Explore Dan Adelman's Healthcare Analytics Laboratory Research
Applications for Winter 2025 open October 30th.
Explore Dan Adelman's Healthcare Analytics Laboratory Research
The Tolan Center for Healthcare Research
Our research explores many aspects of the healthcare system—clinical quality and outcomes, equity and access, innovation, operations, and finance.
The aim of our healthcare research is to inform policymakers, industry leaders, students, and the public.
Our Faculty Research in Healthcare
Our faculty are recognized for their significant contributions to scholarship on the business of healthcare. See recent publications below.
IMPORTANCE Reducing Medicare expenditures is a key objective of Medicare’s transition to value-based reimbursement models. Improving access to primary care is an important way to reduce expenditures, yet less is known about how visits should be organized to maximize savings.
OBJECTIVE To examine the association between Medicare savings and primary care visit patterns.
DESIGN, SETTING, AND PARTICIPANTS This retrospective cohort study used data from a 5% sample of traditional Medicare claims from 2016 to 2019. Participants had at least 3 primary care visits with at least 180 days between the first and the last visit, were not enrolled in Medicare Advantage, did not have end-stage kidney disease, and were not institutionalized. Data were analyzed from June 2022 to April 2023.
EXPOSURES Primary care visit patterns: visit frequency, regularity, continuity of care.
MAIN OUTCOMES AND MEASURES Savings in Medicare expenditures; risk-adjusted Medicare expenditures, number of emergency department (ED) visits, and hospitalizations.
RESULTS Among 504,471 beneficiaries (298 422 [59.16%] women; mean [SD] age, 74.26 [10.41] years), temporally regular visits with higher continuity were associated with the highest savings. For these patients, the savings increased with increasing visit frequencies, with peak savings observed at higher visit frequencies as clinical complexity increased. As regularity and continuity decreased, the association between savings and visit frequencies progressively inverted. The group with a regular and highly continuous pattern was associated with greater savings (175.87%; 95%CI, 167.40% to 184.33%; P < .001), lower risk-adjusted expenditures (−16.61%; 95%CI, –16.73% to –16.48%; P < .001), fewer risk-adjusted ED visits (−40.49%; 95%CI, –40.55%to −40.43%; P < .001), and fewer risk-adjusted hospitalizations (−53.32%; 95%CI, –53.49%to –53.14%; P < .001) compared with the irregular noncontinuous group.
CONCLUSIONS AND RELEVANCE In this cohort study, savings in Medicare expenditures and improvements in acute care utilization were associated with visit frequency, regularity, and continuity in primary care in an interrelated fashion such that optimization of primary care visit patterns along each axis were associated with the largest improvement in outcomes. Demonstrating the magnitude and interdependence of these associations is useful for health care professionals
We study whether and how peer referrals increase screening, testing, and identification of patients with tuberculosis, an infectious disease responsible for over one million deaths annually. In an experiment with 3,176 patients at 122 tuberculosis treatment centers in India, we find that small financial incentives raise the probability that existing patients refer prospective patients for screening and testing, resulting in cost-effective identification of new cases. Incentivized referrals operate through two mechanisms: peers have private information about individuals in their social networks to target for outreach, and they are more effective than health workers in inducing these individuals to get tested.
By January 2022, the COVAX international vaccine collaboration had allocated over a billion vaccines to over 140 countries. We describe and review the allocation process chosen, which reflected both an objective of equitably distributing vaccines across the world and the need to fund that mission. We show how vaccine supply limitations and constraints on some countries' absorptive capacity have affected overall allocative outcomes. We also discuss market design approaches that were considered but not implemented, including the use of an exchange mechanism to better match countries' vaccine allocations to their preferences, as well as a vaccine brokerage under which countries could sell excess vaccines to countries with ongoing need. Our analysis addresses some criticisms of COVAX, and offers suggestions for agencies organizing global vaccine cooperation for future pandemics.
Operating room (OR) teamwork quality is associated with familiarity among team members and their individual specialization. We describe novel measures of OR team familiarity and specialty experience. Surgeon-scrub (SS) and surgeon-circulator (SC) teaming scores, defined as the pair’s proportion of interactions relative to the surgeon’s total cases in the preceding 6 months were calculated between 2017 and 2021 at an academic medical center. Nurse service-line (SL) experience scores were defined as the proportion of a nurse’s cases performed within the given specialty. SS, SC, and nurse-SL scores were analyzed by specialty, case urgency, robotic approach, and surgeon academic rank. Two-sample Kolmogorov-Smirnov tests were used to determine heterogeneity between distributions. A total of 37,364 operations involving 150 attending surgeons and 222 nurses were analyzed. Median SS and SC scores were 0.08 (interquartile range: 0.03–0.19) and 0.06 (interquartile range: 0.03–0.13), respectively. Higher margin SLs, senior faculty rank, elective, and robotic cases were associated with greater SS, SC, and nurse-SL scores (P<0.001). These novel measures of teaming and specialization illustrate the low levels of OR team familiarity and objectively highlight differences that necessitate a deliberate evaluation of current OR scheduling practices.
Having an interpretable dynamic length-of-stay (LOS) model can help hospital administrators and clinicians make better decisions and improve the quality of care. The widespread implementation of electronic medical record (EMR) systems has enabled hospitals to collect massive amounts of health data. However, how to integrate this deluge of data into healthcare operations remains unclear. We propose a framework grounded in established clinical knowledge to model patients’ lengths-of-stay. In particular, we impose expert knowledge when grouping raw clinical data into medically meaningful variables, which summarize patients’ health trajectories. We use dynamic predictive models to output patients’ remaining lengths-of-stay (RLOS), future discharges, and census probability distributions based on their health trajectories up to the current stay. Evaluated with large-scale EMR data, the dynamic model significantly improves predictive power over the performance of any model in previous literature and remains medically interpretable.
It is taken as given by many policy makers that Direct-to-Consumer Advertising of prescription drugs drives inappropriate patients to treatment. Alternatively, advertising may provide useful information that causes appropriate patients to seek treatment. I study this dynamic in the context of antidepressants. Leveraging variation driven by the borders of television markets, I find that a 10 percent increase in antidepressant advertising leads to a 0.3 percent ($32 million) increase in new prescriptions followed by reductions in workplace absenteeism worth about $770 million. I find no effect of advertising on prices, generic penetration, drug switches, adverse effects, non-adherence rates, or therapist visits.
The Centers for Medicare and Medicaid Services (CMS) star rating methodology for publicly evaluating hospitals uses a latent variable model that is based on the pre- sumption of a single, but unobservable, hospital-specific quality factor shared across a group of performance measures. Performance measures are given higher weight if they statistically appear to be more strongly correlated with this hidden factor. We show how this approach, when applied to measures that are weakly or not correlated with each other, can effectively ignore measures and can exhibit “knife-edge” instability, so that even if hospitals improve relative to all other hospitals, they may nonetheless score lower overall because of weight shifting onto different measures than before. In contrast, we provide an approach to scoring and ranking hospitals that, under reasonable conditions, ensures that hospitals that improve relative to all other hospitals obtain higher scores, while also having the capability to autonomously adjust weights as measures are added or subtracted over time. Rather than exploit statistical correlation, we propose a conic optimization framework that offers a new integrated approach in data envelopment analysis for simultaneous efficiency analysis and performance evaluation. We develop theory that explains the behaviour of our approach, in- cluding various properties satisfied by hospital scores at optimality. Using data, we apply our approach to score and rank nearly every hospital in the United States and demonstrate the extent to which it agrees or disagrees with the existing approach to the CMS star ratings.
The Centers for Medicare and Medicaid Services (CMS) star rating methodology for publicly evaluating hospitals uses a latent variable model that is based on the pre- sumption of a single, but unobservable, hospital-specific quality factor shared across a group of performance measures. Performance measures are given higher weight if they statistically appear to be more strongly correlated with this hidden factor. We show how this approach, when applied to measures that are weakly or not correlated with each other, can effectively ignore measures and can exhibit “knife-edge” instability, so that even if hospitals improve relative to all other hospitals, they may nonetheless score lower overall because of weight shifting onto different measures than before. In contrast, we provide an approach to scoring and ranking hospitals that, under reasonable conditions, ensures that hospitals that improve relative to all other hospitals obtain higher scores, while also having the capability to autonomously adjust weights as measures are added or subtracted over time. Rather than exploit statistical correlation, we propose a conic optimization framework that offers a new integrated approach in data envelopment analysis for simultaneous efficiency analysis and performance evaluation. We develop theory that explains the behaviour of our approach, in- cluding various properties satisfied by hospital scores at optimality. Using data, we apply our approach to score and rank nearly every hospital in the United States and demonstrate the extent to which it agrees or disagrees with the existing approach to the CMS star ratings.
The effects of television advertising in the market for health insurance are of distinct interest to both firms and regulators. Regulators are concerned about firms potentially using ads to “cream skim,” or attract an advantageous risk pool, as well as the potential for firms to use misinformation to take advantage of the elderly. Firms are interested in using advertising to acquire potentially highly profitable seniors. Meanwhile, health insurance is a useful setting to study the mechanisms through which advertising could work. Using the discontinuity in advertising exposure created by the borders of television markets, this study estimates the effects of advertising on consumer choice in health insurance. Television advertising has a small effect on brand enrollments, making advertising a relatively expensive means of acquiring customers. Heterogeneous effects point to advertising being more effective in less healthy counties, which runs opposite to the concern of cream skimming. Leveraging the unilateral cessation of advertising by United-Healthcare, evidence is provided that the small advertising effect is not explained by a prisoner’s dilemma equilibrium. An analysis of longer-run effects of advertising shows that advertising effects are short lived, further decreasing the potential of advertising to create long-run value to the firm.
We estimate the benefit of life-extending medical treatments to life insurance companies. Our main insight is that life insurance companies have a direct benefit from such treatments because they lower the insurer’s liabilities by pushing the death benefit further into the future and raising future premium income. We apply this insight to immunotherapy, treatments associated with durable gains in survival rates for a growing number of cancer patients. We estimate that the life insurance sector’s aggregate benefit from FDA-approved immunotherapies is $9.8 billion a year. Such life-extending treatments are often prohibitively expensive for patients and governments alike. Exploiting this value creation, we explore various ways life insurers could improve stress-free access to treatment. We discuss potential barriers to integration and the long-run implications for the industrial organization of life and health insurance markets, as well as the broader implications for medical innovation and long-term care insurance markets.
Craig Garthwaite, John Graves, Tal Gross, Zeynal Karaca, Victoria Marone, and Matthew J. Notowidigdo study the effect of the Affordable Care Act Medicaid expansion on hospital services, with a focus on the geographic variations of its impact, finding that it increased Medicaid visits, decreased uninsured visits, and lead the uninsured to consume more hospital services overall— primarily through outpatient visits to the ED for deferrable conditions. Notably, the authors found significant heterogeneity across Medicaid-expansion states with some experiencing large changes in utilization and others seeing little change.
The U.S. health care system uses commercial algorithms to guide health decisions. Obermeyer et al. find evidence of racial bias in one widely used algorithm, such that Black patients assigned the same level of risk by the algorithm are sicker than White patients (see the Perspective by Benjamin). The authors estimated that this racial bias reduces the number of Black patients identified for extra care by more than half. Bias occurs because the algorithm uses health costs as a proxy for health needs. Less money is spent on Black patients who have the same level of need, and the algorithm thus falsely concludes that Black patients are healthier than equally sick White patients. Reformulating the algorithm so that it no longer uses costs as a proxy for needs eliminates the racial bias in predicting who needs extra care.
During the push to pass the Affordable Care Act, President Barack Obama often described the “crushing cost of health care” that was causing millions of Americans to “live every day just one accident or illness away from bankruptcy” and repeatedly stated that the high cost of health care “causes a bankruptcy in America every 30 seconds.” Stories of illnesses and injuries with financial consequences so severe that they caused households to file for bankruptcy were used as a major argument in support of the 2010 Affordable Care Act. And in 2014, Senators Elizabeth Warren (D-MA) and Sheldon Whitehouse (D-RI) cited medical bills as “the leading cause of personal bankruptcy” when introducing the Medical Bankruptcy Fairness Act, which would have made the bankruptcy process more forgiving for “medically distressed debtors.” But it turns out that the existing evidence for “medical bankruptcies” suffers from a basic statistical fallacy; when we eliminated this problem, we found compelling evidence of the existence of medical bankruptcies but discovered that medical expenses cause many fewer bankruptcies than has been claimed.
We use an event study approach to examine the economic consequences of hospital admissions for adults in two datasets: survey data from the Health and Retirement Study, and hospitaliza- tion data linked to credit reports. For non-elderly adults with health insurance, hospital admis- sions increase out-of-pocket medical spending, unpaid medical bills and bankruptcy, and reduce earnings, income, access to credit and consumer borrowing. The earnings decline is substantial compared to the out-of-pocket spending increase, and is minimally insured prior to age-eligibility for Social Security Retirement Income. Relative to the insured non-elderly, the uninsured non- elderly experience much larger increases in unpaid medical bills and bankruptcy rates following a hospital admission. Hospital admissions trigger less than 5 percent of all bankruptcies.
American hospitals are required to provide emergency medical care to the uninsured. We use previously confidential hospital financial data to study the resulting uncompensated care, medical care for which no payment is received. Using both panel-data methods and case studies, we find that each additional uninsured person costs hospitals approximately $800 each year. Increases in the uninsured population also lower hospital profit mar- gins, suggesting that hospitals do not pass along all uncompensated care costs to other parties such as hospital employees or privately insured patients. A hospital’s uncompensated care costs also increase when a neighboring hospital closes.
Health insurance confers benefits to the previously uninsured, including improvements in health, reductions in out-of-pocket spending, and reduced medical debt. But because the nomi- nally uninsured pay only a small share of their medical expenses, health insurance also provides substantial transfers to non-recipient parties who would otherwise bear the costs of providing uncompensated care to the uninsured. The prevalence of uncompensated care helps explain the limited take-up of heavily-subsidized public health insurance and the evidence that many recipients value formal health insurance at substantially less than the cost to insurers of pro- viding that coverage. The distributional implications of public subsidies for health insurance depend critically on the ultimate economic incidence of the transfers they deliver to providers of uncompensated care.
Transparency of quality in the healthcare sector primarily aims to facilitate patients' care decisions, however, it also provides useful information to competing healthcare providers. We study how competitors respond to increased transparency about rivals' quality by exploiting a regulatory change that initiated disclosure about the quality of all kidney dialysis facilities in the United States. We show that competitors are 27% more likely to open new facilities near low-quality incumbents after the transparency program is implemented. We also show that the effect of transparency on competition is restricted to states without licensing requirements that create barriers to entry. Evidence from patient referrals indicates that the new transparency regime increases the sensitivity of demand to quality and that the increase in competition is costly to low-quality incumbents, as they lose 31% of their new patient referrals—equivalent to a $3.74 million loss of a facility's annual revenue—to higher-quality entrants. Finally, increased competition from rivals leads to improved patient outcomes by reducing hospitalizations, and incentivizes incumbents' investments in patient care through an immediate increase in nurse practitioners and social workers.
Larger regions are more efficient at producing medical services. This leaves policymakers with a trade-off between concentrating medical care production in more efficient large regions and promoting healthcare access in less efficient small regions. Production and travel subsidies can both increase access to healthcare but impact patients, providers, and neighboring regions differently.
We evaluate whether and how branded TV product placement affects sales for cigarette brands. We use data on product placement from TV shows and data on retail sales of cigarettes to estimate a demand model that incorporates the level of product placement exposure for each cigarette brand. We find that product placement has a small yet positive and statistically significant effect on both own-brand sales and competitor-brand sales: both of these elasticities are roughly 0.02. These results indicate that cigarette product placement affects demand for individual cigarette brands and that it also leads to greater overall cigarette use. This issue is of particular importance to policymakers because product placement is one of the few remaining ways that cigarette brands can reach a mass audience. To illustrate how these results could be used by policymakers, we use our model estimates to evaluate how cigarette sales would be affected by two hypothetical kinds of regulations. Limiting brands' ability to be displayed on TV and forcing TV networks to instead use generic, unbranded cigarettes on screen would reduce total retail cigarette sales by only about 2 percent,while forcing TV networks to eliminate all on-screen smoking activity would reduce it by about 7 percent.s.
Explore Dan Adelman's Healthcare Analytics Laboratory Research
In the Media
Our faculty and their research regularly appear in the media. See highlights below.
Read highlights from the work of Associate Professor Jonathan Dingel and PhD student Pauline Mourot (with Joshua Gottlieb and Maya Lozinski) about large healthcare markets that leverage their scale to produce high-quality care that patients travel to obtain.
Listen to Professor Dan Adelman's insights on measuring healthcare outcomes in this podcast episode of the Business of Medicine Series on ENT in a Nutshell, hosted by Dr. Ashley Nassiri.
Financial Times article featuring Professor Notowidigdo's research investigating health outcomes during the Great Recession.
Better Health Economics
Better Health Economics is a warts-and-all introduction to a field that is more exceptions than rules. Economists Tal Gross and Matthew J. Notowidigdo offer readers an accessible primer on the field’s essential concepts, a review of the latest research, and a framework for thinking about this increasingly imperfect market.
Our Data Sources
Below are some examples of data used by our faculty.
- California hospital discharge data
- Centers for Medicare & Medicaid Services (CMS) LDS and RIF data (in partnership with the Health Data Policy Lab and others)
- Clarivate Competitive Intelligence
- Guidestar 990 tax forms from non-profit healthcare organizations (through the Rustandy Center)
- MarketScan® (in partnership with the Booth Center for Applied AI and Center for Health and the Social Sciences)
- Lightcast/EMSI Job Posting Analytics
- U.S. Renal Data System (USRDS)
- Real-world health system and provider data (in partnership with Center for Research Informatics and others).
Contact us for more information about industry data and partnerships, opportunities to combine healthcare with other data, and the request process for faculty and PhD students.
Imagine the Impact You Can Make
Join us as we make a positive impact on the healthcare industry.
Sign up to receive emails from the Tolan Center for Healthcare, including our monthly newsletter and updates on research, training, collaboration, and innovation in the healthcare sector.
YOUR PRIVACY
We want to demonstrate our commitment to your privacy. Please review Chicago Booth's privacy notice, which provides information explaining how and why we collect particular information when you visit our website.
Questions? Here's How to Contact Us.
If you have any questions about Booth’s Tolan Center for Healthcare, please don’t hesitate to reach out. We would like to hear from you.
