Proventys’ Decision-Support Systems Promote Personalized Medicine
An emerging player in the clinical decision-support marketplace is Proventys, whose products fall into 2 distinct categories. The first is a broad, web-based National Comprehensive Cancer Network (NCCN) Guidelines–driven system called Proventys CDS Oncology, the second, called Proventys PDx Oncology, is the first-in-class in a new category of medical devices known as software-based predictive tests.
Proventys CDS Oncology has 2 distinguishing characteristics that may separate this company’s offering from some of the other clinical pathways programs. First, the decision support is based on NCCN Clinical Practice Guidelines in Oncology, which is the result of an exclusive collaboration between Proventys and NCCN.
Second, the company is focused on aligning the interests of oncology providers and payers through an innovative program built on the CDS Oncology system. This program, called AlignQI, is aimed at increasing transparency and collaboration between stakeholders to improve the quality of oncology care.
Proventys PDx Oncology is a predictive medicine system that employs a formal, statistically driven set of models to help plan care; the planned first component of this tool will help assess a patient’s risk of developing febrile neutropenia.
Because of the newness of this modeling approach, the US Food and Drug Administration (FDA) considers this a new category of regulated medical device, and Proventys has been working with the FDA to create a control guidance document that will aid its own approval and help guide future, similar technologies.
The release date for the predictive component is unknown, but Chief Medical Officer Surya Singh, MD, described the process of working with the agency as “very collaborative.”
The Proventys CDS Oncology system (the broad decision-support tool powered by the NCCN) is expected to be released by April 2011. The system fosters the ability to tailor care to the individual patient, and the Proventys notion of personalized care goes beyond advanced biomarkers and genetic testing, said Dr Singh.
A physician has dozens, if not hundreds, of variables to take into account in any particular decision, Dr Singh said, and it’s inevitable that some items will be weighted more than others, depending on personal experience. That personal experience is necessarily less extensive than the experience that would inform and weight a database.
According to an informal survey of his colleagues in Brigham and Women’s Hospital, oncologists take into account 60 to 80 variables in each treatment decision, which “is more than the human cognitive capacity,” Dr Singh noted. As with other pathways programs, the system is constructed to account for these variables.
The program, Dr Singh emphasized, includes the evidentiary consensus behind the guidelines. Not all pathways provide this primary evidence, and it is important to “present all options,” he said.
Incorporating NCCN Guidelines Updates
Proventys is working with NCCN to plan how regular guideline updates will be incorporated into the program. A potential advantage of the system, according to Dr Singh, is that it will help roll out updated guideline information through user prompts.
Incor porating NCCN updates into practice currently requires active information seeking by physicians. “I’m not saying that doesn’t happen,” said Dr Singh, but “it doesn’t happen in a comprehensive and systematic way.”
With the CDS Oncology system, according to Dr Singh, for the first time guidelines will be vetted against realworld practice in real time, without relying on retrospective claims data or electronic medical record (EMR) review. “If somebody deviates from guidelines, and they’re using this tool, we’ll know it.”
The system will also link to specific payer factors (within NCCN guidelines) that may influence a physician’s decision if all clinical factors are equal.
For example, by choosing a particular chemotherapy regimen among 6 recommended by NCCN for a particular condition, a physician may not have to arrange for prior authorization, or will get paid without an appeals process.
Communicating that message without restricting choice is “really important,” Dr Singh suggested. “We’re providing choice within the standard of care.”
Payer variables will not be hard coded into the programming, maintaining the flexibility to customize the program for different users. In addition, the CDS Oncology system will be customizable to integrate with various EMR programs.
Guidelines incorporated initially in the program will include breast, colon, non–small-cell lung cancer, and Hodgkin lymphoma; later in 2011, Proventys plans to add guidelines for non-Hodgkin lymphoma and smallcell lung, prostate, and rectal cancers.
Based on feedback from physicians, the program has been designed with an order entry component that forestalls the use of a separate or a solely paper-based system, and supports electronic ordering for diagnostic tests and supportive medicines. The program allows automated chemotherapy dose calculations and ordering, which is then printed out and signed.
Dr Singh indicated that of the 150 clinicians who have so far been involved in testing the product, only 1 had a negative reaction to it. “I expected that we would see a huge variety of responses on the predictive test side,” but that has not been the case.
There is a growing awareness of the utility of algorithms to support individualized patient care, and deciding to make this the initial topic for decision support was well chosen.
The recent pilot program by UnitedHealthcare to bundle payments for treatment of select cancers may offer additional opportunities for clinical support systems such as this, Dr Singh suggested. With the current disconnect between the patient and physician on the one hand, and the third-party payer on the other, the idea of episodic bundled payments is “very closely related” to these systems.
The economic beneficiary for these types of tools, Dr Singh said, is the payer, and organizations such as the Centers for Medi care & Medicaid Services will likely be very interested in them.
Evolving reimbursement models will necessarily be driven by information technology to increase efficiency through decision support.