Quality & Quantity 36: 379–390, 2002.
© 2002 Kluwer Academic Publishers. Printed in the Netherlands.
The Rate Ratio for the Study of Disorders with
MARIBEL PERÓ CEBOLLERO and JOAN GUÀRDIA OLMOS
Departament de Metodologia de les Ciències del Comportament, Facultat de Psicologia, Divisió de
Ciències de la Salut (IV), Universitat de Barcelona, Vall D’Hebron, 171, 08035 Barcelona, Spain
E-mails: email@example.com and firstname.lastname@example.org
Abstract. Contemporary epidemiology bases its study of the relationship between an exposure factor
and a particular disorder on the ﬁrst occurrence of that disorder. However, many disorders present
more than one episode. This work aimed to develop a strategy for determining the rate ratio (RT) in
the study of disorders with relapses.
The adequacy of the rate ratio was determined by simulation. Six different situations were gener-
ated. In two situations there was no relationship between the exposure factor and the disorder and in
four situations the exposure factor was a risk factor. In order to generate these situations the individual
risk of relapse in the exposed group and in the unexposed group were manipulated. A total of 1000
data ﬁles were generated for each of the six situations. The total rate ratio for the relapse and for each
episode in all data ﬁles was obtained.
A descriptive analysis of the distribution of the rate ratio shows that this index is a good indicator
when there is no relationship. When the exposure factor is a risk factor the data didn’t show the
adequacy of the indicator so convincingly.
Key words: epidemiology, relapse, rate ratio, association study and cohort study
Many authors believe that the most important task in epidemiology is the study of
the ﬁrst occurrence of a disorder (Mac Mahon and Pugh, 1975; Rothman, 1986).
However, many disorders present more than one episode, and the epidemiolo-
gical study of these disorders has also attracted interest. Whilst recognising that
nowadays epidemiology only studies ﬁrst occurrences, Norell (1995) and Villalbí
Hereter (1995) suggest that the number of episodes in the incidence denominator
should be considered. Kleinbaum et al. (1982) and Morgenstern et al. (1980) sug-
gest that the incidence of a disorder in people who have multiple episodes should
be studied, either with respect to each episode or their total number.
Some studies have considered disorders whit relapse. Generally speaking, these
studies have looked at the ﬁrst relapse and tried to ﬁnd factors related with it (Fischl
et al., 1981; Centor et al., 1981; Brown et al., 1990; Tohen et al., 1992; Simkin and
Gross, 1994; Verheul et al., 1998; Fava et al., 1998; Robinson et al., 1999, etc.).