Will JANUS Jeopardize Personalized Medicine?

I have to take slight exception to Peter.   I am afraid that while better post market data is needed to inform product development in  ways that accelerate the targeting of treatment and enrich clinical trials, the JANUS warehouse is not the place FDA should go shopping.  

 

First it is important to know what’s in JANUS and what is not:

 

“The JANUS warehouse was populated with sample synthesized human trial data related to two oncology studies. This data was furnished in standard STDM format and data load scripts were developed to import the data. A sample caCORE application was developed to demonstrate the analytical capabilities of such an application accessing the underlying JANUS repository. The application was modeled against SDTM domain views of the Janus warehouse instead of the warehouse schema itself to alleviate complexity and improve data access efficiency. The sample application was successful in that you could authenticate the users at signon using CSM authentication and then review the clinical trial data in the following ways:

• View patient enrollment by study
• View patient retention by study
• Efficacy reporting by study
• Safety reporting by study

Note the absence of genomic data.

To be sure, according to HHS, “there is a JANUS pharmacogenomics pilot study to identify gaps in the RIM and suggest further improvement for the import of genomics data into the database. “

But the FDA contract does not fund such research.  Moreover, HHS under the Obama Administration is interested in using personalized medicine to save money above other goals. 

“The molecularly informed comparative effectiveness research (NCI) was set up to determine whether molecular information can inform quality, efficacy and cost-effectiveness in health care…The MI-CER project also requires the integration of health care financial data to examine cost-effectiveness."

SO how does the PACES use of Janus fit into all this?

According the FDA solicitation: “The Contractor shall develop innovative clinical trial design strategies for prospective CER clinical trials and analyses of healthcare data including providing formal recommendations for best practices for submission of studies to the FDA when they involve product comparisons. These strategies and recommendations shall be documented in reports and manuscripts suitable for publication in scientific journals.”

·     Conduct a review of a sample of FDA clinical trials performed to assess drugs, biologics and devices.  This review shall include clinical trial data used in pre-market submissions for one class of drugs or devices.

·     Evaluate comparator groups, populations studied, study designs, endpoints, statistical analysis methods, and trial conduct.

·     Include an evaluation of the capability of each study to distinguish differences in effectiveness and safety among different populations, subpopulations and individuals.

·     Provide formal recommendations for best practices for submission of studies to the FDA when they involve product comparisons.  http://tinyurl.com/2fb4fng

 
While a long term goal of analyzing JANUS data is identify robust data standards to capture and exchange clinical genomic data, that goal is not addressed or funded. 

Most troubling is that the vendors interested in submitting bids lack any expertise in achieving this goa.l   Many of them are superb and innovative providers of infrasture or developers of portals for delivering clinical data.  But the only vendor that is in the business of analyzing data has significant conflicts of interest in my opinion.  And that’s The Center for Medical Technology policy run by Sean Tunis.  When Tunis left Medicare in 2005 and set up the Center, the Agency for Healthcare Quality and Research was his first client.   Tunis also served on the Institute of Medicine panel that set the CER agenda for the Federal Coordinating Council on Comparative Effectiveness Research of which FDA is a part. 

Moreover, Tunis has a distinct bias against using any new technology absent the kind of evidence he believes is important.  Consider his self-serving white paper on how Designing More Informative Clinical Trials for Off-Label Uses of Oncology Drugs which has the goal of establishing CMT as the entity that will create standards and conduct studies using clinical data to compare a standard treatment to a newer use.  Tunis wants payors to require such studies before reimbursement is made.  And since CMT would be the standard setter, guess who would get all the CER business?

Tunis makes the same lame claim all CER advocates and contractors use:

“Despite the prevalence of off-label use of oncology drugs and related services, the health outcomes and value of expenditures on these products and services are not well

understood.”

Not understood?  Or does he mean not subject to the sort of one size fits all comparison of treatment A to treatment B that he would conduct as a consultant for reimbursement purposes?

Here’s more:

“Many patients, oncologists and other clinicians, payers and policy makers are dependent on these compendia, yet compendia often are viewed as being “too lenient” in terms of the quality of evidence required for compendia listing. Some studies have raised questions about the rigor and consistency of the compendium process. A 2006 analysis of 14 off-label indications concluded that current compendia “lack transparency, cite little current evidence, and lack a systematic method to review and update generated evidence.”

Let’s set aside the fact that the 2006 analysis was conducted by yet another member of the IOM/CER panel (Harold Sox) who is not an oncologist.  Does the fact that the decline in the rate of death and increase in five year survival rates for most cancers matter to Tunis?  Or how about the fact that by definition “off-label” use is based on clinical insights and forms the basis of common use and further research?  Or that research by Frank Lichtenberg shows about “one-fourth of the mortality decline in cancer is attributable to drug innovation, and 40% of the decline is attributable to (lagged) imaging innovation. ( Life expectancy at birth may have been increased by almost three months between 1996 and 2006 by the combined effects of cancer imaging and cancer drug innovation.)”

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1586687

 

Not at all. 

Does he really care about personalized medicine?

“Off-label oncology CERs should be designed to be generalizable in that they include sociodemographic diverse patient populations as well as patients with common comorbidities that exist among cancer patients and/or are positively or negatively associated with the use of oncology drug.”

http://www.cmtpnet.org/cmtp-research/Off_label_Overview.pdf

Generalizable.  Common.   But off-label use is treating the patient.  CER is used to make a generalization for purpose of coverage and reimbursement.   Tunis calls for prospective studies using clinical data to compare different drugs, not target treatment.  And he would use the JANUS data to advance his agenda. 

The current PACES project can be skewed to carry out such studies and has the potential to do so because of the lack of specific requirements to advance personalized medicine and integrate genomic analysis with clinical data.  

To give a large contract to an organization that has a vested interest in advancing it’s own approach to CER and has a bias towards using CER and consistently avoids establishing whether CER actually improves clinical outcomes.  Rather,  CER and Tunis focus on process and the creation of “Effectiveness Guidance Documents (EGDs)”  that would become the gold standard for CER research throughout the federal government and would include the FDA.

JANUS should be straightforward, not two-faced about personalized medicine.  Given the potential for bias and manipulation of the existing dataset, the FDA should impose additional requirements and different goals on contractors and itself.