Wednesday, June 23, 2010

Pretty data. Useful?

This is an interesting portrayal, provided by GE, of hospital quality metrics. It compiles data from the Joint Commission's 2009 Annual Report on Quality and Safety and presents it in a visually attractive manner that permits a qualitative comparison among hospitals in each state. I provide a partial view above.

So, this is pretty, but is it useful? I don't think so. As I have said before:

There are often misconceptions as people talk about "transparency" in the health-care field. They say the main societal value is to provide information so patients can make decisions about which hospital to visit for a given diagnosis or treatment. As for hospitals, people believe the main strategic value of transparency is to create a competitive advantage vis-à-vis other hospitals in the same city or region. Both these impressions are misguided.

Transparency's major societal and strategic imperative is to provide creative tension within hospitals so that they hold themselves accountable. This accountability is what will drive doctors, nurses, and administrators to seek constant improvements in the quality and safety of patient care.

So, I think a better presentation would be one in which I could compare my own hospital's data year by year using the same kind of side-by-side visual imagery. Then, we could get a quick view of how we have progressed over time -- in essence, competing against ourselves, rather than against our colleagues in the state.


Anonymous said...

The data's usefulness depends on the audience that GE was intending to reach - hospitals, or patients, or insurance companies?

The charts are somewhat useful in that one may conclude, for instance, that BIDMC performs better on these PROCESS measures re surgery than MGH, whereas MGH is probably better in heart failure. But they are, after all, only process measures. Also, I never found a key which defines exactly what the measures are.

Also, I imagine it would be easy for GE to go back and convert the data from previous years' reports for the comparisons you desire.

All in all, I think it's not bad.

Keith said...

Data comparing your performance may be better for you interms of tracking your improvement or lack therof, but I fail to see how this would be more useful to patients trying to guage where is the best place to obtain their health care. If I looked and saw yearly gains in quality at BIDMC, I would next ask what the benchmark is and whether that quality started out very low and how it compares presently to other hsopitals. Both are important parameters to compare, but one is not necesarily more important thanthe other.

~Mary said...

I agree with your concept on each hospital holding itself accountable. When you compete with others, you lower your standards-that is something I firmly believe to be true-especially when it comes to customer service. Each facility has a unique opportunity to capitalize on something different because of it's individuality. Finding what that is should be the administrators top priority. I love your blog-you have your finger on the pulse point of modern health care.

Anonymous said...

These charts are not useful for someone who is red-green color deficient.

Anonymous said...


There is no evidence that consumers use this kind of data.

stevemassi said...

Any traditional business would look at both measures. Internally to view areas of improvement/weakness and externally because you are in a competitive environment where patients can choose.


Cetus said...

I don't see this system catching on. I look at the grids and my brain says, "Huh? What? Where are the numbers?"

Complex litigation has seen the rise of an abusive litigation tactic called "data dumping" in recent years. Party A demands discovery of all electronic records related to a matter. Party B complies by turning over one billion+ pages of documents, unindexed -- just a few days before the discovery period ends. Everyone understands that it is impossible to sift through the data dump in a few days. That's the point.

Before people say, "Oh, it's just those damned lawyers again," folks might consider that the radical accessibility of data has led to the deliberate creation of haystacks of "noise" in which to hide the precious needles of useful data. Sophisticated people know this.

It is the Prisoner's Dilemma: try to maximize public benefit by asking the tough questions internally with integrity -- and leave yourself vulnerable by creating a paper trail -- or play hide-the-ball like it seems everyone else does.

If by "creative tension" you mean building a culture of regularly asking internal questions that make people shift uncomfortably in their seats, I'm all for it! But at the end of the day, as your GE example illustrates, the professional jargon must be translated into lay understanding so that consumers can make informed choices. If BIDMC (or any hospital) had to slap the disclosure equivalent of a Nutrition Chart on your front doors, what would it say? What should it say?

Anonymous said...

Check our website at under Safety and Quality and take a look.

But, really, as I noted above, there is no indication that consumers are (yet) using this kind of information to make choices, even when it is clearly presented. They tend to rely on their primary care MD or other referring doctor as to where to go for hospital services.

Anonymous said...

On further exploration, this appears to be part of GE's new program "Healthymagination", which has a website available through Google. Their presentation of the idea behind this report is here:

Hope this works; I really gotta learn how to embed links.

At any rate, they seem to be indeed trying to compare between hospitals within each state - but I agree with Paul's idea about adding longitudinal data on each hospital. Perhaps you can suggest that to them, Paul. I get the impression they're not quite sure themselves where they're going with this massive project....


Keith said...


That is because thsi data collection is in its infancy and only now becoming readily availible. I can assure you once thsi information is colasced and presented in some intelligible form that provides a comparison, it will indeed be utilized by consumers of health care.

Anonymous said...

As the first poster mentioned, the key question of usefulness needs to be reword as "Useful for who?"
1) the patient? I've never understood how the patient can be considered the optimal decision maker. They cannot be fully informed unless they are medical professionals and then ethically they should neither diagnosis nor treat themselves.

2) the physician? Not likely. Unless the physician is very well versed on what each indicator is, how that indicator relates to the specific patient they are concerned about, and what the over-all result says about the institution in general, I fail to see how it would reliably influence their decision making.

3) Hospital administrators? Probably very useful if used in a co-operative manner. Using pneumonia as a case in point, the four institutions in the example could have the same overall rate. AMH excels at one pneumonia indicator but BWH is the only one below average. However, with another two idicators, AMH and BWH are at or above average and BIDMC and BMC are below. If the four of them got together and said "what do you do differently here, here, and here?" they could potential improve every institution.

There is also the question of "below average" vs "above average" - is a differential of +/- 2% worth worrying about?

Jeff Terry said...

Paul – I’m a long time fan of your blog. I work for GE and while I wasn’t involved in developing this specific tool, I was excited to see you blog about it. My sense is that there is a need for both internal metrics and comparisons between hospitals. That's important for both quality improvement and public transparency. As you noted in a few comments, the evidence suggests that the public does not yet consume this type of data.

We often analyze hospitals’ patient safety and patient flow over a several year period. To your point, this can be powerful. For example, we can see the impact new space, cultural change, new services, new processes, etc. have on the performance of the system as a whole. By the same token, our clients often ask us to compare their performance to their peers – and they usually ask that their peer group be defined in a very specific way.

In any event, thanks again for your blog. If the opportunity presents itself, I’d love your take on other tools that we’re developing.

Managing Principal, Clinical Excellence

Peter said...

I believe that consumers could use this data, but it would be those who are highly aware. However, if this data was marketed and democratized, I believe it could be a useful tool. For example, I suggested my grandfather in Florida check out the local hospitals using the GE data before he had a procedure and he chose the one with the best indicators. I would do the same.

Also, isnt BIDMC doing their own in house evaluations or do you need this data? I assume you are reporting the data already so why the need to have it public if it is not for the consumer?