Big Data, Transparency and Accountability
Robert G. Uzzo, MD, FACS
Physicians have a love-hate relationship with data. We crave clinical and scientific data, yet we grieve the performance-based data increasingly directed toward us.
In her 1969 book On Death and Dying, Elisabeth Kübler-Ross describes the stages of coping with grief. Her model presents an intriguing analogy to physician response to performance data.1
When first faced with publicly reported outcomes, many physicians are unwilling to admit that these results reflect their practice (denial). Next, a degree of resentment at unsolicited dissemination of information is natural (anger).
As medicine is inherently altruistic, many physicians believe such close and public scrutiny is unwarranted, and they argue for further evaluation, improved data collection, and more rigorous severity adjustment prior to public release (bargaining).
The consummation of these processes may lead to concerns regarding the negative impact quality reporting may have on autonomy, referral patterns, financial stability, and future litigation risk (depression). In the final stage, Kübler-Ross describes the individual coming to terms with the irrevocability of his/her own mortality/data (acceptance).
Whether the majority of physicians will accept that active participation in quality evaluation is necessary to improve health care delivery remains to be seen.
This model may help explain the visceral response most physicians felt to the decision of Centers for Medicare and Medicaid Services to release detailed web-based information on utilization by nearly 900,000 physicians and other providers caring for Medicare beneficiaries (data.cms.gov).
In a perspective piece, Brennan et al explain that after an initial request for public comment, which garnered only 130 responses from August 2013 to April 2014, the release was part of a growing commitment toward open data to drive health system improvements.2
So inhale or exhale the data at will, but here is what we physicians should understand. First the data are coming—flaw, limited, and imperfect. We need to engage in the process. Second, guidelines, pathways, and algorithms need not threaten the art of medicine.
Variability is best built around standards, which are ours to help define. Third, the bell-shaped curve is a useful measure of performance as it includes two standard deviations from the median. Significant outliers should be analyzed, as they may teach us the good and the bad. Finally, data always represent opportunity, for the system, the doctor, and most importantly, our patients.
- Smaldone MC, Uzzo RG. The Kubler-Ross model, physician distress, and performance reporting. Nat Rev Urol. 2013;10:425-428.
- Brennan N, Conway PH, Tavenner M. The Medicare Physician Data Release – Context and Rationale. N Engl J Med. 2014 (Epub ahead of print)