Saturday, January 31, 2015

QUALITY METRICS, Probability & GOTCHA

Can the bounded asymmetry of thought align itself to the unbounded reality? That is the question. The premise is: "To drive progress, we are focusing on three strategies. The First is incentives, the Second, improving the way care is delivered and the Third, we aim to accelerate the availability of information to guide decision making through Meaningful Use of EMRs." ( here) The verbiage is enticing and any arguments to the contrary makes one look regressive, obstructionist and non-compliant! What could possibly go wrong if we were to reimburse physicians and hospitals for quality and withhold such payments for lack thereof?


Plenty!

First let me tell you up front about the “probability” function. A lot of it is based on assumptions. When we look at small samples and predict the event onto a larger population, there is always a chance that we will be wrong. As noted by John Ionnadis in his 2005 paper that 54% of all studies could not be validated due to the use/abuse of statistical modeling. Sampling has an inherent bias, even if we cite the Bayesian principle (which simply states that the probability of event B is the sum of the conditional probabilities of event B given that event A has or has not occurred). Here we are using conditional probability of independent events.

Here are a few assumptions:
1.       Patients with multiple (more than one) comorbid diseases have a higher chance of complications. eg: a) A diabetic patient with vascular disease and alcohol related fatty/cirrhotic liver has a higher chance of renal dysfunction. b) A neutropenic patient undergoing chemotherapy has a higher chance of infection with other complications related to the chemicals and biologics on top of other chronic conditions such as obesity, kidney disease, heart disease etc. c) An elderly patient with a coronary artery disease, previous myocardial infarctions has a higher chance of arrhythmia, congestive failure and pulmonary edema. 2 + 2 mostly ends up as 3 or 5 due to confounding factors mentioned above.

2.       Younger patients without comorbid diseases have better outcomes due to their resiliency from age. 2 + 2 indeed can amount to 4 in them. 

3.       The ratio of the healthy-young to elderly with comorbidities is roughly 3:1. This ratio is skewed, to the extent that the numbers vary regionally and community-wide and can be as low as 4:1 or as high as 2.25:1. A similar ratio feeds its way into the healthcare sector.

4.       Using those simple assumptions let us say that the following ACOs and hospitals exist: a) An ACO tied to a tertiary hospital. b) A rural hospital with limited resources. c) A multi-specialty group with a good reputation with privileges in a secondary hospital.

P = ( Probability # of successful hospital outcomes in total # of hospitals in a group)
# of successful hospital outcomes = x
Total # of hospitals in a group = n
Probability of making the successful outcomes = s
Our formula is: P= s^(x (1-s)^(n-x)

Let us look into the magic of probabilities and statistics. Assuming that the Probability of success in achieving the metrics mandated by CMS = 60% and the probability of missing the metrics is (1 – 60%) 40%. We then further assume that 8 out of the 14 hospitals will achieve this success or 14C8 (Out of 14 choose 8) which than gives us the combinatorial formula using factorization (14!/8!(14! – 8!) = 0.20 or 20%. We are giving higher potential of successful outcomes to a select number of hospitals (n=14) and saying that these hospitals had all the right kinds of resources and the right demographics of fewer patients with comorbidities. Then given all the risks and “paper-pushing-and-game-playing,” 8 out of 14 hospitals have indeed the potential of achieving probable success of 60%. The probability of that successful outcome happening is 20%, in other words even though we give this high likelihood of success, the potential probability of achieving that success is only 20%. OR putting it in numbers: The probability of NOT achieving success by these hospitals is ( 1- 20%) = 80%. (Or 80% of the time you are not going to get paid for services).



It is obvious for anyone to see that given the aging demographics of the United States and the retiring baby boomers that what the government keeps telling us suggests that the ratio will be even more skewed against the successful outcomes, unless humans become robots and an oil and lube job keeps them going until rust settles in from disuse.

Now let us take the first set of hospitals (here we take a larger sample n = 30) mentioned above and the ACO combinations:  a) Given the high risk nature of their patients the probability of achieving success in 15 of the 30 hospitals might dip to 40% hence the probability of successful outcomes goes down to 7.8%!



In the second scenario of rural hospitals (n =30) assume the probability of success is at 50% (the toss of a coin), the probability of successful outcomes now is 14.4%. The rural hospitals with a lesser number of complicated cases will have a better chance of achieving the successful outcome; almost double that of the tertiary hospital.



If you think that is questionable, just go back and review the P4P (Pay for Performance) plans conjured up in the dimly lit confines of the bureaucrats. P4P was abandoned as a result!

And with the third scenario, while it may appear good shows that the secondary hospitals and good reputation ACOs with a 70% probability of success in 15 of the 30 hospitals show a probability of successful outcome in 1.06%! 



So where does that leave us?

Actually it paints an interesting picture. Even if we were to double the successful outcomes probability in each case it would still be less than 50%. So here is where that leads us…

If less than 50% are able to play the game and get rewarded for it, then what happens to the rest who cannot? How many will survive without getting paid for services? Something to think about!

Remember the arcades where a quarter will get you a prize if you can steer the picker-upper claws to pick the prize from a large collection of soft teddy-bears and the like and drop it into the basket for you to collect. The only problem is that the picker-upper claws are loose and dangle as they slide over any of the colorful furry toy animals. So you spend and spend but never get a prize. The risk reward appears against the establishment, yet the truth is the consumer gets hosed into spending a lot more than the prize is worth and still never gets any.




We see the same gotcha thought process, only that a few will be able to manipulate their numbers to meet the required demands, but not for long. The evidence is plain to see when it comes to the ACOs how more than half of them shut down because they were unable to make any revenue to keep the doors open. Taking the lead from that concept what could possible go wrong to the entire healthcare system? The agenda to limit expense will be successful, to the extent that monies will not be doled out as mandates and checkpoints will guard the outflow. Success will be writ large in the obedient media. "Money has been saved!" will cry the headlines, and the demise of the scores of shuttered facilities will be ignored by the pundits. Yet with such outcomes the answer will also lie in the liberties of the human element. People who spend their lives acquiring knowledge to better their craft in caring for patients, I feel will find alternate sources of managing their patients and as Denials of Services (DoS) becomes the game du-jour, more patients will run away from the game. The bureaucrats will then come to the table, in order to keep their monopoly, by liberalizing the criteria and a new game will begin for the gullible.

So that leads me to another question. Is the attempt to make people (in this case doctors and hospitals) work to the bone but not claim any remuneration the underlying philosophy? Is hard work related reward now considered greed?  Times are a-changing! Aren't they?

What might happen if we look into the future further? Is the new model to create fewer entities that are able to “play the game” with the government mandates and the rest of the many to forfeit ability? Those that do play well will be glorified and those that don’t will be vilified. In the end the choices will narrow to a few. And with fewer vying for rewards comes better controls and less reasoned critical thinking.

Sounds harsh?

O' I imagine I will be skewered by some experts for being naive and a fool and I admit, I am both. But maybe constructive criticism will open a dialog and some eyes.

Think again! 

No comments:

Post a Comment