Biomedical researchers often need to estimate population parameters from a small part of the population, known as the sample. However, the sample may not fully represent the population. Therefore, during statistical analysis, it is necessary to estimate the range of plausible values for the population parameter. Confidence intervals are a way of estimating this range and assessing how precise the population parameter is.
What are Confidence Intervals in Statistics?
A confidence interval is a range of values that is likely to contain the true population parameter with a certain level of confidence. For example, a 95% confidence interval for the mean blood glucose of a population would be a range of values that is expected to contain the true population mean 95% of the time. The confidence level represents the probability that the interval contains the true parameter, and it is typically set to 95% or 99%, and occasionally 90%.
Uses and Importance of confidence intervals in biomedical research
Confidence intervals are important in biomedical research for several reasons.
1. To determine the precision of the estimate
Confidence intervals allow researchers to assess how precise their estimates are. A narrow confidence interval indicates a precise estimate, while a wide confidence interval indicates a less precise estimate. This information is crucial for determining the clinical significance of your results. For example, if an intervention is found to lower blood glucose levels by 15 mg/dL with a confidence interval of [13, 17], this indicates a precise estimate and suggests that the intervention is likely to be useful clinically since all participants who underwent this intervention achieved clinically meaningful reductions. On the other hand, if the confidence interval is [5, 25], this indicates that the intervention had drastically varying effects on different participants and that in some participants, the decline in blood glucose levels was quite modest. Thus, the clinical importance of the findings may be questionable. In order to ensure that readers can quickly grasp which findings are clinically meaningful, NEJM, one of the top medical journals worldwide, strongly recommends using confidence intervals for effect estimates as well as ratio quantities such as odds ratios and relative risks.
2. To make inferences about the population
Confidence intervals allow researchers to make inferences about the population based on the sample. For example, if a study finds that participants with a history of kidney injury in childhood were more likely to develop end-stage renal disease as adults, with an odds ratio of 4.43 and confidence intervals of [2.21, 6.64], these findings are useful in identifying who among the population needs to be targeted in community measures to prevent end-stage renal disease.
3. To assess statistical significance
Confidence intervals can also be used to assess the statistical significance of the findings. A confidence interval that does not include the null value (0 for differences and correlations or 1 for odds ratios and relative risks) suggests that the findings are statistically significant at the chosen level of confidence. For example, if a study finds that a certain treatment is associated with a reduction in the risk of heart attack with an odds ratio of 0.9 and a confidence interval of [0.85, 0.95], this suggests that the findings are statistically significant at the chosen level of confidence (i.e., 95%, 99%, or 90%). Conversely, if the confidence interval is [0.6, 1.1], this suggests that the findings are not statistically significant at the chosen level of confidence. In fact, the British Dental Journal strongly recommends that authors report results in terms of confidence intervals rather than p-values.
4. To compare different studies
Confidence intervals are a means of comparing the results of different studies. If two studies estimate the same population parameter but have different sample sizes, the study with the larger sample size will typically have a narrower confidence interval and a more precise estimate. This information can be used to assess the consistency of the findings across studies and to inform meta-analyses.
5. To assess the robustness of the results
Finally, confidence intervals allow researchers to assess the robustness of the results to different assumptions and modeling choices. For example, if a study estimates a treatment effect with a certain model and set of assumptions, a sensitivity analysis can be performed by estimating the effect with different models and assumptions and comparing the resulting confidence intervals.
Conclusion
Confidence intervals are an important tool in biomedical research. They enable researchers to estimate the range of plausible values for the population parameter and to assess the clinical importance of their findings. It’s therefore not surprising that leading journals in biomedicine, including Pediatrics and four top urology journals, require authors to report confidence intervals for their main estimates. Therefore, it’s a good idea to include them in your paper, especially for primary analyses, to make your data more credible and enhance your reputation as a researcher.
Would you like advice from an expert statistician on how to calculate and report robust, high-quality data? Editage’s Statistical Analysis & Review services can help!
Comment