Psychosocial Circumstances and Health Status in a Managed Care Population

1
Psychosocial Circumstances and Health Status in a
Managed Care Population

• Project Team
• ER Becker 1, DW Roblin 2, DH Howard 1, PJ Joski
  2, J Ren 2
• 1 Rollins School of Public Health at Emory
  University
• 2 Center for Health Research / Southeast, Kaiser
  Permanente Georgia
• Project Funding
• Centers for Disease Control and Prevention
• NIH 1R01CD000033 (ER Becker, PI)

2
Presentation Outline

• Description of the survey of health and healthy
  behaviors among working age Kaiser Permanente
  Georgia (KPGA) adults in 2005 Roblin
• Associations of patient activation and work
  climate (support/interactions with coworkers)
  with participation in worksite wellness programs
  and activities Becker
• Associations of patient activation, neighborhood
  characteristics, and social climate
  (support/interactions with friends and family)
  with the practice of healthy behaviors Roblin

3
Survey of Health and Healthy Behaviors Among
Working Age Kaiser Permanente Adults in 2005

• Douglas W. Roblin1 and Edmund R. Becker2
• 1 Center for Health Research / Southeast, Kaiser
  Permanente Georgia
• 2 Rollins School of Public Health at Emory
  University
• Project Funding
• Centers for Disease Control and Prevention
• NIH 1R01CD000033 (ER Becker, PI)

4
Background and Objective

• The Chronic Care Model postulates how
• Psychosocial circumstances (social climate, work
  climate, delivery system climate) might influence
  
• Patient activation, which in turn might influence
  
• Lifestyle behaviors and health status
• Objective
• Design a reliable instrument to be administered
  to a targeted, random sample of working age
  adults to measure these influences and outcomes

5
Conceptual Model
6
Study Population

• Kaiser Permanente Georgia (KPGA) members, aged
  25-59, employed by large public agencies or
  private corporations in the Atlanta area.
• Three condition cohorts were sampled
• Low risk adults (no identifiable major
  morbidities)
• Adults with elevated lipids (without acute CAD
  history)
• Adults with type 2 diabetes (without history of
  micro- or macrovascular complications)

7
Survey Instrument Development

• Literature review to identify, brief reliable
  items or scales administered in written surveys
• SF-12 (physical and mental function)
• Trust in physician (PCAS)
• Social climate (MIDUS)
• Work climate (MIDUS)
• Patient activation (PAM-13)
• Physical activity (BRFSS)
• Dietary intake (Block fat, F/V screeners)
• Cognitive pre-testing of draft instrument among 4
  focus groups

8
Survey Administration

• Mixed mode survey (mail or Internet) conducted by
  a professional survey firm from 10/1/05 through
  12/31/05
• 2,224 respondents among 5,309 sampled (42
  response rate)
• Respondents more likely to be female, older
• Diverse respondent sample 60 female, 45
  African American, 18 HS education or less, 31
  household income lt 50,000
• Psychometric properties of previously validated
  scales were similar between these survey
  respondents and respondents to surveys where
  scales were initially used.

9
Cronbachs Alphas for Study Scales
10
Strategies to Address Challenges to the Publics
HealthWorksite Wellness Programs

• Becker ER1, Roblin DW2, Joski PJ2
• 1 Rollins School of Public Health at Emory
  University
• 2 Center for Health Research / Southeast, Kaiser
  Permanente Georgia
• Project Funding
• Centers for Disease Control and Prevention
• NIH 1R01CD000033 (ER Becker, PI)

11
Background and Objectives

• Information on availability of worksite wellness
  programs and activities to employed MCO enrollees
  and their participation in those programs is
  limited.
• Objectives
• Describe the associations of worksite wellness
  programs and activities with employer/worksite
  and employee characteristics.
• Describe levels of participation in worksite
  wellness programs and activities given their
  availability.
• Evaluate the potentially moderating influences of
  patient activation and worksite support/stress on
  participation in worksite wellness programs and
  activities.

12
Methods

• Dependent variables
• Program availability
• Participation in worksite wellness program or
  activity (conditional on availability)
• Program or activity related to physical activity
  or exercise
• Program or activity related to dietary intake,
  meal preparation
• Independent variables (for program
  participation)
• Activation
• Work climate coworker / supervisor support or
  strain
• Ordinary logistic regression (for program
  participation)
• Covariates age, gender, condition cohort, race,
  education, marital status

13
Results Program Availability

• Availability of worksite wellness programs
• 76.9 reported worksites with one or more
  programs or activities promoting exercise
• 31.6 reported worksites with one or more
  activities promoting healthy eating
• Employees with diabetes (plt0.05) or high BMI
  (plt0.10) were less likely to have worksites with
  programs or facilities promoting exercise or
  healthy eating.

14
Results Program Availability
15
Results Program Participation

• Where programs were available
• 22.1 participated in a program promoting
  physical activity
• 15.5 participated in a program promoting healthy
  eating
• Participation did not significantly differ
  between adults with diabetes, elevated lipids, or
  low risk adults
• Participation was significantly more likely with
  associated with
• More coworker / supervisor support, less coworker
  / supervisor strain (physical activity programs)
• Higher activation (physical activity and healthy
  eating programs)

16
Results Program Participation (Physical Activity)
17
Results Program Participation (Physical Activity)
18
Results Program Participation (Diet)
19
Conclusions

• Employees who might clinically benefit from
  availability of wellness programs notably
  adults with diabetes or high BMI were least
  likely to work where supportive programs or
  activities were available.
• Where available, participation in exercise or
  diet programs and activities was low but did not
  differ by clinical condition.
• Participation was primarily affected by employee
  activation and, for exercise, a worksite with
  high levels of support and collegiality among
  coworkers.

20
The Association of Neighborhood Characteristics
and Social Interactions with Physical Activity,
Diet, and Obesity Among Employed Adults

• Roblin DW1, Joski PJ1, Becker ER2
• 1 Center for Health Research / Southeast, Kaiser
  Permanente Georgia
• 2 Rollins School of Public Health at Emory
  University
• Project Funding
• Centers for Disease Control and Prevention
• NIH 1R01CD000033 (ER Becker, PI)

21
Background

• Most existing studies have examined the impact of
  only one of several possible factors contributing
  to physical activity or obesity among urban
  adults
• Neighborhood characteristics Presence/absence of
  sidewalks
• Networks of family and friends Social support /
  strain
• Patient activation
• These factors, however, may have simultaneous,
  independent associations with physical activity
  or obesity.

22
Objectives

• 1. Estimate the independent contributions of
  neighborhood characteristics, social support /
  strain (social climate), and patient activation
  on physical inactivity and obesity among working
  age adults of an MCO.
• 2. Estimate the independent contributions of
  social support / strain (social climate) and
  patient activation on dietary intake in this same
  population.

23
Methods

• Dependent variables
• Physical inactivity (BRFSS)
• Dietary intake
• Percent fat in diet
• Daily fruit and vegetable (F/V) Servings
• Daily fiber intake (grams)
• Obesity (BMI 30 kg / m2)
• Independent variables
• Neighborhood characteristics (for physical
  inactivity, obesity)
• Social climate friend / family support or strain
• Activation
• Ordinary linear or logistic regression
• Covariates age, gender, condition cohort, race,
  education, marital status

24
Results Physical Inactivity

• Physical inactivity was significantly less likely
  with
• Sidewalks in neighborhood
• Adjusted OR 0.992
• Walking or cycling paths nearby
• Adjusted OR 0.724
• Exercise equipment in household
• Adjusted OR 0.726
• More friend / family support, less friend /
  family strain
• Higher activation

25
Results Physical Inactivity
26
Results Physical Inactivity
27
Results Dietary Intake

• Percent fat in diet was significantly lower with
• More friend / family support, less friend /
  family strain
• Higher activation
• Daily F/V servings were significantly higher
  with
• More friend / family support, less friend /
  family strain
• Higher activation
• Daily fiber intake was significantly higher with
• More friend / family support, less friend /
  family strain
• Higher activation

28
Results Dietary Intake
29
Results Dietary Intake
30
Results Obesity

• Obesity was significantly less likely with
• Sidewalks in neighborhood
• Adjusted OR 0.744
• Walking or cycling paths nearby
• Adjusted OR 0.686
• More friend / family support, less friend /
  family strain
• Higher activation

31
Results Obesity
32
Results Obesity
33
Conclusions

• Sidewalks in neighborhoods, nearby walking and
  cycling paths, and household exercise equipment
  decrease likelihood of physical inactivity.
• Supportive networks of friends and family
  decrease likelihood of physical inactivity,
  improve dietary intake, and decrease likelihood
  of obesity.
• Activation decreases likelihood of physical
  inactivity, improves dietary intake, and
  decreases likelihood of obesity.

34
Conclusions

• All 3 factors have independent associations with
  healthy behaviors and ultimately with obesity.