Search

Abstract. Background: Around the globe, 800,000 people die from suicide every year. Despite being one of the leading causes of death, suicide remains a low public health priority. Korea has the second highest total suicide rate among Organisation for Economic Co-operation and Development (OECD) countries. Aims: The aim of this study was to explore how changes of job status influence suicidal risk in Korea, which lags behind other OECD countries in job security because temporary and part-time jobs are more prevalent in Korea. Method: We made use of a large longitudinal dataset, the Korea Health Panel (KHP). Results: Our findings revealed that a negative change in employment status increased the risk of suicide, but only for males. Limitations: Some individuals might intentionally change their job status, but the data do not indicate why the job status of an individual changes. Conclusion: These findings provide useful insights regarding the Korean labor market. In particular, tackling the issue of job stability, providing training polices for the unemployed and under-employed, and considering social insurance schemes may help to reduce suicide risk.

Economists have been researching effects of minimum wages on unemployment, poverty, income inequality, and educational attainment for over 60 years. Epidemiologists have only recently begun researching minimum wages even though unemployment through education are central topics within social epidemiology. Buszkiewicz et al. (Am J Epidemiol. 2021;190(1):21–30) offer a welcome addition to this nascent literature. A commanding advantage of Buszkiewicz et al.’s study over others is its distinction between a “likely affected” group comprised of workers with ≤12 years of schooling versus “not likely affected” groups with ≥13 years of schooling. But there are disadvantages, common to other studies. Buszkiewicz et al. use cross-sectional data; they include the self-employed as well as part-time and part-year workers in their treatment groups. Their definitions of affected groups based on education create samples with 75% or more of workers who earn significantly above minimum wages; definitions are not based on wages. Inclusion of workers not subject to (e.g., self-employed) or affected by minimum wages biases estimates toward the null. Finally, within any minimum wage data set, it is the state—not federal—increases that account for the lion’s share of increases and that form the natural experiments; however, state increases can occur annually whereas the development of chronic diseases might take decades.

BACKGROUND Asynchronous telepsychiatry (ATP; delayed-time) consultations are a novel form of psychiatric consultation in primary care settings. Longitudinal studies comparing clinical outcomes for ATP with synchronous telepsychiatry (STP) are lacking. OBJECTIVE This study aims to determine the effectiveness of ATP in improving clinical outcomes in English- and Spanish-speaking primary care patients compared with STP, the telepsychiatry usual care method. METHODS Overall, 36 primary care physicians from 3 primary care clinics referred a heterogeneous sample of 401 treatment-seeking adult patients with nonurgent psychiatric disorders. A total of 184 (94 ATP and 90 STP) English- and Spanish-speaking participants (36/184, 19.6% Hispanic) were enrolled and randomized, and 160 (80 ATP and 80 STP) of them completed baseline evaluations. Patients were treated by their primary care physicians using a collaborative care model in consultation with the University of California Davis Health telepsychiatrists, who consulted with patients every 6 months for up to 2 years using ATP or STP. Primary outcomes (the clinician-rated Clinical Global Impressions [CGI] scale and the Global Assessment of Functioning [GAF]) and secondary outcomes (patients’ self-reported physical and mental health and depression) outcomes were assessed every 6 months. RESULTS For clinician-rated primary outcomes, ATP did not promote greater improvement than STP at 6-month follow-up (ATP vs STP, adjusted difference in follow-up at 6 months vs baseline differences for CGI: 0.2, 95% CI −0.2 to 0.6; P=.28; and GAF: −0.6, 95% CI −3.1 to 1.9; P=.66) or 12-month follow-up (ATP vs STP, adjusted difference in follow-up at 12 months vs baseline differences for CGI: 0.4, 95% CI −0.04 to 0.8; P=.07; and GAF: −0.5, 95% CI −3.3 to 2.2; P=.70), but patients in both arms had statistically and clinically significant improvements in both outcomes. There were no significant differences in improvement from baseline between ATP and STP on any patient self-reported ratings at any follow-up (all P values were between .17 and .96). Dropout rates were higher than predicted but similar between the 2 arms. Of those with baseline visits, 46.8% (75/160) did not have a follow-up at 1 year, and 72.7% (107/147) did not have a follow-up at 2 years. No serious adverse events were associated with the intervention. CONCLUSIONS This is the first longitudinal study to demonstrate that ATP can improve clinical outcomes in English- and Spanish-speaking primary care patients. Although we did not find evidence that ATP is superior to STP in improving clinical outcomes, it is potentially a key part of stepped mental health interventions available in primary care. ATP presents a possible solution to the workforce shortage of psychiatrists and a strategy for improving existing systems of care. CLINICALTRIAL ClinicalTrials.gov NCT02084979; https://clinicaltrials.gov/ct2/show/NCT02084979.

This literature review analyzes studies from the US, Canada, the UK, and Europe from inception to April 1, 2021 and focuses on treatment designs, health outcomes, demographic categories and data issues. Study designs are classified as treatment-effect-on-the-treated (7 studies), intent-to-treat (37), and what may be called possible-effects-on-anyone (10). Treatment-effects-on-the-treated designs are best for addressing the longstanding question: does income affect health or vice versa? I argue that they are also better for estimating the overall effect of minimum wages on health. Health outcomes are grouped into seven broad categories, such as overall physiological health and behavior, and 33 narrow categories, such as self-rated health and smoking. Demographic categories include gender, race/ethnicity, and age. The preponderance of evidence suggests that studies relying on the treatment-effect-on-the-treated and possible-effects-on-anyone designs find minimum wages improve health; there is no preponderance of evidence for overall health within intent-to-treat designs. With respect to specific health outcomes and demographic categories, there is no preponderance of evidence, except for improving infant and child health. One data issue concerns whether either intent-to-treat or possible-effects-on-everyone studies are reliable given that likely more than 70 % of people in their samples earn substantially above minimum wages thereby favoring the null hypothesis. Treatment-effect-on-the-treated designs are likely the best designs, and findings are largely consistent in showing that minimum wages improve some measures of health, for example, financial anxiety.

BackgroundAsynchronous telepsychiatry (ATP; delayed-time) consultations are a novel form of psychiatric consultation in primary care settings. Longitudinal studies comparing clinical outcomes for ATP with synchronous telepsychiatry (STP) are lacking.ObjectiveThis study aims to determine the effectiveness of ATP in improving clinical outcomes in English- and Spanish-speaking primary care patients compared with STP, the telepsychiatry usual care method.MethodsOverall, 36 primary care physicians from 3 primary care clinics referred a heterogeneous sample of 401 treatment-seeking adult patients with nonurgent psychiatric disorders. A total of 184 (94 ATP and 90 STP) English- and Spanish-speaking participants (36/184, 19.6% Hispanic) were enrolled and randomized, and 160 (80 ATP and 80 STP) of them completed baseline evaluations. Patients were treated by their primary care physicians using a collaborative care model in consultation with the University of California Davis Health telepsychiatrists, who consulted with patients every 6 months for up to 2 years using ATP or STP. Primary outcomes (the clinician-rated Clinical Global Impressions [CGI] scale and the Global Assessment of Functioning [GAF]) and secondary outcomes (patients’ self-reported physical and mental health and depression) outcomes were assessed every 6 months.ResultsFor clinician-rated primary outcomes, ATP did not promote greater improvement than STP at 6-month follow-up (ATP vs STP, adjusted difference in follow-up at 6 months vs baseline differences for CGI: 0.2, 95% CI −0.2 to 0.6; P=.28; and GAF: −0.6, 95% CI −3.1 to 1.9; P=.66) or 12-month follow-up (ATP vs STP, adjusted difference in follow-up at 12 months vs baseline differences for CGI: 0.4, 95% CI −0.04 to 0.8; P=.07; and GAF: −0.5, 95% CI −3.3 to 2.2; P=.70), but patients in both arms had statistically and clinically significant improvements in both outcomes. There were no significant differences in improvement from baseline between ATP and STP on any patient self-reported ratings at any follow-up (all P values were between .17 and .96). Dropout rates were higher than predicted but similar between the 2 arms. Of those with baseline visits, 46.8% (75/160) did not have a follow-up at 1 year, and 72.7% (107/147) did not have a follow-up at 2 years. No serious adverse events were associated with the intervention.ConclusionsThis is the first longitudinal study to demonstrate that ATP can improve clinical outcomes in English- and Spanish-speaking primary care patients. Although we did not find evidence that ATP is superior to STP in improving clinical outcomes, it is potentially a key part of stepped mental health interventions available in primary care. ATP presents a possible solution to the workforce shortage of psychiatrists and a strategy for improving existing systems of care.Trial RegistrationClinicalTrials.gov NCT02084979; https://clinicaltrials.gov/ct2/show/NCT02084979.

We evaluate evidence for the effectiveness of raising minimum wages on various measures of public health within the US, Canada, the UK, and Europe. We search four scientific websites from the inception of the research through May 20, 2018. We find great variety (20+) in measured outcomes among the 33 studies that pass our initial screening. We establish quality standards in a second screening resulting in 15 studies in which we create outcome-based groups. Outcomes include four broad measures (general overall health, behavior, mental health, and birth weight) and eight narrow measures (self-reported health, "bad" health days, unmet medical need, smoking, problem-drinking, obesity, eating vegetables, and exercise). We establish criteria for "stronger" findings for outcomes and methods. Stronger findings include: $1 increases in minimum wages are associated with 1.4 percentage point (4% evaluated at mean) decreases in smoking prevalence; failure to reject null hypotheses that minimum wages have no effects for most outcomes; and no consistent evidence that minimum wages harm health. One "suggestive" finding is that the best-designed studies have well-defined treatment (or likely affected) and control (unaffected) groups and contain longitudinal data. The major methodological weaknesses afflicting many studies are the lack of focus on persons likely affected by minimum wages and omission of "falsification tests" on persons likely unaffected. An additional weakness is lack of attention to how findings might differ across populations such as teenagers, adults, men, women, continuously employed and unemployed persons. Research into health effects of minimum wages is in its infancy and growing rapidly. We present a list of "better practices" for future research.

Few population-based studies have used an econometric approach to understand the association between two cancer risk factors, obesity and stress. This study investigated sociodemographic differences in the association between obesity and stress among Korean adults (6,546 men and 8,473 women). Data were drawn from the Korean National Health and Nutrition Examination Survey for 2008, 2009, and 2010. Ordered logistic regression models and propensity score matching methods were used to examine the associations between obesity and stress, stratified by gender and age groups. In women, the stress level of the obese group was found to be 27.6% higher than the nonobese group in the ordered logistic regression; the obesity effect on stress was statistically significant in the propensity score-matched analysis. Corresponding evidence for the effect of obesity on stress was lacking among men. Participants who were young, well-educated, and working were more likely to report stress. In Korea, obesity causes stress in women but not in men. Young women are susceptible to a disproportionate level of stress. More cancer prevention programs targeting young and obese women are encouraged in developed Asian countries.

Background. Comprehensive economic evaluations have not been conducted on telemedicine consultations to children in rural emergency departments (EDs). Objective. We conducted an economic evaluation to estimate the cost, effectiveness, and return on investment (ROI) of telemedicine consultations provided to health care providers of acutely ill and injured children in rural EDs compared with telephone consultations from a health care payer prospective. Methods. We built a decision model with parameters from primary programmatic data, national data, and the literature. We performed a base-case cost-effectiveness analysis (CEA), a probabilistic CEA with Monte Carlo simulation, and ROI estimation when CEA suggested cost-saving. The CEA was based on program effectiveness, derived from transfer decisions following telemedicine and telephone consultations. Results. The average cost for a telemedicine consultation was $3641 per child/ED/year in 2013 US dollars. Telemedicine consultations resulted in 31% fewer patient transfers compared with telephone consultations and a cost reduction of $4662 per child/ED/year. Our probabilistic CEA demonstrated telemedicine consultations were less costly than telephone consultations in 57% of simulation iterations. The ROI was calculated to be 1.28 ($4662/$3641) from the base-case analysis and estimated to be 1.96 from the probabilistic analysis, suggesting a $1.96 return for each dollar invested in telemedicine. Treating 10 acutely ill and injured children at each rural ED with telemedicine resulted in an annual cost-savings of $46,620 per ED. Limitations. Telephone and telemedicine consultations were not randomly assigned, potentially resulting in biased results. Conclusions. From a health care payer perspective, telemedicine consultations to health care providers of acutely ill and injured children presenting to rural EDs are cost-saving (base-case and more than half of Monte Carlo simulation iterations) or cost-effective compared with telephone consultations.

We investigated Supplemental Nutrition Assistance Program (SNAP) participation among citizen, documented and undocumented immigrant hired crop farmworkers for ten recent years. We analyzed population representative data from the National Agricultural Workers Survey for 2003–2012 (N = 18,243 households). Time-chart, simple mean differences, and logistic regressions described farmworker household participation in SNAP. The 2008 financial crisis almost doubled SNAP-participation by agriculture households (6.5% in 2003–2007 vs. 11.3% in 2008–2012). The increasing SNAP-participation was found for citizen, documented and undocumented immigrant households. We found low participation among documented (OR 0.67, 95% CI 0.56–0.8) and undocumented immigrants (OR 0.63, 95% CI 0.54–0.74) compared to citizens. Low odds ratios (OR 0.70, 95% CI 0.55–0.89) were found for Hispanic-citizens as compared with non-Hispanic white-citizens. Our results may help inform the debate surrounding the effects of the financial crisis on SNAP-participation and on differences in participation among citizens, immigrants, Hispanics and non-Hispanics, the latter suggesting ethnic farmworker disparities in SNAP-participation.

Using longitudinal data from the Panel Study of Income Dynamics for 1997–2013 and difference-in-differences (DD) and difference-in-difference-in-differences (DDD) techniques, we estimate the effects of minimum wages on absence from work due to own and others’ (such as children’s) illnesses. We use person fixed effects within both linear and two-part models, the latter to explore changes at extensive and intensive margins. A lower educated group (likely affected by minimum wages) is compared with higher educated groups (likely unaffected). Within the lower educated group, we find higher minimum wages are associated with lower rates of absence due to own and others’ illness combined and due to own illness alone, but not associated with absence due to others’ illness. A $1 increase in the real minimum wage results in 19 % (in DD model) and 32 % (DDD) decreases in the absence rate due to own illness evaluated at the mean. These findings are strongest for persons who are not employed year-round and among the lowest wage earners. In additional analysis, we show that these effects are likely not due to changes in labor supply or job-related attributes. Instead, we find a possible mechanism: higher minimum wages improve self-reported health for lower educated workers.

Introduction: This study evaluates the financial impact of telemedicine outreach in a competitive healthcare market from a tertiary children’s hospital’s perspective. We compared the number of transfers, average hospital revenue, and average professional billing revenue before and after the deployment of telemedicine. Materials and Methods: This is ar etrospective review of hospital and physician billing records for patients transferred from 16 hospitals where telemedicine services were implemented between July 2003 and December 2010. Hospital revenue was defined as total revenue minus operating costs. Professional billing revenue was defined as total payment received as the result of physician billing of patients’ insurance. We compared the number of transfers, average net hospital revenue per year, and average professional billing revenue per year before and after the deployment of telemedicine at these hospitals. Results: There were 2,029 children transferred to the children’s hospital from the 16 hospitals with telemedicine during the study period. The average number of patients transferred per year to the children’s hospital increased from 143 pre-telemedicine to 285 post-telemedicine. From these patients, the average hospital revenue increased from $2.4 million to $4.0 million per year, and the average professional billing revenue increased from $313,977 to $688,443 per year. On average, per hospital, following the deployment of telemedicine, hospital revenue increased by $101,744 per year, and professional billing revenue increased by $23,404 per year. Conclusions: In a competitive healthcare region with more than one children’s hospital, deploying pediatric telemedicine services to referring hospitals resulted in an increased market share and an increased number of transfers, hospital revenue, and professional billing revenue.

Background:Earlier studies estimated annual income differences across specialties, but lifetime income may be more relevant given physicians’ long-term commitments to specialties.Methods:Annual income and work hours data were collected from 6381 physicians in the nationally representative 2004–2005

To assess potential reductions in premature mortality that could have been achieved in 2008 to 2012 if the minimum wage had been $15 per hour in New York City.Using the 2008 to 2012 American Community Survey, we performed simulations to assess how the proportion of low-income residents in each neighborhood might change with a hypothetical $15 minimum wage under alternative assumptions of labor market dynamics. We developed an ecological model of premature death to determine the differences between the levels of premature mortality as predicted by the actual proportions of low-income residents in 2008 to 2012 and the levels predicted by the proportions of low-income residents under a hypothetical $15 minimum wage.A $15 minimum wage could have averted 2800 to 5500 premature deaths between 2008 and 2012 in New York City, representing 4% to 8% of total premature deaths in that period. Most of these avertable deaths would be realized in lower-income communities, in which residents are predominantly people of color.A higher minimum wage may have substantial positive effects on health and should be considered as an instrument to address health disparities.

Decreasing the use of high-cost tests may reduce health care costs.To compare costs of care for patients presenting to the emergency department (ED) with suspected kidney stones randomized to 1 of 3 initial imaging tests.Patients were randomized to point-of-care ultrasound (POC US, least costly), radiology ultrasound (RAD US), or computed tomography (CT, most costly). Subsequent testing and treatment were the choice of the treating physician.A total of 2759 patients at 15 EDs were randomized to POC US (n=908), RAD US, (n=893), or CT (n=958). Mean age was 40.4 years; 51.8% were male.All medical care documented in the trial database in the 7 days following enrollment was abstracted and coded to estimate costs using national average 2012 Medicare reimbursements. Costs for initial ED care and total 7-day costs were compared using nonparametric bootstrap to account for clustering of patients within medical centers.Initial ED visit costs were modestly lower for patients assigned to RAD US: $423 ($411, $434) compared with patients assigned to CT: $448 ($438, $459) (P0.0001). Total costs were not significantly different between groups: $1014 ($912, $1129) for POC US, $970 ($878, $1078) for RAD US, and $959 ($870, $1044) for CT. Hospital admissions contributed over 50% of total costs, though only 11% of patients were admitted. Mean total costs (and admission rates) varied substantially by site from $749 to $1239.Assignment to a less costly test had no impact on overall health care costs for ED patients. System-level interventions addressing variation in admission rates from the ED might have greater impact on costs.

BACKGROUND:: Whereas national prevalence estimates for workers' compensation benefits are available, incidence estimates are not. Moreover, few studies address which groups in the economy pay for occupational injury and illness when workers' compensation does not. METHODS:: Data on numbers of cases and costs per case were drawn from the Bureau of Labor Statistics and National Council on Compensation Insurance data sets. Costs not covered by workers' compensation were estimated for private and public entities. RESULTS:: Total benefits in 2007 were estimated to be $51.7 billion, with $29.8 billion for medical benefits and $21.9 billion for indemnity benefits. For medical costs not covered by workers' compensation, other (non-workers' compensation) insurance covered $14.22 billion, Medicare covered $7.16 billion, and Medicaid covered $5.47 billion. CONCLUSION:: Incidence estimates of national benefits for workers' compensation were generated by combining existing published data. Costs were shifted to workers and their families, non-workers' compensation insurance carriers, and governments. Language: en

Whereas some studies show statistically significant linear associations between consumption at full-service restaurants and consumer incomes, studies of fast-food restaurants fail to find statistically significant linear associations. In this study, nationally representative data were drawn from the 1994-1996 Continuing Survey of Food Intakes by Individuals and the accompanying Diet and Health Knowledge Survey. The sample contained 4972 individuals who were 21 years of age or older. Dependent variables measured number of restaurant visits on 2 nonconsecutive days. Income was total annual household income. Control variables reflected sociodemographic, economic, lifestyle, and attitudinal variables. To capture possible curvilinear relationships between income and food consumption, we analyzed frequency distributions, regressions on full samples including income squared, and we divided samples into above- and below-average income groups. Zero-inflated negative binomial regressions accounted for excessive zeros within dependent variables. We found that fast-food restaurants were "normal goods" for below-average income, but "inferior goods" for above-average income, whereas full-service restaurants were "normal" for virtually all income levels. Earlier studies were flawed because they only tested for linear associations. Our results have implications for the poverty and obesity debate.

Cost estimates are essential to decision makers attempting to wisely allocate scarce health care resources. Cost-of-illness studies for many diseases continue to proliferate (Foster et al. 2006; Petersen and American Diabetes Association 2008; Rosamond et al. 2007, 2008), with cost estimates for coronary heart disease, stroke, cancer, and hypertension updated annually (Rosamond et al. 2007, 2008). By contrast and despite its importance, the generation of information about the costs of occupational injury and illness has not kept pace, as the most recent comprehensive estimate for U.S. costs applies to 1992 (Leigh et al. 1997). Nevertheless, there are several less than comprehensive and related studies. The National Academy of Social Insurance annually updates its estimates of the costs of workers’ compensation (Sengupta, Reno, and Burton 2009). But workers’ compensation data are incomplete. Bonauto and colleagues (2010) found that workers’ compensation records miss from 23 to 53 percent of all medically attended nonfatal injuries, and Leigh and Robbins (2004) found that workers’ compensation missed at least 91 percent of occupational disease deaths. Although Corso and colleagues (2006) generated national estimates for injuries in 2000, they did not separate those that were job related. Biddle (2009) provided cost estimates for occupational injury deaths, but not for nonfatal injuries or diseases. The National Safety Council (NSC 2009) calculates the costs of occupational injuries but excludes assaults, murders, and all illnesses. Finally, the national cost estimates for circulatory disease, cancer, and chronic obstructive pulmonary disease (COPD) do not estimate the portions of these diseases attributed to job-related exposures. The aim of this study is to estimate the national costs of occupational injuries and illnesses among civilians in 2007. To achieve that aim, I have calculated the numbers and costs of fatal and nonfatal injuries and illnesses. Costs are divided into medical and indirect categories. I use broad methodologies such as the cost of illness, incidence, prevalence, and societal perspective that are standard in studies of nonoccupational diseases and injuries. Numbers and costs for all categories are combined to produce the overall most probable estimate of approximately $250 billion for 2007. Finally, a sensitivity analysis investigates the effects of consequential assumptions. This study introduces numerous methodological advances over one that I and my colleagues conducted earlier (Leigh et al. 1997). For example, first, the previous study estimated injuries for state and local government employees by extrapolating from private-sector employees, whereas this study uses new U.S. Bureau of Labor Statistics (BLS) data from government employees. Second, the current study estimates those injuries for the self-employed and agricultural workers that are not simply averages of those in all other private-sector workers. Third, I rely on recent epidemiologic evidence for fractions of diseases such as cancer and COPD that are attributable to workplace exposures. Fourth, I use hospital costs per hospital stay rather than simply days in the hospital to estimate medical costs. Additional advances are discussed as well. These estimates should help inform the debate about the relative costs associated with occupational injuries and illnesses versus other diseases, as well as estimate costs not covered by workers’ compensation. These estimates may also inform decisions by occupational safety and health stakeholders regarding the allocation of resources to prevent injuries versus diseases.

Background We hypothesized that low socioeconomic status, employer-provided health insurance, low wages, and overtime were predictors of reporting workers compensation indemnity claims. We also tested for gender and race disparities. Methods Responses from 17,190 (person-years) Americans participating in the Panel Study of Income Dynamics, 1997–2005, were analyzed with logistic regressions. The dependent variable indicated whether the subject collected benefits from a claim. Results Odds ratios for men and African-Americans were relatively large and strongly significant predictors of claims; significance for Hispanics was moderate and confounded by education. Odds ratios for variables measuring education were the largest for all statistically significant covariates. Neither low wages nor employer-provided health insurance was a consistent predictor. Due to confounding from the “not salaried” variable, overtime was not a consistently significant predictor. Conclusion Few studies use nationally representative longitudinal data to consider which demographic and job characteristics predict reporting workers compensation indemnity cases. This study did and tested some common hypotheses about predictors. Am. J. Ind. Med. 54:758–770, 2011. © 2011 Wiley-Liss, Inc.

The average costs of Musculoskeletal Disorder (MSD) and odds ratios for filing medical claims related to MSD were examined. The medical claims were identified by ICD 9 codes for four US Census regions within retail trade. Large private firms' medical claims data from Thomson Reuters Inc. MarketScan databases for the years 2003 through 2006 were used. Average costs were highest for claims related to lumbar region (ICD 9 Code: 724.02) and number of claims were largest for low back syndrome (ICD 9 Code: 724.2). Whereas the odds of filing an MSD claim did not vary greatly over time, average costs declined over time. The odds of filing claims rose with age and were higher for females and southerners than men and non-southerners. Total estimated national medical costs for MSDs within retail trade were $389 million (2007 USD).

OBJECTIVE: To estimate the effects of increasing gas prices on mortality. METHOD: We developed a simulation-based partial equilibrium model that estimated the public health effects of a 20% rise in gas prices. Estimates on price elasticity for gasoline, price elasticity of motor vehicle crashes, relations among gasoline use, air pollution, and mortality were drawn from literature in economics, epidemiology, and medicine. RESULTS: For sustained 20% increases in gasoline prices over 1 year, and assuming other prices and factors were constant, we estimated: 1994 (range, 997 to 4984) fewer deaths from vehicle crashes and 600 (range, 300 to 1500) fewer deaths from air pollution. Combining both, we estimated 2594 fewer deaths. A Monte Carlo simulation involving varying assumptions on elasticities and relations indicated that 95% of the combined reduction in deaths was between 1747 and 3714. CONCLUSION: Results suggest that high gas prices have public health implications. Language: en

Physicians, nurses and other healthcare workers (HCWs) are at risk of bloodborne pathogens infection from needlestick injuries, but costs of needlesticks are little studied.We used the cost-of-illness and incidence approaches. We used the perspective of the medical provider (medical costs) and the individual (lost productivity). Data on needlesticks, infections from hepatitis B and C (HBV, HCV) and human immune-deficiency (HIV) among HCWs, as well as data on per-unit costs were culled from research literature, Centers for Disease Control and Prevention reports, and Bureau of Labor Statistics reports. We also generated estimates based upon industry employment and scenarios for source-patients. These data and estimates were combined with assumptions to produce a model that generated base-case estimates as well as one-way and multi-way probabilistic sensitivity analyses. Future costs were discounted by 3%.We estimated 644,963 needlesticks in the healthcare industry for 2004 of which 49% generated costs. Medical costs were $107.3 million of which 96% resulted from testing and prophylaxis and 4% from treating long-term infections (34 persons with chronic HBV, 143 with chronic HCV, and 1 with HIV). Lost-work productivity generated $81.2 million, for which 59% involved testing and prophylaxis and 41% involved long-term infections. Combined medical and work productivity costs summed to $188.5 million. Multi-way sensitivity analysis suggested a range on combined costs from $100.7 million to $405.9 million.Detailed methodology was developed to estimate costs of needlesticks and subsequent infections for hospital-based and non-hospital-based health care workers. The combined medical and lost productivity costs comprised roughly 0.1% of all occupational injury and illness costs for all jobs in the economy. We did not account for lost home production or pain and suffering costs, however, nor did we estimate benefit/cost ratios of specific interventions to reduce needlesticks.

Debate surrounds the provision of Women, Infants, and Children (WIC) benefits to undocumented immigrants. Few studies are available to estimate use of WIC services by documented and undocumented households using nationally representative data. The authors analyzed data from the National Agricultural Workers Survey (NAWS) annual cross-sections from 1993 through 2009 (N = 40,896 person-years). Household documentation status is defined by the status of the adults in the household, not children. Simple mean differences, logistic regressions, and time charts described household participation in WIC over 2-year intervals. Without adjustments for covariates, 10.7% of undocumented farm workers' households and 12.4% of documented households received WIC benefits, yielding an odds ratio of 0.84 (95% confidence interval [CI]: 0.76-0.94). Logistic regressions revealed that for the same number of children in the household, participation by undocumented persons was higher than participation by documented persons. Time charts and logistic regressions with interaction terms showed a stronger correspondence between participation in WIC and number of children

Few US estimates of the economic burden of autism spectrum disorders (ASD) are available and none provide estimates for 2015 and 2025. We forecast annual direct medical, direct non-medical, and productivity costs combined will be $268 billion (range $162-$367 billion; 0.884-2.009 % of GDP) for 2015 and $461 billion (range $276-$1011 billion; 0.982-3.600 % of GDP) for 2025. These 2015 figures are on a par with recent estimates for diabetes and attention deficit and hyperactivity disorder (ADHD) and exceed the costs of stroke and hypertension. If the prevalence of ASD continues to grow as it has in recent years, ASD costs will likely far exceed those of diabetes and ADHD by 2025.

BACKGROUND Despite considerable debate surrounding the use of Medicaid by undocumented immigrants, few studies address the extent of this use or estimate differences in the use between documented and undocumented households. METHOD We analyzed data from the National Agricultural Workers Survey annual cross sections from 1993 through 2009, N = 41,342. Simple mean differences and logistic regressions predicted participation in Medicaid over 2-year intervals. RESULTS Without adjustments for covariates, 12.2% of undocumented farm workers' households and 22.6% of documented households received Medicaid benefits, corresponding to an odds ratio of 0.48 (95% confidence interval, 0.33 to 0.69). By adding only covariates reflecting presence of children in the household, the odds ratio increased to 0.86 (95% confidence interval, 0.73 to 1.02). CONCLUSIONS Undocumented farm workers' households were roughly half as likely to use Medicaid as documented households, and undocumented households' participation was especially responsive to the presence of children.

BACKGROUND: Little is known about cost differences for demographic groups or across occupational injuries and illnesses. METHODS: In this incidence study of nationwide data for 1993, an analysis was conducted on fatal and non-fatal injury and illness data recorded in government data sets. Costs data were from workers' compensation records, estimates of lost wages, and jury awards. RESULTS: The youngest (age /= 65) workers had exceptionally high fatality costs. Whereas men's costs for non-fatal incidents were nearly double those for women, men's costs for fatal injuries were 10 times the costs for women. The highest ranking occupation for combined fatal and non-fatal costs-farming, forestry, and fishing-had costs-per-worker ($5,163) over 18 times the lowest ranking occupation-executives and managers ($279). The occupation of handlers, cleaners, and laborers, ranked highest for non-fatal costs. Gunshot wounds generated especially high fatal costs. Compared to whites, African-Americans had a lower percentage of costs due to carpal tunnel syndrome, circulatory, and digestive diseases. CONCLUSIONS: Costs comparisons can be drawn across age, race, gender, and occupational groups as well as categories of injuries and illnesses.A� Language: en

Objectives This study has ranked industries using estimated total costs and costs per worker. Methods This incidence study of nationwide data was carried out in 1993. The main outcome measure was total cost for medical care, lost productivity, and pain and suffering for the entire United States (US). The analysis was conducted using fatal and nonfatal injury and illness data recorded in large data sets from the US Bureau of Labor Statistics. Cost data were derived from workers` compensation records, estimates of lost wages, and jury awards. Current-value calculations were used to express all costs in 1993 in US dollars. Results The following industries were at the top of the list for average cost (cost per worker): taxicabs, bituminous coal and lignite mining, logging, crushed stone, oil field services, water transportation services, sand and gravel, and trucking. Industries high on the total-cost list were trucking, eating and drinking places, hospitals, grocery stores, nursing homes, motor vehicles, and department stores. Industries at the bottom of the cost-per-worker list included legal services, security brokers, mortgage bankers, security exchanges, and labor union offices. Conclusions Detailed methodology was developed for ranking industries by total cost and cost per worker. Ranking by total costs provided information on total burden of hazards, and ranking by cost per worker provided information on risk. Industries that ranked high on both lists deserve increased research and regulatory attention.

Objective To assess recent trends in professional satisfaction, perceptions of ability to provide high-quality care, and perceptions of ability to obtain needed services for patients in a national sample of obstetricians and gynecologists; to compare obstetrician–gynecologists with physicians in other specialties; and to identify demographic, professional, and practice characteristics associated with high career satisfaction. Methods We used data from the 1996–1997 (n = 12,385; response rate, 65%) and 1998–1999 (n = 12,280; response rate, 61%) waves of the nationally representative Community Tracking Study physicians' survey. The principal outcome measures were one item related to overall career satisfaction, six items measuring physicians' perceptions of their ability to provide high-quality care, and five items measuring physicians' perceptions of their ability to obtain needed services for patients. All results were weighted and adjusted to reflect the complex survey design. Results In 1996–1997, 34% of obstetrician–gynecologists (n = 545) were very satisfied with their careers, and 24% were very or somewhat dissatisfied. Up to 45% perceived significant barriers to the delivery of high-quality care, and up to 58% were unable to “almost or almost always” obtain necessary services for patients. Results in 1998–1999 (n = 484 obstetricians and gynecologists) were similar, except for a deterioration in perceived amount of time with patients and ability to obtain high-quality ancillary services. In comparison with primary care physicians, obstetrician–gynecologists were less satisfied (P = .001); in comparison with both primary care physicians and general surgeons, they had more problems delivering high-quality care (P Conclusion Although most obstetricians and gynecologists are satisfied with their careers, many are experiencing significant professional distress.

The Early Intervention Program (EIP) is California's publicly funded human immunodeficiency virus (HIV) care and treatment program with 30 sites throughout the state. Our objective for this study was to examine the number of days from first HIV-positive result until enrollment into EIP by race/ethnicity, behavioral risk, and other characteristics, with data from clients who enrolled in an EIP site after the availability of highly active antiretroviral therapies. For Model I, logistic regression distinguished clients diagnosed with HIV and enrolled in EIP on the same day (0 days) from those with values of 1+ days; linear regression was then used on the log transformation of days for the majority of clients not diagnosed and enrolled on the same day. For Model II, logistic regression was used to identify client characteristics related to enrollment in EIP over 6 weeks from the date of HIV diagnosis. We found that Latinos were more likely than whites to enroll in EIP on the day they were diagnosed with HIV. For clients not diagnosed and enrolled in EIP on the same day, no differences across racial and ethnic groups were found for days until enrollment in HIV care and treatment. However, clients with a history of injection drug use took longer from the day they were diagnosed with HIV to enroll in EIP. The California EIP represents a model for programs seeking equity in access to HIV care and treatment across racial and ethnic groups. Getting injectors into timely HIV care and treatment represents a challenge.

Objective: To estimate the number of annual deaths, as well as the direct and indirect costs of occupational COPD and asthma, in the United States in 1996. Design: Aggregation and analysis of national data sets collected by the National Center for Health Statistics, the Health Care Financing Administration, and other government bureaus and private firms. To assess mortality, we reviewed data from national surveys and applied a population attributable risk (PAR) of 15% for both asthma and COPD. We use a lower age limit of 35 years for occupational COPD and 20 years for occupational asthma. To calculate costs, we use the human capital method that decomposes costs into direct categories, such as medical expenses, as well as indirect categories, such as lost earnings and lost home production. We calculated proportionately adjusted costs for other plausible PARs. Results: The 15% PARs result in costs of $5.0 billion for COPD and $1.6 billion for asthma. For COPD, 56% of costs were direct and 44% were indirect; for asthma, 74% were direct and 26% indirect. These estimates are conservative since we ignored costs associated with pain and suffering as well as the value of care rendered by family members. The proportionately adjusted costs for 10 to 20% PARs are $3.3 to $6.6 billion for COPD and $1.1 to $2.1 billion for asthma. Conclusions: The estimated $6.6 billion cost of occupational COPD and asthma in 1996 is likely to rise with the increasing prevalence of these diseases and warrants preventive intervention. (CHEST 2002; 121:264 –272)

Objectives. The purpose of this study was to estimate the annual incidence, the mortality, and the direct and indirect costs associated with occupational injuries and illnesses in California in 1992. To achieve this, we performed aggregation and analysis of national and California data sets collected by the U.S. Bureau of Labor Statistics, California Workers' Compensation Insurance Rating Bureau, California Division of Industrial Relations, the National Center for Health Statistics, and the U.S. Health Care Financing Administration. Methods. To assess incidence of and mortality from occupational injuries and illnesses, we reviewed data from state and national surveys and applied an attributable risk proportion method. To assess costs, we used the cost-of-illness, human capital, method that decomposes costs into direct categories such as medical expenses and insurance administration expenses as well as indirect categories such as lost earnings, lost home production, and lost fringe benefits. Some cost estimates were drawn from California data, whereas others were drawn from a national study but were adjusted to reflect California's differences. Cost estimates for injuries were calculated by multiplying average costs by the number of injuries. For the majority of diseases, cost estimates relied on the attributable risk proportion method. Results. Approximately 660 job-related deaths from injury, 1.645 million nonfatal injuries, 7,079 deaths from diseases, and 0.133 million illnesses are estimated to occur annually in the civilian California workforce. The direct ($7.04 billion, 34%) plus indirect ($13.62 billion, 66%) costs were estimated to be $20.7 billion. Injuries cost $17.8 billion (86%) and illnesses $2.9 billion (14%). These estimates are likely to be low because: (1) they ignore costs associated with pain and suffering, (2) they ignore home care provided by family members, and (3) the numbers of occupational injuries and illnesses are likely to be undercounted. Conclusion. Occupational injuries and illnesses are a major contributor to the total cost of health care and lost productivity in California. These costs are on a par with those of all cancers combined and only slightly less than the cost of heart disease and stroke in California. Workers' compensation covers less than one-half of the costs of occupational injury and illness.

This study weighs the risks to workers of cleaning up a Superfund site against the risks to residents of not cleaning up that site. Risks are measured by the number of deaths and disabilities due to injuries and diseases. We posit a cleanup plan involving 31 occupations and 4,581 person-years of work. We posit 20 hypothetical sites and 99 specific sites with varying numbers of residents and levels of cancer death and cancer disability rates. Depending on the number of residents, and the rates, we find that the risks to workers frequently outweigh the risks to residents. We conclude that risks to workers should be accounted for in EPA judgments regarding whether and how a Superfund site should be cleaned up.

Estimates of the value-of-life have relied heavily on labor market studies of compensating wages. These studies typically involve least squares regressions using cross-sectional data. The dependent variable is log of wages and the independent variables are years of schooling, age, age squared, and so on. The key independent variable is a measure of the death rate on the job. Most studies have found a strong partial correlation between log-wages and the death rate and most authors have interpreted this correlation as evidence that the labor market is generating compensating wages for hazardous work. Some economists have used the coefficient from this correlation to estimate a statistical value-of-life. I n t h i s p a p e r , w e f i r s t m e n t i o n s o m e e v i d e n c e t h a t c o r r e l a t i o n s b e t w e e n wages and death rates are fragile. We then discuss a study (Leigh, 1991) that did not find strong wage-death rate correlations using unique data on occupations. Third, we argue that the correlations found in prior literature are more likely the result of historical inter-industry differentials than the result of a compensating wage paid for dangerous work. The fourth part of the paper offers some explanations for why compensating wages fail to appear for the lion's share of jobs in the labor market but do appear for the handful of jobs that are undeniably hazardous. We offer a comment on the academic importance of forensic economists in the fifth part of the paper. A summary concludes our paper. I. Fragile Estimates Ted Miller (1990) reviewed 27 studies and Kip Viscusi (1992) reviewed 30 studies on compensating wages for job-related deaths, most of which found positive and statistically significant estimated coefficients on death rate variables. Virtually all of Miller’s studies were on Viscusi’s list. However, in several of these studies, the estimated coefficients for the death rate variables were either statistically insignificant or significant but close to zero (Dickens, 1984; Dorsey, 1983; Leigh and Folsom, 1984; Leigh, 1987; Moore and Viscusi, 1988; Thaler and Rosen, 1975; Viscusi, 1979; Kniesner and Leeth, 1991). More importantly, these two reviews are incomplete since they do not include at least two prior studies or three subsequent studies that failed to find statisti

This study investigates whether parents' education or unobserved variables partially explain correlations between education and a measure of frailty in adults. Data sets are drawn from the 1986 wave of the Panel Study of Income Dynamics (PSID) — the only wave available that asks questions pertaining to Activities of Daily Living that, in turn, allows the construction of a Disability Index. The Disability Index and respondent's schooling are treated as endogenous. Parental schooling is strongly associated with adult schooling. Mothers' but not fathers' schooling is strongly associated with the Disability Index when adult schooling is ignored. But once adult respondent's schooling is taken into account, parental schooling and Disability Index associations evaporate. Unobserved variables are not of great importance in explaining respondent's education and frailty associations. Respondent's schooling is found to be strongly associated with the Disability Index even after removing the influences of parents' schooling and unobserved variables. [JEL I1, I2]

To describe occupational illness data in two large data sets, two national data sets were aggregated, and the numbers, percentages, and rates of cases of occupational illnesses were determined. Job-related illness data were from Bureau of Labor Statistics documents containing Annual Survey and Census of Fatal Occupational Injury data. A severity index was created to assess the overall burden of a disease. The index multiplies the number of cases times the median days lost. Circulatory disease accounted for 85% of the deaths in the Census and at least 80% in the Annual Survey. More fatal myocardial infarctions occurred on Monday than on any other day. Low-paying occupations had the most myocardial infarctions: operators, laborers, and truck drivers; high-paying occupations had the least: executives, administrators, and managers. Carpal tunnel syndrome and hearing loss accounted for more morbidity, measured by cases and days lost, than any other illness. Persons at great risk for carpal tunnel syndrome included dental hygienists, butchers, sewing machine operators, and dentists. Mental disorders generated more morbidity than is generally acknowledged. Neurotic reactions to stress were highest in the transportation and public utility industries, as well as in finance, insurance, and real estate. Manufacturing contributed far more cases than any other industry. Industries generating significant asbestos-related deaths included construction and boat building. Ninety-three percent of all illness fatalities were among men. Few African Americans died from coal-workers' pneumoconiosis. Illness cases increased much faster than injury cases in recent years. The two data sets provide insights into the incidences and prevalences of occupational illnesses, but underestimate the burden of job-related illnesses.

Professors Baker and Krueger ignore some costs associated with workers' compensation. Because of these costs, the contention that physicians willfully exploit the workers' compensation system for their own gain is questioned.

We sought to determine whether employed patients with HIV infection reported more days unable to work than similar patients without HIV infection. We did not consider whether HIV infection resulted in job loss. Cross-sectional data on 884 employed patients from five physician-sites in California were drawn from the AIDS Time-Oriented Health Outcome Study (ATHOS). An econometric 2-part model with a number of covariates including age, race, education, and an instrumental variable reflecting the probability of any employment was constructed to provide a comparison between three categories of patients, HIV -negative; HIV positive but not yet AIDS; and AIDS. The results suggested no statistically significant difference between the comparison patients (HIV -negative) and the patients with HIV but not AIDS for days unable to work. Employed patients with AIDS, however, reported roughly 5.1 (one week) more days unable to work out of the prior three months than either the comparison patients or the patients with HI...

What accounts for the correlations between schooling and frailty among seniors? Do the correlations differ among women, men, blacks and whites? Data from the 1986 wave of the Panel Study of Income Dynamics (PSID) are analyzed to answer these questions as well as related ones pertaining to the roles of self-selection bias, self-efficacy, risk preference and time preference in explaining the correlations. The correlations between disability and schooling for women were strong after accounting for self-selection bias. The male schooling and exercise correlations were strong after accounting for self-selection, self-efficacy and preferences. Univariate differences in frailty measures between blacks and whites appear to be due to socioeconomic status rather than genetics. No race differences were observed in the correlations between schooling and frailty. The results provide additional evidence that education itself, rather than simply self-efficacy or time or risk preference, acts as preventive medicine.

OBJECTIVE: To determine whether state-mandated minimum nurse-to-patient staffing ratios in California hospitals had an effect on reported occupational injury and illness rates. METHODS: The difference-in-differences method was applied: the change in injury rates among hospital nurses after implementation of the law in California was compared to the change in 49 other states and the District of Columbia combined. Data were drawn from the US Bureau of Labor Statistics and the California Employment Development Department, including numerator estimates of injury and illness cases and denominator estimates of the number of registered nurses (RNs) and licensed practical nurses (LPNs) employed in hospitals. Confidence intervals (CIs) for rates were constructed based on assumptions that favored the null hypothesis. RESULTS: The most probable difference-in-differences estimate indicated that the California law was associated with 55.57 fewer occupational injuries and illnesses per 10,000 RNs per year, a value 31.6% lower than the expected rate without the law. The most probable reduction for LPNs was 33.6%. Analyses of confidence intervals suggested that these reductions were unlikely to be due to chance. CONCLUSIONS: Despite significant data restrictions and corresponding methodological limitations, the evidence suggests that the law was effective in reducing occupational injury and illness rates for both RNs and LPNs. Whether these 31.6% and 33.6% reductions are maintained over time remains to be seen.

Do hazardous working conditions encourage heavy drinking and/or does heavy drinking contribute to job-related injuries and deaths? Simultaneous equations probit and least squares models are constructed to answer these questions. Samples of employed persons are drawn from the National Health and Nutrition Examination Survey II (NHANES II; n = 8,477), and the Quality of Employment Survey (QES; n = 1,393). Heavy total alcohol use is alternatively defined as drinking any alcoholic beverage at least once a day, or frequently drinking three or more drinks at a time. In the QES, heavy use is also defined as drinking on-the-job. Heavy beer, wine, or liquor use is separately defined in the NHANES II as drinking four or more times per week. The endogenous variables reflecting job hazards include subjects' evaluations of the hazardous nature of the job and fatality rates within occupations and industries. Exogenous variables include age, gender, race, marital status, wages or family income, rural residence, Southern residence, years of schooling, union membership, percent of industry unionized, minimum legal drinking age for beer, religion variables, smoking status, and beer tax. Only one robust alcohol finding emerged: Heavy beer use was found to be strongly correlated with the fatality rates within occupations and industries. In a related finding, a disproportionately high number of smokers were found employed in dangerous jobs. Separate analyses of beer, wine, and liquor (spirits) appeared essential to explaining correlations between dangerous jobs and heavy alcohol use. Evidence for a simultaneous relationship between beer abuse and job-related deaths suggests that prior estimates of the effect of alcohol abuse on job-related accidental deaths have been exaggerated.

How much, if any, does smoking contribute to absenteeism? Separate samples of employed men and women are drawn from the 1986 wave of the Panel Study of Income Dynamics (PSID) to answer this question. The 1986 wave is the most recent one with information on smoking. In the first analysis, single equation Tobit regressions are run explaining the absence rate. Independent variables include a binary smoking variable together with others for age, race, marital status, and so on. In the single equation models, smoking appeared to raise absence rates by 42 and 232 percent for women and men, respectively. In the second analysis, decomposition techniques, which also adjust for smoking and employment hazard rates, are applied to four separate samples of smokers, non-smokers, men, and women. The decomposition techniques lowered these differentials. Smoking appears to make a moderate contribution to absenteeism for men, but only a slight contribution for women. The decomposition suggests that it is the observed and unobserved personal characteristics of female smokers that account for most of the simple positive correlations between absenteeism and smoking among women found in the single equation models.

Bureau of Labor Statistics (BLS) and National Institute for Occupational Safety and Health (NIOSH) death rates were combined with three national probability samples to investigate compensating wages for hazardous work. Compensating wages were expected in blue-collar male-only samples, but not in male and female clerk samples. The inter-industry differentials hypothesis was investigated by alternately including and excluding dummy variables for broad industry divisions. The inter-industry differentials hypothesis was supported. The BLS and NIOSH death rates cannot be relied on to produce credible estimates of the value of a statistical life.

Objective: Socio-economic status (SES) is strongly correlated with hypertension. But SES has several components, including income and correlations in cross-sectional data need not imply SES is a risk factor. This study investigates whether wages—the largest category within income—are risk factors. Methods: We analysed longitudinal, nationally representative US data from four waves (1999, 2001, 2003 and 2005) of the Panel Study of Income Dynamics. The overall sample was restricted to employed persons age 25–65 years, n = 17 295. Separate subsamples were constructed of persons within two age groups (25–44 and 45–65 years) and genders. Hypertension incidence was self-reported based on physician diagnosis. Our study was prospective since data from three base years (1999, 2001, 2003) were used to predict newly diagnosed hypertension for three subsequent years (2001, 2003, 2005). In separate analyses, data from the first base year were used to predict time-to-reporting hypertension. Logistic regressions with random effects and Cox proportional hazards regressions were run. Results: Negative and strongly statistically significant correlations between wages and hypertension were found both in logistic and Cox regressions, especially for subsamples containing the younger age group (25–44 years) and women. Correlations were stronger when three health variables—obesity, subjective measures of health and number of co-morbidities—were excluded from regressions. Doubling the wage was associated with 25–30% lower chances of hypertension for persons aged 25–44 years. Conclusions: The strongest evidence for low wages being risk factors for hypertension among working people were for women and persons aged 25–44 years.

In this study, the distribution of inspection laws across states is endogenously determined by the relative strengths of lobbying groups within states. Previous studies that treat the laws as exogenous and find a 5 to 15 percent reduction in fatalities may have produced biased results. A selection bias model is developed in which non-random assignment is taken into account. Two equations are estimated: one explaining how many inspections are required, and the other explaining the effects of the inspections on fatalities per capita. Using single-equation techniques, results are obtained that are similar to prior studies. In the two-equation model that accounts for non-random assignment, inspection laws are not found to significantly reduce fatalities per capita.

Data from the National Health and Nutrition Examination Survey I, 1971–1975 (NHANES I) were used to analyze associations among highest education level and arthritis. The dependent variables indicated whether the respondent had ever been diagnosed with any form of arthritis by a physician (10,678 women and 7243 men) or whether physician X-ray readings suggested arthritis of the knee (3491 women and 3119 men). These variables did not distinguish between osteo- and rheumatoid arthritis. It is likely that the great majority of the sample reporting or diagnosed with arthritis had osteoarthritis. There were strong univariate correlations between answers to the general arthritis question and the knee question on the one hand and gender, age, body mass, schooling, income and employment on the other. Respondents' education level was found to be strongly and negatively associated with self-reported arthritis in the larger samples both before and after controls were entered for employment, income and potential biological risk factors. The association between self-reported arthritis or arthritis of the knees and education was weaker for men, but not for women after employment and income were accounted for. When body mass was accounted for, the association between self-reported arthritis or arthritis of the knees and education was weaker among women but not men. Long-run preventive strategies to combat osteoarthritis ought to consider investments in education.