Showing papers by "RAND Corporation published in 2004"

Least angle regression

Abstract: The purpose of model selection algorithms such as All Subsets, Forward Selection and Backward Elimination is to choose a linear model on the basis of the same set of data to which the model will be applied. Typically we have available a large collection of possible covariates from which we hope to select a parsimonious set for the efficient prediction of a response variable. Least Angle Regression (LARS), a new model selection algorithm, is a useful and less greedy version of traditional forward selection methods. Three main properties are derived: (1) A simple modification of the LARS algorithm implements the Lasso, an attractive version of ordinary least squares that constrains the sum of the absolute regression coefficients; the LARS modification calculates all possible Lasso estimates for a given problem, using an order of magnitude less computer time than previous methods. (2) A different LARS modification efficiently implements Forward Stagewise linear regression, another promising new model selection method; this connection explains the similar numerical results previously observed for the Lasso and Stagewise, and helps us understand the properties of both methods, which are seen as constrained versions of the simpler LARS algorithm. (3) A simple approximation for the degrees of freedom of a LARS estimate is available, from which we derive a Cp estimate of prediction error; this allows a principled choice among the range of possible LARS estimates. LARS and its variants are computationally efficient: the paper describes a publicly available algorithm that requires only the same order of magnitude of computational effort as ordinary least squares applied to the full set of covariates.

Linear Programming under Uncertainty

Abstract: (This article originally appeared in Management Science, April-July 1955, Volume 1, Numbers 3 and 4, pp. 197-206, published by The Institute of Management Sciences.)

Propensity score estimation with boosted regression for evaluating causal effects in observational studies.

Abstract: Causal effect modeling with naturalistic rather than experimental data is challenging. In observational studies participants in different treatment conditions may also differ on pretreatment characteristics that influence outcomes. Propensity score methods can theoretically eliminate these confounds for all observed covariates, but accurate estimation of propensity scores is impeded by large numbers of covariates, uncertain functional forms for their associations with treatment selection, and other problems. This article demonstrates that boosting, a modern statistical technique, can overcome many of these obstacles. The authors illustrate this approach with a study of adolescent probationers in substance abuse treatment programs. Propensity score weights estimated using boosting eliminate most pretreatment group differences and substantially alter the apparent relative effects of adolescent substance abuse treatment.

Interventions for the prevention of falls in older adults: systematic review and meta-analysis of randomised clinical trials

Abstract: Objective To assess the relative effectiveness of interventions to prevent falls in older adults to either a usual care group or control group. Design Systematic review and meta-analyses. Data sources Medline, HealthSTAR, Embase, the Cochrane Library, other health related databases, and the reference lists from review articles and systematic reviews. Data extraction Components of falls intervention: multifactorial falls risk assessment with management programme, exercise, environmental modifications, or education. Results 40 trials were identified. A random effects analysis combining trials with risk ratio data showed a reduction in the risk of falling (risk ratio 0.88, 95% confidence interval 0.82 to 0.95), whereas combining trials with incidence rate data showed a reduction in the monthly rate of falling (incidence rate ratio 0.80, 0.72 to 0.88). The effect of individual components was assessed by meta-regression. A multifactorial falls risk assessment and management programme was the most effective component on risk of falling (0.82, 0.72 to 0.94, number needed to treat 11) and monthly fall rate (0.63, 0.49 to 0.83; 11.8 fewer falls in treatment group per 100 patients per month). Exercise interventions also had a beneficial effect on the risk of falling (0.86, 0.75 to 0.99, number needed to treat 16) and monthly fall rate (0.86, 0.73 to 1.01; 2.7). Conclusions Interventions to prevent falls in older adults are effective in reducing both the risk of falling and the monthly rate of falling. The most effective intervention was a multifactorial falls risk assessment and management programme. Exercise programmes were also effective in reducing the risk of falling.

The Impact of Nearly Universal Insurance Coverage on Health Care Utilization: Evidence from Medicare.

Abstract: The onset of Medicare eligibility at age 65 leads to sharp changes in the health insurance coverage of the U.S. population. These changes lead to increases in the use of medical services, with a pattern of gains across socioeconomic groups that varies by type of service. While routine doctor visits increase more for groups that previously lacked insurance, hospital admissions for relatively expensive procedures like bypass surgery and joint replacement increase more for previously insured groups that are more likely to have supplementary coverage after 65, reflecting the relative generosity of their combined insurance package under Medicare.

Immigrant health: selectivity and acculturation

Abstract: This paper explores some salient issues concerning immigrant health. Ethnic health disparities are inherently linked to immigration since ethnic identities often are traced to the country of origin of one's immigrant ancestors. Two of the central questions that have dominated the medical and social science literature on immigrant health are the central focus of this paper. These issues involve the magnitude and mechanisms shaping health selectivity and the determinants of health trajectories following immigration. As part of this paper, we also developed a theoretical model that attempts to explain the diversity in health selection among immigrants.

Models for Value-Added Modeling of Teacher Effects.

Abstract: The use of complex value-added models that attempt to isolate the contributions of teachers or schools to student development is increasing. Several variations on these models are being applied in the research literature, and policy makers have expressed interest in using these models for evaluating teachers and schools. In this article, we present a general multivariate, longitudinal mixed-model that incorporates the complex grouping structures inherent to longitudinal student data linked to teachers. We summarize the principal existing modeling approaches, show how these approaches are special cases of the proposed model, and discuss possible extensions to model more complex data structures. We present simulation and analytical results that clarify the interplay between estimated teacher effects and repeated outcomes on students over time. We also explore the potential impact of model misspecifications, including missing student covariates and assumptions about the accumulation of teacher effects over time, on key inferences made from the models. We conclude that mixed models that account for student correlation over time are reasonably robust to such misspecifications when all the schools in the sample serve similar student populations. However, student characteristics are likely to confound estimated teacher effects when schools serve distinctly different populations.

Evaluating Value-Added Models for Teacher Accountability

Abstract: Does value-added modeling (VAM) demonstrate the importance of teachers to student outcomes? The authors clarify the primary questions raised by VAM for measuring teacher effects, review the most important recent applications of VAM, and discuss a variety of statistical and measurement issues that might affect the validity of VAM inferences. The authors identify numerous possible sources of error and bias in teacher effects and recommend a number of steps for future research into these potential errors.

Pharmacy Benefits and the Use of Drugs by the Chronically Ill

Abstract: ContextMany health plans have instituted more cost sharing to discourage use of more expensive pharmaceuticals and to reduce drug spending.ObjectiveTo determine how changes in cost sharing affect use of the most commonly used drug classes among the privately insured and the chronically ill.Design, Setting, and ParticipantsRetrospective US study conducted from 1997 to 2000, examining linked pharmacy claims data with health plan benefit designs from 30 employers and 52 health plans. Participants were 528 969 privately insured beneficiaries aged 18 to 64 years and enrolled from 1 to 4 years (960 791 person-years).Main Outcome MeasureRelative change in drug days supplied (per member, per year) when co-payments doubled in a prototypical drug benefit plan.ResultsDoubling co-payments was associated with reductions in use of 8 therapeutic classes. The largest decreases occurred for nonsteroidal anti-inflammatory drugs (NSAIDs) (45%) and antihistamines (44%). Reductions in overall days supplied of antihyperlipidemics (34%), antiulcerants (33%), antiasthmatics (32%), antihypertensives (26%), antidepressants (26%), and antidiabetics (25%) were also observed. Among patients diagnosed as having a chronic illness and receiving ongoing care, use was less responsive to co-payment changes. Use of antidepressants by depressed patients declined by 8%; use of antihypertensives by hypertensive patients decreased by 10%. Larger reductions were observed for arthritis patients taking NSAIDs (27%) and allergy patients taking antihistamines (31%). Patients with diabetes reduced their use of antidiabetes drugs by 23%.ConclusionsThe use of medications such as antihistamines and NSAIDs, which are taken intermittently to treat symptoms, was sensitive to co-payment changes. Other medications—antihypertensive, antiasthmatic, antidepressant, antihyperlipidemic, antiulcerant, and antidiabetic agents—also demonstrated significant price responsiveness. The reduction in use of medications for individuals in ongoing care was more modest. Still, significant increases in co-payments raise concern about adverse health consequences because of the large price effects, especially among diabetic patients.

Watching sex on Television Predicts Adolescent Initiation of sexual Behavior

Abstract: Background. Early sexual initiation is an important social and health issue. A recent survey sug- gested that most sexually experienced teens wish they had waited longer to have intercourse; other data indicate that unplanned pregnancies and sexually transmitted diseases are more common among those who begin sex- ual activity earlier. The American Academy of Pediatrics has suggested that portrayals of sex on entertainment television (TV) may contribute to precocious adolescent sex. Approximately two-thirds of TV programs contain sexual content. However, empirical data examining the relationships between exposure to sex on TV and adoles- cent sexual behaviors are rare and inadequate for ad- dressing the issue of causal effects. Design and Participants. We conducted a national longitudinal survey of 1792 adolescents, 12 to 17 years of age. In baseline and 1-year follow-up interviews, partic- ipants reported their TV viewing habits and sexual ex- perience and responded to measures of more than a dozen factors known to be associated with adolescent sexual initiation. TV viewing data were combined with the results of a scientific analysis of TV sexual content to derive measures of exposure to sexual content, depictions of sexual risks or safety, and depictions of sexual behav- ior (versus talk about sex but no behavior). Outcome Measures. Initiation of intercourse and ad- vancement in noncoital sexual activity level, during a 1-year period. Results. Multivariate regression analysis indicated that adolescents who viewed more sexual content at base- line were more likely to initiate intercourse and progress to more advanced noncoital sexual activities during the subsequent year, controlling for respondent characteris- tics that might otherwise explain these relationships. The size of the adjusted intercourse effect was such that youths in the 90th percentile of TV sex viewing had a predicted probability of intercourse initiation that was approximately double that of youths in the 10th percen- tile, for all ages studied. Exposure to TV that included only talk about sex was associated with the same risks as exposure to TV that depicted sexual behavior. African American youths who watched more depictions of sexual risks or safety were less likely to initiate intercourse in the subsequent year. Conclusions. Watching sex on TV predicts and may hasten adolescent sexual initiation. Reducing the amount of sexual content in entertainment programming, reduc- ing adolescent exposure to this content, or increasing references to and depictions of possible negative conse- quences of sexual activity could appreciably delay the initiation of coital and noncoital activities. Alternatively, parents may be able to reduce the effects of sexual con- tent by watching TV with their teenaged children and discussing their own beliefs about sex and the behaviors portrayed. Pediatricians should encourage these family discussions. Pediatrics 2004;114:e280 -e289. URL: http: //www.pediatrics.org/cgi/content/full/114/3/e280; con- doms, media, sex, television.

Childhood overweight and academic performance: national study of kindergartners and first-graders.

Abstract: DATAR, ASHLESHA, ROLAND STURM, AND JENNIFER L. MAGNABOSCO. Childhood overweight and academic performance: national study of kindergartners and first-graders. Obes Res. 2004;12:58‐68. Objective: To examine the association between children’s overweight status in kindergarten and their academic achievement in kindergarten and first grade. Research Methods and Procedures: The data analyzed consisted of 11,192 first time kindergartners from the Early Childhood Longitudinal Study, a nationally representative sample of kindergartners in the U.S. in 1998. Multivariate regression techniques were used to estimate the independent association of overweight status with children’s math and reading standardized test scores in kindergarten and grade 1. We controlled for socioeconomic status, parent-child interaction, birth weight, physical activity, and television watching. Results: Overweight children had significantly lower math and reading test scores compared with nonoverweight children in kindergarten. Both groups were gaining similarly on math and reading test scores, resulting in significantly lower test scores among overweight children at the end of grade 1. However, these differences, except for boys’ math scores at baseline (difference ! 1.22 points, p ! 0.001), became insignificant after including socioeconomic and behavioral variables, indicating that overweight is a marker but not a causal factor. Race/ethnicity and mother’s education were stronger predictors of test score gains or levels than overweight status. Discussion: Significant differences in test scores by overweight status at the beginning of kindergarten and the end of grade 1 can be explained by other individual characteristics, including parental education and the home environment. However, overweight is more easily observable by other students compared with socioeconomic characteristics, and its significant (unadjusted) association with worse academic performance can contribute to the stigma of overweight as early as the first years of elementary school.

Psychiatric symptoms in methamphetamine users.

Abstract: The Methamphetamine Treatment Project (MTP) offers the opportunity to examine co-occurring psychiatric conditions in a sample of 1016 methamphetamine users participating in a multisite outpatient treatment study between 1999-2001. Participants reported high levels of psychiatric symptoms, particularly depression and attempted suicide, but also anxiety and psychotic symptoms. They also reported high levels of problems controlling anger and violent behavior, with a correspondingly high frequency of assault and weapons charges. Findings continue to support the value of integrated treatment for co-occurring conditions, especially the importance of training counseling staff to handle psychotic symptoms when needed.

The prevalence of antiretroviral drug resistance in the United States.

Abstract: Background: Antiretroviral therapy has dramatically reduced the morbidity and mortality of infection due to HIV. The emergence of drug-resistant virus has limited the usefulness of many drugs. Objective: To determine the prevalence of HIV drug resistance in the population of adults receiving care in the United States. Design and methods: HIV drug susceptibility assays were performed on plasma virus from a random sample representative of the 132 500 HIV-infected American adults who had received medical care in early 1996 yet were viremic with > 500 copies/ml of HIV RNA in late 1998. A blood sample was obtained from 1797 patients who comprised a representative sample of the 208900 adults receiving urban care for HIV infection in early 1996 who survived to late 1998. The sampling procedure permitted weighting each evaluated patient to reflect demographic and other characteristics of the target population. Results: We estimated that 132 500 (63%) of the target population had HIV viremia of > 500 copies/ml. Among viremic patients, an estimated 76% had resistance to one or more antiretroviral drugs. The odds of resistance were significantly higher in patients with a history of antiretroviral drug use, advanced HIV disease, higher plasma HIV viral load and lowest CD4 cell count by self-report. Conclusions: The frequent selection for drug-resistant virus among viremic patients during the first 3 years of widespread use of potent antiretroviral combination therapy has significant implications for HIV treatment and transmission.

A Suggested Computation for Maximal Multi-Commodity Network Flows

Abstract: (This article originally appeared in Management Science, October 1958, Volume 5, Number 1, pp 97-101, published by The Institute of Management Sciences) A simplex computation for an arc-chain formulation of the maximal multi-commodity network flow problem is proposed Since the number of variables in this formulation is too large to be dealt with explicitly, the computation treats non-basic variables implicitly by replacing the usual method of determining a vector to enter the basis with several applications of a combinatorial algorithm for finding a shortest chain joining a pair of points in a network

The role of perceived team effectiveness in improving chronic illness care

Abstract: Background/Objectives:The importance of teams for improving quality of care has received increased attention. We examine both the correlates of self-assessed or perceived team effectiveness and its consequences for actually making changes to improve care for people with chronic illness.Study Setting

Dating violence among adolescents: prevalence, gender distribution, and prevention program effectiveness.

Abstract: Relative to violence among adult intimate partners, violence among adolescent dating partners remains an understudied phenomenon. In this review, we assess the state of the research literature on teen dating violence. Our review reveals that the broad range of estimates produced by major national data sources and single studies make conclusions about the prevalence of teen dating violence premature. Similarly, our review of what is known about risk factors reveals inconsistency among studies. We assess published evaluations of adolescent dating violence prevention programs and discuss their findings and limitations. Finally, we discuss challenges to researchers in this area and suggest that additional investment in high-quality basic research is needed to inform the development of sound theory and effective prevention and intervention programs.

Impact of pain on depression treatment response in primary care.

Abstract: Objective Pain commonly coexists with depression, but its impact on treatment outcomes has not been well studied. Therefore, we prospectively evaluated the impact of comorbid pain on depression treatment response and health-related quality of life. Methods We analyzed data from the ARTIST study, a randomized controlled trial with naturalistic follow-up conducted in 37 primary care clinics. Participants were 573 clinically depressed patients randomized to one of three selective serotonin reuptake inhibitor (SSRI) antidepressants: fluoxetine, paroxetine, or sertraline. Depression as assessed by the Symptom Checklist-20 (SCL-20) was the primary outcome. Secondary outcomes included pain and health-related quality of life. Results Pain was reported by more than two thirds of depressed patients at baseline, with the severity of pain mild in 25% of patients, moderate in 30%, and severe in 14%. After 3 months of antidepressant therapy, 24% of patients had a poor depression treatment response (ie, SCL-20 >1.3). Multivariate odds ratios for poor treatment response were 1.5 (95% confidence interval, 0.8-3.2) for mild pain, 2.0 (1.1-4.0) for moderate pain, and 4.1 (1.9-8.8) for severe pain compared with those without pain. Increasing pain severity also had an adverse impact on outcomes in multiple domains of health-related quality of life. Conclusions Pain is present in two thirds of depressed primary care patients begun on antidepressant therapy, and the severity of pain is a strong predictor of poor depression and health-related quality of life outcomes at 3 months. Better recognition, assessment, and treatment of comorbid pain may enhance outcomes of depression therapy.

All symptoms are not created equal: the prominent role of hyperarousal in the natural course of posttraumatic psychological distress.

Abstract: This 3-wave longitudinal study examined the natural course of posttraumatic stress disorder symptoms using data collected from young adult survivors of community violence. Three key findings emerged. 1. Mean levels of distress for each symptom cluster decreased over time, with reexperiencing decreasing most rapidly. 2. Cross-lagged panel analysis revealed that hyperarousal strongly influences, but is not generally influenced by, other symptoms clusters. 3. Trajectory analysis demonstrated that respondents for whom hyperarousal was the most pronounced baseline symptom showed lower overall symptom improvement relative to trauma exposed counterparts for whom hyperarousal was a less prominent early symptom. Implications for theory, research, and clinical practice are discussed.

Increasing Obesity Rates And Disability Trends

Abstract: Are older Americans becoming more or less disabled? Unhealthy body weight has increased dramatically, but other data show that disability rates have declined. We use data from the Health and Retirement Study to estimate the association between obesity and disability, and we combine these data with trend estimates of obesity rates from the Behavioral Risk Factor Surveillance Survey. If current trends in obesity continue, disability rates will increase by 1 percent per year more in the 50–69 age group than if there were no further weight gain.

Assuring the Health of Immigrants: What the Leading Health Indicators Tell Us

Abstract: Over the past 20 years, the United States has experienced one of the largest waves of immigration in its history. Understanding the health status and needs of immigrants is important because of their growing numbers and their contribution to the health of the nation, but it is challenging because of gaps in national databases, the heterogeneity of immigrant populations, and uncertainty about how migration affects health. Healthy People 2010 outlines the nation's public health objectives for the current decade. It includes ten leading health indicators (LHIs) chosen because of their importance as public health issues, their ability to motivate action, and the availability of data to measure their progress. In this paper, we discuss the health of immigrants from the perspective of these LHIs, as they provide a framework for anticipating some of the future health needs of immigrants and help define priority areas for research and action.

Demand-Side Management and Energy Efficiency in the United States

Abstract: Between 1989 and 1999, U.S. electric utilities spent $14.7 billion on demand-side management (DSM) programs aimed at encouraging their customers to make investments in energy efficiency. This study relies on panel data on 324 utilities spanning 11 years to estimate the effect of DSM expenditures on retail electricity sales. Our estimates imply that DSM had a much smaller effect on retail electricity sales than do estimates reported by utilities themselves over the same study period.

Understanding cancer treatment and outcomes: the Cancer Care Outcomes Research and Surveillance Consortium.

Abstract: The National Cancer Institute (NCI) has funded the Cancer Care Outcomes Research and Surveillance (CanCORS) Consortium as the centerpiece of the NCI's ongoing initiative to improve cancer care. This commentary provides an introduction to the goals and methods of the CanCORS Consortium to the broader community of cancer researchers and clinicians, many of whom are caring for patients participating in this study.

Characterizing Climate-Change Uncertainties for Decision-Makers. An Editorial Essay

Abstract: ROBERT LEMPERT1, NEBOJSA NAKICENOVIC2, DANIEL SAREWITZ3 and MICHAEL SCHLESINGER4 1RAND, Santa Monica, CA 90407, U.S.A. E-mail: lempert@rand.org 2International Institute for Applied Systems Analysis, A-2361 Laxenburg and Vienna University of Technology, A-1040 Vienna, Austria 3Arizona State University, Consortium for Science, Policy, and Outcomes, Tempe AZ, 85287-4501, U.S.A. 4Department of Atmospheric Sciences, University of Illinois at Urbana-Champaign, Urbana, IL 61801, U.S.A.

Five-Year Impact of Quality Improvement for Depression: Results of a Group-Level Randomized Controlled Trial

Abstract: Background Quality improvement (QI) programs for depressed primary care patients can improve health outcomes for 6 to 28 months; effects for longer than 28 months are unknown. Objective To assess how QI for depression affects health outcomes, quality of care, and health outcome disparities at 57-month follow-up. Design A group-level randomized controlled trial. Setting Forty-six primary care practices in 6 managed care organizations. Patients Of 1356 primary care patients who screened positive for depression and enrolled in the trial, 991 (73%, including 451 Latinos and African Americans) completed 57-month telephone follow-up. Interventions Clinics were randomly assigned to usual care or to 1 of 2 QI programs supporting QI teams, provider training, nurse assessment, and patient education, plus resources to support medication management (QI-meds) or psychotherapy (QI-therapy) for 6 to 12 months. Main Outcome Measures Probable depressive disorder in the previous 6 months, mental health–related quality of life in the previous 30 days, primary care or mental health specialty visits, counseling or antidepressant medications in the previous 6 months, and unmet need, defined as depressed but not receiving appropriate care. Results Combined QI-meds and QI-therapy, relative to usual care, reduced the percentage of participants with probable disorder at 5 years by 6.6 percentage points ( P = .04). QI-therapy improved health outcomes and reduced unmet need for appropriate care among Latinos and African Americans combined but provided few long-term benefits among whites, reducing outcome disparities related to usual care ( P = .04 for QI-ethnicity interaction for probable depressive disorder). Conclusions Programs for QI for depressed primary care patients implemented by managed care practices can improve health outcomes 5 years after implementation and reduce health outcome disparities by markedly improving health outcomes and unmet need for appropriate care among Latinos and African Americans relative to whites; thus, equity was improved in the long run.

Natural programming languages and environments

Abstract: Over the last six years, we have been working to create programming languages and environments that are more natural, or closer to the way people think about their tasks. Our goal is to make it possible for people to express their ideas in the same way they think about them. To achieve this, we have performed various studies about how people think about programming tasks, both when trying to create a new program and when trying to find and fix bugs in existing programs. We then use this knowledge to develop new tools for programming and debugging. Our user studies have shown the resulting systems provide significant benefits to users.