Single-subject experimental designs: a practical research alternative for practicing physicians.

Single-subject research designs offer a viable alternative to the more customary group-comparison designs. The flexibility and practicality of these designs make them particularly well suited for practicing family physicians interested in testing their clinical hunches. Three designs are described that are feasible in a practice setting: ABAB (reversal), multiple-baseline, and the alternating treatment design. Either visual inspection or statistical approaches can be used to evaluate these designs. By being aware of their limitations and by following simple practical steps, the practicing physician can use these designs to improve care of individual patients while simultaneously contributing to our general knowledge.