Virtual colonoscopy (VC) offers the potential for effective colorectal cancer (CRC) screening at the fraction of the cost of the more traditional optical method, discussed Bruce Reiner, MD, director of research, VA Maryland Healthcare System, during a university session yesterday at the Society for Imaging Informatics in Medicine (formerly SCAR) annual meeting in Austin, Texas.
The theme of the session – economics of filmless imaging – focused on the economics of VC, as well as natural language processing (NLP) in healthcare and pay for performance (P4P) in medical imaging.
Today, VC is referenced as a possible screening method alternative for colorectal cancer that may improve compliance rates in the target audience – since the procedure is a minimally invasive alternative to optical colonography methods. But does VC yield clinical comparable results to optical colonography, and is it ready for prime time, Reiner posed?
While some say the answer is yes, conflicting results from clinical trials, along with nagging reimbursement issues, has hindered VC’s clinical adoption.
Progress has been made. “In July 2004, the Centers for Medicare and Medicaid Services issued specific category three CPT codes for VC,” said Allyson Mortati, global product manager, virtual colonography, for E-Z-EM. “There are a number of local coverage determinations in 29 states which cover payment of diagnostic VC under certain, specific conditions.”
“Screening VC is a little more difficult to get reimbursement for,” Mortati continued. Barriers for reimbursement include competition between radiologists and gastroenterologists, as well as the need for more studies that support VC.
Reiner further explained that in order for VC to become a cost-effective alternative screening method for CRC, the cost must be less than 60 percent compared to traditional optical methods, and patient compliance rates must be greater than 15 and 20 percent. “VC offers the potential for effective screening at the fraction of the cost of optical colonography, and it also provides the ability to significantly increase patient compliance,” Reiner said.
Natural language processing in healthcare
“There is no single way to approach natural language processing (NLP) because it is an ongoing area of research,” said Mark Morsch, vice president NLP and software engineering at A Life Medical Inc. in San Diego, Calif. “But there are technologies out there, which are not entirely perfect, for information extraction applications, both symbolic NLP and statistical and connectionist NLP.”
NLP, a term in use since the 1980s to define a class of software systems which handle text intelligently, can be used in healthcare for billing purposes but currently has a low penetration rate of 15 percent.
According to Morsch, related NLP technologies include information extraction, text summarization, text mining, text retrieval, machine translation, speech recognition, speech generation and voice identification and or authentication.
Current healthcare applications include computer-assisted coding, which helps assign procedure codes for billing, and outpatient coding and billing. Future healthcare NLP applications include:
- Inpatient coding
- Clinical text mining, including SNOMED-CT and outcomes analysis
- Integration with speech recognition
“Using an NLP engine will help us educate coders and radiologists, and it can also coach physicians on better dictation practices,” said Morsch.
Pay for performance (P4P) in medical imaging
VA Maryland’s Reiner recommended a number of ways in which P4P can be applied to the current digital imaging practice:
- Create quality performance metrics that transcend entire radiology processes.
- Enlist industry support to tie quality performance metrics to imaging modalities.
- Lobby for registration that ties reimbursement to provide education and training.
- Create a third party to establish, oversee and test universal quality assurance standards. “This can be done with real-time quality insurance, creating financial incentives for quality insurance and disseminating quality assurance metrics and results to the public,” Reiner explained.
- Support standardization and integration of referenceable databases.
- Track quality assurance performance from the patient perspective with satisfaction surveys. “Patient feedback data should be made available to the public,” he said.
- Tie medical malpractice rates to individuals quality performance metrics.