Team Work Experience and Team Performance: Evidence from Doctors
Factors driving variation in the quality and cost of care delivered by doctors is a question with important policy implications. Although teamwork between doctors is common in health care provision, most existing policies and research attribute doctor quality and cost of care performance to the quality and incentive of individual doctors. Little attention has been paid to the role of teams in shaping doctor performance. This proposal aims to study the role of team work experience—the amount of experience doctors working as a team have working together—in shaping doctor quality and cost performance in care for the elderly. The empirical strategy exploits quasi-random variation in team work experience among heart attack patients admitted through the emergency department. The main datasets used for analyses are 2008-2016 20% Medicare claims files.
Genetic Cognitive Epidemiology: Using Molecular Genetic Data to Explore the Link Between Cognition and Mortality
In recent years, polygenic scores have proved strong predictors of many outcomes, ranging from height to educational attainment. At the same time, there is growing evidence of a robust relationship between adolescent cognitive ability and health and mortality outcomes across the life course. However, the causal mechanisms linking cognition and health remain underexplored. Using data on a longitudinal cohort of siblings born in the middle of the 19th century in the United States, we decompose the relationship between adolescent cognitive ability and long-term health outcomes. By employing a novel instrumental variable technique using individual genetic data combined with sibling analyses, we separate and quantify the portion of the cognition-health gradient that is caused by early-life environmental stress and trauma from the portion caused by healthy behavior and selection into healthier environments. We also explore mediators and moderators of the cognition-health pathway.
Sleep and Well-being in the Elderly: Is Deterioration Preventable?
Difficulties with sleep increase with age and have a detrimental impact on quality of life. While in the past declining sleep quantity and quality were often seen a “natural” part of aging, accumulating research suggests that sleep problems frequently are associated with the pathological processes that may be preventable. Given the growing size of the elderly population and the new focus on patient-oriented outcomes, the questions are critical. Using the ongoing well-being cohort of the Wellness Living Laboratory (WELL) initiative of the Stanford Prevention Research Center, we aim to elucidate the inter-connections between sleep quantity and quality, lifestyle health behaviors, and multiple dimension of well-being in the elderly. While WELL continues to enroll participants, we will use the 1600 individuals recruited from the Bay Area of whom 20% are 65+ years. By 1/1/19, the expected sample size is expected to be 3,500. While only baseline data is currently available, future follow-up of the findings from this CDEHA project will allow future research focusing on longitudinal data as well as include the international comparison cohorts of WELL Asia.
How does genetic risk of Alzheimer’s Disease affect adolescent and early adulthood cognition?
The APOE gene has been well-established as an important risk factor in Alzheimer’s Disease (AD). APOE ε4 carriers are estimated to have a 3-10 fold increase in Alzheimer’s disease by the age of 751. However, little is known conclusively about the role of APOE in childhood and early adulthood. Do the deleterious effects of APOE ε4 begin in childhood or early adulthood? The proposed study will investigate the effects of APOE, a salient genetic risk factor for AD, on cognition across the lifecourse, in particular in adolescence and early adulthood. To test this, we combine two data sets with complementary strengths – a longitudinal aging cohort, the Wisconsin Longitudinal Study (WLS) and a study focused on transitions from adolescence into adulthood, AddHealth. Using this lifecourse approach, we test for associations between APOE, an important genetic risk factor for AD, with various cross-section and longitudinal measures of cognition, including the use of sibling models. The influence of APOE alleles on cognition at an early age is of particular interest, because it would potentially aid mechanistic understanding of the disease and may provide opportunities to detect or mitigate risks of later-life AZ.
The Long-Term Mental and Physical Health Consequences of Divorce in High-Conflict Families on Aging Parents
Abundant research demonstrates a negative correlation between divorce and individuals’ later mental and physical health. But when it comes to divorce in families with children, most of the existing literature focuses on the impacts on children, rather than the parents themselves. This project will advance knowledge by isolating the long-term mental and physical health impacts of divorce on parents as they age, focusing on those who exhibit high levels of conflict. The project first combines multiple sources of linked administrative population-level data from Denmark and develops a new measure of conflict by identifying parents who go to a civil court to reach a decision about child custody. Then, the project uses the rich panel structure of the data and an event-study framework to compare the health outcomes of high-conflict (i.e., those who have at least one child custody court case) versus low-conflict (i.e., those who are never observed in the court records) parents, before and after divorce. The project will analyze inpatient, outpatient, and pharmaceutical claims for different psychological and physical health conditions measured at various ages throughout the parents’ life cycles, including depression, anxiety, dementia, drug and alcohol abuse, obesity, diabetes, and hypertension. It will also study heterogeneous impacts across different family characteristics, focusing especially on parents who divorce when their children are older, and who are therefore observed into old age. It will also consider heterogeneity based on parental gender, child gender, and parental prior mental and physical health. As Denmark and the U.S. have relatively similar rates of divorce, as well as laws regarding child support and custody, the results should also be broadly relevant to the American setting.
Are the Mentally Ill Rational? An Analysis of Responses to Insurance Design
Many elements of insurance design and health policy contain incentives aimed at encouraging individuals to adopt or avoid specific behaviors. For example, many insurance plans set copayments for prescription drugs to encourage the use of cheaper generics and branded drugs. A crucial assumption underlying these incentives is that patients are rational (from an economic sense) and will respond to these incentives. However, whether people with mental illness respond in the same way to economic incentives as people without mental illness remains unknown. In this study, we will (a) characterize how mentally ill people respond to economic incentives present in insurance design and (b) characterize how any differences affect health outcomes such as emergency room visits and hospitalizations.
Performance Pay and The Principal-Agent Problem
The Hospital Readmissions Reduction Program (HRRP), which penalizes hospitals for high rates of readmissions, is an important performance pay program established by the Affordable Care Act. By tying payments to quality of care, HRRP aims at improving health care quality through hospitals’ financial incentives. However, although physicians are important determinants of health care quality, HRRP puts no direct financial incentives on physicians, creating a potential principal-agent problem where hospitals--the principal--have large financial incentives to reduce excess readmissions, while physicians--the agent--have no direct incentives. This proposal aims at studying the role of principal agent problem between hospitals and physicians in explaining the effectiveness of performance pay.
Preparing for Alzheimer’s Disease: Information and the Use of Formal and Informal Care
Families who provide the care that Alzheimer’s disease (AD) victims need good information about the likely course of the disease. Such information can help families better decide whether and when the AD victim will need long-term care, which in turn can help in financial planning and other measures (such as long-term care insurance) that will be needed to arrange financing for such services. The primary goal of this project is thus to empirically explore whether differences in information explain differences in whether and how people prepare for long-term care in old age. To this end, the study aims to develop new and unique measures of each individual’s stock and flow of information about AD and construct a model of how demand for formal and informal systematically varies with what individuals know about AD.
Medicare Locality Payments and Physician Location Decisions
This study will combine physician practice location data from the American Medical Association’s Masterfile with data on the structure of the Geographic Practice Cost Indices (GPCIs) to accomplish the following specific aims: First, it will characterize the extent to which physician practice location respond to changes in the GPCI. Second, it will determine the extent to which changes in practice location are driven by physician characteristics, such as age and specialty. Third, it will characterize the extent to which changes in practice location are associated with mortality changes.
Did the Korean GI Bill reduce disparities in cognitive function?
Achieving high socioeconomic status (SES) is a strong predictor of level of cognition and slower cognitive decline. Dr. Vable’s prior work revealed that the Korean War GI Bill was associated with the attainment of significantly higher SES for eligible veterans as compared to similar ineligible persons. She hypothesizes that the Korean War GI Bill contributed to higher levels of cognition and slower cognitive decline. To test this hypothesis, this project will use data from the Health and Retirement Study (1992 - 2016), which has sufficient sample sizes of both Korean War GI Bill-eligible male veterans and potential non-veterans to match for comparison. The researchers will compare the analytic samples created by propensity score matching, coarsened exact matching, and un-matched ordinary least squares regression. Level of cognition and cognitive decline among veterans and non-veterans will be assessed using generalized estimating equations in the matched population, and in low and high childhood SES subgroups.
Evaluating Training Preferences and Outcomes from the Residency Match
Very little is known about how the physician training process affects the practice of medicine. In this project, we examine the effect of training during residency on medical utilization decisions. Using data from a large residency program, we exploit the fact that patients, teammates, and supervising physicians are as good as randomly assigned across physician trainees. We find that much of the variation in practice is related to position on the medical team comprised of first-year “intern” trainees and second- or third-year “resident” trainees: trainees who are in their first year of training but close to the end of the first exhibit much less variation in practice than trainees who are in their second year of training but close to the beginning of the second year. We find that learning occurs to a larger extent, among the same group of physician trainees, in environments with more specialized knowledge, such as cardiology and oncology rotations, as opposed to in general medicine. We are currently exploring the effect of training with high-spending supervising physicians on the future spending practices of the trainees.
The Impact of Tuskegee on the Health Outcomes of Black Males
This research study sought to understand the role of mistrust in the healthcare system as a potential cause of historical and contemporaneous disparities using the historical Tuskegee episode as a proxy. The Tuskegee (Alabama) Study of Untreated Syphilis (TSUS) in the Negro Male passively followed black males with syphilis between 1932 and 1972 and failed to provide treatment despite the fact that the men in the study believed they were receiving free medical care. The deception was disclosed by Jean Heller of the Associated Press in 1972 (Jones, 1993). The researchers described the trajectory of mortality disparities for black men and women (as compared to white men and women) before and after the Tuskegee disclosure.
Out-of-Pocket Cost and Utilization of Healthcare Among Elderly and Pre-Elderly Adults
This study brings to bear novel data and quasi-experimental methodology to examine determinants of health-seeking behavior among different demographic groups of elderly and pre-elderly adults, and assess how adopting health-seeking behaviors impacts healthcare costs for these populations. The researchers examined all-payers claims databases (APCDs) data from various states to document how take-up of preventive cancer screenings responded to changes in cost-sharing that occurred as part of the ACA reform of 2010. The study also examined the consumption of prescription pharmaceuticals for chronic conditions among seniors and found that that seniors decrease consumption of prescription drugs when faced with higher prices, and that this decrease is particularly pronounced for so-called “maintenance” drugs that are likely to have delayed impacts on health and well-being.
Exploring the relationship between traffic citation history and crash risk among elderly drivers in Florida
The number of licensed drivers in the US over age 70 years is projected to double over the next 15 years. This poses a substantial challenge for road safety because older age is a risk factor for road traffic crashes, injuries, and fatalities. This project is employing a novel approach to identify elderly drivers at especially high risk of accidents by using information on their citation records. Regulators may then be able to use such “flags” to target interventions, such as driver performance evaluations. The researchers have built a large dataset from Florida, consisting of all citations and crashes that occurred over a 9-year period (2006-14), merged at the driver level. They are exploring the merits of several different approaches to predicting crashes. The statistical analyses for this project is nearing completion, and will be written up for publication during Summer 2017.
Watersheds in Infant Mortality: The Role of Effective Water and Sewerage Infrastructure, 1880 to 1915
This project explored the first period of decline in infant mortality in the U.S. and provide estimates of the independent and combined effects of clean water and effective sewerage systems on infant mortality in Massachusetts, 1880-1915, when state authorities developed a sewerage and water district for municipalities in the Boston Greater Metropolitan area. The researchers found that the two interventions were complementary and together accounted for approximately 44 percent of the total decline in log infant mortality among treated municipalities during the 35 years considered. Considerable research has documented the importance of clean water interventions for improvement in population health, but there is less evidence on the importance of sewerage systems. These findings are directly relevant to urbanization in the developing world and suggest that a dual-pronged approach of safe water and sewerage is important to improving infant and early child survival.
Effect of Social Isolation and Loneliness on Health Care Utilization
The researchers assessed the effect of social isolation and loneliness on healthcare utilization (costs and frequency of care) using longitudinal survey data from the Health and Retirement Survey (HRS) linked to Medicare administrative claims data. This study provides the first representative picture of the correlation social isolation and loneliness have with total Medicare costs. Additional project funding was obtained from the AARP Public Policy Institute. Findings were published in the August 2017 issue of the Journal of Aging and Health and in November 2017 by the AARP Public Policy Institute, including the main finding that social isolation costs the Medicare program $6.7 billion in additional spending every year. This research has also been highlighted in Forbes magazine and in an AARP Public Policy Institute (PPI) sponsored Solutions Forum on Capitol Hill.
Readmissions and Chronic Disease in the Elderly
This study aimed to expand knowledge regarding chronic disease and readmissions in the elderly The researchers conducted a retrospective analysis of California Office of Statewide Health Planning and Development Patient Discharge Data, 2000-2009, which includes all adults age 18 and older with a non-federal acute care hospital discharge (n= 29,009,966 discharges). We identified 7- and 30-day all cause readmissions (ACR) and potentially preventable readmissions (PPR), and then analyzed relationships between index and readmission hospitalizations for each metric. They found that in the eligible samples, the 30-day ACR and PPR rates were 10.0% and 5.6% respectively. Index admissions for Heart Failure had the most 30-day ACRs and highest readmission rate (103,591, 19.8%) among medical APR-DRGs, while Other Vascular procedures had the highest rate (14,635, 14.6%) among surgical APR-DRGs. Overall, relatedness of readmissions did not differ for 7- vs. 30-day ACRs. The PPR algorithm captured fewer unrelated readmissions than did the ACR for surgical but not medical APR-DRGs. Unrelated readmission captured by PPR included both acute and chronic conditions, in particular ambulatory care sensitive conditions. This study demonstrates that the concept of relatedness is similar to, but not identical to “potentially preventable” as determined by the PPR metric. Both the ACR and PPR provide a unique perspective on readmissions and can be useful across applications. The reasons for readmission captured by both metrics can provide guidance for potentially impactful interventions.
Helping Older Adults to Initiate Exercise Habits
The aims of this seed were to 1) assemble a longitudinal dataset by collecting information on individuals’ exercise patterns in the post-commitment contract period, as data on their demographics, randomization to contract duration suggestions, contract choices, and exercise patterns during their contracts have previously been collected, and 2) assess whether longer commitment contracts and consequently more exercise during the contract period increases the formation of exercise habits in the post-contract period and whether this effect is differential in older adults who face more immediate health risks from continued sedentary lifestyles. Data collection has been completed and the researchers are currently finalizing their analysis and write-up of findings. Preliminary findings have been presented in Oslo at international meetings sponsored by the University of Oslo and the University of Bergen.
The Cost Impact of Changing Antihypertensive Medication Prescribing for the Elderly
The aim of this seed is to identify the relative impact of various cost drivers on the increasing cost of hypertension treatment among US elderly. Different drivers may have different implications for cost containment. Thus far, the researchers have analyzed the National Diagnosis and Therapeutic Index (NDTI)-patient encounter data and are now working on the pharmacy dispensing data set to determine the relative changes in aggregate costs over time according to four key cost drivers: population of patients with hypertension seen in primary care, number of drugs prescribed per patient, selection of drugs by drug class and price of drugs by drug class. Following this the researchers will compare analyses for patients 65 year and older versus younger.
Results suggest that over the past 15 years there has been relative stability in the proportion of elderly patients who have visited primary care offices. The number of antihypertensive drugs prescribed these patients at visits has increased, and the proportion with well-controlled blood pressure has also increased. Trends in antihypertensive prescriptions identify increased prescribing of ACE-inhibitors, angiotensin receptor blockers, and beta-blockers and flat or decreased prescribing of diuretics and calcium channel blockers. Of note, sub-set analyses indicate persistent disparities in blood pressure control by race/ethnicity. Remaining analyses include full link-up with NPA data on exact drug dosing and costs.
Effect on Grandparents of Providing Child Care for Grandchildren
This project empirically examined how having daughters versus sons affects the child-care responsibilities of older women in low-income communities. The hypothesis was that mothers often take care of their daughters' children, and that this extra work of providing child care to their grandchildren can be taxing on their health and well-being. The researchers examined these patterns in the Changing Lives of Older Couples survey. The hypothesis was not confirmed, in part due to the small number of observations from low-income households in the Changing Lives data set.
Systematic Review: The Health Benefits and Harms of Gardening Among the Elderly
This meta-analysis evaluated the extent to which gardening among elderly adults is associated with specific health outcomes, including exercise capacity, body weight, serum lipids, fasting serum glucose and insulin, blood pressure, mental status, and gardening-related orthopedic injuries and exposure to infectious diseases. Searches of eight databases, including: PubMed, PsychInfo, AgeLine, ToxLine, Ovid, ERIC, CAB, and Cinhal resulted in identification of over 6,000 articles. Roughly 180 met criteria for full data abstraction. Of these 180 articles that underwent full abstraction, about 30 met the inclusion criteria. Results suggested that gardening interventions increase people's consumption of fruits and vegetables, both in the elderly and in children. Data about harms and other health outcomes have been analyzed. Two poster presentations were delivered to the Bay Area Research symposium.
The Health Effects of Air Pollution: Evidence from Forest Fires
This study examined the link between exposure to particulate matter—a form of air pollution in which fine particles are suspended in the air—and adverse health outcomes for the elderly, who are hypothesized to be affected disproportionately. In order to continue to determine environmental standards for pollution and to design effective public health warnings about pollution, an understanding of the health effects is needed. Such policies are particularly pertinent to the elderly, who might be socially isolated and less able to leave an area during a high-pollution episode. In contrast to previous studies, the researchers did not find an effect of particulate matter (PM) on the health outcomes examined (cardiovascular hospital admissions, all respiratory admissions, and admissions for myocardial infarction). However, the findings do not imply that PM does not affect these health outcomes, as the confidence intervals do not rule out effects comparable to those found in previous research.
Increasing Physical Activity Among the Elderly: A Meta-Analysis of the Effectiveness of Pedometers
This study was the first to synthesize quantitatively the literature on the effectiveness of pedometers to change physical activity and health outcomes among the elderly. Preliminary results were presented at the Stanford Prevention Research Center (March 2007) and at the Northern California regional Society for General Internal Medicine (SGIM) Meeting (March 2007), where it won the award for best presentation. The project was also presented at the International SGIM Meeting in Toronto in April 2007 and received a great deal of media attention. The results of this study were published in the November 2007 issue of JAMA, and Dr. Bravata was also featured as part of the “Author in the Room” series with a web cast on Dec. 19th, 2007.
Effects of Obesity on Employer-sponsored Health Insurance
Drs. Bundorf and Bhattacharya received an RO1 grant (RO1-AG028236) from the National Institute on Health (NIH) to extend their work to look at the external costs of obesity through large public health insurance programs, including Medicare and Medicaid. Results from the seed have been presented at four national meetings. A paper entitled, “The Incidence of the Health Care Costs of Obesity,” was published in Journal of Health Economics.
National Trends and Practice Patterns of Detection and Treatment of Hypertension During Ambulatory Visits in the U.S.
Practice guidelines aim to guide physician practice according to the best available evidence. Data were mixed regarding the impact of practice guidelines on physician prescribing. The researchers analyzed data from three national ambulatory care surveys to depict long-term trends in U.S. antihypertensive prescribing between 1990 and 2004 in relation to JNC guidelines and short-term trends following the 2002 publication of final ALLHAT results. Seed project findings and findings from additional analysis were consistent with those from the NDTI analysis suggesting that both national guidelines and clinical evidence had an impact on antihypertensive prescribing practices but magnitude of impact may be smaller than desired. Results from this seed project have led to one published manuscript, “Screening, Treatment and Control of Hypertension in U.S. Private Physician Offices, 2003–2004”, in Hypertension.