Table 31Summary of studies of the relationship between health literacy and disparities (KQ 1d)

Authors, Year, Study Design, Analysis Sample Size, QualityPopulation and Setting, Health Literacy LevelVariables used in Multivariate AnalysisOutcome Measure

Outcomes By Health Literacy Level
Differences in Results Between Health Literacy Levels
Sentell and Halpin, 2006141

Cross-sectional

N = 23,889

Fair
National sample of adults

Total NALS score
Level 1: 20%
Level 2: 27%
Level 3: 34%
Level 4: 18%
Level 5: 2%
Race
Education
Understand English
Born in U.S.A.
Unemployed
Family income
Income missing
Sex
Age
Married
Get food stamps
Live in Metropolitan
Statistical Area
Region
Self report of physical, mental, or other health condition that keeps respondent from working

Data: NR

Long-term illness (greater than 6 months)

Data: NR
Health literacy mediates the association of black race on having a condition that keeps you from work (adjusted):
Odds associated with black race, not controlling for health literacy: OR 1.54, 95% CI, 1.29-1.84
Odds associated with black race, controlling for health literacy: OR 1.04; 95% CI, 0.85-1.26

Health literacy mediates the effect of black race on having long-term illness (adjusted)
Odds associated with black race, not controlling for health literacy: OR 1.24; 95% CI, 1.03-1.49
Odds associated with black race, controlling for health literacy: OR, 1.07; 95% CI, 0.89-1.30
Howard, 200663 (companion: Gazmararian, 2006;61 Wolf et al., 2007;64 Baker et al., 2007;65 Wolf et al., 2005;66 Baker et al., 2008;67 Howard et al., 2005;68 Baker et al., 200462)

Cohort

N = 3,260

Fair
New Prudential Medicare managed care enrollees in Cleveland, Ohio, Houston, Texas, Tampa, and south Florida (including Ft. Lauderdale and Miami)

S-TOFHLA
By race:
White:
Adequate: 71%
Marginal: 10%
Inadequate: 19%
Black:
Adequate: 36%
Marginal: 12%
Inadequate: 52%
Age
Gender
Race/ethnicity
Education
Income
Site
Morbidity
Smoker
Physical HRQoL mean (SF-12)
White: 44.9
Black: 43.6

Mental HRQoL mean (SF-12)
White: 55.7
Black: 53.0

Self-reported health good or higher
White: 0.39
Black: 0.23

Receipt of influenza vaccine
White: 0.826
Black: 0.701

Receipt of pneumococcal vaccine
White: 0.48
Black: 0.29
Physical HRQoL (difference in scores between white and black, adjusted)
Not controlling for health literacy: 0.1
Controlling for health literacy: -0.5
Difference between models: (0.6, 95% CI, 0.3-0.9)

Mental HRQoL (difference in scores between white and black, adjusted)
Not controlling for health literacy: 0.5
Controlling for health literacy: 0.2
Difference between models: (0.3, 95% CI, 0.1-0.5)

Self-reported health good or higher (difference in scores between white and black, adjusted)
Not controlling for health literacy: 0.8
Controlling for health literacy: 0.6
Wolf et al., 2006157

Convenience

N = 308

Good
Patients with newly diagnosed prostate cancer in 4 outpatient oncology and urology clinics in Chicago area

REALM
Low: 18%
Marginal: 33%
Functional: 50%
Age
Race
Annual income
Marital status
PSA Level > 20 ng/mL
Marginal: 24%
Low: 33%
Functional: 14%

Outcomes by race: NR
Health literacy mediates the association between race (African American versus white) and PSA level (adjusted).

Odds associated with African American, not controlling for health literacy
(OR, 4.6; 95% CI, 2.0- 9.5)
Odds associated with African American, controlling for health literacy
(OR, 3.0; 95% CI, 0.8- 9.1)
Bailey, 200977

Cross-sectional

N = 373

Fair
Adults in Shreveport, LA; Chicago, IL; and Jackson, MI

REALM
Low: ≤ 6th grade: 20%
Marginal: 7th-8th grade: 29%
Adequate: ≥ 9th grade: 51%
Race
Age
Sex
Education
Misinterpretation of medication label instructions

Low: 43%
Marginal: 34%
Adequate: 18%
HL is a mediator between race and gender and misinterpretation of medication instructions

Odds associated with being black vs. white (adjusted)
Not controlling for HL: OR, 1.63; 95% CI, 1.02-2.61
Controlling for HL: OR, 1.22; 95% CI, 0.73-2.04

Odds associated with being male vs. female (adjusted)
Not controlling for HL: OR, 1.67; 95% CI, 1.03-2.72
Controlling for HL: OR, 1.59; 95% CI, 0.97-2.60
Bennett et al., 200985 (companion: White et al. 200886)

Cross-sectional

N = 2,668

Good
Nationally representative sample of US population 65 years and older

NAAL
Below basic: 29.0%
Basic: 29.5%
Intermediate: 38.2
Proficient: 3.3%
Race
Income
Gender
Age
Nativity
NRHL mediates the association between race (black vs. white) and self-reported health status (adjusted)

Odds associated with being black
Not controlling for HL: β, -0.34 (SE, 0.11) (P < 0.05)
Controlling for HL: β, -0.24 (SE, 0.04) (P < 0.05)

Odds associated with being Hispanic
Not controlling for HL: β, 0.02 (SE, 0.14) (P = NS)
Controlling for HL: β,0.21 (SE, 0.07) (P < 0.05)

HL mediates the association between race (black vs. white) and receipt of influenza vaccine (adjusted)

Odds associated with being black
Not controlling for HL: β,-0.24 (SE, 0.10) (P < 0.05)
Controlling for HL: β,-0.18 (SE, 0.04) (P < 0.05)

Odds associated with being Hispanic
Not controlling for HL: β, -0.04 (SE, 0.16) (P = NS)
Controlling for HL: β, 0.08 (SE, 0.07) (P = NS)

HL not found to mediate relationship between race/ethnicity and receipt of mammogram (adjusted, comparison is white)

Odds associated with being black
Not controlling for HL: β, 0.23 (SE, 0.15) (P =NS)
Controlling for HL: β, 0.28 (SE, 0.06) (P < 0.05)

Odds associated with being Hispanic
Not controlling for HL: β, 0.57 (SE, 0.19) (P < 0.05)
Controlling for HL: β, 0.70 (SE, 0.07) (P < 0.05)

HL not found to mediate the relationship between race/ethnicity and dental checkup (adjusted, comparison is white)

Odds associated with being black
Not controlling for HL: β, -0.13 (SE, 0.11) (P =NS)
Controlling for HL: β, -0.04 (SE, 0.04) (P = NS)

Odds associated with being Hispanic
Not controlling for HL: β, 0.19 (SE, 0.14) (P = NS)
Controlling for HL
(β. 0.35 (SE, 0.05) (P < 0.05))
Osborn, 2009171

Cross-sectional

N= 383

Good
Adults with type I or II diabetes

REALM
< 9th grade = 31%
≥ 9th grade = 69%
Age
Year of diagnosed diabetes
Insulin use
African American race
Data NRHL not found to be a mediator of relationship between African American race and HbA1C through structural equation modeling
Sudore et al., 2006167 (companion: Sudore et al., 200695)

Prospective cohort, retrospective analysis

N = 2,512

Good
Seniors (70-79 year old) in Pittsburgh, PA, and Memphis, TN

REALM
Limited: 24%
Adequate: 76%
NRMortality rate
Limited: 20%
Adequate: 11%
Mortality within subgroups comparing limited group with adequate: Interaction between racial group and HL and sex and HL (P > 0.10 for all comparisons implying no moderator effect)
White: HR 2.36; 95% CI, 1.63-3.42
Black: HR 1.66; 95% CI, 1.29-2.29
Male: HR 1.77; 95% CI, 1.20-2.62
Female: HR 2.27; 95% CI, 1.67-3.09
Osborn et al., 200769 (companions: Wolf et al., 2007;70 Waite et al., 200871)

Cross-sectional

N = 204

Fair
Patients at 2 HIV clinics: 1 in Chicago, IL, and 1 in Shreveport, LA

REALM
Low: 11%
Marginal: 20%
Adequate: 69%
Race
Gender
Age
Income
Number of medications in HIV regimen
Non-HIV comorbid conditions
Mental illness
Nonadherence to HIV medications in past 4 days

Low: 52%
Marginal: 19%
Adequate: 30%
HL mediates association of black vs. white race on adherence (adjusted)
Odds associated with being black, not controlling for HL: OR, 2.4; 95% CI, 1.14-5.08
Odds associated with being black, controlling for HL: OR, 1.8; 95% CI, 0.51-5.85
Yin, 2009102

Cross-sectional

N = 6,100

Fair
Parents ≥ 16 years old living in a US household (nationally representative sample)

NAAL
Below basic: 11%
Basic: 18%
Intermediate: 56%
Proficient: 15%
Age
Gender
Number of children living in the home
Educational attainment
Race/ethnicity
Country of birth
English proficiency
Income
Region
Metropolitan statistical area
At least 1 child without health insurance
Below basic: 24%
Basic: 10%
Intermediate: 6%
Proficient: 3%

Self-reported difficulty understanding OTC medication labels
Below basic: 74%
Basic: 43%
Intermediate/proficient: 38%
HL is a mediator between race and health insurance coverage (adjusted)

Race/ethnicity not controlling for HL: P = 0.03
Race/ethnicity controlling for HL: P = 0.08

HL is not a mediator between race and self-report of difficulty understanding of medication labels
Race/ethnicity not controlling for HL: P = 0.04
Race/ethnicity controlling for HL: P = 0.05

CI=confidence interval; Hb=hemoglobin; HIV=human immunodeficiency virus; HL=health literacy; HR=hazard ratio; HRQoL=health related quality of life; N=number; NAAL=National Assessment of Adult Literacy; NALS=National Adult Literacy Survey; NR=not reported; NS=not sufficient; OR=odds ratio; OTC=over the counter; PSA=prostate-specific antigen; REALM=Rapid Estimate of Adult Literacy in Medicine; SE=standard error; SF-12=Short Form 12; S-TOFHLA=Short Test of Functional Health Literacy in Adults; US=United States.

From: Results: Relationship of Health Literacy to Outcomes and Disparities

Cover of Health Literacy Interventions and Outcomes: An Updated Systematic Review
Health Literacy Interventions and Outcomes: An Updated Systematic Review.
Evidence Reports/Technology Assessments, No. 199.
Berkman ND, Sheridan SL, Donahue KE, et al.

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.