Socioec oecon onom omic S Status, Perceptions of of Pain, a and t the Disp spari rity ty i in n SSDI Rece eceipt David M. Cutler, Harvard and NBER Ellen Meara, Dartmouth and NBER Susan Stewart, NBER This research was supported by a grant from the U.S. Social Security Administration (SSA) as part of the Retirement and Disability Research Consortium (RDRC). The findings and conclusions are solely the those of the authors and do not represent the views of SSA, any agency of the federal government, or the NBER Retirement and Disability Research Center.
Large disparity in functional limitations and joint pain by education Share of People Reporting Knee Pain Average Number of Functional Limitations 45% (out of 12) 40% 4 Gap is 5-10 percentage points 35% 3 30% 25% 2 20% 15% 1 10% 5% 0 0% 25 30 35 40 45 50 55 60 25 30 35 40 45 50 55 60 <=HS Some College Coll Grad <=HS Some College Coll Grad Source: NHIS, 2009-16. Functional limitations include walking, climbing, standing, sitting, stooping, reaching, grasping, carrying, pushing, shopping, socializing, and relaxing.
Why is this? Four theories I. It’s in their knees • Knees of less educated people have more structural damage II. It’s in the environment • The tasks required of less educated people are more demanding, and this leads to more pain • BMI differs by education, and this leads to more pain III. It’s in their head • Less educated people have more ‘despair’ and this influences their pain perception and physical functioning IV. It’s in the medicine cabinet • Medical treatments are better for the better educated
Theory I: Is it in their knees? • National Health and Nutrition Examination Survey III (1988-94) • Ages 60-74 • N=3,886 people (~1,578 with x-rays; only during 1991-94) • X-rays to measure knee arthritis. Score using Kellgren-Lawrence (KL) Classification 0=Normal Share of People Reporting Knee Pain 1=Doubtful/Possible by X-ray Finding 100% 2=Mild 75% 3=Moderate Arthritis 4=Severe 50% • Two education groups: ≤ HS, CG 25% • All findings age/sex/race adjusted 0% Normal Possible Mild Moderate Severe
Images of arthritic knees
Images of knees differ only slightly by education. Almost all of the difference is pain conditional on severity KL Score for Knee Images, 1992-94 % Reporting Knee Pain 80% 60% Modest increase in ‘normal’ knees 70% 50% 60% 40% 50% Much lower pain report at every level of knee arthritis 40% 30% 30% 20% 20% 10% 10% 0% 0% Normal Possible Mild Moderate/Severe Normal Possible Mild Moderate Severe KL Score of Knee KL Score of Knee <=HS College Grad <=HS College grad ~85% of the difference in pain is a result of lower pain reports given the degree of arthritis, not the amount of arthritis.
Is it just reporting? Unlikely 20 meter Walking Time • Very specific pain reports 20 • Doesn’t go away at retirement 19 • Self-reported pain tolerance does 18 Seconds not differ by education 17 • Pain report is correlated with 16 physical functioning 15 14 0 10 20 30 40 50 60 Average KOOS Pain Score† † Pain score is subtracted from 100 so that a higher value corresponds to more pain.
Theory II: Environmental characteristics • Continuous NHANES has information on longest job worked • 40 2-digit occupations (e.g., Textile, apparel, and furnishings machine operators) • Matched to characteristics of jobs from 1977 Dictionary of Occupation Titles (England and Kilbourne) • Principal factor from strength, climbing, stooping, reaching
Knee Pain and Physical Requirements on the Longest Job • Job demands are correlated Job Requirements and Knee Pain with knee pain. 35% • Other joints too, but biggest Health service occupations effect is for knee and hip pain. 30% Percent with pain in either knee Other handlers and eqt cleaners 25% • This is NOT true for measures Construction trades of abstract / routine / manual 20% jobs from Autor et al. 15% Other prof Construction laborers specialty • About 1/3 of the difference in 10% Records processing knee pain is a result of Farm and nursery workers 5% differences in physical -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 requirements on the job. Physical Factor Score Data from continuous NHANES, 1999-2004, ages 45-74. Includes people with a longest job that is not in the military.
Obesity Relationship between Knee Pain and Maximum BMI • Knee pain is highly 50% correlated with maximum BMI. • Also current BMI conditional 40% Percent with pain in either knee on maximum BMI 30% • This is independent of the effect of job demands. 20% • About 1/3 of the 10% difference in knee pain by education is due to higher 0% rate of obesity. 15 20 25 30 35 40 45 50 Maximum BMI Data from continuous NHANES, 1999-2004, ages 45-74.
Theory III. It’s in their heads (despair) • MIDUS: Midlife in the US (N~4,000) • Surveyed in mid-1990s (wave A); resurveyed in mid-2000s (wave B) and mid- 2010s (wave C) • Keep people aged 45-74 in last wave. • Dependent variable, Wave C: “Do you have chronic pain, that is do you have pain that persists beyond the time of normal healing and has lasted from anywhere from a few months to many years?” • “Where is your pain primarily located – knees?” • Relate chronic pain in wave C to obesity in wave B, job chars in wave B, and psychological status in wave B
Psychological measures • Life satisfaction (0-10 scale) • Affect: positive and negative (1-5 scale) • Many of these differ by • Control: Personal mastery + perceived education, but the constraints (1-7 scale) relationship with knee pain • Psychological well-being (1-21 scales) is modest. • Only 10% of difference in • Positive relations with others • Self-acceptance knee pain by education is • Autonomy associated with • Personal growth psychological well-being. • Environmental mastery • Purpose in life
Theory IV: It’s in the medicine cabinet • Treatment for knee pain has historically been limited in use or not very effective. • Non-prescription medications (Ibuprofen, Acetaminophen) • Prescription pain relievers (Vioxx, OxyContin) • (Later) knee replacement • NHANES asks about some of these: Any aspirin, Frequent use Any prescription Knee replacement Ibuprofen, (>=10 times)* pain reliever* Acetaminophen* 72% 36% 14% 0.5% *Past month
Treatment rates vary little by education Percent of People Utilizing Indicated Treatment by Education • Treatment rates do not 100% differ greatly by 90% 80% education. 70% 60% • Hard to tell about efficacy 50% because of endogeneity 40% of treatment. 30% • People with more pain 20% use more pain-related 10% care. 0% Any non-Rx med Frequent non-Rx med Rx med <=HS College grad
Summary of results • It’s in their knees • Knees of less educated people have more structural damage • It’s in the environment (~2/3 of the difference in knee pain by education) • The tasks required of less educated people are more demanding, and this leads to more pain • BMI differs by education, and this leads to more pain • It’s in their head • Less educated people have more ‘despair’ and this influences their pain perception and physical functioning • It’s in the medicine cabinet • Medical treatments are better for the better educated • NOT a big deal in this setting.
Implications • For SSDI/SSI • Pain is real but can’t be found by a clinical test • For the future of pain • Work will get more physically demanding over the next decade (home health aides + personal care aides > computer programmers) • Maximum BMI is continuing to rise • For medical care • Perhaps the most important issue for biomedical research
Recommend
More recommend
