Statistical Analysis
Patients aged 60 years and over were identified as older
travelers with an age limit based on that used by many
travel insurance providers to define an older person and
were compared to patients aged 18–45 years as a young
adult reference population. Patients aged 46–59 years
were not included so that the comparison group of adult
travelers would have the greatest probability of differing
from travelers>60 years, in termof physiological status
and behavior during travel. Age groups were defined
prior to the statistical analysis. Data were entered into
and managed in Microsoft Access (Microsoft Corp.,
Redmond, WA, USA). In our evaluation, proportionate
morbidity refers to the number of cases of a specific
diagnosis (or of a group of specific diagnoses within
a syndrome group) compared with all cases of ill
travelers (excluding cases with the specific diagnosis
being investigated) of the same age group that were
seen at GeoSentinel clinics during the same period.
Thus, the proportionate morbidity is not an acquisition
incidence rate of travel-related illness and cannot infer
absolute risk.Differences in the proportions (categorical
variables) were tested using Fisher exact tests, and
Kruskal–Wallis testswereused for continuous variables.
pValues<0.05 were considered significant. Odds ratios
(ORs) (older travelers vs young adult travelers) by
diagnosis were estimated by logistic regression and
adjusted for travel reason, sex, pre-travel advice, region
of exposure, and clinical setting. The Mantel–Haenszel
statistic was used to test for diagnosis trends by age
classes. All statistical tests were two-sided. Percentages
and ORs (with 95% confidence intervals), comparisons,
and graphic analyses were carried out using the R 2.8.1
environment (www.r-project.org).
การแปล กรุณารอสักครู่..
