B-1 THE IMPACT OF HEALTH STATUS ON ANNUAL INCOME IN THE U.S

Monday, October 20, 2008: 1:30 PM
Grand Ballroom B/C (Hyatt Regency Penns Landing)
Patrick Sullivan, PhD1, Michael B. Nichol, PhD2 and Vahram Ghushchyan, PhD1, (1)University of Colorado School of Pharmacy, Denver, CO, (2)University of Southern California, Los Angeles, CA
Purpose:  To examine and quantify the impact of health status on annual income in a nationally representative sample in the U.S.

Methods:  The pooled 2001 and 2003 Medical Expenditure Panel Survey (MEPS) was used.  MEPS is a nationally representative survey of the U.S. civilian, noninstitutionalized population based on self-report which collects detailed information on sociodemographic characteristics, income and health status.  Health status was measured by the Physical Component Scale (PCS-12) and Mental Component Scale (MCS-12) of the SF-12.  In order to address simultaneity between health status and income, a two-step process was used: 1) predicted PCS-12 and MCS-12 scores were derived from results of a linear regression of PCS/MCS-12 scores on age, gender, race, ethnicity, smoking status, education, poverty category and number of chronic conditions; 2) a Heckman selection model was used to regress the log of income on predicted SF-12 scores (from the first step above) controlling for age, gender, race, ethnicity, smoking status, education and marital status.  Smearing retransformation using the naïve assumption was conducted to retransform annual income.  Spline regression was also used to examine non-linearity by categorizing SF-12 scores into quartiles.

Results:  Physical function and mental function had a positive and statistically significant relationship with annual income.  Each point difference in PCS-12 was correlated with an approximate difference of $1,000 in annual income while each point difference in the MCS-12 was correlated with a difference of $1,500 in annual income.  The relationship between PCS-12 scores and annual income did not vary by starting quartile.  However, the relationship between MCS-12 scores and income varied significantly by MCS-12 quartile, with the first and fourth quartile showing the greatest magnitude.

Conclusion:  Physical function and mental function appear to be strongly correlated with annual income in the U.S., with mental function relatively more important than physical function.  The relationship appears to be strongest among individuals with the lowest and highest MCS-12 scores.