|
Introduction to Health Services Research : A Self-Study Course
Module 5: Quality Filtering and Evidence-Based Medicine and Health (page 11 of 15)
Introduction | Sampling | Assignment | Assessment | Analysis | Interpretation | Extrapolation
Extrapolation (to larger groups)
Readers of research studies must decide if the results can be used for groups or in amounts different from the study. Can the tons of heart disease studies be extrapolated from the male participants to women? Can a health promotion study conducted in Asia be applied in the United States? The authors cannot answer these questions, but the consumers of the literature can ask themselves several questions before they "adopt" the study:
- What dose or level of intervention was studied? Will increasing that level yield the same results?
- What time period was studied? If women successfully used estrogen replacement therapy (ERT) for a year without adverse effects, can ERT be extended for many years?
- Does the study population live under similar conditions or share common cultural characteristics as another group? Will urban and rural residents respond to health messages in similar ways?
Generalizations - Sometimes a Problem
Beware of research articles that end with sweeping generalizations to vastly different groups of people. Make sure that the authors or readers don't extend the data farther than they can reach!
Beware of research articles that end with sweeping generalizations to vastly different groups of people. Make sure that the authors or readers don't extend the data farther than they can reach!
The following conclusion demonstrates restraint in extrapolating beyond the study:
These results should not be generalized to rural areas throughout the United States. Although the (study group) is representative of the rural South, it is not representative of one of the major barriers to health care in rural areas in other areas ... The reader must also take caution in generalizing these findings (access to health services between urban and rural elderly people) to younger age groups, especially children ...(Blazer, 1995).
Discussion Questions
- As an expert in quality filtering what questions should you ask before you "adopt" or recommend a study to a consumer, physician or health services researcher?
- What mistakes can readers make when examining a study's data? Why is extrapolating from one study's results a potential problem? Can you identify someone who has taken the data further than it can go? What was the tipoff?
|
