586
ABSTRACTS (ACE) AEP Vol. 13, No. 8
September 2003: 559–596
pattern of difference in SBP mean at age 37 for different levels of
duration of shift (0.61, 0.48, -0.40, 0.56, and –3.08 for Ͻ1 year,
1-4.9, 5-9.9, 10-19.9, & у20 years of age respectively) suggested
some evidence for selection-in and selection out bias. There was
some differences in results of DBP in comparison with SBP results.
doi: 10.1016/S1047-2797(03)00209-6
P076
WILLINGNESS TO PROVIDE BIOLOGIC SAMPLES:
RESULTS FROM A HEALTHY URBAN POPULATION
OF MEXICANS AND MEXICAN-AMERICANS IN
TEXAS
KC Saunders, SS Strom, A Garzon, MR Spitz, ML Bondy,
The University of Texas MD Anderson Cancer Center,
Houston, TX
PURPOSE: To evaluate the willingness of participants in a popu-
lation-based Mexican American cohort study to provide optional
biologic samples during baseline recruitment.
METHODS: Between July, 2001 and July, 2002, 1766 participants
aged 13 - 88 who self-identified as Mexican or Mexican American
were recruited using four methods of recruitment in Houston, Texas:
random digit dial, intercept, blockwalk (or door-to-door), and
network. Informed consent was administered and epidemiologic
data were collected through an in-person interview. Participants
were informed about the prospective nature of the study and given
the option of providing one or more of the following biologic
samples: blood, buccal cells, and/or urine. Participants could choose
to provide no sample and still participate in the cohort study.
RESULTS: Overall, 808 participants (46%) provided a blood
sample, 1537 participants (87%) provided a buccal sample, 1048
participants (59%) provided a urine sample, and 1686 participants
(96%) provided a DNA sample (either blood or buccal). Only 1%
of all participants provided no sample at all. In examining the
data by gender, women were more likely to provide a blood
sample (47% vs. 43% in men) and less likely to provide a buccal
sample (86% vs. 89% in men). There was no difference in their
willingness to provide a urine sample (59% for both men and
women).
CONCLUSION: To our knowledge, this is the first study to
evaluate the willingness of participants to provide a biologic sample
in a population of healthy Mexicans and Mexican Americans
living in the USA. Overall, the percentage of participants willing
to provide optional samples was higher than anticipated in this
underserved urban population, and there was little difference by
gender. The most notable finding was that in participants over
the age of 13, we were able to obtain samples suitable for DNA
extraction from 96% of the study population.
doi: 10.1016/S1047-2797(03)00210-2
OUTCOMES RESEARCH
P077
AMBULATORY CARE SENSITIVE HOSPITAL
ADMISSIONS THAT AREN’T: CARE PRIOR TO
HOSPITAL ADMISSION IN TORONTO, CANADA
RH Glazier, EM Badley, JE Gilbert, MA Agha, MI Creatore,
R Moineddin, Inner City Health Research Unit, St. Michael’s
Hospital, Toronto, Canada
PURPOSE: Ambulatory care sensitive (ACS) hospital admissions
are believed to be related to lack of timely and appropriate primary
care, yet little direct evidence supports this concept. We wished
to determine the relationship between ACS admissions, income
and ambulatory physician visits prior to hospital admission in a
universal health insurance setting.
METHODS: Case-control, case-crossover and retrospective
cohort analyses were used to investigate patterns of physician visits
prior to ACS admission. Mean household income was derived
from the 1996 Canadian census. For the case-crossover design,
the proportion of admitted individuals with a physician visit
prior to, and not one year before, admission were compared with
those without a physician visit prior to, and with a visit one year
before, admission.
RESULTS: Out of 610,880 adult residents (aged 20ϩ) of Toron-
to’s inner city, there were 12,101 hospitalizations for ACS condi-
tions between 1998 and 2001. Age-sex adjusted rates of ACS
admissions were considerably higher in the lowest income quintile
than in the wealthiest (rate ratio ϭ 1.7, 95% CI 1.6, 1.8). Among
those with an ACS admission, few income differences were found
in the proportion with a physician visit prior to an ACS admission
(61% lowest income quintile, 60% highest), with a physician visit
prior to admission but not during the 12 months before the admis-
sion (54% lowest income, 54% highest) or in the number, type
and continuity of physician visits in the year prior to admission.
CONCLUSION: There was an important income gradient in
ACS admissions, yet we found little relationship between income
and ambulatory care prior to admission. Use of ACS admissions
as an indicator of access to ambulatory care needs to be re-exam-
ined and other explanatory relationships sought.
doi: 10.1016/S1047-2797(03)00211-4
P078
THE VALIDATION OF A EUROPEAN PROGNOSTIC
INDEX IN AMERICAN BRAIN TUMOR PATIENTS
CG Lis, JF Grutsch, J Leestma, EH Mkrdichian, LJ Cerullo,
Chicago Institute for Neurosurgery and Neuroresearch, Illinois,
Cancer Treatment Centers of Americaă at Midwestern
Regional Medical Center, Zion, IL, University of Illinois at
Chicago School of Public Health, IL
PURPOSE: The objective of this study was to determine whether
the United Kingdom’s Medical Research Council (MRC) Brain
Tumour Index predicts the survival of American brain tumor
patients. The MRC Index (MRC) was developed and validated
in clinical trials conducted in the United Kingdom during the late
1970’s. We applied the MRC to patients treated in the United
States who did not participate in clinical trials.
METHODS: We used the MRC Brain Tumour Index to risk
stratify the survival of 506 consecutive patients who underwent
treatment at the Chicago Institute of Neurosurgery and Neurore-
search between October 1987 and June 1997. The MRC Index
was calculated from the following: age in years, WHO performance
status, extent of neurosurgery, and history of seizures. A score can
range from 0 (best prognosis) to 38 (worst prognosis). The MRC
Index divided patients into six distinct prognostic categories.
RESULTS: The application of the MRC Index divided glioma
patients into statistically significantly different strata (Wilcoxon
2 tailed p Ͻ 0.0001, chi square 142.2). Median survival declined
in a stepwise fashion with an increasing MRC score (96 weeks for