This study defines vulnerability as a multi-dimensional construct. reflecte
d in the convergence of predisposing, enabling, and need attributes of risk
. Using race, income, and self-perceived health status as indicators and ba
sed on eight rounds of the US 1996 panel of the Medical Expenditure Panel S
urvey, the study examined how the interactions of these vulnerable characte
ristics affect insurance coverage, a critical measure of health care access
. The results of the study demonstrate insurance coverage does vary with th
e extent of vulnerability. While race and income significantly influence in
surance coverage, respectively. there was relatively little disparity in in
surance due to health status. Between race and income, income was a more si
gnificant predictor of lack of insurance coverage since low-income people r
egardless of race and health were significantly more likely to be uninsured
or partially insured. However. it is important to note that minorities wer
e disproportionately over-represented in the low-income or bad health group
s so that any adverse association between income, bad health, and insurance
status would affect minorities significantly more than whiles. Among those
with insurance, the most vulnerable group, the minority-low-income-bad hea
lth group or those with all the three vulnerability indicators, were most l
ikely to be publicly insured. A policy implication is to target limited res
ources on insurance coverage for the more vulnerable groups. those with a c
onvergence or cluster of predisposing, enabling, and need attributes of ris
k. (C) 2001 Elsevier Science Ltd. All rights reserved.