Statistics. A word that can send shivers up even the most seasoned medical writer's spine. But statistics shouldn’t be feared or avoided; in fact, most of the time, it is simply misunderstood.
While you – the medical writer – may not be responsible for conducting the statistical tests or generating the analytical outputs, you are responsible for accurately interpreting and reporting on statistical results. It is your job to create a story around the data and communicate the health messages they contain.
Here you will find a handy cheat sheet covering the 5 essentials of statistics written by a medical writer for medical writers:
1) Measures of central tendencies and dispersion in statistics
Means and medians – both are measures of central tendencies and are important summary metrics for a dataset. The mean provides an average value and a snapshot of the data’s centre point, but outliers can skew it. The median is the middle value of a dataset and offers a more appropriate measure for skewed distributions, which are common in clinical data.
Knowing whether a dataset is normally distributed (symmetrical) or not normally distributed (skewed) is important to understand, as the type of distribution dictates the type of statistical test that is most appropriate.
The standard deviation and interquartile range are measures of dispersion and give insights into the variability or spread of the data.
2) Statistical significance (p-values) and confidence intervals
Statistical significance, commonly represented by p-values, helps determine whether the observed results in a study could have occurred by chance. The commonly used threshold of <0.05 represents a less than 5% probability that the observed results are due to random chance. It is important to remember that a statistically significant result does not necessarily imply clinical significance.
Confidence intervals provide a range of values within which the true risk estimate is likely to be found. Examples of risk estimates (to name a few) are relative risks, hazard ratios, or odds ratios, and are a way to quantify risk in a clinical dataset, i.e., the likelihood of an unfavourable outcome. Confidence intervals offer insights into the precision and reliability of the risk estimate. A narrow confidence interval suggests a more precise estimate, whereas a wide interval indicates more uncertainty.
3) Types of data and appropriate testing
The ability to recognise different data types and the correct corresponding statistical test helps a medical writer interpret data properly and convey the results accurately to their audience. There are two main types of data: numerical and categorical.
The below diagram describes the further sub-categorisation of these and some of the examples of the specific statistical tests that can be used for analysis.
4) Correlation vs causation in statistics
A correlation refers to a statistical relationship between two variables, indicating that they vary together. However, this does not imply that one variable causes the change in the other. This is known as causation, and being able to recognise the distinction is vital in medical research. Misinterpretation of correlation as causation can lead to improper conclusions, for example, about how effective a treatment is, or about the cause of a disease. Certain criteria must be met before a causality is determined, and medical writers must be careful in their language and wary of studies claiming correlation as evidence of causation without proper, rigorous experimental support for any confounding factors.
5) Study design and analysis techniques
Familiarity with various study designs, such as randomised controlled trials, cohort studies and case–control studies, and their analysis techniques is important for medical writers. It helps a writer to understand the strengths, limitations and applicability of a study’s findings, which translates to being able to effectively communicate the key messages generated from that study.
And lastly, one bonus tip to take away from this cheat sheet is awareness of bias in data. Bias refers to systematic errors in data collection, analysis, and interpretation, and can significantly skew research findings and lead to incorrect conclusions. Medical writers, equipped with knowledge of statistical analyses, can critically evaluate research for potential biases, ensuring accurate and reliable communications of health messages.
Here at Elion, we offer an expert team with vast experience in the interpretation of health data. Don’t hesitate to get in touch today to discuss your next medical writing project.
