Issue link: https://www.massageandbodyworkdigital.com/i/918051
intervention's conclusion, the 26 percent improvement in ODI score included benefit in several meaningful life functions such as lifting, sex, and travel. While changes were not as robust for the second example participant, anyone who has experienced chronic pain conditions knows the big impact changes like this make in life from a pain-management perspective. 3. STATISTICAL VERSUS CLINICALLY MEANINGFUL The real-world massage therapy for CLBP study highlighted both statistically and clinically meaningful results, which is sometimes a point of confusion. You may even be asking yourself: What's the difference? To answer this question, I will try not to get too deep into the statistical weeds. Essentially, if there is a large enough sample, small differences between groups can become statistically significant. On the flip side, in small samples, differences between groups have to be really, really large in order for statistical significance to be achieved. This can be problematic for two related, but distinct, reasons. First, if a little change is statistically significant in a large (or small for that matter) sample but doesn't matter in a person's life or in a practical setting (say a 0.05 millimeter decrease in pain score on a 100 millimeter line), who cares if that outcome is statistically significant or not? It makes no real difference in an applied example or for an actual person. Second, exclusive focus on p-values and statistical significance can really miss important, patient-centered outcomes. If a researched intervention consistently demonstrates meaningful change in people's lives or for particular conditions, should the value or worth of said intervention be totally based on whether a study's p-value achieves <0.05? This issue becomes particularly challenging for situations in which research with large samples is limited due to funding, availability of eligible participants, or other research barriers. In your research reading, look for the extent to which study results are framed in terms of established, clinically meaningful outcomes. Or, if clinically meaningful minimums are not available or reported, consider what a study's research outcomes would mean in a person's life if applied at an individual level. This will help contextualize research numbers, especially for those intimidated by statistics. This study addressed the statistical versus clinical significance question in a couple of ways. Researchers conducted a power analysis prior to beginning the study to determine the needed sample size to detect the minimum clinically meaningful change for the primary outcome measure (as noted in the Oswestry Disability Index discussion, a change of 6 percentage points or more is considered clinically meaningful for the ODI). In addition, study analysis results were repeatedly put in the context of the extent to which clinically meaningful change was achieved. These attentions to clinically meaningful change by this research team reflect a patient-centered consideration of study results and provide massage therapists with a practice-relevant metric by which their clients' massage outcomes for pain and disability related to back pain can be compared. 4. CHARACTERISTICS THAT MIGHT INFLUENCE MASSAGE EFFECTIVENESS The study's exploratory analysis uncovered some interesting participant-related factors that may have influenced massage effectiveness. Specifically, analysis determined that after the intervention window, those who were 50 years and older had significantly better outcomes than those who were younger based on mean ODI change scores (4.6 ± 9.7 for younger participants versus 10.3 ± 10.8 for older participants; p=0.01). The astute reader will note that the average ODI change score for younger participants was not clinically meaningful (less than a 6 percentage point change) while those who were older had an average ODI change score well above what is considered 48 m a s s a g e & b o d y w o r k j a n u a r y / f e b r u a r y 2 0 1 8 Definition of P-Value P-value is short for probability value and is used in statistical calculations for hypothesis testing to indicate the data's significance. Many factors play into a statistical test's resultant p-value including sample size, power, and effect size. The p-value indicates the extent to which the compared data is different from each other. A low p-value (usually ≤0.05) indicates statistically significant differences, while those greater than 0.05 indicate any differences found are not statistically significant. The lower a statistical test's p-value, the stronger the data conclusions become from a statistical standpoint because the probability of erroneous conclusions (that there actually are not differences between the compared groups despite the significant p-value) is lower. Thus, a p-value of ≤0.01 indicates there is only 1% or less chance of erroneous conclusions while a p-value of ≤0.001 indicates there is only a fraction of a percentage chance of the conclusion being wrong.