Chapter one investigates whether or not standardizing Medicare HMOs' benefits packages would reduce risk selection. Policymakers have expressed the hope that standardization would reduce health plans' ability to risk select by constructing benefits packages that are differentially attractive to healthy beneficiaries. I find evidence that plans' relative copayment levels, physician network sizes, and quality all influence risk selection among plans. Thus, the standardization of benefits packages would restrict consumers' choices without preventing risk selection. Given this finding, in chapter two I look for ways in which Medicare payments to plans could be "risk adjusted" to reflect the expected costs of plan enrollees. I focus on decedents since they have high costs. I find that while Medicare payment systems could be improved by paying more for beneficiaries with certain terminal illnesses, incentives would remain to select against the terminally ill. Chapter three focuses on methodological issues. It presents and evaluates alternative econometric methods of modeling risk selection and predicted health care cost measures such as the selection measure used in chapter one.

In Paper One, I explored the impact of patients’ experiences during medical care on their trust in physicians using a survey of 1,602 colorectal cancer patients. I used logistic regression on trust scores to assess the association of specific health care domains with the likelihood of having low trust, controlling for patient sociodemographic and clinical covariates. During medical care experiences, problems in access, patient education and particularly coordination and continuity of care seriously affected patients’ trust in their health care providers, suggesting that efforts to improve these aspects of medical services may enhance patients’ trust.

In Paper Two, I studied the effect of patient characteristics on hospital choice. Using data on 38,237 patients, associations of patient attributes with surgical care at high-volume hospitals, teaching hospitals, and hospitals with different level of adjusted mortality were assessed with multivariate regression analysis. Patients with more severe clinical status, of minority ethnicity, and of less privileged socioeconomic status were significantly less likely to be admitted to the high-quality institutions for cancer surgery, suggesting the potential benefit of improving access for patient outcomes.

In Paper Three, I evaluated the utility of cancer registry data for profiling of hospitals by survival of colorectal cancer patients. With 5 years of data from California supplemented with hospital discharge abstracts, I fitted univariate and bivariate hierarchical models with hospital random effects and used a Bayesian algorithm to provide valid estimates of hospital profiles for survival. For larger hospitals, adjusted 30-day survival rate can be estimated with adequate precision to support useful and valid comparisons of hospital quality, especially with additional information from the bivariate-outcome model, suggesting that the already-collected cancer registry data can find a new use for profiling cancer care providers.

The first paper is a longitudinal analysis of the primary drivers of the obesity epidemic in developed countries and the contributions of various markers of development to increased caloric intake. The results indicate that rising obesity is primarily the result of consuming more calories, and that the increase in caloric intake is associated with technological innovations such as reduced food prices as well as changing sociodemographic factors such as increased urbanization and increased female labor force participation.

The second paper uses propensity scores to examine the independent contributions of insurance status (e.g., Seguro Popular vs. uninsured) and health professional supply (e.g., number of doctors and number of nurses per 1000) on coverage of antihypertensive therapy among adults with hypertension in Mexico. The findings suggest that having Seguro Popular (SP) insurance is associated with higher rates of antihypertensive treatment and blood pressure control. Further, Seguro Popular may be most effective in areas with a high health professional to patient ratio. Finally, the results indicate that 3,381 cardiovascular deaths among the uninsured could potentially be avoided through enrollment in SP; approximately six percent of total cardiovascular mortality for the SP-eligible population in 2004.

The third paper uses multivariate regression analysis to assess public trust in scientific experts on obesity and its relationship to both awareness of nutritional recommendations and appropriate behavioral change. This paper also identifies those sociodemographic groups associated with high and low trust in scientific experts. The findings show that trust in scientific experts is the strongest predictor of public attention to nutritional recommendations from scientific experts; that public attention is significantly associated with weight-related behavior; that women and more educated individuals have significantly higher odds of trusting scientific experts; and that Hispanics and older individuate have significantly lower odds of trusting scientific experts.

While the focus and scope of each of these papers is quite different, they each share a common concern for improving our understanding of those factors which may contribute to or reduce the escalation of obesity and its related diseases.

The first paper examines the impact of the number of surgeries at a hospital on the application of new clinical information to hospital practices. Information acquisition is a fixed cost, but benefits accrue per surgery, therefore research adoption is more likely among hospitals performing more total surgeries. The cases of carotid endarterectomy (CEA) and angioplasty during heart attack (PTCA) show that hospitals with more surgeries better adhered to the clinical literature. The data supports the hypothesis that there is more research adoption at high volume hospitals.

The second paper uses parameter estimates from survey, epidemiological and experimental data to model the costs and benefits of new users of anti-depressants due to direct to consumer advertising (DTCA) for the case of depression. This study shows that 94% of new anti-depressant use due to DTCA is from non-depressed individuals. However, the average health benefit to each appropriate new user is 63-fold greater than the average cost per treatment, resulting in a positive net benefit of over $72 million. The cost to benefit ratio of DTCA could be improved through better targeting of advertisements and higher quality treatment of depression.

The final paper examines the determinants of the undermanagement of osteoporosis after hip or wrist fracture in females over age 65. Using a retrospective cohort, I find that patients older than 90, and black patients were less likely than average to be treated for osteoporosis. Female prescribers were more likely than male prescribers to manage osteoporosis and bisphosphonate treatment before fracture was the best predictor of bisphosphonate treatment after fracture. Our findings highlight that prescribers of all experience levels, specialties and locations need to improve osteoporosis management after hip or wrist fracture.

In the first chapter, I evaluated the 2001-04 reforms on patient financial burden taking advantage of its phased implementation using a difference-in-difference approach. The reforms were successful at containing the out-of-pocket payments for hospitalized patients. The difference-in-difference estimator of the reform effect was 400 soms (US$10), equivalent to 29% of the pre-reform out-of-pocket payments in reform oblasts. The reforms were particularly effective at limiting the financial burden among the poorest 40% of the population. At the same time, I observe spill-over effects for visits and outpatient drug purchases reversing the protective effect the reforms exerted for hospitalization. This trend was stronger for the non-poor than for the poor. Combining the distributional impact of public expenditures with the distributional impact of out-of-pocket payments, the reforms redistributed health care resources to the benefit of the poor.

In the second chapter, I examine the socio-economic distribution of informal payments to health care personnel. Price discrimination has been suggested as the model of pricing behavior for informal payments, but it has not been established empirically. I do not find evidence that the poor are less likely to pay informal payment to medical personnel than the non-poor. Conditional on paying, the poor pay 25% less than the non-poor although their per capita annual consumption is 50% less. Additionally, 32% of patients pay less co-payment than they should leading to a revenue loss equal to 43% of co-payment collections. I do not find evidence that the poor are more likely to receive these discounts than the non-poor. Receiving a co-payment discount is associated with a 27% increase in informal payments. Thus, discounts could reflect strategic revenue maximizing behavior as physicians attempt to re-capture informal payment lost after the introduction of the co-payment policy.

In the third chapter, I evaluate three approaches to the measurement of socioeconomic status in the absence of data on consumption or expenditure. I calculate the concentration index for health care visits and hospitalizations with each ranking variable. The choice of welfare indicator has a significant impact on welfare ranking and on the estimated degree of inequality in health service utilization. Based on the findings of this paper, the scale items I validate can be used in future equity-related research in the Kyrgyz Republic.

Using survey, archival, interview and observation data from participants in four collaboratives sponsored by the Institute for Healthcare Improvement, I first assess which features of collaboratives matter most for learning. I find that access to faculty experts knowledgeable about the specified topic and achieving improvement matters most. Second, I capitalize on the existence of collaboratives to investigate interorganizational learning in health care, a process that has received scant attention in the organizational learning and health care management literatures. I propose and test a model of interorganizational learning activity (ILA). Results show that team, organizational, collaborative and task characteristics all influence the use of ILA, which increases organizational improvement. Notably, collaborative identification and team functioning play important mediating and moderating roles, respectively.

Finally, I address the observation that engagement in improvement work beyond collaborative teams is necessary, and that barriers to collaboration for intraorganizational learning exist. I focus on one barrier: the differential status accorded to those in different professions. Survey data collected from another collaborative, Vermont Oxford Network, reveal that profession-derived status is positively associated with psychological safety - a key antecedent of engagement in improvement efforts. Moreover, results show leader inclusiveness - words and deeds exhibited by leaders that invite and appreciate others' contributions - helps overcome inhibiting effects of status differences.

Together, the findings offer insight into antecedents of and strategies for fostering intra- and inter-organizational learning, and ultimately improvement, in health care.

This work applies an interdisciplinary approach, drawing from law, philosophy, sociology, and history, to provide a contextual analysis of privacy interests and the optimal means of regulation. It makes the case for greater consideration of alternative approaches to codification of diminished individual privacy interests. Instead, it advocates increased use of social norms as a way of gaining the considerable positive benefits of genetic advancements without irreversibly sacrificing individual privacy rights.

Chapter 1 explores comparative approaches to unconsented disclosure of genetic test results to relatives. It offers a normative analysis of discretionary disclosure, a policy that would allow the disclosure of test results to relatives for whom there may be an elevated risk of disease. It concludes that this policy is problematic and, ultimately, counterproductive.

Chapter 2 examines shifts in privacy norms occasioned by the advances in genetic technologies and identifies a spillover effect in the form of the inadvertent emergence of new norms. This chapter introduces an original typology developed in response to these new norms regarding privacy. It focuses on the emerging practice of compelling access to genetic information of biologically-related persons in order to gain information about a particular individual. It concludes that greater attention should be given to the spillover effect and the emergence of “shadow norms”.

Chapter 3 explores the force of comparative privacy protections between selected European countries and the United States. I contrast Europe’s application of the proportionality principle with the U.S. model of categorical protections. Does the seemingly more flexible European approach provide lesser or greater protection than the U.S. categorical model? I conclude that architecture alone is not determinative.

Section 1 explores the effect of multidisciplinary primary care teams on patients' experiences with care. Higher physician continuity was associated with more favorable patients' experiences. An exception was patients' assessments of teams, which were better when on- vs. off team visits occurred. For other measures, the decrements associated with discontinuity were the same irrespective of whether discontinuities involved on- or off-team visits. The findings highlight the challenges of incorporating teams into primary care in ways that patients experience as value-added rather than disruptive to primary care relationships.

Section 2 clarifies whether visit continuity influences patients' experiences equally in various clinical situations. Physician-patient interaction quality and organizational access were more strongly influenced by visit continuity among respondents in early stages of a physician-patient relationship and with worse self-rated health. Visit continuity during early stages of a physician-patient relationship may ultimately increase the acceptability of approaches that leverage physician time.

Section 3 assesses which patient, physician, and organizational factors are related to voluntary physician switching among HIV-infected patients. Lower voluntary switching was predicted by patient trust, physician anti-retroviral knowledge, moderate (rather than low or high) HIV patient volume at a care site, and Ryan White Care Act funding. Patients with chronic illnesses may use several markers of specialization and technical quality to make decisions about their care.

Section 4 assesses the effect of care team composition on the quality of HIV care. In adjusted analyses, having a care team composed of three or more clinicians was associated with more consistent prescribing of Pneumocystis carinii (PCP) prophylaxis when medically-indicated. However, patients with multiple physicians generally reported worse care coordination and exhibited less appropriate use of emergency services. These findings highlight that team approaches offer a mix of advantages and disadvantages to patients.

In Chapter 2, the impact of pre-market competition on success in pharmaceutical R&D is estimated using a discrete choice model.Using a large database of molecules in development from 1994-2004, and drawing from the literature on net present value, sequential development projects and real options, I consider the factors that explain molecule transitions through phases of development and examine how the pre-market competitive environment influences the probability that a molecule will move to the next stage. The parameter estimates suggest that the likelihood of a molecule transitioning is influenced by the successes and failures of other molecules in development, findings that are consistent with a model of learning.

In Chapter 3,the economic incentives for research and development of pharmacogenomic therapies are examined through simulations that compare the variables influencing the expected profits for firms considering investing in genomic-based therapies versus investing in non-targeted, or "conventional therapies." The findings suggest that firms face strong incentives to develop conventional therapies, which are more profitable in all of the simulations than targeted therapies. These findings, however, are most sensitive to market share and pricing, which suggest that factors influencing both could make targeted therapies more attractive investments to firms.

Using survey data collected from more than 20,000 individuals in 105 US hospitals, Chapter 2 investigates elements of organizational culture that may impact patient safety in hospitals. I identify a valid and reliable nine-dimension “Patient Safety Climate in Healthcare Organizations” (PSCHO) framework for characterizing safety climate that places constructs into three groups, reflecting organizational, work-unit, and interpersonal contributions to safety.

Chapter 3 recognizes that safety climate may vary within organizations and that understanding variation can facilitate intervention. This research focuses on variation due to an individual’s place in an institution’s management hierarchy. PSCHO survey data confirms large and consistent perceptual differences among senior managers, supervisors, and frontline workers, with senior managers consistently the most positive and frontline workers the most negative. Differences by management level, however, depend on profession, age, and gender.

Chapter 4 undertakes unprecedented research by examining the link between stronger safety climate and better safety performance among hospitals. Drawing on the insight that multiple interacting organizational layers contribute to organizational safety culture, I hypothesize that dimensions related most closely to individuals’ behaviors will be most associated with indicators of organizational safety performance, and I find this to be the case. Building on evidence of large differences in safety climate perceptions by management level, I also find that perceptions of frontline workers relate to indicators of safety performance, but those of senior managers do not. In addition, perceptions of nurses relate to nurse-sensitive indicators while no such relationship exists between physicians’ perceptions and physician-driven indicators.

Together, findings offer insight into strategies for strengthening safety culture and ultimately for improving the quality and cost-effectiveness of healthcare delivery.

The first two papers are based on a new survey of 441 community-dwelling elderly conducted in the greater Boston area. The first chapter examines whether use of various technologies and environmental factors were important determinants of self-reported disability, conditional on objective measures of functioning. This study finds that a majority of respondents who used walking aids did not report disability walking around inside. In addition, increased use of van service, senior housing and ramps may explain up to 19% of the decline in disability grocery shopping between 1982 and 1999.

The second chapter explores whether anchoring vignettes, a new survey technique, identifies systematic differences in thresholds for reporting disability. We hypothesized that respondents who used various technologies and environments may have a higher threshold, i.e. lower propensity, to report disability compared to non-users. In models for walking around inside, we found that respondents using mobility assistance actually had lower thresholds for reporting disability and that random variation across respondents was an important determinant of survey responses.

In the third chapter, we use data from the National Long Term Care Survey, including Medicare-linked files, to evaluate whether improved treatment for acute myocardial infarction led to lowered disability and death among the elderly population over time. We find that increased use of invasive procedures was associated with improved survival and lower disability among survivors, and that increased use of beta-blockers were associated with improved survival. These findings suggest that improved medical care treatment led to greater survival and lower disability.

In the second paper, I examine the impact of the introduction of the Medicaid program on labor force participation among single women. Using variation in the timing of Medicaid implementation across states and in eligibility across demographic groups, I find no evidence that women who were eligible for Medicaid decreased their labor supply relative to women who were not. These results add to an emerging consensus in the literature suggesting that public health insurance programs for low-income parents and children may be able to achieve health benefits and improve access to care without substantial indirect costs from labor supply distortions.

Racial/ethnic concordance between patients and physicians may affect health care disparities by reducing discrimination. In the third paper, I investigate the role of concordance on rates of preventive screening and the length of outpatient, primary care visits. I find little evidence that concordance plays an important role in these outcomes. Physician race tends to be a much more important predictor of these outcomes than patient race or concordance, but the direction of the effect varies. The results highlight the importance of measuring the role of concordance separately from patient and physician race. They also suggest that policies aimed at increasing the number of minority physicians need to be combined with other methods to improve the quality of primary care.

In the first chapter I introduce a method to profile hospitals on the "value" of care for acute myocardial infarction. Using estimates of society's willingness to pay for survival gains to express each hospital's survival benefit in dollars, I combine in-hospital survival and cost outcomes on the incremental net monetary benefit scale to compare the performance of 69 hospitals. While common in cost-effectiveness analyses of new technologies, this approach has not been applied to provider profiling. In the second chapter, I propose and compare three value estimators that make different assumptions about the relationship between cost and quality. The estimators include one based on a regression framework, the cost-effectiveness estimator from Chapter 1, and the prevailing technique used by health plans, which compares performance on both domains independently. I find significant differences in the set of hospitals classified as high value across methods, and recommend against using the most common approach. In the third chapter I present a framework for profiling hospitals using a summary measure of health status following coronary artery bypass graft surgery, quality-adjusted life years (QALYs). I use eleven performance measures to predict two-year survival outcomes, and derive composite health utility estimates based on a patient's exposure to six potential surgical complications, which are used to weight survival time. I then assess differences in average quality-adjusted life expectancy over two years of follow up among 14 Massachusetts' hospitals. These frameworks provide general methods that are applicable to other conditions, and that can also serve as the basis for more detailed models.

In chapter one, we assessed the relationship between race and trust in four national datasets. In these analyses, only one dataset showed a substantial relationship between race and overall trust in physicians. Important racial differences were, however, found in responses to individual items related to the risk taking behavior of physicians ("risk trust") across datasets, potentially as a result of historic mistreatment of minorities.

In chapter two, we estimated the extent to which risk trust explains racial differences in HIV care. We found that such concerns do not mediate racial disparities in the quality of HIV care, but are an important determinant of adherence to antiretroviral medications. Risk trust not only predicts a patient's belief in the worth of these medications, but it may also have an independent effect on whether the patient is adherent.

Finally, in chapter three we evaluated whether the racial mix of patients at HIV sites of care determines the quality of care and mediates racial and ethnic disparities. We found that as the proportion of black patients at an HIV treatment site increases, the probability that a patient will receive highly active antiretroviral therapy (HAART) decreases after adjustment for patient race, age, gender, and CD4 cell count. However, the racial mix of patients at the site was unrelated to the receipt of HAART after adjustment for a number of other patient characteristics. This suggests that although the receipt of HAART is lower in sites with a high proportion of black patients, this difference is explained by other patient factors.

In Chapter 1 we describe the development and calibration of a natural history model of gastric cancer in the two high-risk countries of China and Colombia. Using a likelihood-based calibration approach, we identified multiple parameter sets that provided model outcomes consistent with epidemiological data on the prevalence of precancerous lesions and gastric cancer incidence in each country. The model is capable of evaluating the relative effectiveness of primary and secondary prevention efforts and reflecting the uncertainty surrounding disease progression on important policy outcomes.

In Chapter 2, we use the model to estimate the health and economic consequences associated with Hp screening in China, where over 40% of the world’s gastric cancer cases occur. We estimated that a once per lifetime screening program for 20-year olds may prevent as many as one in every four to six cancers and appears to be highly cost-effective given commonly used threshold heuristic. Given the ease of detecting and treating Hp infection and the poor prognosis and limited treatment options for gastric cancer, it may be reasonable to adopt a more proactive strategy while waiting for more evidence on the benefits of Hp treatment from the ongoing clinical trials.

In Chapter 3, we develop a population model of gastric cancer to project the impact of changing risk factor trends on gastric cancer incidence among men in China. We found that gastric cancer rates have declined as Hp prevalence has fallen, despite the rise in smoking. Even with the continued decline of Hp and smoking, we estimated that the number of new gastric cancers in 2030 will rise substantially as a result of population growth and aging. While targeted Hp reduction programs may prevent a proportion of these cancers, their impact on population gastric cancer risk will not be realized for several decades. Finding effective strategies for primary or secondary prevention of gastric cancer should be a public health priority.

Furthermore, wide variations in safety, efficacy and costs exist across drugs and diseases. In Paper 2, I investigate the cost-saving effect in one clinical area -- atypical antipsychotic agents compared with traditional mood stabilizers as long-term treatment for bipolar disorder. Using commercial insurance claims data from 1998-2001, I test the hypothesis through several econometric methods, including a propensity-score matching, interrupted time series, differencing strategies, and an instrumental variables approach. Each method can identify the cost-saving effect in a different way, but together they can serve as checks for robustness. I consistently find that the new medication does not reduce non-drug medical spending for patients with bipolar disorder.

Nevertheless, the cost-saving effect is plausible both theoretically and empirically. Cost-saving evidence is often shared across jurisdictions and used to guide decisions regarding the adoption and reimbursement of new drugs. However, variations in healthcare rationing policies may complicate the generalizability of such cost-saving evidence. In Paper 3, I propose an economic approach to generalizability of evidence on cost-saving effects for new prescription drugs across health systems with different rationing schemes. I note that the cost-saving effect depends on a new drug's effect on the relative prices of drug and non-drug medical inputs to health outcomes, as well as the substitutability between the two inputs. I also show that quantity constraints on non-drug treatments mainly decrease the cost-saving effect; and for some special cases, overall budgetary rationing reduces the cost-saving effect as well. This paper therefore demonstrates the importance of considering rationing schemes when applying cost-saving evidence across countries.

The first paper uses unique panel data of state spending, politics, and demographics to analyze how state Medicaid spending responds to state fiscal crises and political factors. I find that, after controlling for need, state Medicaid spending changes are not related to state fiscal downturns, though overall state spending declines in these periods. Liberal ideology is a significant, positive predictor of annual changes in state Medicaid spending, while party control and the strength of health interest groups are not associated with short-term changes in state spending. The results suggest that state Medicaid spending follows a unique budget pattern within states.

Since the passage of the State Children’s Health Insurance Program (SCHIP) in 1997, all states have expanded children’s eligibility for publicly-financed health coverage. The second paper investigates whether and how state economic conditions, politics, and need predict state choices in such coverage. It finds that pre-1997 eligibility was driven largely by political ideology and legislative control, as liberal states and states with Democratic legislatures had more generous eligibility criteria. Post-1997 changes in eligibility responded to the party of the governor, rather than legislature, and depended on state poverty levels but not annual budget fluctuations.

Complex dynamics in state mental health systems create tensions in state priority setting for mental health policy. The final paper uses qualitative evidence gathered through semi-structured interviews with state policymakers, advocates, and providers in four states to understand how policy priorities are set in state mental health policy. It finds that a combination of leadership, reliance on structured relationships, and reaction to outside actors or events characterizes priority setting in state mental health policy.

Health services research raises significant questions about the fundamental assumptions of our informed consent laws. After finding that current informed consent laws fail to protect patients’ ability to make informed medical decisions, the first article analyzes the advantages and disadvantages of adopting shared medical decision-making as an alternative to current informed consent requirements. The article concludes that the long-term benefits of shared decision-making and the use of evidence based decision aids to promote patient understanding and inform medical decision making outweigh the costs for both patients and physicians.

Tension surrounding government intervention into the reproductive decisions of individuals has caused lawmakers in the United States to eschew attempts to regulate assisted reproductive technology, including reproductive genetic testing. Advances in genetic technology will soon enable parents to screen their embryos for hundreds of genetic and chromosomal characteristics through preimplantation genetic screening. While these advances promise significant benefits, they also present risks to both individuals and society. The second article proposes the creation of a federal regulatory body to license and monitor assisted reproduction and suggests a balancing framework for addressing conflicting stakeholder interests in reproductive genetic testing.

Mental health courts divert mentally ill offenders from the criminal justice system into court-ordered treatment. Since their creation a decade ago, mental health courts have continued to change court policies, practices and services to improve participant success. The third article first argues that the structure and characteristics of mental health courts facilitated this innovation. Next, it analyzes qualitative interview data on how seven mental health courts have changed since 2003 and the potential impact of those changes on client success and the courts’ capacity for future change. The article concludes that innovation remains key to mental health court success.

In this collection of studies, we address some of the uncertainty surrounding 64-slice CT by analyzing its cost-effectiveness in two important clinical scenarios: the triage of patients with acute chest pain in the emergency department, and the screening of patients with stable chest pain in the outpatient setting. We also examine the overall adoption and diffusion of 64-slice CT in the U.S. to better understand how hospitals manage new technologies with unclear implications for incremental health outcomes and costs.

We find that coronary CT angiography with the 64-slice CT may lead to more efficient triage of acute chest pain patients and could save money in patients with a low coronary artery disease (CAD) prevalence. In undifferentiated patients presenting to their primary care physicians with symptoms suggestive of CAD, we find that coronary CT angiography yields modestly improved health benefits, compared to conventional diagnostic strategies. However, these results are sensitive to radiation risks and incidental findings. Finally, we find that hospitals that treat high volumes of ischemic heart disease patients — a population that could potentially benefit from coronary CT angiography — are more likely to adopt 64-slice CT, as are hospitals that are financially healthy. Our results suggest that the device’s adoption may be related to efforts to improve quality, but the paucity of evidence informing coronary CT angiography’s role in clinical care implies that quality-driven adoption may be premature. Moreover, our finding that adoption is motivated independently by the availability of capital reinforces concerns about haphazard technology acquisition.

My second and third papers investigate atypical antipsychotic agents, a widely-used treatment for serious mental illnesses, including schizophrenia and bipolar disorder. The second paper investigates the temporal relationship between medication nonadherence and hospitalization risk for individuals with schizophrenia. We used a daily measure of medication availability and found individuals in the first 10 days following a missed prescription refill had an over 50% increase in hazard of mental health hospitalization and 77% increase in hazard of schizophrenia-specific hospitalization. Similarly, medication gaps of more than 30 days were associated with 50% increased hazard of hospitalization. These findings suggest clinicians and Medicaid programs might use pharmacy claims to target future adherence interventions.

Finally, the third paper examines the impact of Medicaid prior authorization programs for atypical antipsychotic agents, which have been implemented to reduce costs in more than 10 States. We used interrupted time series analysis to study the impact on market share of non-preferred agents and costs in West Virginia and Texas. These policies reduced market share for non-preferred agents by 14% in West Virginia and 3% in Texas. However, prior authorization did not decrease pharmacy reimbursements in either state. Further evaluation of the clinical consequences resulting from such policies is urgently needed to determine whether their potential clinical risks outweigh their limited cost impact.

Uninsured near-elderly adults experience worse health outcomes than insured adults. However, the health benefits of providing insurance coverage for uninsured adults have not been clearly demonstrated. In the second paper, quasi-experimental analyses of longitudinal survey data were conducted to assess the effects of acquiring Medicare coverage on the health of previously uninsured adults. Relative to previously insured adults, previously uninsured adults, particularly those with cardiovascular disease or diabetes, reported significantly improved health trends after age 65 for a summary health measure and many component measures. Therefore, providing earlier health insurance coverage for uninsured adults with these conditions may considerably improve their health outcomes.

Efforts to improve management of cardiovascular disease and diabetes may or may not reduce disparities in clinical outcomes. Furthermore, the effects of Medicare coverage on these health disparities have not been conclusively demonstrated. In the third paper, serial cross-sectional data were used to assess recent national trends in disease control, trends in disparities in control, and changes in disparities after age 65 associated with near-universal Medicare coverage. Control of blood pressure, glucose, and cholesterol improved substantially since 1999, but racial, ethnic, and educational disparities persisted or widened, suggesting more focused efforts are needed to improve quality of care for disadvantaged groups. Where present, group differences in systolic blood pressure, HbA1c, and total cholesterol were smaller for ages 65-85 than ages 40-64. Thus, expanding insurance coverage before age 65 may reduce disparities in important health outcomes.

To realize potential benefits of coverage choices, Medicare beneficiaries must be aware of Medicare managed care. Paper 2 assessed relationships among beneficiary awareness of the managed care program and the number of Medicare risk plans, managed care penetration, and stability of plans in an area. Using 2002 Medicare Current Beneficiary Survey data, we analyzed 9,277,952 Medicare fee-for-service beneficiaries never enrolled in Medicare managed care but had at least one plan available in their area. Having more Medicare risk plans available was significantly associated with greater awareness, and having an intermediate number of plans (2-4) was associated with more accurate knowledge of Medicare risk plan availability.

Clinical quality of care for Medicare fee-for-service beneficiaries is not better in markets where Medicare beneficiaries receive more intense treatment. To learn if a similar relationship exists in Medicare managed care, Paper 3 examined associations between intensity of service use across hospital referral regions and 11 clinical quality of care measures from the 2003 Medicare managed care Health Care Effectiveness Data and Information Set. Similar to fee-for-service, Medicare managed care beneficiaries in high-intensity markets had a lower likelihood of receiving beta blockers after heart attack, breast cancer screenings, colorectal cancer screenings, and appropriate treatments for depression or osteoporosis after a hip fracture than their counterparts in low-intensity markets. An exception was managed care beneficiaries with hypertension in high-intensity markets were more likely to control their blood pressure than those in low- intensity markets.

The first chapter examines inconsistent reporting of sex and virginity pledge histories in a one year time-frame in the National Longitudinal Study of Adolescent Health (Add Health, n=13,568). More than half of wave 1 virginity pledgers denied at wave 2 having made a pledge; pledgers who initiated sexual activity were 3 times as likely to deny having made a pledge (odds ratio [OR]=3.21; 95% confidence interval [CI] = 2.04, 5.04). Among wave 1 nonvirgins who subsequently took virginity pledges, 28% retracted their sexual histories at wave 2; respondents who took virginity pledges were almost 4 times as likely as those who did not to retract reports of sexual experience (OR=3.88; 95% CI=1.87, 8.07).

The second chapter examines 72 risk behaviors in a reliability study of the Youth Risk Behavior Survey (n=4619) administered at an interval of 2 weeks. It finds that adolescents report their sex and substance use most consistently, and weight control efforts least consistently, possibly due to the different salience of these activities for adolescents.

The third chapter examines whether the sexual behavior and birth control use of virginity pledgers differs from that of non-pledgers. Prior studies use parametric methods which may not adequately adjust for pre-pledge differences between pledgers and non-pledgers. This paper uses exact and nearest-neighbor caliper matching to create a control group of non-pledgers (n=645) comparable to pledgers (n=289) matched on over 100 pre-pledge factors. Five years post-pledge, 84% of pledgers denied having ever pledged. Pledgers and matched non- pledgers did not differ in premarital sex, STDs, anal, and oral sex. Pledgers had 0.1 fewer past year partners but the same number of lifetime sexual partners and age of first sex. Fewer pledgers than matched non-pledgers used birth control and condoms in the past year, and birth control at last sex.

Paper 1 examines problems pertaining to the health and well-being of Army spouses during extensions of deployment, using 2004 survey data. Controlling for demographic and deployment characteristics, we found that spouses who experienced extensions fared worse on an array of measures, including mental well-being (e.g., feelings of depression) and household strains (e.g., problems with household and car maintenance). This suggests that extensions may exacerbate problems related to deployment.

Paper 2 uses national opinion surveys to evaluate proposed disparities in access to and use of comparative quality information. We found that racial/ethnic minorities, those with lower incomes, and those who lack health insurance are less likely to see comparative quality information. Those in rural communities, those with lower incomes and those without insurance are less likely to use the information. Unexpectedly, we also found that among those who have seen comparative quality information, racial/ethnic minorities, those with low income, and those without insurance are more likely to use comparative quality information about hospitals or doctors than their counterparts.

Paper 3 evaluates barriers to seeking and using comparative quality information in context by examining the experiences of women choosing an obstetrician in Massachusetts. Qualitative interviews revealed evidence of barriers identified in the literature: those related to the materials (accessibility); to features of health care decision-making (low salience and immediacy of medical concerns); and to consumer values (mistrust of institutional experts, disconnect with concept of quality, and competing priorities). We also identified a barrier not previously recognized: patients believe they can judge clinical quality without formal information. Patients who were confident in their information seeking and use skills were able to overcome many of these barriers.

The first paper explores the HIV risk associated with concurrency or overlapping sexual partnerships. It finds that concurrency is associated with increased risk for HIV infection among women, but not men, after controlling for key demographic characteristics, risk factors, and sexual health characteristics. The findings underscore the need for frank, concerted messages about fidelity and highlight the need to address contextual issues such as poverty, unemployment, and gender power inequities.

The second and third papers focus on the perceptions of risk that mediate high risk sexual behavior. Virtually all behavioral theories view perceived risk as a principal motivation for behavior change. The second paper explores which risk behaviors are associated with heightened perceived risk and finds that, among women, older sexual partners, multiple partners, and concurrency were not associated with perceived risk for HIV. These findings suggest gaps in current prevention messages and efforts.

The third paper explores whether perceptions of risk are associated with HIV infection. It finds that among young men, but not young women, perceptions of risk are associated with HIV infection after controlling for demographic characteristics, risk behaviors, and sexual health characteristics. The findings suggest that in the context of voluntary counseling and testing, asking young men about their perceived risk of infection could help clinicians and others identify those most in need of testing and counseling. The fact that perceptions of risk and HIV infection are not associated among young women suggests that opt-out HIV testing may be a more appropriate testing model for South Africa.

In my second paper, I evaluate the effect of user fees on the utilization of maternity services in Ghana. User fees reduce demand for health services but are an important source of revenue and provide incentives to health care providers in developing countries. In 2003, Ghana introduced a delivery fee exemption policy, initially rolling the policy out to 4 regions, creating a natural experiment to evaluate the effect of user fees on health service coverage. I find that this policy was effective at increasing the proportion of births supervised by trained medical professionals and delivered in facilities but that the policy may have had an adverse effect on the quality of services delivered. The delivery fee exemption policy was successful at targeting services to pregnant women and may represent an effective strategy where other forms of targeting are more difficult to implement.

In my third paper, I explore the determinants of recent Canadian Physician (CP) migration, by conducting a longitudinal analysis of recent migrants. I find that Canadian trained physicians were less likely to outward migrate and more likely to return home and that many of the departures seen during the 1990s were actually temporary. These findings suggest that, among other factors, additional education and training opportunities abroad may have accounted for a large share of outward migrations seen during the 1990s.

In the second chapter I consider a model of suppliers who are motivated by a desire to perform well in addition to profit. If profit maximization is the only objective of a firm, new information about quality should affect firm behavior only through its effects on market demand. In the alternate model, performance data can alter both pecuniary incentives (i.e. extrinsic motivation) and incentives unrelated to profit (i.e. intrinsic motivation). The introduction of quality “report cards” for cardiac surgery in Pennsylvania provides an empirical setting to test for an effect of new information on quality (mortality) and to isolate the relative role of extrinsic (demand side) and intrinsic (supply side) incentives in determining surgeon response to new information. Using a structural demand system, I estimate the profit incentives facing each surgeon from the introduction of report cards. Extrinsic incentives due to quality reporting led to a .09 percentage point (three percent) decline in mortality. Consistent with a mixed model of objectives, information on performance that was new to surgeons and unrelated to patient demand led to an intrinsic response three times as large as surgeon response to profit incentives.

The third chapter measures the response of consumers to different information sources when choosing health care providers. In the absence of public “report cards” consumers can rely on market based mechanisms such as provider reputation, private reporting (e.g. US News and World Report) or the advice of a referring physician agent or their insurance plan. Such market based learning may be a substitute or complement for publicly released quality information. I estimate a model of consumer demand for surgeon quality (mortality) that integrates unobserved insurance network constraints, agency, U.S. News and World Report rankings and proxies for hospital reputation. Prior to formal information release, market based learning explains almost all of the consumer response to surgeon quality. After public information release demand for high quality surgeons increases relatively. I find that market based learning either substitutes for or is unaffected by public reporting. In particular, a U.S. News and World report ranking reduces consumer response to surgeon quality. Hospital reputation and agency show little differential impact on demand for quality after public reporting.

The first paper examines the impacts of a national health insurance program and a franchise of midwife clinics on achievement of minimum standards for prenatal and delivery care in the Philippines using data from Demographic and Health Surveys. Scale-up of the insurance program was associated with increased odds of receiving 4 prenatal visits, including care in the first trimester of pregnancy. Exposure to midwife clinics did not affect prenatal outcomes. While both programs were associated with slight increases in the odds of delivery in a health facility, these were not statistically significant. Expansion of an insurance program was associated with increases in achievement of minimal standard prenatal care among Filipina women.

The second and third papers use claims data from New Jersey’s Medicaid program and focus on maternal mental health among low-income women. The second paper characterizes the association between diabetes and depression during pregnancy and the postpartum period. Women with diabetes compared with those without diabetes had nearly double the odds of experiencing depression during the perinatal period. A similar relationship was shown among women with no indication of depression during pregnancy. Pre-pregnancy or gestational diabetes was independently associated with perinatal depression, including new onset of postpartum depression.

The final paper investigates the effects of New Jersey’s statewide postpartum depression initiative on mental health care following delivery. Interrupted time series and patient-level longitudinal models are used to estimate policy impacts on initiation, follow-up, and receipt of guideline-consistent depression care. Fewer than 7% of mothers initiated treatment in the 6 months following delivery. Of those who initiated care, 60% received some follow-up, and less than half received guideline-consistent care. There were no policy-associated changes in the level or trends for treatment initiation or follow-up, but there was an increasing trend in receipt of guideline-consistent care following implementation. Postpartum depression is under-recognized and under-treated among low-income women.

Chapter 2 relies on data collected as part of the same Karnataka study and seeks to understand causal relationships underlying comorbidity in physical and mental health. I find that exposure to this acute negative health event increases psychological distress by 1.5 standard deviations on the distress scale one year later. Physical Disability acts as a key mediating mechanism, accounting for 65% of the total observed effect. Indebtedness resulting from health shock did not mediate the observed association between the exposure and distress levels. Physical health events, especially acute conditions that are associated with increased disability levels, have large and significant causal effects on mental health over one year later.

In Chapter 3, I examine the consequences and predictors of onset of chronic conditions in the United States using data from 8 waves of the Health and Retirement Study (HRS). I find that onset of major chronic conditions increases out of pocket expenses by about $2500 in the subsequent wave. Over the 14 year period spanning the 8 waves, the cumulative financial losses associated with a major onset in the second wave is over $68,000. I also find evidence of large reductions in the probability of working following major onsets. Further, I also find that changes in stock wealth as well as baseline asset levels are negatively associated with probability of major onsets of illness, suggesting that baseline SES levels can influence future health outcomes.

Paper 1 assesses trends in White-Black and White-Latino disparities in the use of any mental health care and minimally adequate mental health care. Disparities in the use of any mental health care increased over time, particularly between Whites and Blacks in the general medical (GM) sector and between Whites and Latinos in the specialty mental health sector (SMH) sector. Disparities in the use of minimally adequate mental health care persisted between Whites and Blacks over time, but were not detected between Whites and Latinos in either time period.

Paper 2 examines the role socioeconomic status (SES) plays in changes in White-Black and White-Latino disparities in mental health care. The Institute of Medicine definition of racial/ethnic disparities recognizes the contributing effect of SES to racial/ethnic disparities in health care. Using two different sets of analyses, one where we control for SES and one where we do not, we found that controlling for SES did not influence racial/ethnic disparity estimates at both time periods, nor did it influence how these disparities changed over time.

Paper 3 examines whether racial/ethnic differences in perceived need and perceived barriers to mental health care help explain White-Black and White-Latino disparities in the use of mental health care. In Time 1, Blacks and Latinos were similar to Whites in perceiving a need for mental health care. Among those with perceived need at Time 1, no disparities in the use of care existed. In Time 2, Blacks and Latinos were once again similar to Whites in perceiving need. However at Time 2, racial/ethnic disparities in the use of mental health care emerged among those with perceived need. Among those with perceived need who did not receive care, Blacks and Latinos were more similar to Whites than they were dissimilar in endorsing a perceived barrier to care, with only a few exceptions in either time period.

Following an introductory chapter, in chapter two I examine the emerging perspective that the optimal design of health care organizations depends on the specific characteristics of patient conditions. I do so specifically by examining whether the efficiency benefits of volume (scale) and operational focus (scope of services) depend on the degree of patient comorbidity. Using data from a sample of U.S. hospitals, I find evidence that the average efficiency benefits of volume and focus are diminishing in the level of patient comorbidity.

In chapter three I examine the relationship between operational focus and performance in cardiovascular care, taking into consideration the frequent comorbidities cardiovascular patients have. Drawing on corporate research on related diversification and distinguishing between direct and complementary spillovers, I examine the extent to which the marginal quality benefits of focus in cardiovascular care depend on the degree to which a hospital “co-specializes” in related—based on the frequency of comorbidity—areas (complementary spillovers). The results provide evidence of such complementarities in hospital specialization.

In chapter four I draw on complex systems theory to examine overall patterns of comorbidity in light of the complementarities found in chapter three. Specifically, I conceptualize comorbidity as a network of disease in which network ties are determined by the frequency of co-occurrence. I explore this system using the tools of network analysis and draw two conclusions. First, comorbidities among hospital patients exhibit clear patterns, which may be characterized with a “core/periphery” structure. Second, clustering patterns in the network suggest that even for conditions where comorbidities suggest a broader scope (such as cardiovascular disease), the optimum likely stops short of “everything for everybody” organizations.

Chapter five concludes the dissertation with some thoughts about the implications of the research presented for existing health care organizations, and possible future areas of inquiry.

Paper 1 uses national survey data to assess the relative influences of people’s political ideology and economic self-interest on public opinion about the 2007 reauthorization of the State Children’s Health Insurance Program (SCHIP) and universal health insurance. Data are from national surveys conducted during the 2007 SCHIP reauthorization debate and immediately after the 2008 presidential election. I compare the characteristics of public opinion regarding these two issues and find that ideological orientation and partisan identification had strong influences on Americans’ views about both policies, a common finding in studies of Americans’ views about policy issues. However, self-interest also had a considerable effect on older Americans’ views about these health insurance expansions, and the influence of self-interest among seniors appears to have had implications in the ongoing national debate about health reform as Medicare spending reductions have become a central topic of public discussion.

Paper 2 examines racial disparities in health-related difficulties during major hurricane threats in the Southern United States. I identify several specific problems that African Americans are more likely to face in these situations, and I assess which individual-level factors chiefly contribute to these disparities. The data for this investigation are from a 2007 survey of all coastal counties from North Carolina to Texas, including the areas affected by Hurricanes Katrina and Rita. The results demonstrate that evacuation behavior, receipt of safety information, chronic illness and disability, housing, and socioeconomic status all contributed to racial disparities. These findings should inform future emergency management policies and planning in order to mitigate these health threats among African American communities.

Paper 3 reexamines the effect of self-interest on public opinion about health policy issues and investigates the possibility that meager self-interest effects found in previous research may have resulted from the manner in which these studies have defined self-interest. The analysis also explores the potential moderating effects of the unique political and economic conditions of 2008 on the influence of self-interest. Data for this paper are from a survey of residents of Florida and Ohio, two swing states in which many residents felt the earlier effects of the most recent economic recession. The results indicate that multiple measures of self-interest, including a previously unused measure regarding difficulties paying medical bills, had statistically significant and sizeable effects on attitudes about universal health insurance. The new measure of self-interest and the economic and political environment of 2008 may explain these relatively unique findings.

Chapter 1 explores role-based sources of work process failure arising from interdependencies among health care professionals in an organizational context featuring an electronic medical record (EMR). Interviews with providers and patients revealed professional work routines containing relational vulnerabilities to coordination failure that persisted despite—and, indeed, were potentially exacerbated by—introduction of a technology enabling high informational continuity. These findings demonstrate limitations of information processing as the dominant paradigm through which to improve coordination in complex settings and urge managers and policy-makers to complement information technology with interventions that directly address behavior within institutionalized work roles.

Chapters 2 and 3 employ a mixed methodological approach to examine organizational responses to work process failure at a high-performing teaching hospital. In Chapter 2, I explore predictors of managerial attention to identified failure within the hospital’s adverse event review system. Internalization of external regulatory priorities during review activities shifted managerial attention toward adverse events yielding less new knowledge, such as those caused by involuntary slips during practiced skills, and away from events richer in learning value, such as those involving interdisciplinary process failure. These findings reveal unintended effects of harm-driven event reporting regulation on internal learning processes and suggest the need for separate learning and compliance mechanisms in health care and other high-risk settings.

Chapter 3 examines factors associated with corrective organizational action in response to work process failure. Corrective action following an adverse event was influenced by the dual institutional forces of regulation and professionalism, manifesting in tendencies to intervene on tangible risk factors that did not interfere with autonomous practice. Findings indicate that organizations may respond to failure with varying aggressiveness and success depending on the conditions of error, demonstrating the importance of goal-setting and resource appropriation that are aligned with the challenges confronting change in complex institutionalized environments.

Chapter I present results our analysis of syphilis screening in pregnancy and the potential impact of new screening technologies in resource poor settings. In developing countries, syphilis infection in pregnancy poses significant deleterious consequences to maternal and infant health. While treatment with penicillin is highly effective and relatively inexpensive, identification and timely treatment of infected women often falls short due to logistical and technical obstacles. New rapid testing technologies have become available with the potential to improve diagnosis and treatment of infection during pregnancy. We developed a Markov model of the natural history of syphilis and pregnancy in women in the sub-Saharan Africa setting in order to assess the benefits, cost-effectiveness and policy implications of universal on-site prenatal syphilis screening and treatment using rapid diagnostic tests versus standard testing methods. Our analysis showed that syphilis screening with new rapid tests that allow for same day testing and treatment was highly cost-effective and had the potential to provide substantial benefits by preventing syphilis transmission and adverse pregnancy outcomes.

Chapter II describes our calibration of a 1st order Monte Carlo simulation model of the natural history and treatment of HIV infection, which incorporates components of both infectious and chronic disease modeling. Previously, model natural history parameter estimates were primarily derived from data from the Multicenter AIDS Cohort Study (MACS)—a longitudinal study of HIV/AIDS in gay and bisexual men. Recently, the model was adapted to address questions relevant to U.S. HIV-infected women using data from the Women’s Interagency HIV Study (WIHS). Motivated by parameter changes and assumptions related to use of the WIHS dataset, we assessed the internal consistency of the newly parameterized natural history model by exploring the ability to achieve good visual fits to the empiric survival data for untreated HIV-infected women. We evaluated model performance and characterized uncertainties associated with model assumptions using calibration techniques in addition to exploring model external validity in estimating treatment survival using a unique subset of WIHS empiric survival data for women who received ART.

Chapter III examines the cost-effectiveness and trade-offs between methods for treating panic disorder. Panic disorder is often debilitating, resulting in numerous lost workdays and reduced overall productivity. We developed a 1st order Monte Carlo simulation model of the natural history of panic disorder and treatment to assess the cost effectiveness and the impact of different treatment strategies on burden of disease; this analysis represents a significant extension of modeling research in the area of mental health. Our analysis supports conclusions that treatment of panic disorder is cost-effective even at low cost-effectiveness thresholds and can substantially reduce time spent with disease. However, intervention characteristics, patient heterogeneities and disease course have important implications on treatment choice.

In chapter two, I investigate the effect of the uninsurance rate in a market on access to care for the uninsured and the insured, using a market fixed effects approach. I also consider the role of time-varying market-level factors and instrument for the uninsurance rate using a two-stage residual inclusion approach. I find that the causal effect of the rate of uninsurance in the community on access to care for the uninsured is negative. In contrast, estimates for the insured are small and insignificant. Unlike earlier research, these results imply that there are no spillover effects of uninsurance on access to care for the insured. Thus, policies that expand health insurance coverage may affect access to care for the uninsured but have little effect on access for those who are already insured.

The types of institutions that comprise the health care safety net vary in their structure and function, and there is geographic and temporal variation in the availability of and support for different types of safety net providers. In chapter three, I consider the association between particular types of safety net institutions (specifically, teaching hospitals and federally qualified health centers) and access to and use of care for the uninsured at the market level in both 2003 and in 1996. I find that the presence of teaching hospitals and the presence of health centers are correlated with different aspects of access and use among the uninsured.