ABSTRACT
As shown in Chapter 13, it is easy to parse a year’s worth of deidentified death certificate data contained in one of the CDC public use mortality files.
We have been using the 1999 mortality file, which contains about 2.3 million records. Each record may list up to 20 diseases, representing the underlying and proximate causes of death and any significant additional conditions that the certifying doctor deems noteworthy. When physicians fill out a death certificate, they can choose to be thorough, by listing all diseases that lead to the patient’s death, as well as other significant medical conditions affecting the patient at the time of death. Physicians can also choose to be somewhat lazy, listing the proximate cause of death, and omitting preceding and concurrent diseases. With a little effort, we can count the diseases listed in each cause of death record, and determine the average number of diseases reported for each state.
