affects us even if we don’t take any medicines or need treatment.
For example, outside healthcare, advertisers will try to blind us with numbers and statistics to make us buy. Some countries have tighter regulations over this than others, but if 77% of women who tried SilkiSoft Shampoo liked it, would you buy it? What other questions should you be asking about the research? We’ll come back to that later.
When you read a paper:
Think about the disease being studied.
Now think about the people who get it.
Who are they?
Is the condition common or rare?
If you were a patient, would the information in the paper be useful to you?
Look for the following:
1. The abstract
2. The results/conclusion
3. The type of trial/study
4. The numbers, ages, genders, social class, and ethnicities of people in the trial/study
5. The length of the study
6. Where the study was conducted
7. Who funded the study
8. The journal where it was published.
This is what each of these means.
1. The abstract
This is the first part to the paper. It summarises what the problem is/was, what researchers were trying to achieve, and what the results were.
It’s short – about 200-300 words.
If you’ve done a broad internet search on a topic, the title and abstract tell you if this aspect is one you want to read more about.
2. The results/conclusion (end point)
As the name suggests, this should tell you if the researchers achieved their aim in the study. It should be written clearly so you know ‘yes, they managed it’ or ‘no, they didn’t’.
There may be more than one end point. A primary end point and a secondary one. Sometimes they may achieve one and not the other.
The researchers should also make recommendations for further study here. This is useful because they are acknowledging the shortcomings in their research, and honesty in science is what good science is about.
3. The type of study or trial
There are lots of different types of clinical trials. Here are four common ones:
a) randomised controlled trials (RCTs),
b) cohort studies,
c) case-control studies and
d) qualitative studies.
Read more about these here, but here’s the top-line detail for each.
a) A randomised controlled trial takes a group of subjects and divides them randomly into a group taking a treatment and a group taking either a placebo (sometimes called a ‘sugar pill’) or another, related treatment, or no treatment. The other/related treatment will be either an older or another ‘class’ of drug.
For example, a drug trial for patients with heart failure might look at how effective an ACE inhibitor drug is versus a beta blocker. They may also look at a third group of patients taking an ACE inhibitor and a beta blocker together, to see which is the most effective.
A double-blind RCT means neither the researcher nor the patients know which treatment they’re receiving. This stops any unconscious bias from the researcher influencing the patient and therefore, influencing the outcome of the study.
A double-blind RCT is often thought to be the best type of study, but it depends on what’s being researched and the design of the trial. It’s not always appropriate.
b) Cohort studies are useful to look at how certain factors can influence disease and how disease can influence people’s lives.
A cohort is a group of people who are observed, often for several years. For example, we might look at people who do a certain job where they’re exposed to certain chemicals. We’d want to see what effect the chemicals have on their health.
Listen to this podcast to find out about cohort studies and how useful they can be.
c) Case-control studies look at groups of people who have a condition (the ‘cases’) and those who don’t (the control group).
The cases are matched with people of the same age, gender, race, and other physical similarities who don’t have the condition. This allows scientists to study risk factors for a condition, its symptoms in different sets of people, and the differences in disease trajectory (how it plays out in a person’s life – does it get serious and cause death in some and not others, etc).
For example, is liver cirrhosis more common in men who drink a certain level of alcohol than in men who drink less? Are women with the herpes simplex virus more likely to get cervical cancer than those who don’t have it?
This type of research can involve interviews, questionnaires, or researchers looking at data in patients’ notes and case files.
d) Qualitative studies enable us to learn what the patient experience is like. What’s it like to live with a condition?
This research involves talking to patients and using questionnaires to gather their thoughts and experiences.
The information is analysed and interpreted using different methods.
There are advantages and disadvantages to each method. Some are very expensive to run; sometimes it’s difficult to find the right number of patients to create enough quality data; sometimes a researcher will choose the wrong type of study to find the answer to their question.
4. The numbers, ages, genders, and ethnicities of people in the trial/study
Contrary to what we’re taught in school biology class, humans aren’t all built the same.
Different races have differing risks for different diseases. For example, people of Southeast Asian and Afro-Caribbean backgrounds are more prone to developing Type 2 diabetes, and Chinese people are less likely to get COPD.
Men and women respond differently to different treatments and diseases.
Older people metabolise drugs much more slowly than younger people.
So, you’d want the study test subjects to represent the people with the diseases you’re trying to treat. They need to be the right age, gender mix, and ethnicity. It’s no good studying a new drug for prostate cancer on a group of white men in their 50s. Prostate cancer affects men of all ethnicities, mainly after the age of 65.
And to be sure to get enough information in the study, it needs to include enough people. If the condition we’re studying is rare, that’s hard, but in a common condition like diabetes, COPD, or autism, there’s no excuse.
Remember the SilkiSoft Shampoo ad from earlier? If 77% of the women sampled said they liked it, is that a good recommendation? If it’s 77% of 100 women, no, it might be. But what if 10 women didn’t know the criteria to use for ‘liking’ SilkiSoft? Is it the feel, smell, how it lathered, how their hair felt? What if 20 women didn’t return their questionnaires?
There’s plenty of room for the numbers to lie to us when they’re small.
And millions of women use shampoo – if the numbers are going to reflect what women think, they need to ask more women. A better number for such a common product would be 10,000 women. That way, the little quirks of maths and statistics get ironed out.
5. The length of the study
Some studies are appropriately short. Some are long, even multigenerational. Some are too short to show statistically significant results.
A study looking at the effectiveness of a new antibiotic would be quite short. We want to know how long it takes to cure the infection. We might also want to know if there are relapses or problems after the treatment is complete, and this would be a secondary endpoint for the study.
A study looking at a long-term condition, like Parkinson’s Disease or diabetes, needs to be long enough to get numbers that reflect the amount of time a patient might use the treatment.
These studies are often done over multiple months. The results are extrapolated (an educated prediction as to what might happen in the future) for the patients taking it for the long term.
6. Where was the study conducted?
You’d be more likely to trust information from a prestigious university than if someone had been doing experiments in their garden shed.
Is the study centre accredited – does it have an external body approving its activities?
Who were the researchers? Are they doctors, scientists and people qualified to look at the topic?
7. Who funded the study?
Drugs and other treatments can take a decade or more to come onto the market. That means an investment of 100s of millions of $, £ or any other currency. A company funding the research won’t see a penny of its investment until people are buying the product.
The drug treatments we use are mainly developed by pharmaceutical companies. They answer to shareholders, and they need to justify the millions they spend on research and development (R&D).
There are strict rules around what pharma companies can and can’t do in their research.
Research that’s been done by independent bodies (universities, for example), without any commercial sponsorship, may be less biased. Read more here.
8. Which journal was the research published in?
Just as you’d trust research from a reputable research centre, you’d also want to see it published in a serious journal, not the National Enquirer.
Reputable journals like JAMA, BMJ, The Lancet etc, have expert reviewers to cast a critical eye over what researchers send them.
To find whether a study has been published in a reputable journal, look at PubMed®.
PubMed® is a service of the US National Library of Medicine. You can look up a database of research papers for free. This is the abstract in medical, nursing, dental, veterinary, health and preclinical science journals.
In 2025, there are questions about how reliable PubMed® will be, given the political interventions it’s facing.
Alternatives include the UK’s version: UKPMC, Europe’s PubMed® and Lens.org. Brown University lists more here.
The nurse’s role in helping patients decide on treatment
Science should be unbiased and reproducible; however, we all have beliefs and biases we carry with us. If you believe HIV/AIDS is a curse from god, vaccines are harmful, abortion is wrong, or that contraception should only be available to married people, as a healthcare professional, it’s your job to keep your beliefs to yourself.
It’s harsh, but we are here to care for the patient according to current evidence-based guidelines. We use these to help them make wise health and treatment decisions based on the best available science.
In other words, your beliefs can guide your health and treatment decisions, but they must not interfere with the patient’s.
Until next week,
All the best,
Elspeth.