8. Limitations
The present study has limitations. The most significant
threat to validity is the number of patients lost to followup
in the present sample. These losses were more frequent
in the control group than the experimental group, and
more common in less educated subjects. The study sample
was therefore more educated, and more willing to
participate in the study, and thus possibly more receptive
to the intervention. The loss of patients to follow-up,
especially in the control group, attenuates the efficacy of
randomization in a clinical trial. Therefore, a comparison of
baseline characteristics was carried out which revealed
four differences between the two groups based on pvalues
< .10 or clinical importance: driving status, ambulance
transportation to the hospital, mean number of
medications and gender. After controlling for each of these
variables (one at a time) the secondary outcome results did
not change.
Another limitation of the study is the fact that the study
hospital is dedicated to cardiac care. Although one-third of
the study patients were consulting the emergency
department for non-cardiac problems, this nevertheless
limits generalizability of the results. Other limitations to
generalizability are that the sample was restricted to
French- or English-speaking patients, as well as patients
discharged directly home from the emergency department.
Finally, the sample size was based on the primary
outcome of emergency department revisits, and no power
analysis was performed for the secondary outcomes
examined in the present paper.
Strengths of the trial include a large sample size,
allocation concealment, and blinding of the analyst.
Additionally, randomization was supervised by an external
coordinating center, and CONSORT guidelines were
followed throughout the study.