SelfRated Health in Epidemiological Surveys as a Predictor of Disability and Mortality

1
Self-Rated Health in Epidemiological Surveys as a
Predictor of Disability and Mortality

• Ellen Idler, PhD
• Institute for Health, Health Care Policy and
  Aging Research
• Rutgers University, NJ, USA

2
Ellen L. Idler, Ph.D.

• Ellen Idler is Professor and Chair of the
  Department of Sociology, Rutgers University, New
  Brunswick, NJ, US. She has been interested in
  self-rated health since graduate school when she
  read the original Mossey and Shapiro article
  (AJPH 1982). She has received multiple grants,
  including a 5-year FIRST Award, from the National
  Institute on Aging for studies of self-rated
  health, mortality, and disability. In 1999 she
  was a visiting professor at the University of
  Copenhagen, Denmark. Self-ratings of health are
  an appealing research topic because they support
  the importance of the lay persons perspective in
  health.

3
Learning objectives

• To understand that self-ratings of health (SRH)
  have been studied for decades
• To trace the history of the identification of SRH
  as a predictor of mortality
• To report new findings on SRH as a predictor of
  both mortality and ADL/IADL disability
• To suggest new directions for research on the
  mechanisms through which SRH affects health
  outcomes

4
Self-Ratings of Health (SRH)

• All in all, would you say your health is
• Excellent, Good, Fair, Poor
• How would you rate your health at the present
  time? Excellent, Good, Fair, Poor, Bad
• How is your health, compared with others your
  age? Better, Same, Worse

5
Duke Longitudinal Study of Human Aging, 1962 -
1973

• Consistent differences between SRH and physician
  (MD) rating
• Differences tend toward higher SRH than MD rating
• Highest SRH (compared to MD rating) among the
  most elderly
• SRH appears to predict future MD ratings better
  than MD ratings predict future SRH

6
Self-Ratings Predict Mortality

• In 1982 a Canadian study of a large and
  representative sample of elderly residents of
  Manitoba found that SRH was among the strongest
  predictors of mortality over 7 years, second only
  to age.
• The analysis adjusted for individual health
  status obtained from medical records and
  self-report of conditions.
• Even after adjustment for covariates, respondents
  rating their health Poor were 2.9 times as likely
  to die as those rating their health Excellent.

7
SRH - Mortality Studies since 1982

• Over 50 prospective, population-based studies to
  date
• From Canada, US, Poland, Israel, England, France,
  Hong Kong, Sweden, Wales, Netherlands, Australia,
  Japan, Lithuania, Finland, Denmark, Italy, China,
  Korea

8
SRH - Mortality Studies since 1982

• Sample sizes N421 to 7725
• Follow-up times 2 to 18 years
• Health status covariates MD Exams, Chronic
  conditions, Symptoms, ADL disability,
  Medications, Weight, Blood pressure
• Significant OR or HR for Poor vs.
  Excellent 1.4 to 93.5

9
Survival, Functional Limitations, and Self-Rated
Health in the NHANES I Epidemiologic Follow-Up
Study, 1992

• Ellen Idler, Louise Russell, Diane Davis
• Institute for Health, Health Policy and Aging
  Research
• Rutgers, The State University of New Jersey
• American Journal of Epidemiology 2000 152874-83

10
NHANES-I Epidemiologic Follow-Up Study(NHEFS
Data)

• General Medical History Supplement subsample
  N6913
• Complex sample design, weighted
• Ages 25-74 at baseline
• Follow-up 1971-1992
• 3.5 of subsample lost to follow-up

11
NHEFS Data

• N6641, complete data for mortality analysis
• Dependent variable Time-to-death in days
• N4136, complete data for ADL/IADL
  
  limitations analysis
• Dependent variable Scale of 23 ADL/IADLs
• Stanford Health Assessment Questionnaire
• Cronbachs alpha .96 (1982), .92 (1992)
• Assessed 1982 and 1992 only

12
NHEFS data

• Self-reported data
• Chronic conditions
• 42 items
• Symptoms
• 22 items
• Health practices
• 6 items

• Observed data
• MD examination
• 17 ICD-8 categories
• Clinical measurements
• 4 blood, urine tests
• blood pressure
• height, weight

13
Mortality Hazard Ratios (plt.05)
Males
Females

• Age
• Overweight
• SBP gt160 mmHg
• Heart attack
• Stroke
• Protein, sugar in urine
• Shortness of breath
• Current smoker
• No exercise
• Self-rated health (SRH)
• Excellent .52
• Very good .56
• Good .68

• Age
• MD Circulatory disease
• Underweight
• Hematocrit gt43
• SBP 140-159 mmHg
• SBP gt 160 mmHg
• Protein, sugar in urine
• Current smoker
• No exercise

14
ADL/IADL limitations analysis (plt.05)

• Age
• MD Circulatory disease
• MD Musculoskeletal disease
• Overweight
• Arthritis
• Diabetes
• Heart attack
• Cough
• Pain in legs
• Wheezy chest
• Drinks weekly (-)
• Self-rated health (SRH)
• Excellent -8.1
• Very good -7.1
• Good -8.1
• Fair -3.9

• Age
• Bronchitis
• Heart attack
• Hernia (-)
• Hives (-)
• Cough
• Chest pain
• Pain in legs
• Self-rated health (SRH)
• Excellent -5.8
• Very good -5.7
• Good -5.4
• Fair -4.1

Females
Males
15
Conclusions

• Data quality
• includes both self-report and MD exam
• unlikely to be surpassed in US studies in future
• Multiple endpoints for analysis
• Mortality - includes entire sample
• ADL/IADL limitations - discriminates among
  survivors
• Gender differences
• implications for future research