How well are we doing addressing disability in America? Examining the status of adults with chronic disabling conditions, 1995 and 2005.

Bolin, Jane Nelson

ABSTRACT

Despite laws like the Americans with Disabilities Act (ADA), (1992), and The Ticket to Work Act, (TTWA), (2001), working age adults who develop chronic disabling conditions often find themselves faced with a choice of leaving the labor market in order to qualify for public health insurance or continuing to work, often on a "hit and miss" basis, disqualifying them from employee health benefits. Federal and state policy makers continue to struggle to find solutions addressing the needs of working age adults with disabling conditions and illnesses. In this study we examine the work status of working age adults using two National Health Interview Surveys conducted a decade apart (1995 & 2005) to investigate and compare adults who have chronic, disabling conditions and self-reported rates of work. Specifically, this research investigates whether reported work activity among working age adults who report chronic health conditions has improved in the decade between 1995 and 2005. The effects of racial/ethnic differences, age, and region of residence on one's work status are also examined.

BACKGROUND

It has been over 15 years since passage of the Americans with Disabilities Act (ADA) (42 [section] U.S.C. 12100 et seq.), and nearly five (5) years since passage of the Ticket to Work Act, (TTWA) ((Public Law 106-170), yet federal and state policy makers still struggle to find solutions addressing the needs of working age adults with disabling conditions and illnesses who choose to leave employment because of chronic mental and physical health conditions. Despite federal and state programs providing disability assistance, work retraining, and health care services, these programs have often been hampered by rising costs and budget cutbacks as well as poor participation levels (Ticket to Work Advisory Panel, 2003). The Social Security Administration has reported that it has been difficult to increase poor employment rates among Social Security Beneficiaries despite targeted programs such as the Ticket to Work Act (SSA, 2005)

Against this backdrop, a strong and consistent literature stream documents increasing rates of chronic, disability producing conditions such as diabetes, congestive heart failure ["CHF"], chronic obstructive pulmonary disease, ["COPD"], asthma and arthritis, particularly among African Americans and Hispanics living below the federal poverty level (Eberhart, Ingram, Makuc, et al 2001; Pope, Tarlov, 2001; Kaiser Commission on Medicaid and the Uninsured, April, 2001; Black, Ray, Markides, 2001; Fried, Prager, Mackay, Xia, 2003).

According to the 2000 U.S. Census, nearly 50 million men, women and children have a chronic condition or illness capable of affecting work and functioning (Census, 2005). While estimates of the prevalence of reported work-disability or functional limitation vary, the Social Security Administration currently estimates that nearly 18 million Americans between the ages of 25 and 64 has one or more work-related disability but that approximately 27 percent of these persons are currently employed (U.S. Census Bureau, 2005).

Title II of The Social Security Act was passed nearly five decades ago in order to address the public welfare of disabled Americans establishing the disability safety net we now know as Social Security Disability Insurance, ["SSDI"]. Through periods of recession and affluence in our Nation's history, the Social Security Disability Insurance system has served as an important financial safety net for millions of working-age Americans who become unable to work because of disabling conditions or injuries. An important public benefit, considering that approximately 12 million workers and their families receive either SSDI or SSI benefits at a cost of $54 billion for SSDI and $23 billion for SSI (Benitez-Silva, Buchinsky & Rust, 2002; Massanari, June, 2001). For the majority of working-age Americans, SSDI is their sole source of disability protection (GAO-01-35). In 1996 only 26 percent of private sector workers had commercial, self-purchased, or private disability insurance making Social Security Disability ["SSDI"] and Supplemental Security Income ["SSI"] the most important source of financial assistance and medical insurance for disabled working-age Americans (GAO-01-35). The Social Security Advisory Board has estimated that over the course of work life, "nearly one out of three young men, and nearly one out of four young women who are now age 20 will become disabled before reaching age 67" (1998, p.1). When considering that disability claims are highest for workers over the age of 50, and that this is our nation's fastest growing segment of the population, this disability "safety-net" is especially important (Massanari, June, 2001),. Over 30 million people between the ages of 21 and 64 self-report a disability that interferes with work (Census, 2000). Of those reporting a work disability, fifty-seven percent (or 17.4 million) continue to work, while 13.2 million (46 percent) reported no employment (Census, 2000). These statistics support national and state policy efforts towards insuring that working-age individuals with chronic illnesses and impairments be supported in their efforts to continue to work. Nevertheless, existing research also demonstrates that a significant percentage (46 percent) of persons with chronic illnesses end up leaving the work force and applying for public disability benefits (GAOHEHS-98-39; Stapleton, Livermore, Scrivner, Tucker, 1998; Bound 1989, Baldwin & Johnson, 1994; Bolin, 2002). While access to public disability benefits is critically important to the lives of millions of Americans (Massanari, 2001), qualifying for public disability benefits is not an easy task (Bilder & Mechanic, 2003; Mashaw & Reno, 1996, Rosenheck, Dausey, Frisman & Kasprow, 2000; Wunderlich, Rice & Amado, 2002; Benitez-Silva, Buchinsky & Ruse, 2002). Applying for disability benefits can be intimidating and unpredictable (IOM, 2002; Hoffman, 2000; Anderson, 2000; Benitez-Silva, Buchinsky & Ruse, 2002).

In sum, the conundrum faced by disabled adult workers historically has required a choice between working or not working contingent upon receiving health insurance through public or private mechanisms. The TTWA's laudable intention was to address this untenable choice faced by workers, giving them a substantial "grace period" of remaining qualified for public health benefits while they underwent retraining and reentered the job marker. What has been unclear is whether the TTWA has made a difference in the employment rates of adult workers who report one or more chronic condition. This research addresses this gap in the disability literature stream by examining whether adult workers have improved rates of reported employment in the decade from 1995 to 2005.

RESEARCH QUESTIONS AND HYPOTHESES

The research questions and hypotheses addressed in this study are:

Research Question 1: Are there observable differences in rates of employment among working-age adults who have one or more chronic mental or physical condition in two national surveys (1995 and 2005 NHIS) conducted 10 years apart?

H1 (null): There will be no differences in rates of work activity observed in 1995 compared to 2005 using a national sample of working-age adults reporting one or more chronic condition.

Research Question 2: How do employment rates differ between 1995 and 2005 among adults who have physical or mental conditions?

H2: Rates of reported work activity among adults reporting a physical or mental condition will be higher in 2005 than rates reported in1995.

Research Question 3: How do reported rates of work among adults with chronic conditions differ across regions of the United States?

H3: Rates of work activity among adults reporting one or more chronic condition will not vary across regions of the United States.

METHODS AND DATA

Data Sources

The data for this study comes from two large national surveys, the 1995 and 2005 National Health Interview Surveys (NHIS). The 1995 and 2005 NHIS surveys are large and complex surveys consisting of a household survey, individual person survey and several supplements. Only the person-level surveys for 1995 and 2005 were used to address the research questions in this study.

The 1995 NHIS survey is larger than the 2005 NHIS survey with a total of 102,467 person records, while the 2005 NHIS survey has 92, 148 person records. Both samples were limited to adults ages 25-64 resulting in a total sample of 47, 865 adults for 1995 and 52,496 adults for 2005. The decision was made to strategically limit the analyses to adults with one or more chronic condition (lasting three months or longer). Hence, both data sets were further reduced to include only adults between the ages of 25 and 64 who report one or more chronic condition resulting in a sample of 18, 879 for 1995 and 6,148 for 2005. Chronic conditions include mental illnesses and related conditions such as depression. We noted the differences in size of the samples of adults with chronic conditions. This is ostensibly attributable to the disability focus of the 1995 NHIS survey. In order to address this difference we used the national weights and conducted our analyses using population weights and strata. The NHIS employs a national sample design with primary sampling units (PSUs); strata and clustering. Hispanics and blacks are over sampled at rates of approximately two and 1.5 times the rate of other households.

Key Variables

Key dependent and independent variables are described in Figure 1. For the multivariate analyses the primary outcome variable is reported work activity within the past two weeks.

United States Our model (Figure 2) is stated formally as: Reported Work (Working)= [dependant upon] (1) Type of condition, (2) Race/Ethnicity, (3) Age, (4) Marital Status, (5) Region of the U.S. and (6) Years of education Completed (1).

Logistic regression is carried out on a combined sample of males and females, with male/female included as a separate predictor. We acknowledge and control for well-documented differences between male and female labor force participation and wage earnings (Chirikos & Nestel, 1985; Luft, 1975, Stern, 1996), however, we were not able to conduct separate analyses on males and females using the 2005 survey due to small cell sizes of some of the chronic conditions. Marital status is included as a separate predictor because of previous research suggesting that married persons who are chronically ill may not work as much as their unmarried counterparts, which include those who are separated, divorced, widowed or never married, because of the availability of financial support from a working spouse.

Statistical analyses were performed using Stata 9, (2005). We also employ Stata's survey software option allowing for analyses using national weights and strata. Descriptive statistics were first calculated to describe the population in terms of sociodemographic status, health and health care status. Chi-square coefficients were employed to determine the existence of significant relationships among categorical variables.

In order to evaluate the effect of chronic conditions on the likelihood of work activity, we elected to use multivariate logistic regression (Hosemer and Lemeshow, 2001; Long and Freese, 2001) using Stata 9 survey command options in order to derive national estimates using the NHIS assigned final weights, strata and PSU's.

RESULTS

The results of statistical analyses confirm the importance of type of condition, race/ethnicity and age on the probability that a given individual is employed. We also noted differences in percent sizes across age categories that we attribute to the "graying" or aging of baby boomers across America. We also observed changes in the percent reporting types of chronic diseases from 1995 to 2005.

Table 1: Differences Among Adults Reporting Chronic Conditions.

Between 1995 and 2005, the average age of working age adults reporting a chronic condition increased by 3-4 years, from a mean of 44.3 years for males and 44.1 for females in 1995 to a mean of 48 years for male respondents and 49 years for females in 2005. Table 1. Examining reported employment by age category reveals that the percent of adult respondents in 1995 reporting employment in the younger age categories (Age 25-44) was significantly higher for both males and females than those observed in 2005. Rates of employment reported in the 1995 NHIS survey remained higher for males up to retirement age, whereas it was equal for females in the age categories 55-64 in 1995 and 2005. Figures 3 and 4.

Shifts in the percents in age categories confirm that adults in the older 45-65 age group categories have increased. In 1995, 22 percent of males age 25-34 with a chronic condition and 23 percents of females reported employment within the past two weeks. In contrast, in 2005 the percentage employed in this age group dropped to 14 percent for males and 13 percent for females. These differences are significant at p <.05. The downward trend in the younger age categories is likewise seen in the age 35 -44 age group, with 29 percent respondents reporting employment in 1995 compared to 22 percent and 21 percent for females in 2005. The trend then appears to reverse itself for the older age categories--with larger cell sizes observed in 2005 for those in the 45 to 54, 55 to 61 and 62-64 age groups, compared to 1995. These age differences are significant at p <.05.

Differences in employment of adults with chronic conditions, by racial category, also revealed some significant shifts between 1995 and 2005. The percent of whites who reported working dropped from 78 percent for males and 64 percent for females in 1995 to 64 percent for males and 54 percent for females in 2005, while employment rates for blacks dropped from 57 percent to 45 percent for black males and from 55 percent to 47 percent for black females. Differences shown in the tables are significant only for whites and blacks. (p < .05)

Comparison of employment by marital status revealed that the percent of those reporting that they were currently married had dropped from 74 percent for males and 64 percent for females in 1995 to 52 percent for males and 46 percent for males in 2005.

Examining employment by reporting of chronic disease revealed that employment stayed fairly stable for those reporting diabetes, renal problems neurological and cancer problems, but that the rates of those with cardiovascular disease had increased significantly from 7 percent for males and 6 percent for females in 1995 to 14 percent (males) and 12 percent (females) in 2005, (p < .05). This shift in the percent reporting cardiovascular disease corresponds to other trends reported elsewhere [cite]. Likewise, those reporting any mental condition (includes depression, bipolar disorder, schizophrenia, paranoia, etc) increased by three percent between 1995 and 2005. (p < .05).

Table 2 Comparing Work Activity By Race/Ethnicity And Region

Differences in employment by region of residence of adults with disability producing conditions were also noted across region of the United States. We chose to analyze employment rates across regions by race and ethnicity. Table 2 shows that reported rates of employment for white females and males is highest among white males in the Midwest followed by similarly high employment rates in the Northeast and West.

Overall, we note a downward trend in 2005 across all regions in reported employment rates for all race and ethnic groups for adults with chronic, disability associated conditions, with the exception of females in the "other racial/NonHispanic" group category, which increased by four (4) percent between 1995 and 2005.

Table 3: Comparing Rates Of Work Activity Among Adults Reporting Mental Or Physical Conditions

In 1995 adults reporting chronic physical conditions had higher employment rates than adults reporting physical conditions in 2005. Males with physical conditions reported employment at a rate of 72 percent, compared to 59 percent in 2005. Females in 1995 report a 59 percent employment rate, compared to 52 percent in 2005. These differences are significant at p < .05.

Mental conditions continue to be extremely disabling with 44 percent of males in 1995 reporting employment and just 33 percent of males in 2005. The figures are even more dismal for females, with just 38 percent of females with a chronic mental condition reporting employment in 1995 and an six (6) percent drop to 32 percent in 2005. These differences are significant at P < .05.

Table 4: Multivariate Analysis: Effects Of Disease And Disability On Employment Outcomes

Table 4 provides the results of multivariate logistic regression examining the likelihood of employment given type of condition, age, marital status, race and ethnicity and region of the country. The effects of chronic conditions and impairments, particularly mental conditions, on the probability of work are particularly noteworthy. Table 4 reveals important evidence of the work-disabling effects of certain health conditions and impairments providing the odds ratios, confidence intervals and significance levels, as well as marginal effects which provide additional useful evidence when interpreting logit or probit coeffients or odds ratios. Each individual variable's marginal effect provides the probability of a given outcome at the (variable's) mean resulting in more specificity of the importance of each individual predictor on the outcome variable (Stata, 2005). Marginal effects (Dx/Dy) represent discrete change of the dummy variable from 0 to 1, or the effect of a one unit increase in an independent variable on the overall probability of the dependent variable's change in odds or probability.

* 1995 NHIS Survey Multivariate Analyses:

In 1995, the odds that a female with a chronic condition reported working within the past two weeks was 48 percent less than for a male, age 25-34. Table 4.

This outcome is significant at p < .05. The effects of chronic conditions on the probability of work are significantly negative with the exception of respiratory conditions which are slightly positive, but only at the p < .10 level. If a person has a sensory disorder (difficulty with hearing or seeing), they are .20 percent less likely to work, while musculoskeletal disorders are 30 percent less likely to work. Most notably, persons with mental disorders are 73 percent more likely not to work than persons without a mental disorder. In 1995 a person with diabetes was 54 percent less likely to work than a person who did not report having diabetes. However, somewhat surprisingly, persons with respiratory conditions are 10 percent more likely to work, but at a significance level of p<.10. Both blacks and Hispanics are 44 percent less likely to work than whites, while we observed no significant effect for other racial categories (Asians, Indian, and Pacific Islanders).

Examining the effect of age we see that in 1995 the risk of a person with a chronic condition not working rose with age. Those in the 45-54 years category are 19 percent less likely to work at (p< .05), while those who are 55-61 years old were 63 percent less likely (P<.05). Most notably, individuals ages 62-64 with a chronic condition are 86 percent less likely to work (p < .05).

Region of the United States also appear to play a role in the likelihood of working. Here the referent category is Midwest. Persons with chronic conditions residing in the Northeast are 16 percent less likely to work, while persons residing in the South or West are 15 percent less likely. (P < .05).

2005 NHIS Survey Multivariate Analyses

Examining the outcomes of the multivariate logistic regression using 2005 NHIS person survey we also observed that females were 34 percent less likely than males to report work (P < .05). Reported work rates for persons with cardiovascular disease stayed nearly the same, as did sensory and digestive disorders. Odds ratios for persons with respiratory conditions fell substantially from 1995. In 2005 a person reporting a respiratory condition was 38 percent less likely to work than others in the sample, while persons with caner were 72 percent less likely to report working than those who were surveyed in 1995 (P < .05). Persons reporting chronic renal conditions were 69 percent less likely to report working than others included in our 2005 sample (P < .05). The odds of reported employment also fell by region of the U.S. as compared to 1995. Differences in odds of reporting employment by age also changed--to more positive likelihood as compared to the reference category, 25-34 years (p < .05). This trend was noted across all age categories and is attributed to the changing age average of baby-boomers across the United States.

DISCUSSION OF LIMITATIONS AND CONCLUSIONS

Somewhat paradoxically this analyses reveals good news and bad news. While overall employment has improved somewhat in the older age categories, the bad news is that these age groups have greater frequency of chronic conditions. Conditions such as cancer, diabetes, cvd, musculoskeletal conditions and sensory impairments are markers for "age" and reflect the progression of baby boomers heading towards retirement age, and do not necessarily reflect improvement in disability outcomes or TTWA retraining programs. While reported employment for all racial and ethnic categories fell--the percentage of black and white males with chronic conditions reporting employment fell by 12 percent, but the same magnitude drop in reported employment was not observed for Hispanics or other racial categories--nor was it observed to the same extent in female respondents.

As to differences observed for specific conditions, overall employment worsened for persons with cardiovascular disease and mental conditions. However, it is not known whether these observations are due to overall exacerbation of disease states commonly observed as one advances in age.

This research also documents changing employment trends across regions of the United States. It is not clear whether reduced employment is an outcome of more favorable public assistance programs in certain states and regions of the country, or whether it was merely more difficult to find employment.

Particularly noteworthy are the sharp downward differences observed in employment among those reporting mental conditions. These outcomes are noteworthy for policy makers considering continued mental health parity laws to provide those with mental conditions with continued health insurance coverage.

Limitations of This Study:

This study, though national in scope is not without its limitations. Though suggestive of an overall downward trend in employment for persons with chronic conditions, we note that the 1995 and 2005 NHIS surveys were somewhat different in their focus, with the 1995 survey specifically targeting disabled or chronically ill individuals. The same was not true for the 2005 sample. Moreover, changing economic conditions between 1995 and 2005, diversion of national resources to respond to international demands rather than domestic programs makes it more difficult to compare these two national samples. Finally, we note that the employment trends may be attributable to the increasing average age of Americans. As baby boomers age it may not only be more difficult to find work, it also may become more difficult to perform work tasks. These trends and limitations are noteworthy and particularly worthy of further research.

REFERENCES

Americans with Disabilities Act, (ADA), 2002. 42 U.S.C. [section]12100 et seq.

Anderson, C.L. (2000). Deserving disabilities: Why the definition of disability under the Americans with Disabilities Act should be revised to eliminate the substantial limitation requirement, Missouri Law Review 65: 83-150.

Baldwin, M., & Johnson, W. G., (1992). Estimating the employment effects of wage discrimination. The Review of Economics and Statistics, 446-455.

Black, SA, Ray LA, Markides, KS The prevalence and health burden of self-reported diabetes in older Mexican Americans: findings from the Hispanic established populations for epidemiologic studies of the elderly, American Journal of Public Health 89(4) 546-552 1999.

Bolin, J. (2002). Diseases and Disability: Important Factors Affecting Work, Job Based Insurance and Hourly Wages. Dissertation. Pennsylvania State University, PA.

Bound, J., (1989). The health and earnings of rejected disability applicants. American Economic Review, 79, 482-503

Chirikos, T. N., & Nestel, G. (1985). Further evidence on the economic effect of poor health. The Review of Economics and Statistics, 67, 61-69.

Eberhart, MS, Ingram DD, Makuc, DM et al. Urban and Rural Health Chartbook. Health United States, 2001. Hyattsville, Maryland: National Center for Health Statistics. 2001.

Fried, VM, Prager, K, Mackay AP, Xia, H Chartbook on trends in the health of Americans. Health, United States, 2003. Hyattsville, Maryland: National Center for Health Statistics. 2003.

GAO-01-35 (2001). Social Security Reform: Potential effects on SSA's Disability Programs and Beneficiaries.

GAO/T-HEHS-96-62, (1996). SocialSecurity: Disability Programs Lag in Promoting Return to Work.

GAO/HEHS-98-39, (January 12, 1998). Social Security Disability Insurance: Multiple Factors Affect Beneficiaries' Ability to Return to Work.

Hosmer, D. W., and S. Lemeshow. Applied Logistic Regression, 2nd Ed. New York, NY: John Wiley & Sons Inc. Publishing.

Kaiser Commission on Medicaid and the Uninsured (April, 2001). Medicaid's role for the disabled population under age 63. Accessed at www.kff.org, September 16, 2003.

Long, J.S., Freese J., (2001). Regression models for categorical dependent variables using Stata, Stata Press Publication, STATA Corporation, College Station, TX.

Luft, H. (1975). The impact of poor health on earnings, The Review of Economics and Statistics, 57, 43-57. In Thompson-Hoffman, S. Storck, I. F. (1991). Disability in the United States: A portrait from national data; Springer Publishing Co. New York, NY.

Massanari, L. (June 28, 2001). Statement of Larry G. Massanari, Acting Commissioner, Social Security Administration Testimony Before the Subcommittee on Social Security of the House Committee on Ways and Means Hearing on Social Security Disability Program's Challenges and Opportunities.

Pope AM, Tarlov AR. Disability in America: toward a national agenda for prevention. Washington, DC: National Academy Press, 1991.

Stapleton, D. Livermore, G. Scrivner, S. Tucker, A. The Lewin Group, Inc. (1997, Oct). Exploratory Study of Health Care Coverage and Employment of People with Disabilities: Literature Review, Report of DHHS and The Lewin Group, (found at) http://aspe.os.dhhs.gov/daltcp/reports/eshcclit.htm.

StataCorp. 2005. Stata Statistical Software: Release 9.0. College Station, TX: Stata Corporation

Stern, S. (1996). Semiparametric estimates of the supply and demand effects of disability on labor force participation, Journal of Econometrics 71: 49-70.

Stoddard, S., L. Jans, J. Ripple, and L. Kraus. Chartbook on Work and Disability in the United States, 1998. An InfoUse Report. Washington, D.C.: U.S. National Institute on Disability and Rehabilitation Research, 1998.

Ticket to Work Act (TTWA) (1999) Public Law 106-170. Effective January 2001.

Ticket to Work Advisory Panel. Available at http://www.ssa.gov/work/panel/. Accessed May 19, 2006.

U.S. Census Bureau, (2005). Disability: Labor Force Status--Work disability status of civilians 16-74 years old, by educational attainment and sex: 2005. Available at: http://census.gov/hhes/www/disability/cps/cps205.h tml Accessed 5/22/06.

ENDNOTE

(1.) The model used in these analyses does not include years of education because of reporting differences in the 1995 and 2005 surveys.

JANE NELSON BOLIN

Texas A&M Health Sciences Center

Table 1
Comparison of population means and percents using two national
samples of adults with one or more chronic, disabling condition.
(1995 NHIS and 2005 NHIS)

 1995 NHIS 2005 NHIS
 Males Females Males Females

Mean Age 44 yrs 44 yrs 48 yrs 49 yrs

Percent who report 74% 62% 62% 52%
employment (all)

Percent who report
employment (by age)

Age 25-34 82% ** 67% ** 77% 55%
Age 35-44 82% ** 70% ** 70% 62%
Age 45-54 76% ** 66% ** 67% 60%
Age 55-61 62% ** 45% 53% 45%
Age 62-64 36% ** 26% 32% 26%

Percent who report
employment by race/ethnicity

White NonHispanic 78% 64% 64% 54%
Black NonHispanic 57% 55% 45% 47%
Hispanic 67% 50% 64% 45%
Other NonHispanic 74% 65% 64% 55%
Marital Status
Married 74% 64% 52% 46%
Separated/Divorced 11% 18% 21% 26%
Widowed 1% 6% 21% 26%
Never married 13% 12% 18% 16%

Percent with selected chronic
diseases

Diabetes 4% 5% 5% 5%
Sensory problem 16% ** 7% ** 3% 3%
Skin and skeletal 19% ** 23% ** 16% 16%
problem
Cardiovascular 7% ** 6% ** 14% 12%
disease
Mental condition 7% ** 7% ** 10% 11%
Renal problem 1% 2% 1% 1%
Digestive problem 6% ** 6% ** 2% 2%
Neurological 5% 9% 5% 8%
Cancer 1% 2% 2% 1%
Respiratory 13% ** 17% ** 6% 7%

Source: Author's calculations using the 1995 and 2005 National
Health Interview Surveys (person file)

Table 2
Comparing employment rates by region of residence among adults
reporting chronic conditions by race and gender (United States,
1995 and 2005)

 Males 1995
 White Black Hispanic Other

Northeast 77% 48% 61% 77%
Midwest 80% 52% 72% 64%
South 74% 62% 66% 77%
West 77% 58% 69% 75%

 Females 1995
 White Black Hispanic Other

Northeast 65% 49% 66% 43%
Midwest 67% 53% 59% 56%
South 61% 58% 60% 52%
West 62% 48% 68% 50%

 Males 2005
 White Black Hispanic Other

Northeast 63% 49% 49% 79%
Midwest 72% 29% 57% 60%
South 55% 50% 64% 67%
West 67% 44% 71% 61%

 Females 2005
 White Black Hispanic Other

Northeast 57% 39% 35% 47%
Midwest 59% 52% 61% 60%
South 48% 46% 46% 56%
West 54% 52% 46% 54%

Table 3: Comparison of work activity "in the past two weeks" by
persons with chronic mental or physical conditions, by gender

 Chronic Mental Chronic Physical
 Condition Condition
 1995 2005 1995 2005
Employment
Males: All Ages 44% 33% 72% 59%

Employment
Ffemales: All Ages 38% 32% 59% 52%

Source: Author's calculations using the 1995 and 2005 National
Health Interview Surveys.

Table 4: Multivariate Logistic Regression comparing likelihood of
work using the 1995 and 2005 NHIS Surveys

 1995 NHIS
 Marg.
 Odds P Eff
Working Ratio Value dy/dx Cl

Female 0.52 *** -0.139 0.48 -- 0.56
Sensory 0.79 *** -0.051 0.72 -- 0.88
Musculoskeletal 0.71 *** -0.075 0.66 -- 0.77
CVD 0.44 *** -0.191 0.38 -- 0.52
Mental Condition 0.27 *** -0.316 0.22 -- 0.29
Diabetes 0.46 *** -0.183 0.38 -- 0.55
Respiratory 1.11 * 0.022 0.99 -- 1.25
Cancer 0.78 ** -0.053 0.62 -- 0.99
Neurological 0.55 *** -0.138 0.48 -- 0.62
Digestive 0.85 ** -0.036 0.72 -- 0.99
Renal 0.64 *** -0.102 0.51 -- 0.81
Hispanic 0.56 *** -0.131 0.50 -- 0.63
Black, NonHispanic 0.56 *** -0.132 0.51 -- 0.62
Other, NonHispanic 0.97 -0.007 0.78 -- 1.20
Age 35-44 1.06 0.012 0.96 -- 1.17
Age 45-54 0.81 *** -0.045 0.73 -- 0.91
Age 55-61 0.37 *** -0.233 0.33 -- 0.41
Age 62-64 0.14 *** -0.459 0.12 -- 0.16
Widowed 0.84 * -0.037 0.70 -- 1.01
Separated/Divorced 1.03 0.006 0.93 -- 1.14
Never Married 0.81 *** -0.046 0.72 -- 0.91
North East 0.84 *** -0.039 0.73 -- 0.96
South 0.85 *** -0.034 0.76 -- 0.97
West 0.85 ** -0.034 0.74 -- 0.99

y = Pr(working) 0.69
* p<.10
** p<.05
*** p<.01

 20055 NHIS
 Marg.
 Odds P Eff
Working Ratio Value dy/dx Cl

Female 0.66 *** -0.10 0.58 -- 0.75
Sensory 0.83 -0.05 0.58 -- 1.19
Musculoskeletal 1.11 0.03 0.94 -- 1.32
CVD 0.45 *** -0.20 0.36 -- 0.55
Mental Condition 0.35 *** -0.26 0.28 -- 0.44
Diabetes 0.53 *** -0.16 0.38 -- 0.73
Respiratory 0.62 *** -0.12 0.48 -- 0.82
Cancer 0.28 *** -0.30 0.15 -- 0.52
Neurological 0.57 *** -0.14 0.44 -- 0.73
Digestive 0.85 -0.04 0.44 -- 1.65
Renal 0.31 *** -0.28 0.16 -- 0.59
Hispanic 0.74 *** -0.07 0.61 -- 0.90
Black, NonHispanic 0.65 *** -0.11 0.53 -- 0.79
Other, NonHispanic 0.94 -0.02 0.59 -- 1.51
Age 35-44 1.03 0.01 0.82 -- 1.29
Age 45-54 0.93 -0.02 0.75 -- 1.16
Age 55-61 0.50 *** -0.17 0.39 -- 0.63
Age 62-64 0.19 *** -0.38 0.15 -- 0.25
Widowed 1.01 0.002 0.74 -- 1.38
Separated/Divorced 1.06 0.01 0.90 -- 1.24
Never Married 0.77 *** -0.07 0.63 -- 0.93
North East 0.77 *** -0.07 0.63 -- 0.93
South 0.72 *** -0.08 0.62 -- 0.85
West 0.83 ** -0.05 0.69 -- 1.00

y = Pr(working) 0.57
* p<.10
** p<.05
*** p<.01

Figure 1

Key Variables

Independent Variable
 * Working Worked in the past two weeks

Dependant Variables
 * Gender Males, Female [0,1]
 * Chronic Conditions Sensory, musculoskeletal, CVD,
 Mental, Diabetes, Pulmonary,
 * Race/Ethnicity Cancer, Neurological, Digestive and
 * Age Renal.
 * Marital Status White, Black, Hispanic & Other
 25-34, 35-44, 45-54, 55-61,
 * Region of the U.S. 62-64
 Married, Separated/Divorced, Never
 married, Widowed.

Figure 2

Multivariate model

 Multivariate Logistic Regression Model

 (Survey Logistic Regression)

Logistic Regression Type of Conditions +Race/Ethnicity/Age+
(Any Work Activity) Marital Status + Age + region of the
 country + Education + Variables for
 Weights & Strata

Figure 3

Comparing Reported Employment By Age Category: 1995 vs. 2005 (Females)

 1995 2005

Age 25-34 67% 59%
Age 35-44 70% 62%
Age 45-54 66% 60%
Age 55-61 45% 45%
Age 62-64 26% 26%

Note: Table made from bar graph.

Figure 4

Comparing Reported Employment By Age Category: 1995 vs. 2005 (Males)

 1995 2005

Age 25-34 82% 77%
Age 35-44 82% 70%
Age 45-54 76% 67%
Age 55-61 62% 53%
Age 62-64 36% 32%

Note: Table made from bar graph.