Clinicians receive data and interpret it to understand how to treat diseases. However, how does one determine if a study treatment helps with a disease? To do this, researchers use measures of clinical and statistical significance. Through them, the study’s relevance can be established, and the results explained. The two criteria are different because they present different results about the studied phenomenon. There are difficulties in interpreting statistical data, while clinical significance is a more precise and firmer criterion. Sometimes only clinical significance can be used in projects because it reflects the extent of the effect.
Comparison of Clinical and Statistical Significance
Statistical significance reflects confidence in the authenticity of the result obtained. It means that the event could not have occurred by chance, or such probability is minimal. The statistical data obtained are processed using several criteria: the main one is the P-value. It reflects the degree of confidence in the result obtained: a value below 0.05 is reliable, and the conclusions can be trusted (Pantaleon, 2019). However, higher values raise doubts and additional questions, so one should be critical of such results. Statistical significance is necessary to confirm that the effect under study was present. It works only with quantitative data, while clinical significance reflects the reality of practice. It registers results unrelated to the success of the study.
Clinical significance reflects the effect relative to the endpoints, takes into account the social attributes of the study subject, and registers the overall safety profile of the study. Many expert clinicians choose this criterion for decision-making because it allows them to assess differences in effect size (Pantaleon, 2019). The Jacobson-Truax approach assesses the change in reliability index (RIC) and indexes two main criteria: patient status before therapy and the change itself. The Edwards-Nunnally calculation method is a more detailed and critical approach, whose outcome scales are more multifactorial. Clinical significance, unlike the statistical system, is practice-oriented and provides insight into the extent to which the sample deviates from expectations.
Using Clinical Significance without Statistics
Any study always includes a sample on which an experiment is conducted. The size of the model and its qualitative characteristics can affect statistical significance. For example, the sample was only ten people: the results were satisfactory and showed a positive trend. However, can these results be applied to groups with thousands of people? Suppose the sample was low-variability, so the statistical significance showed reliability. However, when attempting to use another group, no result was obtained. It is appropriate to exclude statistical significance in both examples because they did not reflect facts but only showed cases in the sample studied. Thus, statistically significant data obtained do not consistently demonstrate effects in practice, and it is worth using only clinical significance.
Clinical significance can be used separately from statistical validity if it does not explain differences in patient information. Moreover, if the statistics differ little from each other, it probably will not reflect the need for an intervention. Armijo-Olivo discusses why clinical significance is more reliable than statistical significance. The author points out that statistics greatly limit research in physical therapy, so the severity of the effect is more important (Armijo-Olivo, 2018). Schober et al. (2018) analyze the reliability of the p-value for a study of patient temperature after cavity surgery (Schober et al., 2018). The authors emphasize that significant differences in the sample do not allow an unambiguous interpretation of the result.
Statistical significance represents the results of quantitative data processing and demonstrates their validity. Clinical significance focuses on determining the magnitude of the effect, so it is more subjective. However, the limitations imposed by statistical significance allow clinical significance to be used separately. It may be the case in projects that lack sample variability or show too slight or too much variation in statistics. Thus, clinical significance can act as an independent criterion in practice.