CHAS 2007-2008 Seed-Grant Awards

G. Caleb Alexander, MD, MS

Exploring the impact of the Medicare Prescription Drug Benefit (Part D)

The Medicare Modernization Act Prescription Drug Benefit (Part D) established a prescription drug benefit for all 43 million Medicare beneficiaries. This represents the single largest change to Medicare since the program’s inception decades ago. Although subject to tremendous scrutiny, early evaluation of Part D has been scarce. For example, thus far, analyses have not examined the simultaneous impact of Part D on drug utilization and expenditures, explored the combined experiences of enrollees and nonenrollees alike, or examined the differential impact of the drug benefit for key subpopulations such as those dually eligible for Medicare and Medicaid. We propose a focused research agenda that builds upon a promising academic-corporate partnership between the University of Chicago and the largest retail pharmacy chain in the country. Specific aims are: (1) to compare the demographics, heath, past prescription drug usage, and out-of pocket costs between those who enrolled in Part D and those eligible subjects who did not, (2) to estimate the effect of Part D on aggregate prescription drug use and out-ofpocket costs for Medicare beneficiaries, regardless of whether they enrolled in Part D, (3) to estimate the effect of the Part D on changes in prescription drug use across therapeutic classes for Medicare beneficiaries, regardless of whether they enrolled in Part D, and (4) to estimate the association between generosity of pharmaceutical coverage, benefit exhaustion, and non-adherence among a sample of beneficiaries. These analyses, if conducted expeditiously using rigorous pharmaceconomic and pharmacoepidemiological methods, will have important implications for patients, physicians, researchers, and policy makers while creating a new knowledge base and a broader understanding of health care and health policy for disadvantaged populations such as those dually eligible for Medicare and Medicaid.

Yoosun Choi, PhD

The Korean American Families (KAF) Project

Immigrants generally report better health and mental health outcomes than nonimmigrants but those better outcomes deteriorate as they stay longer in this country (i.e. as they are acculturated and/or in later generations of immigrants). It is critical to understand what aspects of their adaptation are linked to the deterioration. Studies (e.g., Choi & Lahey, 2006) show that the deterioration may be the greatest among Asian Americans as an aggregated group. There was also an indication that youth outcomes of Korean American in particular may get worse faster than other subgroups (Choi, revise and resubmit). The Korean American Families (KAF) Project is to study Korean American families with early adolescents with the following Specific Aims to inform policy and culturally relevant interventions to prevent the deterioration and occurrence of negative health, mental health and negative behavioral outcomes. The KAF is to survey about 300 families (850 individuals: youth and two parents/primary caretakers) to:

    * Investigate the rates and patterns of youth health, mental health and other behavioral outcomes among Korean American youth. Youth outcomes include academic achievement, school problems, aggressive and nonaggressive delinquent offenses, substance uses (initiation and current use of alcohol, tobacco and other drugs) and depression.
    * Examine the relationships between correlates and causes and youth outcomes, specifically:
    * To what extent are risk factors (e.g. poor family management, differential levels of acculturation, low income status and experience of racial discrimination) predictive of adolescent positive and problem outcomes?
    * To what extent do protective factors (e.g., family bonding, strong ethnic identity, and culture-specific or bicultural parenting) enhance resilience among Korean American youth?
    * To what extent are the effects of these risk and protective factors consistent across gender and age?

Sarah K. Hoehn, MD Mbe

DNAR Focus Groups to Improve Communication with Minorities

The broad, long-term objective of this project is to understand the viewpoint of an underserved population of parents on Do Not Attempt Resuscitation (DNAR) orders in different situations. This work will illuminate the parental viewpoint of DNAR in different environments, which is essential to providing optimal care at the bedside. The University of Chicago Hospital system is the perfect place for this type of study, as the perspective of racial and ethnic minorities on end of life decision making is underrepresented in the literature. The first specific aim is to understand the influence of different factors on DNAR decisions, such as chronic illness, technology dependence and location. The second specific aim is to understand parental factors, such as race, religion and religiosity that impact DNAR decisions. This project proposes to conduct focus groups with parents of children who are both inpatients and outpatients. The focus groups will discuss the perception of DNAR in different situations. The three hypothetical vignettes will present a child whose family wishes to make him DNAR. The first vignette is a critically ill child in the PICU, the second is a technology dependent child who lives in a chronic care facility, and the third vignette is a child with chronic illness who would like his school to honor DNAR. Dr. Hoehn has a unique experience to conduct this project, given her background in both pediatric intensive care and palliative care. She has the qualitative research skills to conduct this project. This project is relevant to CHAS in promoting health policy research in disadvantaged populations by describing and sharing the viewpoint of parents who are racial and ethnic minorities on DNAR in different situations..

David Meltzer, MD, PhD

Using Multiple Imputation to Correct for Rater Unreliability in a Diagnosis of Post-Traumatic Stress Disorder

Advances in community-based research are highly dependent on the quality of the data that are collected. Given the challenges inherent in community-based research, such as cultural/linguistic differences among participants and clinicians, widely variant preferences for treatment, cost, losses to follow-up, and the need for short assessments, methods are needed that can handle data that are incomplete or subject to measurement error. A motivating example is provided by the WECare study, a large longitudinal treatment study of low-income mostly minority women with depression in the Washington, DC metropolitan area. Participants were randomly assigned to a treatment group and then assigned a clinician based on their treatment assignment. Therefore, clinician assignment itself was randomized. At baseline and prior to treatment, the participants in the study were assessed for PTSD during a clinical interview. The rates of PTSD assessed by trained psychotherapists (Ph.Ds in Clinical Psychology) were almost half that of psychiatric nurses (34.1% versus 60.6% respectively) due to the fact that the psychotherapists used more stringent criteria when diagnosing PTSD.

The specific aim of this proposal is to develop a method for correcting for rater unreliability in psychiatric diagnoses using multiple imputation. Using the WECare data, we propose to correct for the imbalance by clinician type by treating some of their diagnoses as censored and then estimating the latent probability of having PTSD using other covariates collected in the WECare study. Since the psychotherapists used a higher threshold when diagnosing PTSD, if they gave a participant a positive diagnosis, then that patient most certainly had PTSD. Conversely, since the nurses used much less stringent criteria than the psychotherapists, if they determined that a patient did not have PTSD then the patient most likely did not show any symptoms of PTSD. Positive diagnoses by nurses and negative diagnoses by psychotherapists are censored in the sense that a positive diagnosis by a nurse might not have been a positive diagnosis by a psychotherapist and a negative diagnosis by psychotherapists might have been a positive diagnosis by a nurse. Censored observations are imputed by drawing imputations from a Bayesian ordinal latent regression model. We will also redo a previous analysis of the WECare data that looked at the effects of PTSD on depression without adequately correcting for the imbalance in PTSD rates between the two types of clinicians.

Monica Peek, MD, MPH

Enhancing Diversity in Academic Medicine: An exploratory study

This project’s goal is to look at issues of diversity within academic medicine at both the local and national level, in order to gain insight that has relevance to the University of Chicago and the communities it serves, and to learn about strategies to enhance diversity that are applicable to institutions throughout the country. There is evidence that a racially/ethnically diverse medical faculty can ultimately improve the quality of care that is the delivered to racial/ethnic minorities and reduce the existing gap in health outcomes within such populations. For example, racial/ethnic minorities are more likely to be uninsured and live in a medically underserved area; conversely, minority physicians are more likely to care for such vulnerable patient populations. There is also strong evidence that a racially/ethnically diversity academic faculty can reduce health disparities.

Towards this end, this research seeks to explore mechanisms that can enhance the representation of under-represented minorities within academic medicine. The specific aims of the study are two-fold: 1) To characterize the diversity climate within the Department of Medicine at the University of Chicago, and explore factors associated with the diversity climate, 2) To identify factors that may influence the success of efforts to enhance physician racial/ethnic diversity within departments of medicine. The proposed research consists of 2 separate studies: 1) a web-based diversity climate survey distributed among all physicians (faculty and non-faculty) within the Department of Medicine at the University of Chicago, and 2) a series of in-depth interviews with key informants (chairs of the department of medicine, chiefs of sections of general internal medicine, and diversity leadership) at academic medical centers across the U.S. whose levels of racial/ethnic diversity are in the highest quartile or lowest quartile nationally. These interviews will probe priorities, successes and lessons learned about enhancing academic diversity. Specific analyses will be conducted to identify distinctions in themes between upper and lower-quartile departments, and “diversity best practice” scores will be compared between these groups using a 2-sample t-test.

Lisa Shah, MD

Consistency of Preparedness of Hospitalized Patients to Quit Smoking and its Relationship to Outcomes

Smoking cessation is a major cause of cardiovascular morbidity and mortality in the United States. Given the large number patients suffering from cardiovascular disease, hospitalization becomes a natural opportunity to advise patients on the perils of smoking. One reason for this is that patients see smoking as a cause of the disease that has placed them in the hospital, thus influencing their desire to quit. In addition, the “smoke free” nature of hospitals provides a supportive environment within which to begin a cessation attempt. Furthermore, promoting smoking cessation in the hospital has been adopted as a measure of quality of hospital care for cardiac patients. While the preparedness of hospitalized smokers is poorly understood, it remains integral to the effectiveness of smoking cessation interventions (i.e. a program offered to a smoker who is not prepared to quit will be costly, ineffective, and likely result in poor compliance). Our preliminary work suggests that over three quarters of hospitalized smokers state that they prepared to quit. However, this number may be artificially high due to a socially desirable response bias (i.e. appearing health conscious while hospitalized) or because acute illness prompts patients to reevaluate their behavior. It is unknown whether or not a smoker’s reported level of preparedness to quit in the hospital is representative of their level of preparedness when they leave the hospital. This could be important for the success of smoking cessation interventions, which require long-term participation after discharge. Therefore, it is important to understand whether reports of preparedness during hospitalization accurately reflect patients’ preparedness to quit after they return to their home environment. This project aims to assess the consistency of patient preparedness to quit smoking during hospitalization and after discharge and its relationship to outcomes. By identifying a population with a consistent level of preparedness to quit, intensive interventions can be better targeted.

Dana Suskind , MD

Project ASPIRE*: Closing the Gap on Pediatric Cochlear Implant Habilitation" (* Achieving Superior Parental Involvement for Rehabilitative Excellence)

Cochlear implantation has dramatically improved the quality of life of deaf children and resulted in significant societal cost savings[1] . Nonetheless, introducing the deaf child to the world of sound is only the first step in a much longer process. Ultimately, it is the carefully structured post-operative rehabilitation which provides the skill-set necessary to mainstream the child into society. Success is invariably the domain of children whose backgrounds and parental involvement provide the necessary tools for rehabilitation[2]. Urban children of lower socioeconomic status (SES) have demonstrated significantly poorer outcomes after implantation than children of higher SES [3, 4] . Anecdotal evidence suggests that this disparity is related to two essential factors: a lack of access to critical rehabilitative services and parents who lack the tools to navigate the complex medical system. Although habilitation is theoretically provided in public schools, the schools these children are likely to attend are often unprepared for taking on yet another task. As a result, while their more-affluent peers attend public or private schools with sophisticated programs and are being supplemented with privately-funded, after-school habilitation with well-trained professionals, poorer, inner-city youngsters have neither the opportunity nor the benefit of such programs. The consequence is a disparity in success with negative long-term consequences, both to the individual and society at large. The question which remains is how to solve this problem in a way that assures success while being sensitive to real economic constraints.

We hypothesize that a web-based, professionally-supervised, parentally driven (PSPD) rehabilitation-enrichment program for lower-SES children would provide an efficient and cost-effective intervention that will result in post-operative rehabilitation comparable to the results in more affluent children. The primary aim of PSPD is the mainstreaming of post-cochlear implant patients into society. To this end, in addition to a hearing/speaking rehabilitation program, we will add a developmental-enrichment program. In addition, by utilizing parents as collaborators in the process, our program will help lower-SES parents establish increased health literacy and empowerment to positively affect the lives of their children.

Mishka Terplan, MD, MPH

Patterns of “Compassionate Coercion”: Factors Associated with Court Mandated Drug and Alcohol Treatment in Pregnancy

Pregnant women occupy a unique place in drug treatment. In the late 1980’s, the war on drugs and the growing debate on abortion collided in the image of the “crack baby.” Since then, there has been a gradual shift in the cultural discourse on drug policy away from harm-reduction and voluntary treatment and towards a greater criminal justice involvement. Because drug and alcohol dependence is relatively common in pregnancy, many women are potentially subject to what the government has called “compassionate coercion” -- court mandated treatment referral. In spite of the increasing presence of the criminal justice system in drug and alcohol treatment, there is little evidence to support its efficacy over that of voluntary treatment. In fact, outcomes from coerced treatment may be worse. Unfortunately, coercive treatment has not been well studied in pregnancy. We do know that treatment in pregnancy has only limited success when compared with the general population. However, before we can evaluate the efficacy of coercion versus voluntary treatment in pregnant women, we need to describe the landscape of coercive treatment referrals in pregnancy. Thus we propose a secondary data analysis of the Treatment Episode Data Set (TEDS), a national data set of treatment admissions from 1994 through 2005, to answer the following two questions: 1) Among pregnant women in drug treatment, how do those with criminal justice referrals differ from those with other referral sources? 2) Over the last decade, what are the trends in criminal justice referral? The characteristics of the two groups (women with criminal justice referrals versus other referral sources) will be described. Logistic regression will be used for analysis. Effect measure modification will be assessed by the Breslow-Day test of homogeneity of the Odds Ratio (OR), and confounding will be evaluated via both a backwards elimination strategy and the use of a Directed Acyclic Graphs. Results will be reported in OR with 95% confidence intervals. The ORs will also be plotted by year to illustrate change in time. Pregnant women with drug and alcohol problems are often stigmatized and consequentially marginalized because of their dependency. Treatment in pregnancy is critical but interventions should be based on evidence not ideology. In order to evaluate the efficacy of coerced treatment in this vulnerable population, the factors associated with court mandated treatment need to be illuminated first.