Evidence Base Lacking for Almost 1 of 5 Metastatic Colon Cancer Regimens

June 2011, Vol 2, No 3 -


Chicago, IL—Treatment regimens that are not based on evidence and therefore not included in treatment guidelines are used “with some frequency” in patients with metastatic colon cancer, according to academic investigators who teamed up with UnitedHealth Group to examine claims data. They presented their findings at ASCO 2011.

The average patient with a diagnosis of metastatic colon cancer had a 19% chance of receiving a regimen not backed by evidence. In the case of bevacizumab alone, this amounted to $1.3 million in costs for United – Healthcare.

Jonas A. de Souza, MD, a hematologist/oncologist at the University of Chicago Medical Center, examined the use of various regimens that include panitumumab (Vectibix), cetuximab (Erbitux), bevacizumab (Avastin), oxaliplatin (Eloxatin), capecitabine (Xeloda), and irinotecan (Camptosar) for metastatic colon cancer. Since the early 2000s, he said, the newer agents have “strikingly improved the treatment of colon cancer; however, they have increased the price of treatment.”

Median survival has increased from approximately 18 months in the early 2000s to more than 25 months today. The availability of cetuximab, bevacizumab, and panitumumab is largely responsible for this improvement.

One of 7 patients with cancer now spends >20% of household income on healthcare and insurance and often must choose less-effective treatments. “Not uncommonly we see patients referred to us having received substandard care,” he said.

Given the growing number of drugs used for this type of cancer, it is a “given” that some percentage of treatments will not be based on evidence, which has clinical and cost implications for payers, patients, and society. “It is important to identify low-value care,” Dr de Souza maintained.

“Our hypothesis was that non–evidence-based regimens for metastatic colon cancer are both commonly prescribed and reimbursed,” he said.

3 Regimens Singled Out
The primary objective was to estimate the utilization of 3 regimens addressed in the National Com – prehensive Cancer Network guidelines as not recommended in patients with metastatic colon cancer:

  • Single-agent capecitabine as a salvage therapy after failure on a fluoropyramidine- containing regimen (ie, 5-fluorouracil, capecitabine)
  • Bevacizumab combined with a second- line regimen after progression on a bevacizumab-containing firstline regimen (ie, continuation of bevacizumab after progression)
  • Panitumumab after clinical failure on cetuximab or cetuximab after failure on panitumumab (ie, sequential use of an epidermal growth factor receptor [EGFR] inhibitor).


The UnitedHealthcare claims database represents >26 million members and 700,000 physicians. The analysis included 7642 patients with incident colon cancer diagnosed between January 2007 and July 2010. Of these patients, 1041 (14%) received ≥1 agent used only in the metastatic setting.

Of 864 patients using bevacizumab, 90 received the drug beyond progression. This included patients receiving bevacizumab/oxaliplatin who progressed and went on to use bevacizumab plus irinotecan or vice versa. Of 121 patients receiving capecitabine, 49 received single-agent capecitabine after having used it in combination with standard regimens. For the EGFR inhibitors, 144 patients received cetuximab and 38 received panitumumab; 6 patients received panitumumab after progressing on cetuximab.

In total, non–evidence-based regimens constituted 10% of bevacizumabbased treatments, 40% of capecitabinebased treatments, and 16% of EGFR-based treatments. The average patient had a 13% chance of receiving a regimen for metastatic colon cancer that was not evidence-based.

A sensitivity analysis based on treatment and the original diagnosis identified 917 patients, and 600 patients were identified based on the drug regimen only. “We considered these true cases of metastatic colon cancer,” he said. Examination of this group showed that 139 (19%) patients had received ≥1 treatment regimens not based on evidence.

What Did This Cost Payers? The cost of inappropriate prescribing of antineoplastic agents was considerable. Bevacizumab was used non–evidence-based in 636 claims, meaning that for a 70-kg man receiving 5 mg/kg at a cost of $5.96/mg (average sales price January 2011), UnitedHealthcare paid out $1,326,696, Dr de Souza said.

Capecitabine was non–evidencebased in 218 claims, resulting in $621,463 in reimbursements. Panitumu – mab had 19 unsupported claims, resulting in $69,665.

Dr de Souza acknowledged that this was a retrospective analysis of administrative data, and no data on clinical outcomes were available.

Patient demographic information was limited, and investigators made assumptions regarding the diagnosis of metastatic colon cancer. They did, however, take into consideration that some patients may have been switched to unsupported regimens based on toxicity with standard treatment. If this occurred within 90 days, they were not counted in the analysis; the analysis included only those patients switching to unsupported regimens beyond 90 days of the initial therapy.

Dr de Souza speculated that physicians may prescribe regimens that are not backed by evidence because they have no knowledge of their lack of benefit, they have financial incentives for doing so, or they simply lack alternatives for these very ill patients. “Our findings support the development of clinical pathways and new payment methods that limit the use of non–evidence-based regimens and that promote high-value care,” Dr de Souza concluded.