In the Journals

Use of ICDs not meeting Medicare criteria declined after DOJ investigation

Paul A. Heidenreich, MD, MS, FAHA
Paul A. Heidenreich

After a U.S. Department of Justice investigation into implantable cardioverter defibrillators, there was a decrease in the volume of ICDs for primary prevention and in the proportion of devices that did not meet CMS National Coverage Determination from 2007 to 2015.

This decrease was substantially larger at hospitals that reached settlements from the investigation performed by the DOJ, according to the analysis published in JAMA.

ICD registry

Nihar R. Desai, MD, MPH, assistant professor of cardiology and in the Institution for Social and Policy Studies at Yale University School of Medicine, and colleagues analyzed data from the National Cardiovascular Data Registry ICD registry regarding the implantation of 300,151 ICDs for primary prevention in patients with ischemic or nonischemic cardiomyopathy at 1,809 hospitals from 2007 to 2015.

During this time, an investigation by the DOJ began in 2008, which focused on ICDs that did not meet CMS National Coverage Determination criteria. Settlements were announced in October 2015.

The primary outcome was the proportion of initial primary prevention ICD implantations in patients who were Medicare beneficiaries that did not meet National Coverage Determination criteria at 6-month intervals. Other outcomes of interest included the proportion of initial primary prevention ICDs that did not meet the criteria in patients who were not Medicare beneficiaries and the volume of initial ICDs for primary and secondary prevention over time.

At the beginning of the analysis, 25.8% of hospitals who later reached settlements from the investigation (95% CI, 24.7-26.8) and 22.8% of those that did not reach settlements implemented initial ICDs that did not meet criteria (P < .001).

Decreases in implementation

Over time, there was a 16.1% absolute decrease (95% CI, 14.8-17.5) and a 62.7% relative decrease (95% CI, 59.2-66.1) in ICDs that did not meet criteria in hospitals that reached settlement (P < .001; P for interaction < .001). There was a 12.1% absolute decrease (95% CI, 11.2-13) and a 53.2% relative decrease (95% CI, 50.4-56) for hospitals that did not reach settlement (P for interaction < .001).

After the announcement of the investigation by the DOJ between June 2010 and June 2011, trends significantly differed between both hospital groups, with hospitals that reached settlement having larger and more rapid decreases (P for interaction < .01).

During the study, there was an absolute decrease of 1,703 ICDs (95% CI, 1,520-1,886) and a relative decrease of 32.8% (95% CI, 29.9-35.7) in the volume of ICDs implanted for primary prevention at hospitals that reached settlement (P < .001; P for interaction < .001). Hospitals that did not reach a settlement had an absolute decrease of 1,495 ICDs (95% CI, 1,249-1,741) and a relative decrease of 17.4% (95% CI, 14.8-20; P for interaction < .001). Secondary prevention ICD volume had more modest decreases or slight increases.

Patients who were non-Medicare beneficiaries had similar patterns for ICD utilization.

“Understanding potential unintended consequences of the DOJ investigation remains an important area for further study,” Desai and colleagues wrote. “The analyses of secondary prevention ICDs do not suggest that access to necessary procedures was negatively affected by the investigation and observed changes in secondary prevention ICD volume preceded the announcement. These analyses offer some reassurance, but further research into hospital responses to the investigation could offer additional insights about possible unintended consequences.”

In a related editorial, Paul A. Heidenreich, MD, MS, professor of medicine and health research and policy and vice chair of clinical, quality and analytics in the department of medicine at Stanford University School of Medicine, wrote: “Many inappropriate procedures were likely avoided, although future work is needed to determine whether unintended consequences resulted. With big data available to an increasing number of analysts, sizable financial incentives and growing reliance on appropriate use criterion, more whistleblower complaints may be anticipated. However, physicians and hospitals can lead (rather than be chased) by developing appropriate use criteria and participating in registries tracking appropriate care.” – by Darlene Dobkowski

Disclosures: Desai reports he is the recipient of a research agreement from Johnson & Johnson through Yale University to develop methods of clinical trial data sharing and works under contract with CMS. Heidenreich reports no relevant financial disclosures. Please see the study for all other authors’ relevant financial disclosures.

Paul A. Heidenreich, MD, MS, FAHA
Paul A. Heidenreich

After a U.S. Department of Justice investigation into implantable cardioverter defibrillators, there was a decrease in the volume of ICDs for primary prevention and in the proportion of devices that did not meet CMS National Coverage Determination from 2007 to 2015.

This decrease was substantially larger at hospitals that reached settlements from the investigation performed by the DOJ, according to the analysis published in JAMA.

ICD registry

Nihar R. Desai, MD, MPH, assistant professor of cardiology and in the Institution for Social and Policy Studies at Yale University School of Medicine, and colleagues analyzed data from the National Cardiovascular Data Registry ICD registry regarding the implantation of 300,151 ICDs for primary prevention in patients with ischemic or nonischemic cardiomyopathy at 1,809 hospitals from 2007 to 2015.

During this time, an investigation by the DOJ began in 2008, which focused on ICDs that did not meet CMS National Coverage Determination criteria. Settlements were announced in October 2015.

The primary outcome was the proportion of initial primary prevention ICD implantations in patients who were Medicare beneficiaries that did not meet National Coverage Determination criteria at 6-month intervals. Other outcomes of interest included the proportion of initial primary prevention ICDs that did not meet the criteria in patients who were not Medicare beneficiaries and the volume of initial ICDs for primary and secondary prevention over time.

At the beginning of the analysis, 25.8% of hospitals who later reached settlements from the investigation (95% CI, 24.7-26.8) and 22.8% of those that did not reach settlements implemented initial ICDs that did not meet criteria (P < .001).

Decreases in implementation

Over time, there was a 16.1% absolute decrease (95% CI, 14.8-17.5) and a 62.7% relative decrease (95% CI, 59.2-66.1) in ICDs that did not meet criteria in hospitals that reached settlement (P < .001; P for interaction < .001). There was a 12.1% absolute decrease (95% CI, 11.2-13) and a 53.2% relative decrease (95% CI, 50.4-56) for hospitals that did not reach settlement (P for interaction < .001).

After the announcement of the investigation by the DOJ between June 2010 and June 2011, trends significantly differed between both hospital groups, with hospitals that reached settlement having larger and more rapid decreases (P for interaction < .01).

During the study, there was an absolute decrease of 1,703 ICDs (95% CI, 1,520-1,886) and a relative decrease of 32.8% (95% CI, 29.9-35.7) in the volume of ICDs implanted for primary prevention at hospitals that reached settlement (P < .001; P for interaction < .001). Hospitals that did not reach a settlement had an absolute decrease of 1,495 ICDs (95% CI, 1,249-1,741) and a relative decrease of 17.4% (95% CI, 14.8-20; P for interaction < .001). Secondary prevention ICD volume had more modest decreases or slight increases.

PAGE BREAK

Patients who were non-Medicare beneficiaries had similar patterns for ICD utilization.

“Understanding potential unintended consequences of the DOJ investigation remains an important area for further study,” Desai and colleagues wrote. “The analyses of secondary prevention ICDs do not suggest that access to necessary procedures was negatively affected by the investigation and observed changes in secondary prevention ICD volume preceded the announcement. These analyses offer some reassurance, but further research into hospital responses to the investigation could offer additional insights about possible unintended consequences.”

In a related editorial, Paul A. Heidenreich, MD, MS, professor of medicine and health research and policy and vice chair of clinical, quality and analytics in the department of medicine at Stanford University School of Medicine, wrote: “Many inappropriate procedures were likely avoided, although future work is needed to determine whether unintended consequences resulted. With big data available to an increasing number of analysts, sizable financial incentives and growing reliance on appropriate use criterion, more whistleblower complaints may be anticipated. However, physicians and hospitals can lead (rather than be chased) by developing appropriate use criteria and participating in registries tracking appropriate care.” – by Darlene Dobkowski

Disclosures: Desai reports he is the recipient of a research agreement from Johnson & Johnson through Yale University to develop methods of clinical trial data sharing and works under contract with CMS. Heidenreich reports no relevant financial disclosures. Please see the study for all other authors’ relevant financial disclosures.