Data, use and misuse- Contrawise: Confessions of A Data Contrarian

“Contrawise,” continued Tweedledee, “if it was so, it might be, and if it were so, it would be, but as it isn’t, it ain’t. That’s logic.”

Lewis Carroll, 1832-1898, Alice in Wonderland, “The Jabberwocky”

As I write, I’m swimming upstream against downstream logic. The logic goes like this: If data says it’s so, it must be. Data doesn’t lie. Trust intuition: All others use data, please. Data isn’t a question of would be, or might be, it’s so.

Yet reasons exist for doubting raw application of health information technology and data as keys to effective quality improvement.

1. In human endeavors, psychology and supply and demand trump data, charts, statistical trends or analyses. Look no further than stock markets, “laws of economics, ”or gas prices.

2. Vagaries of human interaction and returns of patients to bad health habits are more likely to determine outcomes than adhering to data-based protocols or what occurred during doctor visits or hospital stays.


3. For 30 years, John Wennberg and followers at Dartmouth and Harvard have lamented regional data care variations without notable effect. Culture, not medical malfeasance, dictates the differences. Data has not “corrected” or “narrowed” these differences.

4. Pick of the litter medical centers – Johns Hopkins, Mayo, Cleveland Clinic, Harvard Hospitals, Columbia Presbyterian – vary greatly in lengths of stay, number of specialists consulted, and outcomes. These factors are more a function of patient bases, severity of illnesses and institutional and regional environments, than data deviations.


5. Procedure use often depends on subjective factors – such as demands for knee or hip replacements to maintain an active life style. Deploying data to justify these procedures is unlikely to curtail demand or to stop specialists from doing procedures. Patients will shop around. They will find someone who will do what they think needs to be done. There are always ways around data, as in cash only practices or cosmetic procedures not covered by insurance.

6. It takes money to make data – as much as $50,000 per doctor per year to create, maintain, and monitor data. It is said ubiquitous EMRs and robust data-based infrastructures will save 20% of our national health bill. The flip side of this cost equation is seldom mentioned.

7. Controlling medical decision making is a tricky and complicated proposition. Data does always reflect this “trickiness” and “complicatedness.” Medical care is full of paradoxes and tensions, “right degrees” of information flow, diversity and differences, connections inside and outside the practice, and between patient and doctor. These all involve balancing intuition and data, safety and risk. Government and management cannot control or even track the 2 billion health marketplace transactions.

8. Supply and demand for innovative technologies, hospital “marketing wars,“ media reports of technology wonders, consumers seeking miracle cures or relief, are difficult to regulate, assess, and suppress. Some policies makers say government should require technological assessments before market introductions. It will not work. You cannot bottle up or block entrepreurship, innovation, and sheer human inventiveness. There’s always a better way, and people will find it.

9. Data-advocates should be more modest in over-estimating their limitations and their abilities to reduce “unnecessary procedures,” enhance necessary clinical conduct, and transform human nature at the point of care. We live in a button-up, not a top-down, society.

10. Some approaches to quality improvement are worthwhile. Data can be useful when shared, compared, aggregated, analyzed, sliced, and diced, but data has profound limitations. It is but one small arrow in the quality quiver and has limits in deciphering deficits and exposing excesses/

David Nash, MD, professor and chairman of the Department of Health Policy at Jefferson Medical College. is an articulate spokesman for the value of data-gathering technologies. He edits two publications, Health Policy Newsletter and Prescriptions for Excellence in Health Care, the latter in collaboration with Eli Lilly and Company. These newsletters champion performance improvement methodologies, evidence-based medicine, and comprehensive data-driven support systems to establish comprehensive clinical safety and quality. I salute his efforts, particularly his promotion of the “scholarship of quality.” W need more quality scholars.

Data systems may increase safety and quality, but they are expensive and limited in their capacity to change fundamental human nature or outpatient medical practices. Most physicians to whom I have spoken say that quality control measures have yet to significantly influenced practice. That influence may come, but we are not there yet.