when minutes count

Doc, I am a stickler for methodology because of my background and some work that I do that looks at a lot of variables and often has people making wrong correlations.

I only gave that release a once over, but a few questions came to mind. It looked like they looked at emergency admissions for 11 conditions, but the conclusions didn't necessarily evaluate the emergency care per se, as much as the overall hospital care. They really evaluated the hospital-wide care using mortality as the criteria, without finding factors that could single out to what extent the emergency care contributed. Am I correct about that?

In another words, they didn't come up with some way to evaluate the extent to which the emergency care contributed to low mortality rates for those 11 conditions as opposed to the care those patients may have received after being admitted. I know it's safe to assume that good emergency care is critical and has a pronounced effect, but if they are going to make conclusions and attribute the low mortality rates to good emergency care, it just seemed to me they should have methodology that isolates the emergency care. And they didn't do that.

Can you think of any way to possibly correlate the emergency care specifically to survival that could do what I am suggesting?

Also, I'd think a serious analysis would have to distinguish even within those 11 cohorts as they called the conditions, in terms of seriousness of the illnesses, and it didn't look like they did that. Not all strokes are the same, for example, right? Some are more serious than others, I would think. That study doesn't say if they did any kind of regression to factor in things like how severe the illness was at the time the person showed up at the hospital or the age of the patient or any other conditions in addition to the cohort being evaluated that put them at higher risk of mortality.

I don't mean to entirely pick apart the study. The study would seem to be a good indicator of how good the hospitals acknowledged provided hospital-wide care for certain emergency conditions, which is something I'd want to know as a patient going in (but of course, no hospitals in New York, where I am, made their cut). But using the methodology they described, they don't really evaluate the emergency departments themselves. That methodology reads more like they are looking at 11 specific emergency situations and evaluating the overall hospital care, and then singling out some hospitals as better, based on lower than average mortality outcomes. How much those lower mortality rates are attributable to the emergency care as opposed to other factors may be a subtle distinction, and maybe it's something only I care about. But am I wrong about that? I may be misunderstanding.

 

Doc, I understand your point. From a patients' perspective, it probably doesn't matter if it was an emergency room that saved their life or the overall quality of the hospital. And as you pointed out, obviously, when it comes to many diseases such as stroke, the immediate care is critical -- although even there that study looked at just mortality and a very small percentage of people die in the hospital. Yet, your quality of life after a stroke can vary dramatically, based on the emergency care you receive. A study that could somehow look at a broader array of outcomes than just mortality, which is very rare (only 3 percent of patients die in hospitals, according to what I just read) might be interesting. But your point makes sense to me. Maybe my questions missed the important point that it doesn't matter how you emerged from the hospital alive (emergency care or otherwise), just that the hospital overall did a good job.

That said, while you responded, I was looking at the full report and even though one of my questions was answered, it raised new questions. The cohorts WERE risk adjusted. I'll assume it was done using things that have been proven to increase mortality risk, such as age, gender and other conditions that are present. But that may be problematic in itself because of the source of the data. They use Medicare data, because it's a nationally-available database that gives them an identical source for their clinical data across all hospitals. However, based on what I was just reading, Medicare patients only account for slightly half of all hospital admissions and they exclude children and young adults. Also, Medicare coding (what they used to risk adjust their data) is apparently unreliable. Hospitals are in a constant battle to get maximum reimbursement from Medicare, and some hospitals --especially those that have a disproportionate number of Medicare patients -- may inflate the billing codes to make a patient's condition appear more complicated so the hospital can get a bigger Medicare reimbursement. Other hospitals are understaffed, so their coding is not done well. A hospital that is more forthright might actually be penalized for coding their data better than another hospital.

I also came across a few other things that were disturbing when looking at this. I had no familiarity with Healthgrades, but apparently there is a conflict of interest. Looking at their site and their last few SEC filings just now, there is a conflict of interest in their claim of being an independent evaluator. They charge hospitals to use their excellence awards for advertising and publicity purposes. It's actually a significant source of revenue for the company, according to their filings. I would think you have to be somewhat skeptical of a for-profit company that charges hospitals to use awards they reward for excellence, while the company is selling itself to the public as an independent evaluator.