Friday, January 09, 2009

Sharp numbers

I cut out an article last March from one of those free magazines you get while flying. This was on a flight from London in a magazine called BusinessLife, and the piece was by Tony Thorne at King's College London. The title was "Sharp Practice". (Sorry, I can't find a link.) I had meant to write about it at that time, but I just found it buried on my desk.

This is about precision heuristics. One application is in the pricing of goods and services.

More generally, we tend to think that "claims expressed in numbers ('78.6 per cent effective') appear to be objective -- based on empirical data -- while claims expressed in words ('finished to the highest technical standards') tend to be judged as subjective. The same distinction operates between round figures or round numbers, often suspected of being guesstimates, and sharp numbers, assumed to show verifiability."

I can already hear some of my colleagues in health care saying, "I told you that transparency of clinical outcomes is misleading! You give the impression of precision when the measurement of the quality of medical care is inherently subjective."

Sorry, guys, don't even try it. I have addressed that before, saying:

We have learned from studying other industries that have engaged in and achieved process improvement that such improvement requires an approach to the organization of work that is very different from that seen in most hospitals. But it also requires measurement and transparency. While even the best calculations and data don't tell all, they do tell a lot, and they are the only way we have for an organization to hold itself accountable.

But those in the medical profession sometimes fall into the trap of believing that because measurement is an inherently reductionist and mechanistic act, it can never be sufficiently accurate to reflect the overall realities of patient care. The paradox is that without it, we can inadvertently fall into the trap of self-congratulatory statements about our good intentions. Only with it can we demonstrate that we actually have a "relentless determination to do best by each patient."


So, sure, let's watch out for the traps of precision heuristics, but let's also watch out for the traps of imprecision, unsupported generalization, and unsubstantiated reputation.

3 comments:

Anonymous said...

Speaking of attempting to measure claims expressed in words, you might be interested in an article from your very own Harvard School of Public Health (et al) in the New England Journal:
"Patients' perception of hospital care in the United States." NEJM 2008:359;1921-31.
Among other things, it found that patients generally gave higher marks to those hospitals demonstrating higher quality of care as measured by the Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey. It also said that whether the hospital was teaching or non-teaching had no effect on patient ratings. Maybe somebody forgot to tell Boston patients that? Also, I read in another blog recently that the Partners' CEO called you a "whiner"???!!!!! LOL - is he back playing in the schoolyard or what??!!

nonlocal

Anonymous said...

Great post, sir! I work in a hospital in Ireland and we face exactly these same issues with regard to transparency and clinical quality. On a visit to a Belfast Hospital some years ago, the Medical Director told me that in his experience, leaving clinicians to audit themselves lead them to tow things: 1. Audits that showed how wonderful they were 2. Audits that proved they needed more new equipment/facilities/staffing. In that hospital, they combined a top-down approach with a bottom-up - allowing clinicians to pursue their own audit interests, but also insisting that as part of the deal, the Hospital must be provided with a data set of clinical performance indicators that were driven by international standards and patient concerns, and with a stipulation that performance must be improved if it is shown not to be at the required standard.

Anonymous said...

Correction on my last comment: it was Partners' Board Chairman, not the CEO. Have to keep MY facts straight, speaking as a scientist.

nonlocal