Previous Next

CONCLUSION

Clinicians use statistical methodology both to analyze their own data and to assess the value of others' scientific reports. Because data analysis is generally the application of one of a number of commercial computer packages, each with its own menus, formatting, and subtle caveats,


890
the most valuable suggestion to the budding statistician is to know the program, read the manual, and plot the data. The human visual system is well adapted to see groups and trends—if the conclusions do not look like they fit the data, caution and skepticism are prudent.

When assessing the value of a scientific report, a careful clinician reads the "methods" section and asks the questions suggested earlier: Are the subjects selected to minimize bias? Are the statistical methods appropriate? Is the study large enough to be statistically powerful? Are the conclusions warranted by the evidence? Can the conclusions from this study be extended to cover more general cases or do they apply only to the limited situation under study?

Finally, despite the best statistical approaches and the most careful experimental procedures, a wise clinician recognizes that even some apparently well done experiments can yield erroneous conclusions. A study that produces an important clinical conclusion, especially one that seems to change tradition or violate previous theoretical understanding, bears confirmation by another group using different patients in a different setting. Healthy skepticism is the basis of good statistics and good science.

Previous Next