Determinant, mediator, moderator? How does Health Literacy ... · NVS -.081 IE.365 -.168 -.112.071...

Jürgen M. Pelikan

Em. Prof. Institute for Sociology, University of Vienna

Director, WHO-CC Health Promotion at Gesundheit Österreich, Vienna / Austria

& Kristin Ganahl

MA, PhD Student at Gesundheit Österreich, Vienna / Austria

8th Annual Health Literacy Research Conference,

Bethesda, October 13th to 14th, 2016

Determinant, mediator, moderator? How does Health Literacy influence self-assessed health? Results from the HLS-EU-study

Pelikan & Ganahl HARC 2016 1

1. PROBLEM & RESEARCH QUESTION


There is growing evidence for the relevance of health literacy (HL) for health, but what are the causal mechanisms at work?

» It has been demonstrated quite well that functional HL has relevant effects on health related indicators for patients and health/disease care related outcomes (cf. Berkman et al. 2011)

» But there is much less research on these relationships in general populations and/or with a wider more comprehensive definition and measure of HL, e.g.

» For community samples of older citizens or patients & mortality (e.g. Baker et al. 2007, 2008, Bostock & Steptoe 2012, McNaughton et al 2015)

» For self-rated health status for aged US adults (Bennett et al. 2009, Sentell, et al. 2014), Taiwanese adults (Lee et al. 2010), for Japanese adults (Furuaya et al. 2015), for elderly Chinese (Liu et al. 2015), for immigrants in Canada (Omariba & Ng, 2011)

» There are only a few studies looking explicitly at mediating effects of HL, e.g.

» between education & self-reported health for Dutch adults (van der Heide et al. 2013)

» between education & health behaviors for Danish adults (Friis, et al. 2016), or heath-insured Belgiens (van den Broucke & Renwart, 2014)

» between race & patient activation for elderly urban minority patients (Gwynn et al. 2016)

» Even less on moderating effects of HL, e.g.

» between age & health related quality of life responses to chronic disease in Chinese rural women (Wang et al. 2015)

» Data of the HLS-EU survey provide an excellent opportunity for testing the relationship of HL and health for general populations in different countries.


Research questions HL as a determinant

HL as a mediator

HL as a moderator

1. In how far is functional or comprehensive health literacy (HL) associated with self-assessed health (SAH)?

2. How does HL influence SAH?

As a determinant,

or a mediator,

or a moderator of health,

or a combination of these mechanisms?


Deter-minants

Indicators of Health

(SAH)

HL

Deter-minants


(SAH)

HL


(SAH)

HL

Deter-minants

2. CAUSAL MODEL & MEASURES OF VARIABLES


A causal model of limited HL for patients in health care


Paasche-Orlow, M.K., Wolf, M.S.

(2007): The Causal Pathways Linking

Health Literacy to Health Outcomes.

Am J Health Behav; 31 (Subbl 1):s19-

s26.

The Vienna Model of determinants & consequences of HL in general populations based on the HLS-EU model (Pelikan et al 2014)

0. Situational Determinants (Country, Province, District, Urban/Rural, etc.)

1. Personal Determinants

2. Individual Health Literacy

3. Health Behaviors

4. Health Status

5. Illness Behaviors


The Vienna Model of determinants & consequences of HL in general populations based on the HLS-EU model (Pelikan et al 2014)

0. Situational Determinants (Country, Province, District, Urban/Rural, etc.)

1. Personal Determinants

Gender

Age

Education

Financial Depr.

Social Status

2. Individual Health Literacy

Functional HL

Comprehensive HL

3. Health Behaviors

Physical activity

BMI

4. Health Status

Long term illness

Self-assessed health

5. Illness Behaviors


Causal ordering & measures of variables used

1. Gender (0 = male, 1= female)

2. Age (years of age, or grouped)

3. Education (ISCED: from 0= lowest to 6=highest level)

4. Financial deprivation (Factor score based on 3 questions from low deprivation to high)

5. Self-assessed social status (from 1=lowest to 10=highest place in society )

6. NVS (6 questions on ice crème label,0-6 points (6= best), 3 levels)

7. Comprehensive HL (HLS-EU-Q47, Index 0-50 = highest HL, 4 levels)

8. Physical Activity (1=almost every day to 4=no physical activity)

9. BMI (Index of weight-for-height; kg/m2; Low to high BMI Index)

10. Long-term illness (Minimum European Health Modul, from 1=No long-term illnesses, 2=Yes, one, 3=Yes, more than one)

11. Self-assessed health (SF36 item, from 1=very good to 5=very bad)


HLS-EU Survey Overview: Sampling, Data collection

Countries Austria (AT), Bulgaria (BG), Germany (DE) (only NRW), Greece (EL) (only Athens +), Spain (ES), Ireland (IE), Netherlands (NL), Poland (PL)

Survey Institut TNS Opinion on behalf of the HLS-EU Consortium

Survey Periode Summer 2011

Target Population, Population Coverage EU citizens (!) aged 15 years and over (Euro-barometer Methodology)

HL Instrument HLS-EU-Q86 (including HLS-EU-Q47 and NVS Test)

Data collection by computer-assisted personal interviewing technique (CAPI) (BG, IE = PAPI)

Sampling design Euro-barometer Methodology Stratified probability sampling (multistage random sample): • National sampling points selected randomly (applying random-walk procedure) after

stratification for population size and population density (metropolitan, urban and rural areas).

Response Rates Austria (67%), Bulgaria (75%), Germany (DE) (53%), Greece (65%), Spain (62%), Ireland (69%), Netherlands (36%), Poland (67%)

Sample Sizes Austria (1015), Bulgaria (1002), Germany (DE) (1057), Greece (1000), Spain (1000), Ireland (1005), Netherlands (1023), Poland (1000)

Weights National samples were weighted by gender, age group and size of locality, based on national census data Country size was not used as a weighting criterion for the analyses of the total sample. Total sample values therefore represent a ‚country average‘ where all countries are represented with equal weights regardless of their population size.

Pelikan - HL Geneva 7-10-2016 10

3. RESULTS


Spearman Correlations for all used variables

(for Total, N=8000, HLS-EU 2012)


Gender Age Education Social Status Fin. Dep. NVS Comp. HL

Physical Activity BMI

Long-Term Illness SAHS

Gender 1

Age ,066** 1

Education -,018 -,204** 1

Social Status -,007 -,137** ,336** 1

Fin. Dep. ,050** ,036** -,200** -,422** 1

NVS ,011 -,290** ,347** ,296** -,196** 1

Comp. HL ,046** -,121** ,239** ,291** -,300** ,245** 1

Physical Activity

,030** ,110** -,151** -,205** ,196** -,170** -,189** 1

BMI -,149** ,336** -,103** -,113** ,092** -,113** -,066** ,121** 1

Long-Term Illness

-,073** -,425** ,152** ,129** -,120** ,165** ,156** -,109** -,205** 1

SAHS ,062** ,443** -,213** -,254** ,188** -,231** -,274** ,200** ,256** -,589** 1

There are significant correlations between HL indicators & health related outcome

indicators, between different health related indicators, but also with socio-

demographic & socio-economic confounders!

There is a gradient: The better health literacy, the better Self-Assessed Health (SF-36) (for Total Sample) (HLS-EU 2012)

10,8 10,8 15,2 17,7 24,1

34,0 35,2 43,8 18,6

25,1

32,7 40,2

43,7

42,4 42,0 37,4

31,4

35,9

35,3

31,4

25,9 18,7 19,4 16,1

26,4

22,4

14,7 9,3 5,6 4,4 2,8 2,7

12,7 5,8

2,2 1,3 0,7 0,5 0,7 0,1

0

10

20

30

40

50

60

70

80

90

100

<15 15-20 20-25 25-30 30-35 35-40 40-45 45-50

Percentages of Categories of

Self-assessed Health Status

Grouped Scores of Comprehensive Health Literacy Index

Very bad

Bad

Fair

Good

Very good

<15[N=102]|15-20[N=259]|20-25[N=600]|25-30[N=1348]|30-35[N=2185]|35-40[N=1531]|40-45[N=1048]|45-50[N=704]| TOTAL[N=7777]

The gradient for self-assessed health holds true for each country, but on a different level and differing correlation,

(for 8 Countries & Total Sample (HLS-EU 2012)

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

< 5 5 - 10 10 - 15 15 - 20 20 - 25 25 - 30 30 - 35 35 - 40 40 - 45 45 - 50

Percentages of Respondents Reporting

Good or Very Good

Health

Grouped Scores of Comprehensive Health Literacy Index

Austria (r=-.314)**

Bulgaria (r=-.302)**

Germany (r=-.227)**

Greece (r=.-.379)**

Spain (r=-.295)**

Ireland (r=-.230)**

Netherlands (r=-.172)**

Poland (r=-.336)**

Total (r= -.294)**

(only levels with N > 10 are illustrated)

HEALTH LITERACY AS A DETERMINANT OF SELF-ASSESSED HEALTH, WHEN RELEVANT CONFOUNDERS ARE CONTROLLED FOR


Self-assessed health by 5 social and 2 HL determinants (Beta Weights and Adjusted R-Square for Total Sample and Countries) (HLS-EU 2012)

NL .144 AGE .496 EL

BG -.103 CHL -.205 AT

AT -.020 SOC.STATUS -.158 NL

IE .053 FIN. DEPRIV. .212 DE

PL -.027 GENDER .077 ES

NL -.007 EDUCATION -.131 IE

AT .000 NVS -.081 IE

.365

-.168

-.112

.071

.039

-.036

-.026

italic…not significant on the 0,05 level

Adj. R2= .268

(NL.88 / .452 EL)

There is a social gradient for SAH,

CHL is the second strongest predictor

and NVS has no significant influence!


Self-assessed health by 5 social, 2 HL determinants and 3 risk factors

(Beta Weights and Adjusted R-Square) (NVS were not significant!), for Countries and Total Sample (HLS-EU 2012)

ES -.369 LT-ILLNESS -.556 IE

NL -.039 AGE .279 EL

EL -.063 CHL -.176 ES

ES/IE -.009 SOCIAL STATUS -.147 NL

NL .013 PHYS ACTIVITY .135 DE

EL -.029 BMI .131 NL

IE -.001 NVS -.041 EL

PL -.061 GENDER .073 ES

-.459

.161

-.125

-.095

.075

.066

-.027

Adj. R2= .447

(NL .330 / .630 EL)

.026

EL .032 FIN DEPRIVATION .132 PL

.018


The direct effect of CHL on SAH stays on

as third strongest, even when LTI

and 2 life-style indicators are included!

italic…not significant on the 0,05 level

HEALTH LITERACY AS A DETERMINANT & MEDIATOR OF HEALTH RELATED INDICATORS – DIRECT AND INDIRECT EFFECTS OF HEALTH LITERACY


Path Model for all variables explaining variation of self-assessed health (Beta weights for TOTAL, HLS-EU 2012) Explained Variance of SAH by CHL: direct .133, indirect .63, total .196

Gender

Age

Educ.

Fin. Dep.

Soc. St.

NVS CHL PHA BMI LTI SAH -.458

-.133

.066

.180

.030

.078

-.021

.030

-.102 .123 .158

-.058

.236 .135 -.188 .098 -.092

-.056

.030 .062 -.119

-.249 -.037

-.242

.102

.090

-.373 .250

-.123


Direct & indirect effects of CHL on SAH and explained variance of dependent health related indicators in path models (Beta weights and adjusted R2, Total & 4 countries, HLS-EU 2012)

Total NL ES PL EL

Effects of CHL on SAH

- direct -,133 -0.113 -0.188 -0.123 -0.054

- indirect -,063 -0.040 -0.010 -0.053 -0.069

- total -,196 -0.153 -0.197 -0.176 -0.124

R-Square

- SAH .454 0.315 0.412 0.555 0.616

- CHI .196 0.119 0.187 0.291 0.281

- BMI .087 0.059 0.146 0.110 0.117

- PHA .063 0.019 0.016 0.035 0.083

- CHL .189 0.081 0.102 0.258 0.293

- NVS .220 0.366

0.229 0.211 0.348


HEALTH LITERACY AS A MODERATOR OF HEALTH RELATED INDICATORS


Probability of having fair, bad or very bad self-assessed health for age groups moderated by HL levels (for Total Sample, N= 6923, HLS-EU 2012) > With older age the probability of having fair, bad or very bad self-assessed health is rising steadily, but much less, when one has better HL!

Logistic Regression controlled for: long-term illness (*), social status (*), exercising (*), BMI (*),

Gen-HL-Raw (ns), age (*), gender (female ↑) (*), financial deprivation (*), demographic

standard weights (ns),) – means by HL-levels and age groups

Nagelkerke/pseudo R²=0,46

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

25 or younger between 26 and 35between 36 and 45between 46 and 55between 56 and 65between 66 and 75 76 or older

inadequate HL

problematic HL

sufficient HL

excellent HL

total


Probability of having fair, bad or very bad self-assessed health, for education (ISCED-Levels) moderated by HL levels (for Total Sample, N= 6903, HLS-EU 2012) > With better education the probability of having fair, bad or very bad self-assessed health is decreasing steadily, but much less when one has better health literacy!

Logistic Regression controlled for: long-term illness (*), social status (*), exercising (*),

BMI (*), Gen-HL-Raw (*), Education (ns), Gender (female ↑) (*), financial Deprivation(*))

– means by HL-Levels and Education-Levels

Nagelkerke/pseudo R²=0,44

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

Level 0 Level 1 Level 2 Level 3 Level 4 Level 5 Level 6

An

teil

Pe

rso

ne

n m

it s

chle

chte

m

Ge

sun

dh

eit

szu

stan

d (

pro

gno

stiz

iert

er

We

rt)

Bildungsgrad (ISCED)

inadäquate GK

problematische GK

ausreichende GK

exzellente GK

TOTAL


Summary & discussion

» Summary

» For a comprehensive measure of HL (CHL) and for general populations in 8 European countries it could be demonstrated

» That CHL is a relevant social determinant of health (SDH) in multivariate models

» That CHL & NVS are mediating social determinants on health and their impact on health is also mediated by health behavior indicators or LTI

» That CHL is moderating the relationship of age and education on self-assessed health

» Explained variances & coefficients of effects vary by country.

» Limitations

» Cross-sectional study!

» Limited number of variables included

» No model combining mediating & moderating effects

» Our results partly depend on the model we have chosen for analysis!


References » Baker, D.W., Wolf, M.S., Feinglass, J., Thompson, J.A., Gazmararian, J.A., Huang, J. (2007) Health Literacy and

Mortality among Elderly Persons. Arch Intern Med, 167 (14).

» Baker, D.W., Wolf, M.S., Feinglass, J., Thompsons, J.A. (2008): Health Literacy, Cognitive Abilities, and Mortality among Elderly Persons. J Gen Intern Med, 23 (6).

» Bennett, I.M., Chen, J., Soroui, J.S., White, S. (2009): The contribution of health literacy to disparities in self-rated health status and preventive health behaviors in older adults. Ann Fam Medd; 7(3):204-11.

» Berkman; N.D., Sheridan, S.L., Donahue, K.E., Halpern, D.J., Crotty, K. (2011): Low Health Literacy and Health Outcomes: An Updated Systematic Review. Ann Intern Med, 155: 97-107.

» Bostock, S., Steptoe, A. (2012): Association between low functional health literacy and mortality in older adults: longitudinal cohort study. BMJ 44:e1602. doi: 10.1136/bmj.e1602.

» Friis, K., Lasgaard, M., Rowlands, G., Osborn, R.H., Maindail, H.T. (2016): Health Literacy Mediates the relationship between educational attainment and Health Behavior: A Danish Population-Based Study. Jounral of Health Communication, 21: sup2: 54-60.

» Furuaya, Y., Kondo, N., Yamagata, Z., Hashimoto, H. (2015): Health literacy, socioeconomic status and self-rated health in Japan. Health Promotion International. doi:10.1093/heapro/dat071.

» Gwynn , K.B., Winter, M.R., Cabral, H.J., Wolf, M.S., Hachate, A.D., Henault, L., Waite, K., Bickmore, T.W., Paasche-Orlow, M.K. (2016): Racial disparities in patient activiation: Evaluating the mediating role of health literacy with path analyses. Patient Education and Counseling, 99(6): 1033-1037.

» HLS-EU Consortium (2012): Comparative report on health literacy in eight EU member states. The European Health Literacy (Second Revised and Extended Version, Date July 22th, 2014), online Publication: http://lbihpr.lbg.ac.at.w8.netz-werk.com/sites/files/lbihpr/attachments/neu_rev_hls-eu_report_2015_05_13_lit.pdf

» Lee, S., Tsai, T., Tsai, Y., Kuo, K. (2010): Health literacy, health status, and healthcare utilization of Taiwanese adults: results from a national survey. BMC Public Health, 10: 614.

» Liu, Y., Liu, L., Li, Y., Chen, Y. (2015): Relationship between Health Literacy, Health-Related Behaviors and Health Status: A Survey of Elderly Chinese. Int. J. Environ. Res. Public Health, 12, 9714-9725. doi:10.3390/ijerph120809714.

» .


http://lbihpr.lbg.ac.at.w8.netz-werk.com/sites/files/lbihpr/attachments/neu_rev_hls-eu_report_2015_05_13_lit.pdf





References

» McNaughton, C., Cawthon, C., Kripalani, S., Liu, D., Storrow, A.B., Roumie, C.L. (2015): Health Literacy and Mortality: A Cohort Study of Patients Hospitalized for Acute Health Failure. J Am Heart Assoc. 4:e001799 doi: 10.1161/JAHA.115.001799.

» Omariba, D.W, Ng, E. (2011): Immigration, generation and self-rated health in Canada: on the role of health literacy. Can. J. Public Health, 102(4):281-5

» Sentell, T., Zhang, W., Davis, J., Baker, K.K., Braun, K.L. (2013): The influence of community and individual health literacy on self.reported health status. J Gen Intern Med; 29(2): 298-304.

» Sørensen K, Van den Broucke S, Fullam J, Doyle G, Pelikan JM, Slonska Z, Brand H, (HLS-EU) Consortium (2012): Health literacy and public health: A systematic review and integration of definitions and models. BMC Public Health, 12 (80).

» Sørensen K, Van den Brouck S, Pelikan JM, Fullam J, Doyle G, Slonska Z, Kondilis B, Stoffels V, Osborne RH, Brand H (2013): Measuring health literacy in population: illuminating the design and development process of the European Health Literacy Survey Questionnaire (HLS-EU-Q). BMC Public Health, 13, doi:10.1186/1471-2458-13-948

» Sørensen K, Pelikan JM, Röthlin F, Ganahl K, Slonska Z, Doyle G, Fullam J, Kondilis B, Agrafiotis D, Uiters E, Falcon M, Mensing M, Tchamov K, Van den Broucke S, Brand H (2015): Health literacy in Europe: comparative results of the European health literacy survey (HLS-EU). Eur J Public Health.

» Van den Broucke, S. & Renwart, A. (2014): Health literacy mediates the relationship between education level and health behaviour. European Journal of Public Health 24 (Suppl. 2):202-202.

» Van der Heide, I., Wang, J., Droomers, M., Spreewenberg, P., Rademakers, J., Uiters, E., (2013): The relationship between health, education, and health literacy: results from the Dutch Adult Literacy Skills Survey. Health Communication, 8(Suppl 1): 172-1184.

» Wang, C., Kane, R.L., Xu, D., Meng, Q. (2015): Health Literacy as a moderator of health-related quality of life responses to chronic disease among Chinese rural Women. BMC Womens Health, 15:34.


Contact

Jürgen M. Pelikan

Stubenring 6

1010 Vienna, Austria

T: +43 1 515 61-380

F: +43 1 513 84 72

E: [email protected]

www.goeg.at


Thank you for your attention!

Determinant, mediator, moderator? How does Health Literacy ... · NVS -.081 IE.365 -.168 -.112.071...

Documents

Transcript of Determinant, mediator, moderator? How does Health Literacy ... · NVS -.081 IE.365 -.168 -.112.071...