Hello, and I hope that you are all well during this time. My name is Oliver Tang. I'm a medical student at Brown University, and I'd like to thank the AANS for the opportunity to present my research with Dr. Toms at the Rhode Island Hospital on how interhospital competition impacts neurosurgical outcomes. We do not have any disclosures. Practice consolidation is a trend that is reshaping neurosurgical care. Over the past decade, around 90 hospital consolidations have occurred annually, and 20% of all hospitals are projected to seek a merger in the next two years. A big driver for this change in practice pattern is the growth of pay-for-performance reimbursement models, as well as an overall increased emphasis on the management of large populations. While conventional economic theory suggests that competition raises quality because institutions signal quality as a way to attract patients, a more complicated picture exists in the literature. For example, Duran and colleagues found that more competitive markets had higher complications in length of stay for lumbar fusion patients, and Cola and colleagues have similarly documented poor outcomes for hip fracture and stroke patients. However, the relationship between interhospital competition and inpatient outcomes has not been studied in the setting of cranial neurosurgery. As a result, this was the aim of our study. To this end, we analyzed data from the National Inpatient Sample, the largest all-pair inpatient database in the US, representing a 20% stratified sample of all community hospital discharges. In 2006 and 2009, we identified all cranial neurosurgery patients using 11 DRG codes corresponding to these procedures. We further classified patients into five subspecialties, cerebrovascular, CSF diversion, tumor, neurotrauma, or functional. The outcomes that we analyzed included mortality, favorable discharge disposition, which was discharged to home or a short-term hospital, development of an inpatient complication, and length of stay. To study interhospital competition, we used a variable called the Herfindahl-Hirschman Index, or the HHI. It is a validated economic metric that quantifies the competitive intensity of a given hospital market. For example, it is used in antitrust legislation. For a given market, the HHI is equal to the sum of the squared market share for each hospital in that market. To summarize what this means in real-world terms, the HHI ranges continuously from zero to one, with lower values indicating a more competitive market and higher values indicating a more monopolistic market. To give an example, if you consider a market with 10 hospitals, the market on the left where each hospital has an identical 10% share of patients has an HHI of just 0.1, but the market on the right where one hospital dominates and has a share of 80% of that market's patients has an HHI of 0.64, signifying that market's more monopolistic nature. When evaluating the relationship between interhospital competition and inpatient outcomes, we used multivariate regression to adjust for 13 confounding variables, including patient demographics, case severity, and hospital characteristics. Our study population was composed of 472,938 admissions for cranial neurosurgery in 2006 and 2009. Cerebrovascular was the most common subspecialty at 37%, whereas functional was the lowest at 2%. Admissions were an average age of 56 years old and most were male, white, and on Medicare. Our study population was treated at 896 unique hospitals, with a mean HHI of 0.305. The HHI within our study population ranged from 0.099, very competitive, to 0.724, monopolistic. Importantly, all subsequent results that will be reported in this talk correspond to a negative 0.10 change in HHI, or in other words, an increase in competition. We first looked at how HHI was associated with case complexity for neurosurgical admissions, and we found that severity of illness scores were higher in more competitive markets for all admissions, as well as cerebrovascular and CSF diversion patients. We also found that risk of mortality scores were higher in more competitive markets for cerebrovascular patients. These two scores were part of the multivariate model we used when it came to severity adjustment. Now we looked at how HHI was associated with inpatient outcomes. We found that tumor admissions exhibited decreased mortality in more competitive markets. However, HHI was not associated with inpatient mortality for the overall study population or any other subspecialty. To dissect this relationship with tumors further, we looked at three subtypes of tumors, primary tumors, meningioma, and metastasis. We found that while all three subtypes exhibited decreased mortality in more competitive markets, this relationship only reached significance for meningioma admissions. Next, we found that HHI was not associated with rates of favorable discharge disposition. In other words, there was no difference between competitive and less competitive markets. We similarly found that inter-hospital competition was not a significant predictor for the odds of developing an inpatient complication following neurosurgery. However, an interesting result we found was that more competitive markets exhibited higher lengths of stay for all admissions as well as every subspecialty studied except for functional neurosurgery. For each negative 0.10 change in HHI, there was a 6% increase in a patient's average length of stay. The following two slides summarize results from the study population that have been presented earlier. We've also found that admissions in hospital markets with higher competition have higher inpatient charges, which is the amount that patients are billed for. For every negative 0.10 change in HHI, there was an 8% increase in average charges, or around $4,900 more per patient. This relationship holds true for all five subspecialties, ranging from $3,900 more for each tumor admission to $7,200 more for each functional admission. While we found a robust relationship between HHI and inpatient charges, we did not find such a relationship for inpatient costs, which is the actual resource use during admission. While increased inter-hospital competition was associated with higher costs for CSF divergent patients, this relationship was not significant for the overall study population or any other studied subspecialty. All in all, our primary takeaway is that in the setting of cranial neurosurgery, higher competition was associated with lower mortality for tumor admissions, as well as elevated length of stay for overall admissions. However, we did not find evidence that competition was associated with discharge disposition status or inpatient complications. Nevertheless, it is important to bear in mind that our study population also demonstrated evidence of increased case complexity in more competitive markets. Finally, we've also shown that charges, the amount that patients are billed for, are elevated in more competitive markets, without an analogous increase in costs. While one might expect competition based on quality to become more prevalent as public reporting on hospital outcomes becomes more readily available, other studies suggest differently. Below and colleagues in a study this year looking at 246 hospital acquisitions over the past decade found that mergers had no impact on quality metrics like mortality or readmissions. They hypothesized that clinical outcomes like readmissions are still not attentively tracked by patients and thus have minimal impact on the competitive landscape of hospitals. Kola and colleagues looked at 38 hospital quality indicators and found that at higher competition, 8 improved but 8 decreased, with the metrics decreasing being ones that were the least visible or apparent to patients, such as those related to systems of care instead of individual physician error. These results collectively suggest that hospitals may prioritize competing for patients based on metrics that patients can easily observe and evaluate, rather than the most clinically meaningful metrics. For example, marketing and advertising expenses, as well as amenities and quote-unquote hotel aspects of care have all been found to be elevated in more competitive markets. Another example of this that has been commonly discussed is the medical arms race, where hospitals facing more competition invest more intensively into newer treatment modalities and technologies to attract patients. For example, it has been shown that a hospital is more likely to acquire a surgical robot if a neighboring hospital purchases one. Studies have raised concerns that the medical arms race may drive spending in areas that don't impact patient outcomes or may result in unnecessarily intensive care, with evidence demonstrating this in the setting of lumbar fusion, stroke, and aneurysm surgery. With this in mind, our results raise some policy and practice implications. Competition may improve outcomes in certain areas like brain tumor mortality, which is a finding that warrants further study. We hypothesize this may be due to improved technology adoption or the predominance of large academic institutions in competitive markets. However, this finding seems to be the exception and not the norm, with most associations like complications being insignificant, and in the case Lent-to-Stay, competition was associated with lengthier hospitalizations, which may be due to the medical arms race and more intensive levels of care in competitive markets, but there may have been residual confounding due to case complexity. We also found evidence of higher charges for neurosurgical patients in competitive markets despite no differences in outcomes or underlying costs, which may reflect a predominance of what is called non-price competition through factors like patient amenities. In an environment where one-third of Americans report medical access limitations due to costs such as delay in care, this is critical to address. Finally, our results importantly suggest that measures may need to be taken to realign hospital competition to be based on meaningful quality metrics, like mortality and complications, in order to ensure that competition can drive a race to the top. A continued endeavor to improve public reporting of institution and certain outcomes is one promising path towards this. I would like to thank Dr. Toms and Chrystia for their mentorship, without which this project would not have been possible, and I would be extremely eager to answer any questions that you may have on this email address on this slide. Thank you so much for your time.

Oliver Tang

medical student

Brown University

interhospital competition

neurosurgical outcomes