Retrospective Study Protocols

A retrospective study in the medical field involves documenting past experiences through interviews, reviews of medical records and other relevant data, in order to study a particular disease, condition, or factors involved in a problem. Each study establishes its own protocols. These protocols, or procedures to be followed, give the study its structure and direction. They provide guidance on how to proceed in reviewing the past.

Review of Medical Records

Establishing an outcome for the retrospective study requires determining the study's purpose. For example, a study might have the intention of determining lifestyles that contribute to a certain disease. One protocol or methodology might dictate reviewing medical records of individuals with the disease, looking for specific kinds of data. The protocol gives directions as to the type of data to be extracted from the records. In this case, annotations on the chart related to the patient's diet, exercise, work, stress, marital status would be recorded, with the understanding that they may or may not contribute to the disease.

Personal Interviews

Personal interviews involve speaking with subjects with the disease in order to supplement their medical charts with information not found in the charts. Since these interviews involve recollections over a period of years, the interviewer may deal with patients suffering loss or blurring of memory, or with the phenomenon of "selective memory", that is, the patient may only remember the favorable parts of their lifestyle, and not the part the interviewer is seeking to discover. Overcoming this problem involves a careful review of medical records prior to the interview and good questions designed to elicit the information needed. Any bias needs to be avoided in the questioning, with the questions being phrased to obtain key facts.

Sample Size

Determine at the outset of the study how many medical records and individuals are going to be examined. Too small a sample size can distort the findings. One guideline developed, according to an article about Methodology in "Conducting Chart Review for Retrospective Studies," is ten samples for every predictor or factor to be studied. For example, if the study determines that there are six lifestyle factors that lead to the disease, then sixty records would be reviewed, with corresponding interviews for as many of the individuals as feasible.