Presenter: Dr Zaid MuslimView Large Image Figure ViewerDownload Hi-res image Download (PPT)Dr Erkmen. I have a question about how you stratified the institutions into low, middle, and high. How did you get a significant difference between just the very slight number of Medicare, Medicaid, and uninsured patients? The differentiation is between 8% and 12%, but you still found a significant difference. What would you say is accounting for that? How will you account for institutions like mine that have 75% to 80% of people who are underinsured and poorly insured?Dr Muslim. Sure. I think that's a good question and something that we've given a lot of thought to. I think the logic behind categorizing them this way was again splitting the distribution into an equally spaced threshold just to generate 4 comparison groups that represent the spectrum of hospital burden. As to why there's a significant difference with such a small difference in the percentage of uninsured patients, it's a good question. I think we certainly see that patients on the extreme ends of the spectrum have a greater magnitude of differences in outcomes versus the patients who are closer together on the distribution. So essentially, there is a pattern of or a relationship that appears to be between hospital burden and outcomes. But again, the study does a better job of outlining that there is a potential problem. Essentially, we need more granular data and more prospective studies to hone in on why that is.As for your second part, the hospitals with the larger percentage of uninsured or underinsured patients, we need to recognize that not all hospitals are the same. I think it's easy to extrapolate from this data that we're generalizing and calling all high-burden hospitals bad, but even high-burden hospitals can have good outcomes. And we saw in our study that academic hospitals are more likely to be high burden, but we also saw in our study that they're more likely to have better outcomes. So, that clearly shows that in academic hospitals, there are characteristics and qualities that allow them to have better outcomes despite seeing a large number of underinsured or uninsured patients, and we need to study why that is. I think it's an interesting point to consider. Presenter: Dr Zaid Muslim Dr Erkmen. I have a question about how you stratified the institutions into low, middle, and high. How did you get a significant difference between just the very slight number of Medicare, Medicaid, and uninsured patients? The differentiation is between 8% and 12%, but you still found a significant difference. What would you say is accounting for that? How will you account for institutions like mine that have 75% to 80% of people who are underinsured and poorly insured? Dr Muslim. Sure. I think that's a good question and something that we've given a lot of thought to. I think the logic behind categorizing them this way was again splitting the distribution into an equally spaced threshold just to generate 4 comparison groups that represent the spectrum of hospital burden. As to why there's a significant difference with such a small difference in the percentage of uninsured patients, it's a good question. I think we certainly see that patients on the extreme ends of the spectrum have a greater magnitude of differences in outcomes versus the patients who are closer together on the distribution. So essentially, there is a pattern of or a relationship that appears to be between hospital burden and outcomes. But again, the study does a better job of outlining that there is a potential problem. Essentially, we need more granular data and more prospective studies to hone in on why that is. As for your second part, the hospitals with the larger percentage of uninsured or underinsured patients, we need to recognize that not all hospitals are the same. I think it's easy to extrapolate from this data that we're generalizing and calling all high-burden hospitals bad, but even high-burden hospitals can have good outcomes. And we saw in our study that academic hospitals are more likely to be high burden, but we also saw in our study that they're more likely to have better outcomes. So, that clearly shows that in academic hospitals, there are characteristics and qualities that allow them to have better outcomes despite seeing a large number of underinsured or uninsured patients, and we need to study why that is. I think it's an interesting point to consider.