By Marie Johnston and Derek Johnston, University of Aberdeen, Scotland
Practitioners frequently want the answer to a problem which concerns one person, one health care team, one hospital or one region etc. For example, it may be important to know how often an obese man snacks, when and where he snacks and if stress makes it worse. Or you may wish to find out how often members of the healthcare team omit hand hygiene, if it is worse when they are under-staffed and if ward adverts improve it. Or you may be investigating sources of clinical errors to check if they are more common on some wards or for some grades of staff. Or, at a policy level, it might be valuable to investigate whether a new regulation, such as a smoking ban in public places has affected smoking rates.
You might try to answer these questions by asking people what they think or remember but it would be better to ask or observe at the critical times and places to avoid problems of bias and forgetting. Recent technological advances such as digital monitoring using smartphones make it easier to track what is going on in real time and an n-of-1 study might allow you to answer your question.
N-of-1 studies are possible when the problem can be assessed repeatedly to look at change over time. Then one can describe the problem and examine whether it is better or worse under some conditions. Or one may introduce a new intervention or treatment and assess whether it is having the proposed effect.
The simplest evaluation of the data collected is the observation of trends on a graph as in the illustrations below. This is an essential step in any n-of-1 analysis and can be sufficient. Additionally, there are methods of statistical analyses for n-of-1 studies. More complex methods continue to be developed (e.g., methods for assessing dynamic change ).
(more…)