Why are White Nursing Home Why are White Nursing Home Residents Twice as Likely Residents Twice as Likely as African Americans to as African Americans to Have an Advance Directive? Have an Advance Directive? Understanding Ethnic Understanding Ethnic Differences in Advance Care Differences in Advance Care Planning Planning Jennifer L. Troyer Jennifer L. Troyer Departments of Economics and Health Behavior and Departments of Economics and Health Behavior and Administration Administration University of North Carolina at Charlotte University of North Carolina at Charlotte William J. McAuley William J. McAuley Departments of Sociology and Gerontology and Communication Departments of Sociology and Gerontology and Communication Center for Social Science Research Center for Social Science Research George Mason University George Mason University
23
Embed
Why are White Nursing Home Residents Twice as Likely as African Americans to Have an Advance Directive? Understanding Ethnic Differences in Advance Care.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Why are White Nursing Home Why are White Nursing Home Residents Twice as Likely as African Residents Twice as Likely as African Americans to Have an Advance Americans to Have an Advance Directive? Understanding Ethnic Directive? Understanding Ethnic Differences in Advance Care PlanningDifferences in Advance Care PlanningJennifer L. TroyerJennifer L. TroyerDepartments of Economics and Health Behavior and AdministrationDepartments of Economics and Health Behavior and AdministrationUniversity of North Carolina at CharlotteUniversity of North Carolina at Charlotte
William J. McAuleyWilliam J. McAuleyDepartments of Sociology and Gerontology and CommunicationDepartments of Sociology and Gerontology and CommunicationCenter for Social Science ResearchCenter for Social Science ResearchGeorge Mason UniversityGeorge Mason University
IntroductionIntroduction
Providers, family members, and patients Providers, family members, and patients often spend a considerable amount of time often spend a considerable amount of time weighing treatment options when a patient weighing treatment options when a patient is seriously ill and nearing the end of life.is seriously ill and nearing the end of life.
Patients may make wishes about treatment Patients may make wishes about treatment decisions known through advance care decisions known through advance care planning.planning.
Patient Self-Determination Act of 1991Patient Self-Determination Act of 1991
Purpose of StudyPurpose of Study
The rise in the adoption of advance directives has not The rise in the adoption of advance directives has not been equal across ethnic groups.been equal across ethnic groups.– African American nursing home residents are much less African American nursing home residents are much less
likely to have an advance directive.likely to have an advance directive. Question remains unaddressed by researchers:Question remains unaddressed by researchers:
– Can we explain part of the ethnic gap in advance care Can we explain part of the ethnic gap in advance care planning by considering group differences in the following?planning by considering group differences in the following?
Personal factorsPersonal factors Micro-environmental characteristics of the facilityMicro-environmental characteristics of the facility Social and economic environment represented by the Social and economic environment represented by the
county in which the facility is located county in which the facility is located
LiteratureLiterature
Many studies find ethnic differences in advance Many studies find ethnic differences in advance care planning when controlling for some resident care planning when controlling for some resident characteristics.characteristics.– Demographic characteristicsDemographic characteristics– Health status characteristicsHealth status characteristics
One study (Castle and Mor 1998) looks at how One study (Castle and Mor 1998) looks at how facility characteristics influence advance facility characteristics influence advance directive adoption.directive adoption.
LiteratureLiterature
Few studies consider geographic variation.Few studies consider geographic variation.– Castle and Mor (1998) – 10 statesCastle and Mor (1998) – 10 states– Kiely et al. (2001) – 4 statesKiely et al. (2001) – 4 states– Levin et al. (1999) – 3 regionsLevin et al. (1999) – 3 regions– Buchanan et al. (2004) – rural/urbanBuchanan et al. (2004) – rural/urban
Primary Research ObjectivePrimary Research Objective
To determine the extent to which ethnic To determine the extent to which ethnic differences in advance care planning are differences in advance care planning are attributable to differences in attributable to differences in – the personal characteristics of African the personal characteristics of African
American and White nursing home residents,American and White nursing home residents,– the facilities in which they reside,the facilities in which they reside,– and the counties in which the facilities are and the counties in which the facilities are
located. located.
Data Data
Medical Expenditure Panel Survey – Nursing Medical Expenditure Panel Survey – Nursing Home Component from the Agency for Home Component from the Agency for Healthcare Research and Quality Healthcare Research and Quality – The data include a nationally representative sample of The data include a nationally representative sample of
individuals who were nursing home residents as of individuals who were nursing home residents as of January 1, 1996.January 1, 1996.
– Include information on facility characteristicsInclude information on facility characteristics– Matched to Area Resource FileMatched to Area Resource File– 93% White or African American93% White or African American– Use 2,665 (of 3,209) residents in 730 nursing homesUse 2,665 (of 3,209) residents in 730 nursing homes
DataData
Measuring the presence of an advance directive– At least one of four types of advance care
plans:Living willDo not resuscitate orderDo not hospitalize orderLimits on feeding, medication, other
Geographic (Market) Characteristics – County– Urban/Rural– Per Capita Income– % High School– % Poverty– % Black– % 65 or older
Data: Group Differences in Data: Group Differences in Mean CharacteristicsMean Characteristics
63.4% of White residents have advance directive vs. 27% of African-American residents
Resident Characteristics– Differences in education – whites more likely to have high school
diploma– Differences in presence of living child – whites more likely to have a
living child Facility Characteristics
– African-American residents more likely to be in a for-profit facility– African-American residents in facilities with higher proportion of
Medicaid-funded residents County Characteristics
– African-American residents in counties with lower rates of high school completion
MethodsMethods
Probit Estimates: Full SampleProbit Estimates: Full Sample– The effect of ethnicity on the probability of having The effect of ethnicity on the probability of having
an advance directive is estimated, controlling for an advance directive is estimated, controlling for resident, facility, and county characteristics.resident, facility, and county characteristics.
– Probability of Having an Advance Directive = Probability of Having an Advance Directive =
f ( ethnicity, resident characteristics, facility f ( ethnicity, resident characteristics, facility characteristics, county characteristics)characteristics, county characteristics)
– Gives us the effect of ethnicity on the Gives us the effect of ethnicity on the probability of advance care planning, when probability of advance care planning, when controlling for other factors.controlling for other factors.
MethodsMethods
Probit Estimates Using Sub-Samples: African American and White
– Estimates are done using two samples: African American residents White residents
– Probability of Having an Advance Directive = Probability of Having an Advance Directive = f (resident characteristics, facility characteristics, f (resident characteristics, facility characteristics, county characteristics)county characteristics)
– Estimates show whether characteristics have the same impact on the likelihood of having an advance care plan for the two groups.
MethodsMethods
Using Probit Estimates from Sub-Samples: African American and White
– Estimates may be used (with sub-sample means) to determine how much of the difference in the probability of having an advance advance directive directive between the two ethnic groups may be attributable to differences in average group characteristics.
– Probability Gap Between the Two Groups = (Portion Explained by Differences in Group Characteristics) + (Portion Unexplained by Group Differences)
– Portion Explained by Differences in Group Characteristics can be broken down further to look at:
How much resident, facility, and market characteristics help to explain the probability gap.
How much each measured characteristic helps to explain the probability gap.
MethodsMethods
Estimated with whole sample and two sub-samples Using the estimates of for the African American
sample ( ) and White ( ) sample, the vectors of African American (XiB) and White (XiW) characteristics, and the size of the African American (nB) and White (nW) samples, the predicted probability of having and advance directive may be computed for each of the two samples, S=B and S=W:
iS iS iSPr( AdvanceDirective 1| X ) ( X )
B W
∑1
ˆ1ˆSn
iSiS
SS X
nP
MethodsMethods
Then, the predicted difference in the probability of advance directive adoptions between the two groups is:
-W B
ˆ ˆ ˆG P P
MethodsMethods
Using a weighted average of the estimated coefficients for African American and White residents, *, the degree to which is explained by differences in the measured characteristics of African American and White residents is:
G
W Bn n
* *iW iB
i 1 i 1W B
1 1ˆ ˆExplained X Xn n
Marginal effects presented in tables.– For continuous variables in Xi, such as per capita income in
the county, the marginal change in the kth continuous variable, Xk, on the probability of having an advance directive
– For binary variables in Xi, the effect of switching the kth binary variable from 0 to 1 is
MethodsMethods
ˆˆˆ
XX
XM k
k
iik
)ˆ()ˆ(
ˆ01
ikikk
iik XX
X
XM
ResultsResults
Full Sample:Full Sample:– When controlling for resident, facility, and market characteristics, When controlling for resident, facility, and market characteristics,
African American residents are 23% less likely to have an advance care African American residents are 23% less likely to have an advance care plan.plan.
Sub-Sample Estimates:Sub-Sample Estimates:– For both groups:For both groups:
Alzheimer’s Alzheimer’s more likely to have advance directive more likely to have advance directive High poverty and more 65+ High poverty and more 65+ more likely to have advance more likely to have advance
directivedirective– Differences:Differences:
Long stay (2+ years) increases the probability of having an advance Long stay (2+ years) increases the probability of having an advance directive for African American residents but not white residentsdirective for African American residents but not white residents
Urban location and more poverty Urban location and more poverty less likely to have advance less likely to have advance directive for African Americans, but no effect for white residentsdirective for African Americans, but no effect for white residents
ResultsResults
Table Three: Explained vs. Unexplained Differences in Advance Directives Between African-American and White Residents
Probability of Advance Directive: White 63.66%
Probability of Advance Directive: African-American 27.03%
Percentage Point Difference in Probability of Any Directive 36.63%
Decomposition of Gap in Probability of Advance Directive Total Gap
Percent of Gap
Explained by Differences in Group Characteristics 16.02% 44%
Unexplained by Differences in Group Characteristics 20.61% 56%
Percentage Point Difference in Probability of Any Directive 36.63% 100%
ResultsResults
Contribution of Resident, Facility, and County Characteristics to Explaining Gap (44% Explained)
Portion of Gap Explained by Resident Characteristics: Demographic 8.68% Portion of Gap Explained by Resident Characteristics: Health Status 3.54% Portion of Gap Explained by Facility Characteristics 7.75% Portion of Gap Explained by County Characteristics 24.03%
Important Resident CharacteristicsImportant Resident Characteristics
– High-school diploma, presence of living child, age 85+High-school diploma, presence of living child, age 85+ Important Facility CharacteristicsImportant Facility Characteristics
– Higher proportion of Medicaid-funded residentsHigher proportion of Medicaid-funded residents Important County CharacteristicsImportant County Characteristics
– Higher proportion of individuals in poverty and metropolitan Higher proportion of individuals in poverty and metropolitan areaarea
ConclusionsConclusions
Consistent with prior research, we find a significant and Consistent with prior research, we find a significant and large difference in the probability of advance care large difference in the probability of advance care planning for white vs. African American residents.planning for white vs. African American residents.
Contribution of our paper:Contribution of our paper:– Nationally representative sampleNationally representative sample– Consider three levels of explanatory variablesConsider three levels of explanatory variables– Determine the relative impact of variables on the African Determine the relative impact of variables on the African
American-White difference in advance care planningAmerican-White difference in advance care planning Nearly half of the ethnic gap in advance care planning is Nearly half of the ethnic gap in advance care planning is
attributable to differences in group characteristics.attributable to differences in group characteristics. Facility and market characteristics do play an important Facility and market characteristics do play an important
role. role.
ConclusionsConclusions
Suggestions for Future Research and Suggestions for Future Research and LimitationsLimitations– Location-specific MeasuresLocation-specific Measures– Facility-specific MeasuresFacility-specific Measures– More resident Health Status controlsMore resident Health Status controls– More information on religious values, More information on religious values,
attitudes, expectations of residentsattitudes, expectations of residents