Photo: Hoag Levins
Charles Ornstein, a senior reporter at ProPublica and Penn alumnus who began his journalism career as student editor of the Daily Pennsylvanian, spoke about clinician outliers at Penn’s Leonard Davis Institute on November 3.

Although “outliers” are often discarded from academic studies as a matter of balance, they have recently become an intense target for one of the country’s most aggressive investigative journalism organizations, ProPublica’s Charles Ornstein told a University of Pennsylvania audience of health services researchers.

Speaking at a seminar at Penn’s Leonard Davis Institute of Health Economics (LDI), the Putlitzer Prize-winning journalist and senior ProPublica reporter noted that much of his newsroom’s success has come from a focus on clinician outliners found in various Medicare data sets.

Clinician outliers
Outliers are physicians or other health care professionals whose practice patterns — like drug prescribing or treatment coding — rank near the top or bottom of their overall peer group.

“Outliers can tell you a bigger story that can lead you to ask very interesting policy questions,” Ornstein said as he launched into the first of several examples of database outlier discoveries that resulted in high-profile ProPublica health care exposés.

Treatment Tracker database
Using the “TreatmentTracker” online tool created with a year’s worth of Medicare Part B billing data from more than 986,000 health care providers, Ornstein and his colleagues looked to see which doctors jumped out from the data as very different.

One example of what they found was a Michigan family physician who ranked 135th out of 3,200 doctors in terms of services performed (an average of 42.5 per-patient vs. an average of 5 per-patient by most other doctors). In the amount-per-patient spent, this same physician’s data showed an average of $3,800 per-patient compared to other doctors’ $208 average per-patient.

When the same physician was looked up in the ProPublica online Prescriber Checkup tool that contains searchable Medicare Part D data, the reporters found that 86 percent of his patients who filled their prescriptions received a Schedule II controlled substance — compared to an average of 4 percent among other physicians.

Ornstein pointed out that although it was the agency’s own data, Medicare had not noticed or further investigated this outlier’s details until they were published in a national news story. He also noted that ProPublica was not using the sophisticated algorithms that Medicare developed for identifying potential fraud but rather doing “simple arthimetic” based on data set sorting.

ProPublica, which became the first online news organization to win a Pulitzer Prize for investigative reporting in 2010, has subsequently won two more in 2011 and April of this year.

Ornstein’s own 2005 Pulitzer Prize for Public Service was for a Los Angeles Times series documenting gross medical insufficiencies and racial discrimination within a public hospital. 

Health care focus
Funded by various philanthropic organizations including the Knight, MacArthur, Pew, and Ford foundations, and staffed with 45 full-time investigative reporters, ProPublica’s hard-hitting reports have targeted a broad range of subjects from Wall Street finance  to Law Enforcement. But throughout its nine years of existence, ProPublica has consistently taken a particular interest in things health care.

Some of its recent shorter reports in this area include an analysis of the negative impact of a March Supreme Court ruling prohibiting states from requiring the submission of health care claims to all-payer databases; a look at how the nation’s opioid epidemic is morphing into something much larger; and an exposé on how nursing home employees across the country have routinely been taking dehumanizing photos of their patients and posting them in social media galleries.

Big data
ProPublica’s larger “big data” journalism projects often go on for years and three of them — Surgeon Scorecard, Dollars for Docs and Red Cross — have broken new ground at the same time they’ve created national controversies.

Surgeon Scorecard is a big data project based on Medicare records that rates surgeons according to the post-surgical complications their patients experience. It allows prospective patients to search for local surgeons who specialize in eight elective procedures such as knee replacements, lumbar spinal fusion or prostate problems. This online tool was hailed by health services researchers such as Harvard Professor of Health Policy Ashish Jha, MD, MPH for its innovative effort to bring a new sort of transparency to the consumer-doctor relationship. Simultaneously, it was faulted for its methodology by researchers like LDI Fellow Karen Lasater, PhD, MS, BSN, RN, of Penn School of Nursing’s Center for Health Outcomes and Policy Research.

Dollars for Docs is a ProPublica online database tool that enables consumers to search by name to pull up a list of the payments physicians have received for corporate consulting, guest appearances and other paid activities outside their practice.

Red Cross is another ProPublica project that has been going on for two years and continues to investigate the global activities and finances of the Red Cross. The 42 articles to date fill their own website-sized collection of stories including “Red Cross Failures in Haiti,” “The Corporate Takeover of the Red Cross,” and “Red Cross: How We Spent Sandy Money Is a ‘Trade Secret’.”