APPLICATION OF THE NON-PROPORTIONAL RATES MODEL TO RECURRENT EVENT DATA: ANALYSIS OF RISK FACTORS FOR PRE-SCHOOL ASTHMA - Quantitative Medical Data Analysis Using Mathematical Tools and Statistical Techniques - page 189

Table 2
Risk factor prole of 1984/85 Manitoba birth cohort (n=16,207)

Characteristic Variable n %

Low birth weight (< 2500 g) LBW 703 4.3

Respiratory distress syndrome RDS 247 1.5

Transient tachypnea TTN 326 2.0

of newborn

Birth asphyxia
mild-to-moderate

ASPH.mod 918 5.7

severe ASPH.sev 307 1.9

Male gender MALE 8,357 51.6

birth cohort is provided in Tables 3 and 4. A total of 376 hospitalizations
and 1,377 hospital days were experienced by the 207 children hospitalized
for asthma during age 0 to 4. Among those hospitalized, 65% were hospi-
talized only once. The mode of the length-of-stay distribution was 2 days
(28%); almost 5% of hospitalizations were for 10 days or more.
Results of the non-proportional rates models are listed in Tables 5 and 6
for asthma-attributable hospitalizations and days hospitalized, respectively.
Signicant increases in hospitalization rates were associated with low birth
weight (Rate Ratio(RR)=2.03), respiratory distress syndrome (RR=3.65),
severe birth asphyxia (RR=3.41) and male gender (RR=1.89, at t =2.5
years), while an important but not statistically signicant increase was
associated with mild-to-moderate birth asphyxia. As indicated in Table 6,
all covariates having a signicant eect on mean number of hospitalizations
were also associated with signicant increases in mean days hospitalized.
For each parameter estimate, there was great disparity between the naive
and robust standard error estimates indicating strong intra-subject event
correlations, particularly for the days-hospitalized model (Table 6).
As indicated in both Tables 5 and 6, signicant departure from pro-
portionality was detected for the gender eect. The trend in the mean