My experience as a public health director has taught me that solving complex
health problems requires identifying useful interventions. Th is book is my
intervention to make it easier (and more fun) to conduct intervention studies.
1
Many  different types of interventions can improve health and health
care, including drugs and medical devices, laws and changes in medical and
organizational practices. To enable you to evaluate a wide range of interventions, this book explains the advantages and disadvantages of many different
approaches: randomized and nonrandomized studies; prospective and retrospective studies; planned clinical trials and observational studies (e.g., studies
using administrative databases, disease registries); evaluations of investigatorinitiated interventions as well as evaluations of interventions created by others
(e.g., a law or a system change).
By covering both study design and statistical analysis in one book I hope to
save you from having to go back and forth between a textbook on study design
and one on statistics. Also, for many intervention studies, especially nonrandomized designs, understanding the statistical approach to analyzing the data
is key to judging the robustness of the design. Within the statistics section I
have organized the material the way that researchers think: what is the best
test to perform to answer a particular question? For readers who wish to learn
more about particular tests (e.g., McNemar test, mixed-eff ects model) I have
provided references in the footnotes.