Washington, D.C.
The Subcommittee on Standards and Security of the NCVHS was convened on May 25, 2004 at the Hubert H. Humphrey Building in Washingon, D.C. The meeting was open to the public. Present:
ACTIONS
The Subcommittee approved a letter to the Secretary addressing the DSMO's request that NCPDP 5.1 be adopted as a HIPAA standard for electronic billing for supplies by a retail pharmacy.
E-Prescribing Networks
Mr. Hutchinson detailed the e-prescribing process steps. NCPDP SCRIPT was not an undue administrative burden; he noted potential enhancements. The provider ID worked as the pharmacy identifier code set. He recommended the NDC number for packaged drug code sets. He advocated the MMA grant program for pharmacies and utilizing pilots to ensure MMA worked and was not a burden in the private sector. He encouraged NCVHS to define appropriate formulary/clinical alert messages, consider models in the banking industry, and building upon physician/pharmacist relationships to achieve patient safety.
RxHub divided the work flow process into: “pre-scribing,” scribing, delivery and fulfillment with real-time clinical decision support (CDS) and a rational set of choices. Mr. Bradley stressed a lack of infrastructure. Challenges were the patient ID and establishing the enrolled population and directory process. RxHub's directory service has 150 million members and utilizes algorithms and elements from drivers' licenses to identify individuals, coordination of benefits (COB) opportunities, and find drug coverage. RxHub delivers medication history in five-to-ten seconds. Benefit Coverage included formulary, preferred drug status, and preferences. Mr. Bradley emphasized national adoption and noted lack of an adequate business model.
Uniquely identifying sender/receiver was troublesome. ProxyMed recommended NCPDP as the pharmacy facility ID. Physician office vs. physician ID was an issue. Mr. Guinan stressed standardizing code sets. He approved of RxNorm” but said it had to be finalized soon. SIG lacked a viable code set. NCPDP SCRIPT was established for new prescriptions and refill authorizations, but content for some fields was proprietary and/or free text. He noted gaps in NCPDP SCRIPT and other standards and suggested building from NCPDP's message format.
Findings from RAND on E-Prescribing
Douglas Bell, M.D., Assistant Professor in Residence, RAND/UCLA
Dr. Bell noted many e-prescribing options were available with a variety of aims. Pfizer funded a study developing 60 recommendations for e-prescribing rated on: patient safety and health outcomes, helping patients manage their cost, maintaining patient privacy, and promoting clinician acceptance. Recommendations mapped to four federal goals in MMA: safety, quality, efficiency, and inter-operability. Some 29 products met inclusion criteria.
Patient Safety and E-Prescribing, David Bates, M.D.
Dr. Bates noted studies indicating: 6.5 adverse drug events (ADE) per 100 admissions, three potential ADE/near misses per preventable. A third of the ADEs were preventable. A fifth of the patients had a non-preventable and over ten percent a preventable ADE. Computerizing prescribing was the highest yielding intervention strategy in the outpatient setting, improving drug ordering by streamlining/structuring the process, providing information, and performing checks. Allergy detection, notification about drug-drug interactions, and dose suggestions were most important. He cited a 55 percent decrease in the serious medication error rate. He noted key CDS areas to work on, future directions, recommendations and issues.
State of the Art of E-Prescribing Implementation
The Council on Affordable Quality Healthcare's (CAQH) Formulary DataSource was a free searchable Web site where vendors could access a centralized formulary data file. Integrating Preliminary results of a one-year pilot indicated: drug interaction warnings were heeded four percent of the time; six percent of prescriptions were modified after allergy warnings. A fourth of the exclusions were modified to formulary. Generics could be substituted for over 75 percent of brand-name drugs. Ms. Thomashauer emphasized linking formularies to the patient. Challenges included: inconsistencies in traditional methods of data communications, complex prescription benefit offerings, and unavailability of pharmacy coverage information at point of care (POC). CAQH currently focused on providing a comprehensive centralized source and formulary check in the provider workflow.
Dr. Medvedeff provided an overview of the Florida Medicaid e-prescribing program focused on impacting over-prescribing with: drug information explaining appropriate use and expected outcomes, patient information, and checks against fraud and abuse. A clinical drug database was solely dedicated to clinical reference. The turnkey project with PDA and Web site support was offered to the 1,000 highest prescribing Medicaid doctors in Florida that represent 14 million scripts and one billion dollars per year. He noted a four percent reduction in significant clinical drug interactions in the physician's office and minimized fraud and abuse.
Dr. Mandel detailed a Blue Cross Blue Shield Massachusetts (BCBSMA) pilot program offering free high-speed access to eligibility and medication history. Incentives included receiving one dollar per-member-per-month for prescribing half or more of their prescriptions electronically. He noted physicians' interest in new technologies and low tolerance for difficulties. They wanted ROI that brought efficiency and productivity, recognized patient safety issues, but felt they were not making mistakes and staff could handle phone calls. Monitoring patient compliance seemed an overload. Noting a one percent shift in generic prescribing for BCBSMA represented a ten-million-dollar-savings, he said the savings was enough to support adoption.
Ms. Adams said prescribers in the Rhode Island Quality Institute (RIQI) statewide e-prescribing project found the value of the automatic renewal function unquestionable. Physician workflow and portability were a problem. The biggest struggle was with technologically sophisticated physicians' parallel systems. Seventy percent of Rhode Island's pharmacies were connected. Some 250 physicians were activated; 1,100 would sign on in the summer 2004. She emphasized collaboration and working together to fix integration problems. RIQI's next project is to imbed e-prescribing into EHR and CDS and connect them. RIQI will pay 200 physicians $25,000 each per year to participate. Predicting 30 percent of the waste could be eliminated, Ms. Adams said capturing 11 percent would produce a 77 percent ROI.
Open Microphone
ExpressScripts cautioned about solving a problem with OTCs or going to EMRs and requirements without a supporting infrastructure. Testifiers advised creating a standard that “allowed for perfect but provided good enough” or adopting existing standards and not forcing what would not work. Terri Byrne from RxHub noted downloading information to vendor applications and real-time transactions depended upon keys to find patients, the plan or group. Ross Martin from Pfizer suggested couching comments in terms of ultimate benefit; he supported tiered implementation strategy focused on what was important for the patient. Dr. Bizzaro expressed concern about classification systems that were hierarchical, unique and proprietary to compendia. Kathy Hollinger from FDA noted research on package inserts found no single classification system worked. Dr. Whittemore said the percentage of unfilled prescriptions was closer to 20 percent than half. Anne Canfield with Rx Benefits Coalition suggested preemption was so broad that what happened at the state level was irrelevant.
Subcommittee Discussion
Dr. Huff asked if there was a benefit in tackling non-controversial issues sooner. He suggested organizing recommendations into: identifiers; messages; codes; rules, knowledge representations and classifications; functional requirements; and digital signatures. Members noted some things recommended for applications might not be fit for national adoption. Members will hear in July and August from software developers', private-sector drug knowledge bases' (KB), and providers' perspectives, and then take a first-cut at “the low-hanging fruit.” Standards organizations will provide feedback in September. The Subcommittee will determine interim recommendations in November. Members will move forward as appropriate; hoping to announce final standards earlier than the April 2008 date.
E-Prescribing: The Software Vendors Perspective
Mr. Doerr said the electronic problem list and (ICD-9-CM) diagnosis codes were needed for prescription. He noted benefits of diagnosis-driven workflow, said standards enabling messages about lower cost alternatives would be most effective when related to diagnoses treated, and identified trade offs between efficacy, cost, and contraindications. A study showed payers' savings-per-doctor-per-month increased 8.3 percent ($584) the first six months and 15.7 percent ($1,101) the next six months. Payer ROI was 6-to-1. He said requiring diagnosis codes would not impede workflow; interfaces between e-prescribing systems and practice management systems (PMS) were the biggest hurdle. NCPDP did not resolve barriers to interoperability caused by prescribing apps, networks, PBMs and PMS using competing code sets for names. State requirements were barriers. Standardization at the national level, standardized tables for managing formulary information, and policies defining formulary were needed.
Dr. Faughnan noted McKesson's relationship with First DataBank (FDB), use of SNOMED CT for patient conditions that might impact e-prescribing, and plan to use LOINC. McKesson did not find ICD-9-CM suitable for checking and used NCPDP SCRIPT and other messaging structures for formularies. RxNorm was supported as an external norm. McKesson looked forward to SIG. He advocated addressing process and structural issues with NDC in an evolutionary way and encouraged SNOMED/FDB collaboration for patient conditions. He noted code gaps and adoption issues, underscored the importance of incentives for collaboration, and suggested a trusted entity monitor all code sets.
DrFirst supported connectivity, RxNorm for naming, and a standard SIG. Dr. Kaufman called for a unified naming system (e.g., RxNorm), standard classifications (e.g., RDF-NT), and mature descriptive language with SIG (e.g., SCRIPT). Continuity of care (CCR) was an important early step. He illustrated the back-end connectivity network and complexity of e-prescribing, stressing connectivity. He said physicians' main concern (that technology was not fast enough) was ungrounded and that they would sign up, once they understood the benefits.
Ms. Helm said funding and incentives increased physician participations. Increasing generic and driving formulary compliance increased cost effectiveness. Automated prescribing and refill halved the phone calls. Health plan IDs, a core set of form attributes, data element and term standards, and support for multiple standards were needed. NCPDP SCRIPT was solid and expandable; challenges stemmed from lack of appropriate national identifiers and lapses in operational processes. NDC was useful for pharmacies communicating dispensed medication information to payer/PBM. She emphasized that standards had to drive measurable outcomes for efficient connectivity, flexibility, and a natural transition path guided by provider incentives.
Mr. Nayak emphasized understanding workflow and adoption issues. The physician identifier was needed and encouraged standardizing legal/compliance matters with pharmacy boards. He cautioned that a manual process for controlled substances would limit efficiencies. He noted MPI could link the formulary list and patient for real-time eligibility for medication insurance. He suggested notifying the physician when dispensing a drug would help make this a “true” medication history. He cautioned about inserting too much information into physicians' workflow and advocated incremental steps, creating system flexibility, and focusing on the physician and what was not a hindrance or took more time in the prescription cycle.
Mr. Nutkis said a third of the health plans and BASF use Zix secure messaging. Zix worked around a lack of standards; delivered services via Web, PDA, cell phone or wireless; provided real-time lab results and linked to diagnosis code at point of prescribing. Lack of deployment stemmed from uncertainty about emerging technology. Zix used NCPDP SCRIPT, X12 270/271 and HL7 standards; interfaces were “heavy lifting.” Lack of incentive, dollars and sophistication were issues. Nationwide adoption and lack of a standard for vetting and authentication were issues. Organizations wanted embedded e-prescribing. He emphasized understanding where formularies “were going” and SDOs' role in DUR. Time requirements for point-to-point were a challenge. Incentive programs compensated on a pay-for-performance basis for e-prescribing. Zix believed Medicare would be the de-facto standard.
E-Prescribing: The Drug Knowledge Base Vendors
KB vendors collected, codified, linked to and distributed applications useful at POC. They initially supported claims and inventory management; used NDCs, UPC, and health-related ID; and linked pricing and images. Clinical content, aggregated to formulation level, is linked to NDC. Therapeutic classes embedded in Medicaid/Medicare were based on common class descriptions of drugs; nomenclature for hierarchical conditions was embedded in the drug indications database. A name-based drug identifier or NDC is used to place ingredients in the context of the patient allergy. Allergen information had no standard; SNOMED-CT was considered for interoperability. POC and navigation was supported by grouping vocabulary content into concepts, allergens, disease and lab. A subset, proprietary disease vocabulary, expressed pharmaceutical concepts with flexibility and linked standard terminology. KB vendors welcomed structured labeling, hoping to gain internal efficiencies. Complementary content would come from evidence-based approaches. The goal was for the end-user to write as always, with converting to RxNorm concept “behind the scenes.” FDB was open to a non-competitive form aggregating this information.
Dr. Bizzaro urged NCVHS to delineate between e-prescribing and “screening” functions. RxNorm was a name-based identifier representing how physicians think; but pharmacy needed NDC. Gaining consensus for a standard signature was “thorny”: systems writing orders had numerous proprietary code sets, cross-referencing was difficult. Codified, retail pharmacy might still require textual representation as a quality check. Noting duplication of clinical screening efforts, he suggested messaging could explain responses to alerts. He cautioned about requirements so complex it was difficult to implement and retain acceptance.
Dr. Millikan called NCPDP SCRIPT a solid foundation; nationally adopted code sets for drug products, descriptions and instructions were needed. He supported: RxNorm for identifying drug products; text fields and proprietary code sets; standards for instructions (SIG); incorporating drug-drug warnings, cautions and dosage alerts and a summary of DUR actions. Medi-Span recommended SCRIPT, necessary code sets and RxNorm for interoperability between proprietary drug terminology code sets. NCPDP's strengths were widespread use and acceptance; codified or partial codification of SIGs would enhance it. HL7 and NCPDP were creating structured standards. Gaps include SIGs and a code set for drug allergens.
Multum was a reference terminology. Patient medical history existed in some prescriber settings and could enable logical choices determining appropriate drug therapy. Multum had various severity levels of interactions; 96 percent could be overridden. Structured product labeling (SPL) would help identify active ingredients for allergy checking and FDA-warnings. Mr. Hobson urged FDA to consider inconsistencies provided by manufacturers and implications of drug actions. He supported NCPDP and SNOMED terminology for conditions.
Follow-up on HIPAA Implementation Issues
WEDI consolidated their recommendations into those needing immediate attention and ongoing and strategic longer-term issues. Mr. Jones emphasized (1) continued focusing on full compliance while maintaining contingency plans with timelines to “move beyond,” (2) renewed effort to identify and promote benefits of other transactions and move toward them (contingency plans had to be addressed and data content standardized), (3) a requirement for use of a standard acknowledge (e.g., the 997) and more definition on area reporting, and (4) comprehensive review of COB issues, especially the health plan identifier.
Other recommendations regarding COB include instructions on populating crossover items and addressing problems crossing between payers when doing COB at line or claim levels. Ongoing implementation issues include developing standardized test data, expediting change standards, addressing timely data content issues, restructuring to annual code set maintenance, and specifying code values on instructions. Priority strategic implementation issues were: ways to draw providers into standards development; balanced DSMO representation; and boosting participation in standards development. Pilot standards need to identify changes needed before adoption and validate costs and benefits prior to the final rule. WEDI recommended enhancing the implementation process.
Follow-up on DSMO Request - Donna Pickett, NCHS, CDC
The request concerned use of the standard for building and supplies and use of the code sets in relationship to supplies. Universal Product Codes (UPC) and Health related item (HRI) codes were external code sets used within the NCPDP standards. UPC was assigned to OTC products; HRI code sets identified and numbered device packages. Neither UPC or HRI codes were precise, unambiguous or met all guiding principles under HIPAA. Product-driven, the codes would not be updated in the open process vital to HIPAA. Ms. Pickett noted Ms. Gilbertson referenced consequences the industry would face if it could no longer use them.
Development of Potential Letter - Simon Cohn, M.D.
Members reviewed the draft letter to the Secretary addressing the DSMO's request that NCPDP 5.1 be adopted as a HIPAA standard for electronic billing for supplies by a retail pharmacy. (See Actions, above.)
E-Prescribing: The Provider Perspective
Ms. Narcisi noted many specialties had to be considered and flexibility/scalability was needed. RxNorm might be suitable for VA, but not the private sector. Demonstrations had to address workflow and business rules in various settings. Noting the many interchanges in e-prescribing, she said standards were needed. Flexibility could enable doctors to consider critical factors without undue financial burden. PMS and EHR must be accommodated and the physician–patient relationship enhanced. AMA sought a private-sector approach with federal participation. An extensive rule-making process should not hinder maintenance/modifications. State issues must be reconciled. Clinical quality improvement and patient safety were drivers. New technology had to be reliable; false-positive alerts minimized. She emphasized lessons learned from HIPAA. Development and pilot testing were essential, but there was a lack of interested early adopters. AMA estimated the percentage of physicians using e-prescribing in the single digits. User authentication was a key enabler: the process had to be simple, quick, reliable and flexible. The Florida Medicaid program used the AMA Internet ID for pharmacy transactions. NCVHS and HHS had to understand the future Medicare environment: anticipating an annual five percent reduction in Medicare reimbursements from 2006 to 2012 and cost of care exceeding inflation, CMS estimated physicians pay will be cut 40 percent.
Dr. Hale emphasized e-prescribing must be useable in multiple settings, interoperability, freedom of choice for physicians and patients, incentives, and that everything must roll toward EHR. She stressed a need for: universal coding and nomenclature, simple easy-to-implement standards, adaptable e-prescribing scaleable to EHR and conforming to DEA/HIPAA security standards, transparent systems enhancing physician/patient relationships and helping overcome acquisition costs, and pilot testing. System features and functionalities should: prove e-prescribing speedy and efficient, integrate, utilize other KBs for CDS, enable prior authorization integral to the process, accommodate choices, and be free of commercial bias while providing pricing information enabling best choices for patients. Code set gaps include standards reconciling state requirements; standards for doctor-level drug name and ordering format; standards for SIG needs; format, vocabulary and structured formulary data and alerts; and messaging for all entities. Incentives suggested: grants for regional health information networks, grants for pilot studies assessing implementation in multiple practices and methodologies for implementation, and Medicare pay-for-performance and reimbursement differentials. Incentives should encourage a focus on EHR.
Dr. Kibbe reiterated the need for common nomenclature and expressed the American Academy of Family Physicians' (AAFP) concern that standards not hamper progress toward EHR. Implementation had to be easy for small/medium-sized offices. E-prescribing should be voluntary, free to patients and physicians, and without unfair incentives for commercial pathways or forced surrender of data for commercial use. AAFP wanted information from all sources. He said CCR might help EHRs exchange medication lists physician-to-physician and noted the full history was critical to making the promise of EHR real.
Ms. Schumann urged NCVHS to look beyond the physician and consider: participants in the process, settings and workflow, minimum data set components, regulatory issues, and confirm educational system content. She noted vulnerable populations (e.g., K-12 schools and college health, reproductive health, mobile clinics, gerontology patients, self-pay and indigent care, and hospice) and issues.
Open Microphone
Mr. Martin from Pfizer emphasized SDOs' understanding providers who could not always defend their interests, structuring RxNorm to capture differentiating features, and articulating a hierarchy of needs for RxNorm in a structured way. Testimony from ExpressScripts stressed gathering pre-data to measure after implementation and proposed standardized extracting methodology for getting at PMS data. A testifier cautioned about restricting system capabilities so much that possibilities for cost savings were lost and noted driving from one drug to another could benefit patients. Dr. Bizzaro contended the marketplace drove best systems forward. Mr. Martin said clinical information provided at POC for prescribing should not be created and managed by anyone with financial interest. Dr. Cobuzzi from VA said e-prescribing standards should be as broad as possible and functionality of collaborative practice agreements between pharmacies and physicians worked well in some states. Mr. Robinson pointed out there was no proposal for a centralized patient record and the lack of a patient-centric and holistic, disease management approach. He suggested a CMS initiative for case management fees as a way for gain an agent for the patient. Ms. Cronin said FDA was considering pay-for-performance with e-prescribing.
Wrap Up/Next Steps
Members will consider this hearing at the June break-out session. If enough Subcommittee members are available the Quality Workgroup June 24-25 meeting reviewing changes to billing and data collection practices related to HIPAA administrative and financial transactions might become a joint hearing. Hearings scheduled for July 29-30 and August 17-18 might be consolidated into a July 28-29-30 meeting.
The details of all presentations and letters drafted can be found in the meeting transcript posted on the NCVHS Web site, www.ncvhs.hhs.gov
-DAY ONE-
E-Prescribing Networks
Mr. Hutchinson said the National Community Pharmacists Association and National Association of Chain Drugstores were creating open, neutral and secure system compatibility. Over 60 percent of retail pharmacies tested on network by the fourth quarter 2003. Electronic-prescribing (e-prescribing) was activated in eight states.
He noted numerous steps in the e-prescribing process: reviewing the patient's chart, patient examination, diagnosis, writing the prescription and noting it in the chart, and refill/renewal. Pharmacy tasks include checking for allergies and other conditions, insurance information, input into the pharmacy record with DUR check, billing to insurer with additional DUR check, resolution of issues, dispensing and patient counseling. He cautioned against focusing too narrowly; the goal was to improve the overall process.
Mr. Hutchinson said NCPDP SCRIPT did not present an undue administrative burden on providers or pharmacists. Further enhancements could include: eligibility, compliance information, lab data, diagnosis, disease management protocols, drug therapy profiles, and transfers. He noted collaboration with HL7 to enable SCRIPT to be more widely used in inpatient and long-term care.
He said the provider ID worked well for decades as the pharmacy identifier code set and had no weaknesses. He recommended the NDC number for packaged drug code sets, noting its strength was universal acceptance. NDC's weakness was that it did not specify ingredients and there were no standards for ingredient terms. Noting location was a key deficiency, Mr. Hutchinson said SureScripts created its own SBI, route and location ID. Other gaps were lack of a nationally accepted SIG database, standard MPI/patient ID, or cross-reference between proprietary drug ingredient databases.
Mr. Hutchinson recommended expanding the MMA grant program to pharmacies. He advocated utilizing the pilots to ensure the MMA did not impose undue administrative burden and worked in the private sector. He said formulary management was important and should be supported, but contended NCVHS's biggest challenge was defining appropriate formulary or clinical alert messages: emphasizing that physician and patient choice should not be altered. He encouraged the Committee to review the banking industry for a model and the industry to build upon physician/pharmacist relationships to achieve patient safety.
E-Prescribing Networks
Mr. Bradley asked everyone to consider the history of drug benefits that went from adverse selection through bellwether events (e.g., physicians adopting technology, a drug benefit for senior citizens providing sufficient mass for the physician's office). RxHub dissected the work flow process into: “pre-scribing” (decision support on the front-end of prescribing), scribing (actual writing of the order utilizing front-end information), delivery (order sent to pharmacy of choice), and fulfillment (filling, generating label with instructions, refilling information, and dispensing). He noted the process included access to CDS, sharing patient history, facility integration, and an efficient business framework. He emphasized that determining eligibility for the drug benefit required exchanging information on an interactive real-time basis and a rational set of choices. Mr. Bradley cautioned that, without this structured approach to benefit and decision making, the benefit could bankrupt Medicare.
He stressed the lack of necessary infrastructure, noting the patient ID was the biggest challenge. Mr. Bradley was cynical about this being accomplished before the benefit rolled out. The next challenge was establishing the enrolled population and directory process necessary to keep the system from breaking down.
RxHub's solution to MPI was a directory service containing 150 million members that utilized sophisticated algorithms to ensure correct selection (no false positives). In under a quarter of a second, five elements from drivers' licenses identified individuals and COB opportunities. While the system was not a database and did not house patient eligibility or medication history records, it pointed to drug coverage. RxHub recommended the X12 270/271 transaction for physician offices for eligibility. Mr. Bradley said NCPDP eligibility was harder to adopt and not recommended. He said RxHub could deliver physicians' first interest, medication history, in between five-and-ten seconds with a standard already in the NCPDP adoption process. He noted that combining the payer and dispensing pharmacy databases would add over-the-counter (OTC), co- and self-pay drugs to what RxHub had.
Benefit Coverage included formulary, preferred drug status, and preferences. Mr. Bradley believed national adoption was essential. He noted the prescription process was complex and included new, refill, and change requests. RxHub planned to take standards lacking for many aspects through the NCPDP process.
He reported an uptake in technology vendors resulting from the passage of MMA, and pointed out a lack of an adequate business model for physicians to adopt these products. The commercial side was exploring incentives from payers for physician adoption and, potentially, lower reimbursement when not adopted.
Mr. Bradley emphasized standards were needed, including for affordability. Patients had to be identified and an efficient market where decision makers had information to make rational clinical and economic decisions had to be created. RxHub was willing to adopt other standards if necessary; it was time to collaboratively build-in capabilities everyone needed.
E-Prescribing Networks
Mr. Guinan said ProxyMed served 150,000 physicians and 400 PMS, handling a billion transactions per year. Market segments include: routing new and refill authorizations, providing patient medical histories and patient eligibility, influencing which drugs were prescribed to facilitate a higher level of patient safety, benefit plan compliance and/or cost savings, generic utilization, formulary compliance, and ADE avoidance.
EDI routing of transactions was similar for claims, ERA, prescriptions and lab results. Uniquely identifying sender and receiver was troublesome. ProxyMed recommended continuing to use NCPDP as the pharmacy facility ID. Mr. Guinan noted physician office vs. physician ID was an issue; pharmacies used the office phone number for routing. Currently, proprietary numbers were assigned by networks and facility and provider IDs were proprietary.
Mr. Guinan emphasized standardizing code sets. All systems used DEA for identification of authorizer, but not all prescribers had one. DEA “frowned upon” this use. NPI was a viable though still unavailable alternative. Proprietary ID filled the gap. He noted that NDC code sets used for claims did not support advanced software features and were not viable for prescription; RxNorm was “going in the right direction” but had to be finalized quickly. SIG also lacked a viable code set; currently all systems used free text which also did not support advanced features. Content for some fields of NCPDP SCRIPT was still proprietary and/or free text. Mr. Guinan said NCPDP SCRIPT was well established for new prescriptions and refill authorizations; all aspects of the marketplace actively participated in the SCRIPT work group and adopting another standard for moving these messages would be a burden and introduce greater risk. Noting gaps in NCPDP SCRIPT and additional standards that were addressed but not finalized, he suggested building from NCPDP and its message format.
Discussion
Mr. Blair commended the testifiers for the value they set forth before the government endorsed national standards. Mr. Hutchinson said, currently, preauthorization checks were done primarily at the pharmacy level by calling the physician's office.
Mr. Bradley said RxHub developed a point-to-point encrypted transaction used to exchange preauthorization or other messaging in an encrypted envelope. Processes could not be invasive or stop work: e.g., status of drugs ordered and consequences of alternatives ought to be apparent in “one mouse click.” He noted this facilitated step therapy. He emphasized preserving patient and physician choice. Medication histories varied depending on the depth of the application in the physician's office. RxHub downloaded when appointments were made or the patient checked in; each time required the patient's permission. DUR, drug-drug, drug-disease and other contraindications were screened. RxHub offered either a text report or a transaction inserted through the ADT system.
Mr. Guinan agreed it was critical that the physician could select and receive CDS concisely. Databases with formularies and pre-authorizations distributed via CDs on a quarterly basis were burdensome, had inadequate version control, and expensive. He recommended online transactions allowing for a request response and noted RxHub was putting forth a standard. Mr. Guinan noted physicians found medical history and DUR valuable and both lacked a standard. NCPDP had done much work in this area that was not formalized but could be the “jump-off point.” However EMR vendors were used to representing medical information with HL7 interfaces. He emphasized that either could be used, but a decision had to be made.
Mr. Bradley said medication history in ambulatory was evolving from NCPDP SCRIPT without much conflict; but the clinical side had to “learn to speak HL7” because of the magnitude of the problem in the inpatient setting. Mr. Beckley noted pharmacy could communicate back to the prescriber reactive, the issue was becoming proactive. Mr. Hutchinson emphasized that information from many sources must be combined. SureScripts viewed DUR as a multi-stage process: physician-level for drug checking, pharmacy against their store and payer/PBM.
Mr. Bradley said RxHub developed a minimum dataset to solely perform patient ID and point to drug coverage without replicating information contained by competing organizations. RxHub identified the organization and their unique ID to track individuals. Mr. Guinan said holding the routing element to the facility level and only including the prescriber in the record would help secure IDs when submitting prescriptions. Mr. Hutchinson disagreed; SureScripts implemented directly to the provider, allowing physicians mobile access in multiple settings.
Mr. Guinan said lack of continuity between ambulatory care, hospital and discharge was a prime driver of patient medication errors that had to be addressed. An ambulatory prescribing process upon discharge had to be equally aware of medication history, DUR and benefit rules. Mr. Hutchinson noted that NCPDP script was prevalent with prescribing vendors in the ambulatory environment, but HL-7 was standard of choice for more established electronic medical record (EMR) companies. Many wanted to send to pharmacy. SureScripts was looking at a partnership between HL-7 and NCPDP for some level of translation.
Noting tension between some CHI and PMRI initiatives' recommendations (NDC and clinical drug code sets expressed by RxNorm), Dr. Steindel asked if fully populating RxNorm into UMLS and tying it into NDF-RT would meet the needs of medication coding and if anyone had looked at SNOMED for SIG. Dr. Whittemore said SureScripts had not looked at SNOMED but was analyzing RxNorm: if connected to NCPDP SCRIPT, SureScripts could possibly “make it work.” Mr. Bradley said RxHub concurred with standardizing the ordering process; the challenge was physicians' reluctance to deal with complexities of NDC coding down to the component level. Physicians saw RxNorm as the answer. Pharmacy still used NDC for dispensing and the two would be tied. Claim would close the loop with NDC. Clinicians saw SNOMED as the answer for coding for SIG. Mr. Guinan agreed with RxNorm's goals, but noted pharmacy could not manage inventory without an interface to NDC and that HIPAA mandated using NDC on their claims.
Mr. Bradley explained that eligibility was payment while formulary was a different transaction and could be in real time, yet was invasive and slowed the work process. Batch formulary loads with indexing structure allowing a pointer better suited most applications' work flow. The processes should be combined but were different. COB opportunity existed, but required data not included in some standards. Noting human intervention in COB was not efficient, he recommended 270/271 for COB.
Mr. Beckley said SCRIPT had much of the foundation and could be used; speed was the issue. Noting eligibility requests were often made in batch and everything came back except prescription issues, Mr. Guinan suggested the prescription eligibility in 270/271 be included.
Asked how Medicare Part D and COB would work, Mr. Hutchinson said public-private partnerships would be “coming down the road” with Part D. Asked about the scope and flow of clinical information, he noted HL7 was default standard for exchange of clinical information, even between inpatient and outpatient. Payer to PBM to Physician and DUR in pharmacy did not use HL7. History was not delivered to pharmacy; lab/diagnosis information would help community pharmacy counsel patients. Standards should address how many multiple copies exist with primary care and referring physicians. Calling physicians caused inefficiencies.
Mr. Bradley noted that medication history has been downloaded from the payer database to the physician's office and stored with patient permission. Lab values could flow through the same pipe, but only represent 30 percent of “the world.” What has made medication history effective is consolidation of PBM: three companies reach 70 percent of the commercial world.
Mr. Guinan noted HL7 was primary transport for Lab data and NCPDP SCRIPT was the primary format for e-prescribing. Lab and medical history had to be the same. NCPDP SCRIPT was not too difficult to implement and could have separate transport protocol. Acknowledging that one request from the physician would be nice but might not happen soon, he suggested that not “marrying” the two would not be an impediment.
Mr. Bradley commented on the patient's role in fleshing out the medical history by contributing OTC and alternative medicines. A number of organizations advocated a patient-controlled-and-contributed record and there was a movement within the health insurance industry towards more consumer-directed care.
Mr. Guinan agreed that not mentioning the patient sooner in that morning's discussion was a significant omission. Patients wanted to know what the physician saw and give or withhold permission. They did not always want physicians to know everything (e.g., the stigma against alternative medicine).
Dr. Whittemore explained that prescribing and dispensing medications was regulated primarily (DEA governed controlled substances) at the state level and e-prescribing was not always recognized or the technology understood. A number of issues remained. He noted the National Association of Boards of Pharmacy changed their view and advocated eliminating handwritten scripts; the Model Pharmacy Practice Act will be changed.
Findings from RAND on E-Prescribing - Douglas Bell, M.D., Assistant Professor in Residence, RAND/UCLA
Dr. Bell noted many e-prescribing options were available with a variety of aims: workflow, increasing formulary adherence, and need to ensure the patient's interests were represented. Pfizer funded a study to develop recommendations for e-prescribing that promote patients' best interests, do not hinder adoption, ensure privacy, and are supported by a rigorous objective multidisciplinary process. A summary of evidence drawn from a literature review was fed into a Delphi expert panel process that produced recommendations published that day in Health Affairs. A field study assessed current systems in comparison with the recommendations and looked at how they were being implemented in the marketplace.
The literature review published in Journal of the American Medical Informatics Association looked at specific features of e-prescribing and effects that might be anticipated from those features: 1) menus for dosage selection and calculations reduced dosage errors, 2) safety alerts reduced ADE, 3) automated orders for monitoring tests reduced monitoring errors, 4) select medication by diagnosis improved guideline adherence, 5) and alerts increased formulary adherence. Dr. Bell noted the literature was weak, yet emerging evidence suggested under use of highly effective medications might be bigger than ADE.
The expert panel and project staff developed 60 recommendations that the panel rated on four dimensions: patient safety and health outcomes, helping patients manage their cost, maintaining patient privacy, and promoting clinician acceptance (achievable in the physician's office within three years). Recommendations were organized into categories. The first seven correspond to steps a prescriber would follow in caring for a patient: patient ID, access to patient history, medication selection, alerts and other, patient education, data transaction, and monitor. Three categories were broader recommendations applying across caring. Dr. Bell noted patient safety was high, cost mostly moderate, and clinician acceptance mainly high. Only recommendations specific to privacy impacted it.
Initial results of the field study found 58 e-prescription products from 51 of 129 companies (software as used by end-users) screened. Some 29 products from 26 companies met inclusion criteria: outpatient, significant adoption (1,000 users at 50 sites or more based on vendor claims), 10 representative product sites (five EMR, five handhelds; four Web ASPs, six locally-installed client server systems) were selected. Dr. Bell reported half the standards were implemented. EMRs implemented more than non-EMR products. Noting the degree of variance (30 percent to 77percent) among systems, Dr. Bell suggested one interpretation was that the marketplace did not yet create convergence on patient-centered features.
Dr. Bell said the panel's 60 recommendations to be used in creating standards mapped to four federal goals in MMA: safety, quality, efficiency, and inter-operability. He noted commercially available systems did not consistently implement important features. Recalling MMA requires evaluation of standards in 2006, he said more study was needed where evidence was limited. Issues to explore include: how e-prescribing could reduce underused, was transparency enough to deter biases or should third-party sponsorship be banned and an audit mechanism look for potential biases, was there room for human judgment, should unintended consequences be monitored, and the need to regularly update standards.
Discussion
Dr. Bell said RAND's recommendations were intended as non-technical standards that might be useful in establishing both bounds for competition among systems and "aspirational capabilities." RAND saw E-prescribing as a subset of EMR functionality that some companies broke off into a separate product.
Asked if RAND's study could identify points of diminishing return for acceptance within a three-year timeframe, Dr. Bell cautioned that potential safety benefits would require changes in what doctors did daily and too much alteration of workflow would result in pushback. Dr. Cohn said realizing reality had not come as far as their vision was helpful and noted the need to ensure practicing physicians' input on acceptance.
Asked how faculty knew about work flow and what was taught about the work flow technology would require, Dr. Bell said there was some awareness on medical education list serves and in literature, but he was not aware of any evidence or studies.
Asked about transparency and accountability recommendations, Dr. Bell noted concern that a third-party sponsor could bury responses in CDS rules. None were found, but there was concern that the prescribing process could be biased in ways not readily observable. Decision support rules had to be monitored to help eliminate bias. He noted display of options in a commercial manner was prohibited.
Patient Safety and E-Prescribing - David Bates, M.D.
Dr. Bates presented findings of a study of two large hospitals in Massachusetts indicating 6.5 ADE per 100 admissions. About a third was preventable. There were three potential ADE or near misses for every preventable ADE. Some 62 percent occurred in ordering and transcription. Another study in two general care and one intensive care units found 530 medication errors in 10,070 orders, 35 potential and five preventable ADE. Some 18 percent of 2,000 outpatients in Boston clinics surveyed in another study reported problems with their medications. Some 35 percent of the medications were not changed, 20 percent of the symptoms lasted more than three months. Three percent had ADE on chart review and five percent required hospitalizations. Thirteen percent were preventable. Dr. Bates said the study did not determine the “true” frequency of ADE. He emphasized that ADE had an impact on outpatient results, effecting both patients and the system (48 percent of the patients sought medical attention, 49 percent worried). A forth study of 629 patients published in The New England Journal of Medicine indicated that 20 of every 100 patients had a non-preventable ADE; 11 patients had a preventable ADE. Dr. Bates emphasized the importance of monitoring in nursing homes, noting a study published in The American Journal of Medicine found 1.9 ADE per 100 resident months. Little data on admissions due to ADEs was available. Some 1.4 percent of admissions to one hospital were due to ADE. All originated in outpatient; 78 percent were severe. Incidence and severity of ADE after discharge were 19 percent in another study.
Dr. Bates said computerizing prescribing improved drug ordering by: streamlining/structuring the process, providing information (e.g., relevant labs, guidelines, guided dose algorithms), and performing checks in background (drug-allergy, drug-drug, dose ceiling, drug-patient, drug-lab). Data from an interventional trial that required all orders be completed in minimized transcription including drug-allergy and the top ten drug-drug checking saw a 55 percent decrease in the serious medication error rate.
Computerizing prescribing was the highest yielding intervention strategy in the outpatient setting: allergy detection, notification about drug-drug interactions, and dose suggestions were most important. Better follow-up tools were pivotal; many patients do not let physicians know their problems.
Key areas of CDS to work on include: requiring complete prescriptions, default doses, renal dosing, drug-allergy checking, drug-drug interactions (checking, geriatric dosing, drug-lab checking), and dose ceilings. Drug-allergy checking was complex and needed standards. Too many warnings on drug-drug interactions impeded adoption. Drug-allergy checking was inconsistent across clinical applications. Most drug-allergy recommendations were overridden; many new reactions were not entered.
Future directions include: determining better when/how to alert, addressing cognitive and workflow issues; determining how to get people to document better, using standards for representation of reactions, determining the information most useful to clinicians and how to deliver more complex decision support. Recommendations regarding allergies include: consistent decision support across applications, reducing unnecessary alerting, adjusting notification based on severity, improving allergy documentation, and including food-drug allergies
Dr. Bates noted issues regarding variability in IV medications: most CPOE applications do not use dose limits, wide variation among institutions in dose limits, and variation in general for infusion therapy. He remarked that medication error rates were different than ADEs and more difficult to determine. Standards were needed for rules, alerts and reactions.
Discussion
Dr. Bates advocated reaching consensus on a set of important alerts. He suggested better interoperability would eliminate some redundancy in checking and that agreement was needed on how to represent rules. He thought of Arden syntax in terms of a set of standards for rules and a KB when considering looking at drugs. The two interacted, but he focused on Arden Syntax (level of notification or response). He noted a good standard representation of reactions was needed.
State of the Art of E-Prescribing Implementation
Ms. Thomashauer said CAQH was a not-for-profit alliance of health plans for development of non-proprietary standards development and support of physicians in quality of care. Challenges faced by CAQH's simplified prescribing initiative included: inconsistencies in traditional methods of data communications, more complex prescription benefit offerings, and the unavailability of pharmacy coverage information at POC.
Health plans knew the benefits of e-prescribing and realized that the best way to provide assistance was to promote formulary check as a component of e-prescribing. Barriers to health plan participation in e-prescribing included the evolving market and lack of interoperability. Early efforts had inconsistencies in data format and content, inaccurate or limited data, lack of communication between and among stakeholders, and vendors were not yet convinced doctors were interested in formulary at POC.
The initiative's objectives were to simplify the formulary process and improve interaction with formulary management programs, reduce physicians' burdens, and help consumers maximize the value of coverage. The Formulary DataSource was developed as a free searchable Web site where vendors could access a centralized formulary data file. Integrating this information required stakeholder engagement, standards and a quality assessment process.
Areas of standardization included: an Internet-based process; data content (drug name at NDC 11 level, on/off formulary, brand/generic, modifiers); file format (provided vendors in their preferred format), therapeutic classifications (crosswalked FDB and Medi-Span to create own database as the CAQH standard). Ms. Thomashauer emphasized the need for a single standard.
Preliminary results of a one-year e-prescribing pilot involving 100 providers in D.C. indicated: drug interaction warnings were heeded four percent of the time. Six percent of prescriptions were modified after allergy warnings. Some 25 percent of the exclusions were modified from non-formulary to formulary. Over 75 percent of brand-name drugs could be substituted with generics. Ms. Thomashauer noted that in a fragmented market 70 percent of the health plans were needed to influence physicians to use formulary. She emphasized linking formularies to the patient as well as plan level, noting it was a long-term, costly, but desirable goal. Adoption required significant resources (information technology [IT], training, and workflow adjustment).
Noting the Formulary DataSource Web site alone was insufficient, she said CAQH focused on promoting formulary check as an integrated part of e-prescribing. CAQH partnered with RxHub to provide a more comprehensive centralized source and formulary check in the provider workflow that was accurate, timely and quality-checked.
State of the Art of E-Prescribing Implementation
Dr. Medvedeff said Gold Standard Multimedia (GSM), which implemented the Florida Medicaid Project, developed a clinical drug database solely dedicated to clinical reference. He provided an overview of the Florida Medicaid program, noting a four percent reduction in significant clinical drug interactions in the physician's office and minimized fraud and abuse.
He said Medicaid had set up “huge hassles” to reduce cost. Formulary went from open to controlled access with preferred drug lists, clinical PAs, step therapies, fail-first, brand and monthly prescription limits. Simultaneously, political pressures mounted in Florida (e.g., media reported that five Florida patients died daily due to prescription drugs). Dr. Medvedeff noted, nationally, the rate is 200 patients daily.
Medicaid adopted PDA-based e-prescribing focused on impacting over-prescribing with: drug information explaining appropriate use and expected outcomes, patient information, and fraud/ abuse checks. Dr. Medvedeff noted a “DUR button” on the physician's PDA messaged Medicaid when the physician needed to talk directly about opportunities to improve utilization.
Dr. Medvedeff described a turnkey project with PDA and Web site support offered the 1,000 highest-prescribing Medicaid doctors in Florida that represent 14 million scripts and one billion dollars per year. Some 3,000 physicians (80 percent of Medicaid providers) are approved. A doctor given but not using the PDA and application is offered technical assistance; if it still is not utilized, the opportunity is given to one of 500 physicians on a waiting list. Lab results will soon be incorporated. Dr. Medvedeff said everything done on the clinical side with the proprietary database, which uses NDC and NCPDP SCRIPT, was “very transferable.”
State of the Art of E-Prescribing Implementation
Dr. Mandel said BCBSMA and their vendor, Zix, believed establishing standards would increase the likelihood that any system a physician adopted would provide the functionality beneficial to everyone. BCBSMA implemented a pilot program with Zix offering prescribers free high-speed access (via handheld, local server or wireless) to eligibility and medication history. The collaborative program is ASP through access points or cell phone. Demographic data is pre-loaded from the physician's practice. New patients are loaded in both PMS and e-prescribing (five data elements).
He noted lessons learned. Interest does not translate into adoption; physicians are interested in new technologies, yet have low tolerance for difficulties. One benefit, reduced phone calls from pharmacies and patients, is not recognized immediately. Incentives for primary care physicians include receiving one dollar per-member-per-month for implementing and continuing to prescribe at least half of their prescriptions electronically. Dr. Mandel noted there was a waiting list for the devices. He emphasized that many adopters were not aware of much of the functionality and needed more “hand holding” to support adoption. The cost of the technology was an issue; physicians felt they could not afford e-prescribing due to limited increase in revenue, increased malpractice premiums, and decreased reimbursement. Physicians recognized patient safety issues, but felt they were not making mistakes and staff could handle the phone calls. Physicians wanted ROI that brought efficiency and productivity.
Dr. Mandel said physicians initially interpreted “compliance” to mean whether they used the device and were not pleased about having to monitor patient compliance with medication. He noted reports that half the patients do not fill their prescriptions.
He pointed out that vendors needed similar functionality and communication standards and that having them would make vendor selection simpler. He said reducing confusion in the marketplace was a driving force in the collaboration. Eighty percent of their networks overlapped. Over time, the bar would be raised on incentives. Physician feedback was that no cost was necessary but not sufficient given other changes. Noting a shift of one percent in generic prescribing for BCBSMA represented a savings of ten million dollars, Dr. Mandel emphasized that there were significant savings in both commercial and Medicare/Medicaid populations and enough money to support adoption.
State of the Art of E-Prescribing Implementation
Ms. Adams said the five principles of RIQI (a collaboration of hospitals, providers, organized medicine and nursing, pharmacies and academics) were: collaboration, a requirement for real improvement, win-win solutions, focus on system improvements only achievable collectively, and involvement of top leadership. Strategic focus was on: patient safety, healthcare Information infrastructure, and evidence-based medicine.
Noting that going “application by application” will not be sustainable, Ms. Adams said RIQI gained skills initiating e-prescribing that equipped them for larger scale projects. RIQI chose to work with SureScripts on their e-prescribing initiative because competitors (large chains and independents) collaborating fit their principles. SureScripts acted as RIQI's clearinghouse rather than as a software vendor. They had an open-platform agreement. The value proposition for RIQI's adoption strategy was the significant efficiency gains seen in Rhode Island from the ability to automate refill information.
Ms. Adams described the statewide e-prescribing project comprised of a professional advisory team (Rhode Island's most respected clinicians) and 2,000 active prescribers and their interactions with the Board of Pharmacy and Dept. of Health, emphasizing the significance of engaging all stakeholders (e.g., legalizing e-prescriptions in the state).
She reported that, despite barriers, the six-week beta test was successful. Focus groups found the value of the renewal function unquestionable, noting it: simplified a cumbersome process, gained huge efficiencies, improved quality and freed-up time for clinically relevant staff functions. Reactions to the new prescription function were mixed; most pharmacies, still in early retraining, returned faxes in response to electronic requests. Physician workflow and portability were a problem; often a physician's office had only one computer. Ms. Adams agreed that adoption would continue to pose problems. Focus groups found that the application was good with specific enhancements and choices (e.g., formulary, med history). The biggest struggle was with parallel systems technologically sophisticated physicians already had. Focus groups sought more pharmacy participation. Currently, 70 percent of Rhode Island's pharmacies were connected. Two-thirds of the independents still had to be connected. Some 250 physicians were activated; 1,100 will sign on in the summer 2004. An impact study is surveying the effect on efficiency and prescriber, staff and patient satisfaction.
Ms. Adams said critical enablers included: involving staff, engaging major stakeholders, broad participation in design, flexibility, and heavy support. Challenges included: early workflow design where pharmacies were unaware of prescriptions, database inaccuracies keeping physicians from being recognized, workflow integration in sophisticated systems, pharmacies not fully “up,” and how the benefits varied. Ms. Adams noted desired features: real-time formulary information, drug prices available, a community-wide drug database, and notification about whether the script was filled.
She mentioned anecdotal evidence that physicians found quality benefits: physicians liked a one-step pharmacy (treated as phone-in) and felt safer with e-signature and legible prescriptions, efficiencies (e.g., one office had 1.1 fewer RN FTEs and 50–80 fewer phone calls a day). Ultimately, patients called the pharmacy for refills and were connected electronically to the physician. Ms. Adams emphasized that collaboration worked and was essential. Collaborators agreed to acknowledge that integration problems were inevitable and worked together to fix them. They avoided workarounds and found solution. Training had to be done immediately prior to implementation for physicians and pharmacists; otherwise adoption was hindered by forgetfulness. Noting Blackford Middleton's research indicating 11 percent ROI, she said RIQI's next project was to imbed e-prescribing into EHR and CDS, and connect them. RIQI will install a demonstration for 200 physicians and connectivity to the entire community (payers, purchasers, pharmacists, providers). RIQI plans to pay participating physicians $25,000 per year. Capturing 11 percent of the waste would produce a 77 percent ROI. Ms. Adams predicted 30 percent of the waste could be eliminated.
Discussion
Asked what would happen if a patient did not “show” after an electronic transmission, Dr. Mandel said that was the physician's responsibility. Dr Medvedeff explained that pharmacies' automated phone-out systems pulled in patients.
Gwen Lowe, Project Directory for CAQH's formulary database project, said $1,200,000 was spent on CAQH's Web site. The plans use Medi-Span and FDB, which classify drugs differently. The plans' pharmacy directors crosswalked. The Web site received 7,000 hits per month. CAQH subsequently concentrated on vendor distribution, partnering with RxHub. Mr. Thomashauer noted physicians had to understand whether the patient had a two- or three-tier benefit and granularity enabled conveying more information about eligibility benefits. The same formulary applied irrespective of tier, but third-tier required a larger member contribution for costly drugs and generic or preferred brand were preferred. Ms. Lowe noted EHR vendors would have to do costly rebuilding of their systems in order to accept patient-specific information. Mr. Thomashauer pointed out that standards also would help bring down the cost of systems to those currently unable to afford them, encouraging vendors into the market.
Testifiers were asked the percentage of scripts done with fixed desk tops, laptops and PDAs. Ms. Adams said vendors did not adequately support handhelds; physicians either had to hand write or enter directly into the system. Dr. Mandel said physicians used handhelds and staff used the browser version. Dr. Medvedeff said the Florida Medicaid program used AMA internet PKI through PDA by physician. Ms. Lowe said the DC pilot used desktops and handhelds. Testifiers agreed the standard should include any device.
Open Microphone
An unidentified testifier from ExpressScripts cautioned about going to EMRs and requirements where supporting infrastructure did not exist. Providing the relative cost of a drug might be easier than calculating actual cost. Pharmacies had to submit a claim to know member responsibility for “shoe box claims” where co-pays were higher than the drug cost, but the history level was good. The testifier cautioned about solving a problem with OTCs when the infrastructure did not support capturing that information. The testifier advised creating a standard that “allowed for perfect but provided good enough,” noting solving too many problems could create lengthy transactions, costly systems and resistance.
Terri Byrne from RxHub advocated adopting existing standards and not forcing those that would not work (e.g., SCRIPT would not work for inpatient). She noted tech vendors communicating between in-house and retail pharmacies use different standards. Ms. Burns pointed out that different plans had benefit information at different formulary levels; both downloading information to vendor applications and real-time transactions depended upon keys to find patients and the plan or group. She clarified that RxHub had not mapped for new or refill. RxHub created a transaction utilizing SCRIPT segments for medication history and would use an NCPDP request to the PDM for information regarding an HL7 request from the hospital vendor, translating that input into an HL7 response. Another participant noted RxHub would translate between all standards.
Noting he had observed the burden of collecting data for both payors and pharmacies, Ross Martin from Pfizer suggested couching some comments in terms of ultimate benefit. The payment structure was complex and MMA could be altered by the drug plan. He supported tiered implementation strategy focused on what was most important for the patient, rather than concentrating on an individual participant's relative hardship. He cautioned that having many classification schemas would confuse patients. MMA let schemas be varied and changing; providers would have to think clinically and differently in varying prescribing environments. Dr. Bizzaro expressed concern about eight classification systems that were hierarchical, unique and proprietary to compendia. Each had different specificity. Drugs within certain classes had to be covered by PDPs; with drugs classified differently it would be unclear whether they belonged and reimbursement would pressure manufacturers to get drugs into certain classes. He noted these were not indications databases and that therapeutic classification systems should not be used in that manner. Classifications did not impact CDS. Kathy Hollinger from FDA noted that research on package inserts found no one classification system would work because classification had to be done in various ways (e.g., by mechanical action, physiologic effect, therapeutic intent). Dr. Whittemore said SureScripts had checked a number of sources and the percentage of unfilled prescriptions was closer to 20 percent than half. Anne Canfield with Rx Benefits Coalition suggested that preemption was so broad that what happened at the state level was irrelevant.
Subcommittee Discussion
Harry Reynolds remarked that the industry seemed interested in making a difference until they discussed therapeutic classifications. Noting David Brailer, the new EMR leader, wanted e-prescribing faster than legislation, Dr. Huff asked if there would be a benefit in tackling non-controversial issues sooner in stages. He suggested organizing their recommendations into: identifiers; messages; codes; rules, knowledge representations and classifications; functional requirements; and digital signatures. NCVHS will not make recommendations about: VAN or medication networks, and prescriber or pharmacy applications and systems.
Members noted that MMA had a list of requirements for standards and another for attributes; some things called for might be for applications and not have to be recommended for national adoption. They had heard that day from network providers and early adopters and would hear other perspectives from software developers, private sector drug KBs, and providers via professional associations. Members agreed that, with all those perspectives, the list they were building could be a solid first-cut at the low-hanging fruit.
Mr. Reynolds proposed that they crossfoot all the reports, indicating what was already in place and being used. Mr. Blair observed that areas the MMA asked NCVHS to look at might or might not need to be recommended for national adoption, but in the private sector they would find a different perspective.
Additional hearings in July and August will listen to other stakeholders specified in the law. Standards organizations will provide feedback in September. Based on that feedback, the Subcommittee will pull together interim recommendations in November. The desire was to announce the final standards earlier than the April 2008 stipulated in the regulation; members will move forward as seems appropriate.
-DAY TWO-
E-Prescribing: The Software Vendors Perspective
Mr. Doerr emphasized that the electronic problem list and (ICD-9-CM) diagnosis codes were needed for prescription (with signed HIPAA permission from patient). Diagnosis-driven workflow could: improve auto checking drug-diagnosis and omitted appropriate drugs at several points in the prescription process; improve communication between PBMs and payers, preventing unnecessary phone calls; and lower treatment costs by intelligent messaging about pharmacotherapy. He noted e-prescribing standards accommodating messages about lower cost alternatives would be most effective when they related to the diagnosis being treated and trade offs between efficacy, cost, and contraindications. Content must be evidence-based, therefore diagnosis-based. He noted a study in Missouri found payers' savings-per-doctor-per-month increased in the first six months by 8.3 percent ($584) and by 15.7 percent ($1,101) in the next six months. Payers' ROI was 6-to-1. Some 65 percent were written for generics compared to 50 percent nationally. A study in Maine was consistent. Mr. Doerr said requiring diagnosis codes (already used for billing purposes and associated by many EHRs in products) would not impede workflow.
He considered interfaces between e-prescribing systems and PMS the biggest hurdle in adoption. Without an interface between scheduling and billing, double data entry (too time consuming to be practical) was required. HL7 v3 required conformance testing and did not address exorbitant costs, mostly a result of anti-competitiveness.
Mr. Doerr noted NCPDP did not resolve the barriers to interoperability caused by prescribing apps, networks, PBMs and PMS using competing code sets for names. RxNorm promised a solution. Noting the incompatibilities blocked some of Wellinx's sales, he emphasized that resolution was needed. State requirements were also barriers and standardization at the national level and standardized tables for managing formulary information and policies defining formulary were also needed.
E-Prescribing: The Software Vendors Perspective
Dr. Faughnan said McKesson had a lengthy relationship with FDB, used SNOMED CT for patient conditions that might impact e-prescribing, and planned to use LOINC. They frequently “had to deal with” ICD-9-CM but did not find it satisfactory for checking. NCPDP SCRIPT and other messaging structures were used for formularies. RxNorm was supported as an external norm; McKesson looked forward to having SIG included.
He called for addressing process and structural issues with NDC, which are deeply imbedded in medication transactions, in an evolutionary way that preserved NDC. He recommended a true dispensable identifier rather than NDC.
Dr. Faughnan encouraged collaboration between SNOMED and FDB for patient conditions. Prescribing vendors typically represented patients in an internal terminology optimized for prescribing use. Alignment was needed.
He said code gaps included acknowledgment of receipt, clinical SIG, text messages, encoding of prescriber review of alerts and formulary notices, formularies and tiered formularies, medication history messaging, a politically acceptable method for patient ID, e-signature, reason code was rejected, and DME code consistent with RFID standards.
Dr. Faughnan noted four adoption issues: customer readiness (vendors did not choose to be too far ahead of customers), message could dictate application structure, implementing some powerful standards required reengineering, and coordination and collaboration between messaging and terminology standards, especially where partitioned by domain.
Emphasizing that licensing cost and business models must work for the SDOs, Dr. Faughnan said incentives should be predictable over the long term and sufficient to keep SDOs healthy but not so high that it was passed to customers. He called for consideration of international licensing costs and said incentives should be graduated for partial development. He conveyed McKesson's support for patient, provider and dispenser communications and conviction that this flow of data was manageable. Applauding NLM's major role with RxNorm. NCPDP Script and SNOMED-CT medication data, he underlined the importance of incentives for collaboration and suggested that a trusted agency or organization monitor all code sets.
E-Prescribing: The Software Vendors Perspective
Dr. Kaufman said DrFirst supported connectivity, RxNorm for naming, and a standard SIG. He pointed out that: even with e-prescribing systems, patients and pharmacies still often ended up faxing; about 25 percent of prescriptions went unfilled. Noting the levels of e-prescribing in the EHI report, he said he knew of none truly at the full connectivity of level 5. He predicted the next level would involve critical connectivity across the continuum of care between inpatient and outpatient and different providers and health systems. He called for a unified naming system (e.g., RxNorm), standard classifications (e.g., RDF-NT), and mature descriptive language with SIG (e.g., SCRIPT). He emphasized CCR as a small and important early step, noting the number of organizations signed on. He illustrated the back-end connectivity network and complexity of e-prescribing, noting connectivity was the key: when standardized, the number of connections, while not trivial, was less important.
Dr. Kaufman said: DrFirst used HL7 but it had too many optional fields. NCPDP SCRIPT was “great” but needed a way to convey how medication should be taken. NDC was not doctor friendly. RxNorm was designed to specify meds as described by the physician but did not include radiographic materials; OTCs were being added. NDFRT specified contents of medications and other features/classifications and might simply formulary management; drug-drug and drug-allergy interactions and insurance formulary lists will be aided. Dr. Kaufman recommended a uniform format specification, including a standardized tier-rating scheme and exclusion list. He noted the SIG standard was needed for describing directions. E-prescribing systems had to be flexible to allow free text for outliers; most content should be shared as XML, increasing CCR value by allowing native display of data.
Noting the media continued to indicate high pricing for e-prescribing, Dr. Kaufman said many applications, even with hardware, were under $1,000 per-provider-per-year. Physicians wanted simple, powerful and fast technology. He said once they understood the benefits they would sign up. He contended that physicians' main concern (that the technology was not fast enough) was ungrounded.
E-Prescribing: The Software Vendors Perspective
Ms. Helm said Allscripts had offered e-prescribing software and medication management solutions since the early 1990s and created and routed millions of prescriptions a year. Allscripts provided a full range of sizes and standalones as well as fully integrated EMRs. She said funding and incentives increased physician participations, adding that free “wasn't good enough.” She noted that increasing generic and driving formulary compliance increased cost effectiveness and that automated prescribing and refill reduced phone calls by 50 percent.
She noted that member IDs and other identifiers were challenging to collect and enter in determining eligibility with health plan drug benefits. Health plan IDs would be helpful. Prescription plan identifiers were not always clear on cards. RxHub helped X12/HL7 exchange. Standard formulary information and presentation were needed. RxHub, other health plans, third parties, and PBMs offered a variety of formulary designs but a core set of form attributes were needed. Prior authorization should be automatically communicated to the prescriber.
Ms. Helm said the medical history jump-started the physician experience. Noting the challenge of reconciling multiple sources of information to a single medication list, she emphasized that multiple standards had to be supported and the need for data element and term standards. She reported that NCPDP SCRIPT was solid and expandable. Some challenges stemmed from the lack of appropriate national identifiers and lapses in operational processes.
She said NDC was useful for pharmacy to communicate dispensed medication information to payer or PBM. Therapeutic classifications were a more meaningful classification for physicians than drug classifications based on mechanisms of action or chemical structure. Drug databases had to include medical devices and DME. RxNorm (a clinically orderable drug compendium) and NDFRT (detailed facts about ingredients) provided clinically relevant information. Allscript identified pharmacies via the NCPDP SCRIPT number. DEA, even with its limitations, was the most prevalent identifier for prescribers. She emphasized the need for the NPI.
Regulatory challenges included: a variety of prescription formats based on states (and in-state differences for printed, handwritten, and faxed), DEA lack of recognition for e-signatures for controlled substances, and state requirements (e.g., approval of e-prescribing systems).
Ms. Helm noted an evolutionary aspect to e-prescribing. E-prescriptions had to be comparable to other mechanisms, not held to a different standard. Standards had to create value to drive measurable outcomes for efficient connectivity, flexibility, and a natural transition path guided by provider incentives for adoption.
E-Prescribing: The Software Vendors Perspective
Mr. Nayak viewed e-prescribing as “management of prescribing inside a physician's office” and emphasized understanding workflow and adoption issues. Pointing out that DEA did not identify setting, important for two-way communication, he noted the physician identifier was needed. Citing legal and compliance issues with pharmacy boards in many states as a detriment to widespread adoption, he encouraged standardization. He said RxHub and SureScripts helped make e-prescribing possible, but he cautioned that efficiencies would be limited if a manual process had to be inserted for controlled substances. He noted it would be useful for the pharmacy to know when the physician overrode a drug contraindication and that NCPDP SCRIPT's ability to incorporate allergy information was seldom used. Noting the different codes for various packages, he observed that NDC codes selected by a physician usually would not be those dispensed. Remarking on the extent of past development, he called for gradual modification.
Noting the formulary list could not be linked to the patient, Mr. Nayak reported that MPI made that work for real-time eligibility for medication insurance. He also noted a couple mechanisms existed to do this with medication history.
Mr. Nayak said SureScripts provided retail pharmacy with bi-directional communication, particularly for renewals. He suggested that a notification back to the physician when dispensing a drug would help make medication history a “true” history. Mr. Nayak emphasized that integrated e-prescribing in conjunction with the EMR systems and levels 5/6 provided value that physicians were willing to pay for.
He said applications needed to be integrated into the physician's workflow, but cautioned about providing too much information. He advocated incremental steps. Noting “no two physicians did things the same way,” he encouraged creating flexibility in systems and focusing on the physician and what would not become a hindrance or take more time in the prescription cycle.
E-Prescribing: The Software Vendors Perspective
Mr. Nutkis said about a third of the health plans and BASF use Zix secure messaging. Zix e-prescribing services are used in about 20 states. Zix connects through RxHub and goes direct to some other PBMs. Some data for PBMs is stored in a Zix data center. He reported that Zix found a way to workaround a lack of standards, emphasizing this lack was not a substantial impediment.
Zix delivered services via the Web, PDA, cell phone or wireless and could provide real-time lab results and link to diagnosis code at point of prescribing. Noting functionality was evolving at a relatively rapid pace, he expressed confidence it would keep pace with technology. He noted that requirements from the plans and PBMs to keep up was feverish.
Mr. Nutkis said they had heard that any lack of implementation or limitations around deployment was not related to standards: the pain in the marketplace stemmed from questions about emerging technology.
Zix used NCPDP SCRIPT, X12 270/271 and HL7 standards. Zix faced about 120 EMR or PMS systems and interfaces were “heavy lifting.” Mr. Nutkis said lack of incentive, dollars and sophistication were often the issues at the practice site, not the interface. Many small providers were not under maintenance and had no resources to turn to in trying to work with the standard and interface.
He said nationwide adoption was an issue, especially in border states. Authentication was a significant problem. Seeking more integration and seamless access, organizations wanted e-prescribing more embedded in what they did. The AMA Verisign digital identity program and others were used ands there was no standard for vetting and authentication.
Mr. Nutkis emphasized the need to understand where formularies “were going” and SDOs' role in DUR. DUR messaging standards were needed along with standards for adverse events reporting. A consistent issue in operating Zix's certified data centers was the need for standards for appropriate transactions sets. He noted the time for point-to-point requirements was a challenge: after 36 seconds the device was “put down.”
He remarked on a number of incentive programs compensating on a pay-for-performance basis for e-prescribing, noting the question would be the limitations put on them. The proposed basis for incentives included e-prescribing, formulary compliance and educational programs. Zix believed Medicare would become the de-facto standard.
Discussion
Ms. Helm explained that Allscripts would like expand the data standards to accommodate what was ordered in a prescription, including devices in ambulatory care as part of a treatment plan or to administer medications (e.g., home health, oxygen, DME).
Testifiers summarized the strengths and weaknesses of three standards for prescriber identifiers. DEA was currently used, but not everybody had one. Emerging NPI was intended for every prescriber entity, but did not represent a site/facility where the patient's chart was located and would not be fully available for a few years. HCIdea was similar to NPI, could identify site and have multiple IDs per site, but data was not fully supportive. Message structure and use cases would determine additional identifiers needed beyond the prescribing agent.
Asked about speed of entry, the pyramid of sophistication providing more information, and a threshold, Dr. Kaufman said additional information included weight, age, diagnosis, drug contraindications and allergy. Doctors were concerned about both safety and time. Level five might be easier to reach than level six, but the full continuum should always be considered. Mr. Nutkis noted a wide variety of messages (e.g., including co-pays) increased the time to receive the data and the likelihood physicians would “walk away.” Physicians were willing to wait for real- time prior authorization (X12 278). The goal was five seconds, 22 seconds was the upper level of threshold. Dr. Doerr suggested the market would find “sweet spots” visa vie features, functionalities and tradeoffs. Decision support should be hooked to diagnosis (level five plus). He said physicians drove 80 percent of expenditures in the healthcare market but had little time to “do it right; the payer needed to share benefits above and beyond the cost of the system. Dr. Kaufman observed that once people start missing and discarding spurious and unwanted alerts the decision-support system stopped being effective. He said clinical terminology (e.g. SNOMED CT) and integrating knowledge could reduce false alerts. He also noted that no one was comfortable turning off black box warnings (FDA category one) on labels, which might not be correct anymore, producing more unwanted alerts. Dr. Doerr said it took $10,000 in new revenue to get physicians to change their work flow. Larry Trusky, Vice President Product Development, Allscripts remarked that some formulary checks were just too hard locally and needed network access. Mr. Nayak said “a release valve” was needed so that there was flexibility to do SIG how one wanted despite a standard.
Participants noted low-hanging fruit on the pyramid were not in a linear path line but, except for the mid-to-upper levels, arrayed along levels of functionality that were not necessarily a continuum. Level one especially stood apart. Some gradations were independent of standards. Certain attributes of different standards might “play well” in particular instances but not be an appropriate trade-off for the sake of uniformity (e.g., ICD-9-CM had features needed for billing, but did not necessarily have to be the only identifier.
Dr. Faughnan said: NDC11 worked but could be extended and grow, NCPDP SCRIPT had a “good starter” and process, ICD-9-CM could send diagnostic information but not contraindication checking, X12 handled demographics. Standards on the formulary transactions and DUR would speed implementation. Mr. Nayak noted RxHub worked to normalize formulary files between their three main PBMs. Dr. Kaufman suggested allergy information, drug interactions, and diagnosis as low-hanging fruit in terms of patient safety. An issue was how to get patient medication lists shared among providers while valuing privacy. He emphasized providing step values with formularies and a standardized format for the prior authorization system. Mr. Nutkis noted DUR was needed and formulary inconsistencies. He observed that the criteria came down to complexity/comprehensiveness versus speed in developing standards along with the cost of reimplementation.
Dr. Faughnan remarked that standards reduced the cost of implementation and interoperability and enabled doing more advanced things “down the road.” Diagnosis codes worked in selection of drugs but not for contraindication checking. RxNorm was about interoperability. McKesson used NDC “as a fudge” when it worked with Multum instead of FDB. Mr. Trusky said NDC was not a good identifier for e-prescribing. He noted one could get to level four with “heavy lifting;” identifiers were needed beyond that. Mr. Nutkis noted reference guides were difficult to manage in multiple form. Dr. Kaufman advised against doing workarounds with NDC: RxNorm was nearly done and, when it came to NHII, e-prescribing “had to be done right.” Dr. Doerr said there was compelling justification for a nationwide standard for SIGs and presenting prescriptions and benefit in standardizing formulary expression. InfoSpan was an alternative with formularies. Physician identifiers were problematic. Dr. Faughnan emphasized e-signatures and unique identification of patients. He noted RxHub used algorithms. Mr. Nayak noted the general response: workarounds had been done a long time and people did not always want standards to reengineer.
Dr. Kaufman explained that controlled substances written on e-prescribing had to be printed out (New York required a triplicate form) and hand it to the patient. He expressed concern that infrastructure might not be in place for upcoming FDA requirements for authentication: PKI and biometrics. Mr. Nutkis noted the physicians' perceived that the telephone was more secure. He emphasized security mechanisms (how to vet individuals and ensure their identity) and noted the Federal bridge project and a VHA pilot. The VHA's main concern was not opening a portal for fraud.
Ms. Gilbertson said NCPDP SCRIPT could do a compliance transaction. She noted there were “different flavors” to prior authorizations, depending on what the health plan set up. Ms. Gilbertson urged those with ideas about enhancements to submit changes for SCRIPT. Participants noted there were different reasons and philosophies for prior authorizations. Mr. Blair said the Subcommittee would discuss at another time whether barriers were appropriate or standards were needed to address them.
E-Prescribing: The Drug Knowledge Base Vendors
Mr. Robinson said KB vendors collected, codified, and linked to and distributed applications useful at POC. They relied on product labeling information from manufacturers, information related to OTC and alternative medications and some supplies, as well as primary literature. They initially supported claims and inventory management and use NDCs, UPC, and health-related ID as well as linking pricing and images. Clinical content is aggregated to a formulation level and linked to NDC. He noted name-based concepts were needed in the clinical environment.
Versions of therapeutic classes embedded in Medicaid and Medicare were based on how drugs were commonly described by class. Nomenclature for describing hierarchical conditions was embedded in the drug indications database. Either a name-based drug identifier or NDC is used to place ingredients in the context of the patient allergy. There was no standard for allergen information; SNOMED-CT was being considered for interoperability. POC and navigation was supported by grouping vocabulary content into: concepts, allergens, disease and lab. A small controlled subset, proprietary disease vocabulary, expressed pharmaceutical concepts with flexibility and linked to standard terminology. Lab-based content was similar to diagnosis, only concepts needed in combination with the drug might give a false reading. Drug therapy assessment was the other link. Drug-lab assessments required the vocabulary codes used in the lab system.
Dosage screening used proprietary drug concepts and possibly diagnosis. Proprietary methods were used for dose string and parsed elements. Options provided included application program interfaces (API) tools and object technology. Vendors or others provided the user interface. Pricing could be delivered daily, weekly, monthly or quarterly. Vendors pushed the data integrated into their internal database structure to customers.
Mr. Robinson said KB vendors welcomed structured labeling coming from FDA and hoped to gain internal efficiencies. Complementary content (e.g., off-label uses) would come from evidence-based approaches. He noted an opportunity for links from KB to label for reference purposes. He saw no change with the NDC with drug information codification beyond possibly reestablishing some links to RxNorm clinical codes. With drug identifiers, he believed a link could be published from FDB to generic semantic from RxNorm. A crosswalk could be provided for the ingredient between the FDA UNII and FDB. Links could be expected for translation purposes among TPs. Therapeutic classification was an unknown. USP was challenged to rate some classifications, but comparison was difficult. Allergen information was an open issue from a standards perspective; FDB would cooperate. FDB planned to link disease-based content to SNOMED-CT; an issue was CAP's willingness to modify their editorial policy and introduce pre-coordinated terms. He noted the dilemma of post-coordinating qualifiers or active mapping to join multiple SNOMED concepts. He said FDB was already committed to LOINC for laboratory and willing to cooperate with any third-party on a standard SIG for dosage information and screening. API tools only needed a code to handle standard terminology as an input property in lieu of propriety code. FDB was open to delivery of standard terminology with proprietary content though he noted this might be a dilemma for customers to coordinate and a drug KB might choose to redistribute some in a terminology. FDB hoped to do all this transparently for customer who would still need to code to send a code, using the standard as a user interface. The goal was for the end-user to write as previously, and then convert to RxNorm concept “behind the scenes.” Mr. Robinson said FDB would be open to NLM or another non-competitive form aggregating this information.
E-Prescribing: The Drug Knowledge Base Vendors
Emphasizing the complexity of what database providers provided customers and time issues, Dr. Bizzaro urged NCVHS to consider how much was e-prescribing and how much was a separate “screening” function. He noted problem areas. RxNorm is a name-based identifier and represents how physicians think; pharmacy needs NDC and it “will not go away.” He noted gaining consensus for a standard signature would be a thorny problem: any system writing orders had numerous proprietary code sets that made cross-referencing difficult and errors in cross-referencing a liability. Dr. Bizzaro said, even if codified, retail pharmacy would probably always require textual representation as a quality check. Noting a duplication of clinical screening efforts in office, pharmacy and PBM, he suggested that messaging could explain responses to alerts. He agreed there was industry acceptance of e-prescribing, but cautioned about requirements that became so complex it was be difficult to implement and retain acceptance.
E-Prescribing: The Drug Knowledge Base Vendors
Dr. Millikan said NCPDP SCRIPT was widely used and a solid foundation to build upon, though enhancement was needed. Nationally adopted code sets for drug products, descriptions and instructions were needed to facilitate interoperability. Medi-Span supported: adding RxNorm for identifying drug products; SIG; inclusion of text fields and proprietary code sets that would accelerate adoption and provide a final quality check with narrative text; incorporating drug-drug warnings, cautions and dosage alerts including a summary of DUR prescribing; dispensing, or reporting actions; DUR results between pharmacy, health plans and prescriber; the SCRIPT version should either include a drug product description longer than 35 characters or mandatory dosage form and strength; and NCPDP's standard for transmitting an entire formulary should be rolled into SCRIPT.
Medi-Span recommended use of SCRIPT, necessary code sets and RxNorm for interoperability between proprietary drug terminology code sets. Medi-Span used NCPDP and SCRIPT; other code sets were customer driven. Dr. Millikan said NCPDP's strength was its widespread use and acceptance; it could be enhanced by codified or partial codification of SIGs. He noted that HL7, in coordination with NCPDP, was creating structured standards. Gaps included SIGs and a code set for drug allergens.
E-Prescribing: The Drug Knowledge Base Vendors
Mr. Hobson said Multum was a reference terminology. He noted patient medical history already existed in some prescriber settings and could enable the prescription writer to use logical choices to determine appropriate drug therapy. Pediatric and renal disease patients were special cases for help. Multum had various severity levels of interactions; all but four percent could easily be overridden. Cerner Multum used NDC, which he noted was not readily available from a single source. SPL would help identify active ingredients for allergy checking and FDA-warnings. He advised FDA to consider inconsistencies provided by manufacturers and implications of various drug actions in creating SPL. Cerner Multum supported NCPDP and SNOMED terminology for conditions.
Discussion
Asked the status of standardizing DUR message standards, Dr. Bizzaro said NCPDP Telecom and SCRIPT had qualifiers identifying source of information and levels of interaction specific to that proprietary source. DUR code and free text field were available. Messages were time-consuming. He suggested that the bar code got through rulemaking quickly because it was simple and had value. He recommended keeping e-prescribing simple and adding value after people were doing it.
Dr. Bizzaro said the only discussion AMCP had with others about standards for distributing KB was about a list of drug interactions and their severity; the compendium decided to create a methodology to create proprietary database. Timeliness was an issue for nationwide distribution. Mr. Robinson said everyone was open to common messaging standards or criteria, but wanted to reserve rights to intellectual property assets and distribute in a competitive environment. Mr. Hobson said Cerner Multum agreed that standards would be helpful.
Dr. Bizzaro said he did not know how to reduce useless information without a liability issue: different practitioners had different experience; doses in a hospital with an oncology unit would be outliers compared to a typical retail pharmacy. He noted four areas of benefit to the physician with e-prescribing: eligibility, access to formulary for patient, DUR, and e-scripting.
Asked the timeframes for using the mapping from RxNorm to Drug KB, Mr. Robinson said that depended on NLM's progress with RxNorm. RxNorm focused first on the generic/semantic drug level and relationships to brands as synonym, and then it focused on branded names. He noted they would still have to map back to some concepts. Commonly prescribed devices and DME presented another level of complexity. FDB was a drug data base that included 1,000s of other materials. Dr. Millikan said Medi-Span was evaluating mapping to RxNorm in 2005. Mr. Hobson said Multum Cerner was mapping concepts to the product level and would look at branded levels. AHRQ provided six million dollars to work on the issue. Members noted AHRQ would provide testimony in September which could include discussions with drug KB vendors about whether there was enough time to have all those mappings in place by January 2006.
Dr. Bizzaro explained that how drugs and dosages were handled in transaction was more an issue for vendors using Medi-Span content within their applications and then sending information into Script. Most pharmacy vendor products only accepted 28-30 characters. NCPDP's newest version had a longer capability but was optional because of pharmacy vendor capabilities. Dr. Huff said the fundamental complexity was that in ordering a drug one had to allow many things either whole or “in pieces”: flexibility was needed, but everything had to be computable, preferably without redundant work effort.
Follow-up on HIPAA Implementation Issues
WEDI sent a letter in March to the Secretary indicating challenges and successes of implementation of the transaction and code set standards noted by testifiers at a recent WEDI hearing. Asked at the March 29 Subcommittee meeting to prioritize the recommendations, WEDI consolidated them into: issues needing immediate attention; ongoing issues; and strategic longer-term issues. Mr. Jones highlighted four implementation issues. First priority: continue focusing on full compliance while maintaining contingency plans with timelines to “move beyond.” He said progress was being made. Mr. Jones emphasized the need for a definitive, authoritative source for frequently asked questions (FAQ). Second: a renewed effort to identify and promote the benefits of the other transactions, and move toward them. He stressed that contingency plans had to be addressed and data content had to be standardized. Third: a requirement for use of a standard acknowledge (e.g., the 997) and more definition on area reporting. Fourth priority was a comprehensive review of COB issues, especially the health plan identifier (without it one could not move from plan to plan efficiently in electronic formats).
Follow-up on HIPAA Implementation Issues
Mr. Daley said other recommendations regarding COB included clear instructions to Medicare contractors on how to populate crossover items in a standard way. Another COB issue was associated with serving various claim form types: whether COB was done at line or claim levels caused problems with many claims when trying to cross from one payer to another.
Ongoing implementation issues include: developing the standardized test data, expediting change standards (WEDI recommended that CMS, WEDI, and the DSMO consider the regulatory process and stress the standards development process in determining whether an expedient change process should be developed), address in a timely manner data content issues, restructure code set maintenance to once a year where appropriate, and specify code values on instructions.
Priority strategic implementation issues were: ways to include providers in standards development; balanced representation on the DSMO; and encouraging decision-making and participation in standards development. Mr. Daley emphasized that balanced industry representation was needed to come up with standards workable for everybody. He said pilot standards that would indicate changes needed before adoption and validate costs and benefits were needed prior to issuing the final rule. WEDI also recommended looking at enhancing the implementation process. He encouraged looking at a realistic time frame for all aspects (e.g., vendor products had to be designed, developed, rolled out and published in reported provider, clearinghouse and payer sites for anyone needing the crossover. He cautioned that once everyone believed they were compliant, they still had to migrate all the different trade department connections. Even providers had dozens or hundreds of connections that had to be rolled into the process
Discussion
Dr. Cohn reminded everyone that implementation sometimes took longer than initially anticipated: HIPAA was supposed to be completely implemented by 1999. He emphasized focusing on how to monitor and promote implementation of all the regulations. Dr. Steindel noted it was suggested at the recent NUCC meeting that CMS track more vigorously, initiating an effort for COB similar to the one for the 837 with a weekly list of compliance and a notice of gains. He noted other claims also involved COB and probably would go into effect at the same time. He cautioned that some plans believed they already hit 90 percent or better on many compliance conditions and were considering discontinuing contingency plans. Mr. Schuping said eyes would be on CMS as the big player; CMS putting “the stake in the sand” would create momentum for full implementation.
Mr. Daley noted that the Administrative Procedures Act was not a useful guide when technology moved so rapidly that the market outpaced the ability to promulgate regulations. He said the important thing begun in 2003, particularly on the clinical side, was the federal government taking initiative through a variety of vehicles (e.g., EHR, SNOMED-CT, CHI initiatives) to use its market power in the health care field as an impetus to move forward. He predicted that, over time, the benefits would come from the provider initiatives and the data that flowed through those clinical processes in a practice or hospital and populate the administrative transactions. Mr. Daley emphasized that education and outreach should push these HIPAA transactions.
Mr. Schuping encouraged industry and government to work together to have a central point for the FAQs and “a reasonable chance” of getting some of them answered more timely. Ms. Friedman noted that many people wanted the FAQs to solve their business decisions. Mr. Jones noted in many situations it took time to develop appropriate answers. Rather than policy, business or regulation questions, sometimes these inquiries were about interpreting an IG.
Responding to the issue of complete compliance of the 837, Ms. Trudel reported the Medicare compliance statistics for the previous week were about 85 percent of the inbound claims. Compliance continued to inch up weekly. She said the Medicare interim steps effective July 1 paying electronic non-compliant claims more slowly (as though they were paper) was having an effect. She said this interim measure was meant to be less dire than cutting off the contingency altogether. Payers were striving to get “over the hump” with the claims, and working with clearinghouses, vendors and providers to get everyone into production, so they could turn to other transactions. CMS was tracking percentages on the remittance device; out and inbound bound claims had already identified COB issues that left trading partners unable to process files, and had put in system changes that should improve that percentage. She urged everyone to keep “chipping away” at the claims status and eligibility, ensure providers understood the benefits of using the transactions, and consider doing outreach or putting incentives into place.
Mr. Reynolds said other transactions were slow in BCBSNC's environment, but there was more action. Providers asked most about the claims process for their plan. BCBSNC's second largest submitters, two major clearinghouses, had yet to submit an 837; they could not get everyone around the country going because everything was a bit different. Noting 80 percent of the COB transactions were probably Medicare-related, Mr. Reynolds said once some COB issues were settled, the industry would probably make an 80 percent move without having to wait for the health plan identifier. He noted state prompt pay laws required certain messaging back to the provider on remittance devices. WEDI was eager to collaborate in establishing the code sets' meanings for the 835. He noted that BCBSNC would be required to send a paper notification of payment, as well as an 831, in at least two or three prompt-pay states.
Mr. Jones said getting a remittance and the money side working was the key to making people move forward. Ms. Trudel observed that having the standard transactions, code sets and data content enabled people to realize there were nuances and differences in business practices from plan to plan that were not necessarily relevant to their actual operation. She said it was an opportunity to clarify and make the eligibility business process more consistent. When they made commercial sense, transaction standards served as harmonization processes. She said CMS was clearing a frequently asked Q/Q that addressed how lifting the Medicaid contingency plan would leave physicians' other plans. Mr. Daley said there was much uncertainty about that and dealing with it would be helpful. He emphasized that the health care industry was confused about what the standards were; many believed much of the electronic processes, data systems and software had a competitive advantage. Noting counter examples (e.g., airport kiosks for ticketing, how ATM networks were globally seamless, how cell phone technology evolved), he stressed conveying that a tool was being created to bring benefits to investments and that other business-related factors impacted on competitiveness.
Mr. Daley reported that WEDI was prepared to assemble task groups to work on incentives on both the provider and payer side to foster faster compliance. WEDI will talk with David Brailler about how to focus on this issue and incentives to bring to the table. Noting another issue was diminishing uncertainty about contingency plans, he suggested that ordering priorities and a range of time frames would focus industry attention. With NCVHS's blessings, Mr. Schuping said WEDI would take a lead role in making any or several of the collaborations with entities and industry/government that stakeholder requirements called for. Mr. Jones added that WEDI would like NCVHS to endorse the recommendations and noted they consistently heard that anything the Committee did to enhance outreach would be beneficial.
Mr. Reynolds said all the recommendations were good and about business process that already had ways to do them. But he doubted most the private sector would do a dramatic link when Medicare “pulled off” compliance if their own client level was not sufficient to keep their business going in a reasonable way. Emphasizing that messaging was key, he encouraged WEDI and DSMOs to team up with CMS on FAQs, so their message reached all the industry.
Dr. Cohn emphasized the strength in working with the industry and suggested it should be possible to resolve the issue of external code sets and other things working directly with the DSMOs. If not, than CMS and NCVHS might need to get involved, but he said this was not something they should approve or disapprove. Ms. Trudel said HHS was already looking at a report and analyzing recommendations, irrespective of the Committee's recommendations. Mr. Jones noted progress on some of these items, including a white paper on the announcement on error reporting. Mr. Blair said WEDI had done an outstanding job in a tough situation shepherding adoption of HIPAA standards.
Ms. Gilbertson noted NCPDP maintained Version 5 Editorial, a document containing editorial corrections, clarifications and examples on their Web site. An industry task group answered questions from the public. Editorial was updated at least quarterly. She said the DSMOs discussed at length with the Office of HIPAA standards information for the next NPRM regarding doing a modification/maintenance process as rapidly as possible within the constraints of the regulations. She suggested the post-certification maintenance should be restructured for a version release scheduled only once a year, cautioning that the industry would be shut down if NDC was entirely maintained that way.
Mr. Daily pointed out that outreach to physicians required totally different marketing than targeting their office managers (e.g., physicians were reached on the weekend). He said WEDI was doing a lot of that outreach in collaboration with CMS.
Dr. Cohn emphasized that the Subcommittee would be working more than full-time on e-prescribing for the next three-to-four months. The main goal, however, was to determine a robust, efficient way to monitor the industry and understand how it was moving forward on all this. Noting they repeatedly heard that many of the benefits of these transactions did not go to those having to implement them, he noted they ought to look at the significant value to be derived by the industry as a whole.
Follow-up on DSMO Request - Donna Pickett, NCHS, CDC
Dr. Cohn noted a request from the DSMOs related to use of the NCPDP transactions in building and supplies that the Subcommittee deferred was recently brought back to them. Two issues were involved: use of the standard for building and supplies and an issue related to use of the code sets in relationship to supplies. NCVHS had been waiting for HHS's further evaluation, assistance and guidance on the code sets' issues.
Ms. Pickett explained that both UPC and HRI codes were external code sets used within the NCPDP standards. UPC was assigned to OTC products (e.g., vitamins, St. John's Wort). Companies seeking a UPC code joined the Uniform Code Council, the not-for-profit organization administering the UPC. Guidelines for using UPC codes were on the UCC Website. UPC contained a labeler portion. HRI code sets identified and numbered device packages (e.g., syringes, band-aids). HRI were a 10-digit code that appeared in 4-6 or 5-5 format. Manufacturers seeking a HRI number for unique product identification applied to FDA for the labeler code forming the initial portion of the HRI. Manufacturer and distributor identified their device with a sequential number completing the code and were responsible for maintaining it. FDA voluntary compliance regulations prescribed how to govern use of the product package sections of NDC HRI codes.
Ms. Pickett noted neither UPC nor HRI codes were precise or unambiguous or met all the guiding principles under HIPAA. These supply codes were product-driven and would not be updated in the open process essential to HIPAA; however the codes were used successfully for years and at the March 30-31 meeting Ms. Gilbertson referenced consequences the industry would face if it could no longer use them.
Development of Potential Letter - Simon Cohn, MD
Dr. Cohn noted the Subcommittee had discussed looking at the issue of supply codes late in 2004 or early 2005 and waited on bringing forward the draft letter on recommendations until they had more information on the various code sets. Recognizing the need for further evaluation, the draft gave the Department latitude in considering possible rulemaking. Members reviewed and amended the draft and voted unanimously to approve it.
-DAY THREE-
E-Prescribing: The Provider Perspective
Ms. Narcisi noted broad-scale e-prescribing barely existed in ambulatory settings. Some hospitals and large practices had closed systems, but two-way, interoperable and secure open systems were rare. Vendors had to address this setting's needs. She said a single set of standards would be a significant undertaking. Many specialties had to be considered; flexibility/scalability was needed. Standalone systems had to be accommodated along with integration with EHRs. There was no single solution for medical vocabulary deficiencies. Terminology should be comprehensive. RxNorm provided standards names, doses and links to active ingredients as well as some brand names. NDC were stored as attributes in RxNorm, but nomenclature had to account for details. Ms. Narcisi said RxNorm might be suitable for VA, but not necessarily the private sector. Terms differed between hospital and ambulatory settings. Standards for intent of prescription were needed. She noted state form issues. Demonstration requirements should address workflow and business rules in a variety of settings (e.g., prescriber had to send payer/PBM formulary, than pharmacy, and should not have to default to paper) Funding models were needed.
Ms. Narcisi noted there were many interchanges in e-prescribing and standards were needed. Currently, two messaging standards supported e-prescribing: NCPDP SCRIPT and HL7. She said NCPDP SCRIPT should address eligibility, formulary information and medical history. Workflow and business requirements had to be considered and all parties' needs defined. Prior authorization was necessary. HL7 was used for administrative and clinical transactions including medication therapy, patient ID, clinical notes, and lab results.
Noting physicians wanted standards for e-prescribing to accommodate a wide variety of clinical settings and specialties, Ms. Narcisi urged NCVHS to recommend flexible, scalable standards that encouraged adoption from small to large practices and low-to-high volume prescribing physician specialties. Flexibility would allow physicians to consider critical factors without undue financial burden. PMS and EHR must be accommodated. She emphasized the physician–patient relationship had to be enhanced, not impaired. Physicians needed safety and efficiency and time was a precious commodity. Vendors needed a platform that created technology solutions that saved time for refocusing on care.
Ms. Narcisi said AMA wanted a private-sector approach with federal participation. Maintenance and modifications should not be hindered by an extensive rule-making process. State issues must be reconciled. Standards, terms and structures had to be unified; without standards, systems would be complex and costly. AMA called for further enhancement and testing of RxNorm and NDF-RT. Ms. Narcisi said NDC, which worked for pharmacy, was too granular for physicians and the free-form text submitted instead was not effective. Drug dictionaries had to be consistent to enable interoperability between systems using different commercial systems. All vocabularies and coding systems referenced should have an open (though not necessarily ANSI-accredited) updating process.
She said clinical quality improvement and patient safety were drivers for physician adoption. E-prescribing provided promise, but new technology had to be reliable: (e.g., the Maryland Medicaid pharmacy system went down, preventing dispensing for days). False-positive alerts must be minimized. Lessons learned from HIPAA should be considered. The critical factor in the protracted implementation had been an inability to upgrade PMS systems. Vendors need to see a commitment from the federal government before modifying their systems. DEA was developing an e-signature for Schedule II, yet for six years had not released the NPRM. Staggered implementation dates might be necessary.
Ms. Narcisi noted development and pilot testing were essential to determining standards work and there were relatively few interested groups of early adopters. She estimated that the percentage of physicians using e-prescribing was in the single digits. A study AMA was conducting with Forrester Research will be available in July. Ms. Narcisi said user authentication would be a key enabler; the authentication process had to be simple, quick, reliable and flexible. AMA worked with VeriSign to deploy a form of electronic ID. GSM was using the AMA Internet ID for pharmacy transactions in the Florida Medicaid program.
She said NCVHS and HHS had to understand the future Medicare environment to successfully implement voluntary e-prescribing in the Medicare program. Anticipating an annual five percent reduction in Medicare reimbursements from 2006 to 2012 and that the cost of care would exceed inflation, CMS estimated physicians would be paid 40 percent less. Ms. Narcisi noted matching grants had been approved but were not yet allocated.
E-Prescribing: The Provider Perspective
Dr. Hale said her rural, upstate New York practice twice failed to implement e-prescribing. She noted five key areas: e-prescribing must be useable for a physician office in multiple settings, interoperability was critical and required much information in usable form, freedom of choice for physicians and patients about drugs and pharmacies, incentives (monetary and workflow improvements, patient care and safety), and everything done should roll toward EHR.
Dr. Hale noted: the need for universal coding and nomenclature as well as standards that were simple and easy to implement, e-prescribing should be adaptable and scaleable to EHR and conform to DEA and HIPAA security standards, transparent systems were needed that enhanced the physician/patient relationship and helped overcome acquisition costs, and pilot testing was essential. She said the system's features and functionalities should: help e-prescribing prove itself speedy and efficient, be integratable, utilize PEER and other KBs for CDS, provide and not obstruct choices, enable prior authorization as a “smooth” part of the process, be clear of commercial bias while providing pricing information physicians need to make the best choices for patients.
She said RxNorm, NCPDP SCRIPT and NDC were important code sets. HL7 and CCR were also important for messaging. Most physicians doing e-prescribing used NCPDP SCRIPT, but Dr. Hale noted a wide variation in drug databases and dictionaries. HL7 and some of the more sophisticated systems were used for messaging.
Dr. Hale noted NCPDP SCRIPT had different proprietary codes and needed a single system. NDC was too granular. Multiple competing proprietary drug vocabularies were difficult to use in systems and not necessarily compatible. Script had to be supplemented with standards for format and vocabulary. Most drug and SIG messages were in free text, which did not help CDS. Prior authorization was only covered between prescriber and pharmacy; the loop had to be completed to health plans and PBMs. Formulary had to be presented in a neutral way. RxNorm did not go far enough in documenting critical details of the drug of choice. State regulations and the ordering of controlled substance had to be unified. HL7 and CCR needed to collaborate on the complete flow of messaging between parties.
Noting more communications occurred, Dr. Hale said nationwide adoption of standards and code sets was necessary. Code set gaps included standards reconciling state requirements; standards for doctor-level drug name and ordering format; standards for some SIG needs; format, vocabulary and structured formulary data and alerts; and messaging for all entities.
Dr. Hale emphasized the time and effort it took to develop detailed, unifying standards. Other barriers were the difficulty of gaining upfront support to overcome proprietary drug dictionaries. ACP believed the effort should be done similar to LOINC standard development.
Incentives suggested included: grants for pilot studies looking at implementation in multiple practices and development of methodologies for implementation, grants for regional health information networks, and Medicare pay-for-performance and reimbursement differential for utilizing the system. Dr. Hale emphasized incentives encouraging physicians to focus on EHR.
E-Prescribing: The Provider Perspective
AAFP agreed with 99.9 percent of AMA's and ACP's testimony, reiterating the need for a common nomenclature. Dr. Kibbe expressed AAFP's concern that standards not hamper progress toward EHR. E-prescribing was complicated because all parties had to communicate and connectivity was needed. Potential for increased safety and productivity had been identified, especially in practices currently using earlier generations of e-prescribing. Savings were available from workflow and process improvements (e.g., reduced calls, ease of storage and access to information). Safety, quality, and efficiency were integrated and integral.
He emphasized the context in which EHR occurred was critical and that comprehensive practice settings had to be considered in the prescribing process. Family physicians preferred for e-prescribing to be within the context of fully featured EHR, not through single purpose or “sliver” applications providing electronic connection to the pharmacy or PBM. Small and medium-sized offices, where most health care occurred, had to be able to easily implement the standards. E-prescribing should be voluntary and based on choice; not federal mandates. E-prescribing should be free to patients and physicians and without unfair incentives for commercial pathways or forced surrender of data for commercial use. Family physicians wanted information from all sources, especially for Medicare patients (two-thirds had multiple chronic conditions). Noting most EHRs could not exchange medication lists from physician-to-physician, Dr. Kibbe suggested CCR might help. He said the promise of e-prescribing was to correct this problem and the full history was critical to making that promise real.
E-Prescribing: The Provider Perspective
Ms. Schumann broadened the perspective to include dentists, psychiatric and mental health practitioners, integrative health practitioners, infusion services, pain management and ambulatory surgery centers, public health practitioners, practitioners in schools, home care, long-term care, skilled nursing care, assisted living facilities, and patients. She encouraged looking at: participants in the process, settings and workflow, minimum data set components, regulatory issues, and confirming educational system content. She noted vulnerable populations included: K-12 schools and college health, reproductive health, mobile clinics, gerontology patients, self-pay and indigent care, and hospice.
Ms. Schumann noted other issues: coding for complementary therapies; samples and other components not officially dispensed; compounded pharmaceuticals and those engineered and tailored for genetic therapies; e-prescribing for others resources, diets, PT and OT; supplies; authentication and credentialing, second signatures, unclaimed scripts; links with diagnosis, outcomes, and other interventions including nursing practice concepts; drugs lists from various sources; record retention, disposition, and discovery; and downtime and recovery procedures. She urged NCVHS to look beyond the physician and typical office settings.
Discussion
Dr. Kibbe said standards for SIGs might require a concerted effort but were necessary and attainable. It would be difficult and required collaboration but with the right people this could happen fairly quickly. Dr. Hale agreed these standards were important. Dose-checking would be difficult without standardization. She noted everything did not have to be done at once. Ms. Narcisi cautioned that testing was essential before requiring the standards.
Asked about issues contributing to the two failures in her rural office, Dr. Hale said one was a build-out of a home-grown EMR, the other a company that folded. Some issues were technology (i.e., the human/computer interface). The biggest problem was interoperability with the PMS. Multiple small health plans did not have resources to often provide updated formularies. Connection to pharmacies that did not leave their faxes on was difficult.
AMA estimated an average cost of $140,000 for a physician to implement e-prescribing. Dr. Hale noted the hard part covering changes in workflow, education and downtime during implementation made this less budget neutral. Ms. Schumann said everyone connected was impacted. She noted there were always questions about user friendliness, whether “real people” could use e-prescriptions or if they were relatively efficient. Experience with implementing systems suggested no real-time or cost saving. Dr. Kibbe said better data from the industry was needed. Large scale transformation was being created with considerable alignment among many parties and significant collaboration, but elementary projections or gauging progress could not be done without baseline information on the use, kind and cost of IT and what was needed in the next time period. Dr. Kibbe suggested this was a piecemeal way to view the emerging market. He cautioned about standalones that did not fit the total system and added costs to the overall industry.
Dr. Hale noted suggestions of upfront loans and others paid back out of ROI. Physicians lacked confidence in meeting economic demands on a year-to-year basis; implementing IT always cost more, and no one was sure of ROI. Ms. Narcisi said AMA surveys indicated that access to the Internet was high for physicians, but not their staff. Dr. Kibbe noted many studies focused on physician adoption of EHR locked on opinion rather than reality. None surveyed the 267 EHR vendors AAFP identified being used the way other industries provided data on production. He said this could not be done on the basis of a single medical society.
Noting RxNorm lacked detail needed in the private sector, participants said all potential users had to come to the table. Dr. Hale clarified that “commercial bias” referred to the risk that a physician's decision might be pushed by information provided (e.g., requiring “a number of clicks” to go beyond preferred choice, highlights, or embedded lists). Dr. Kibbe said he did not mean e-prescribing had to be totally free to the physician, but that there should not be direct or indirect transactions fees in the exchange or data supplied for commercial gain. Data need to be provided ethically and physicians had to have a say in how it was used. Ms. Schumann described how, depending on the state and class of drugs, various practitioners needed a second signature. Participants noted facility/actual provider and authentication issues and that the physician was critical for narcotics. Ms. Narcisi said AMA worked to develop an authorization for the physician's office staff, but no database similar to the physician master file existed for them and permission to use digital certificate had not been worked out.
Noting Dr.Kibbe had said access to the medication list was important but that he did not want e-prescribing required of physicians, he was asked about physicians who wanted the data when prescribing, but colleagues might not be participating. Dr. Kibbe said most prescribing information on paper was available electronically from pharmacy, PBM and health plan systems could be faxed if the provider did not e-prescribe. Dr. Hale noted the medication list was different than having CDS (clinical diagnosis/billing and intent for prescription).
Asked about overlap between prescribing in ambulatory and inpatient and integrating these environments, Ms. Narcisi said proprietary standards were being used and outreach was needed. HL7 handled the EHR ballot by going to different regions of the country. Dr. Hale noted hospitals used a separate formulary and drug dictionary; synchronization was lacking from inpatient to outpatient and errors could be created on discharge. Ms. Schumann said discharge instructions were unclear and patients and families were uncertain whether to continue previous or new meds. Patient/family advocates were needed. Dr. Kibbe noted divides existed between IT in big hospitals, integrated delivery systems and academic medical centers and the relative lack of IT in medium and small practices. The 267 vendors were small; HIS were huge vendors and did not understand ambulatory care, though this was changing. He suggested working with medical specialty societies.
Dr. Hale said commercially available systems for drug knowledge were not adequate, but with unification they could have 70-80 percent of prescribing information most commonly used. Ms. Narcisi emphasized testing and noted DEA had not resolved the controlled substances issue. Noting many parties were willing, Dr. Kibbe agreed good pilots and demonstrations under controlled circumstances were needed.
Asked the scope of standards for decision support and needs from the user perspective, Dr. Kibbe recommended AMA's Physician Consortium of Quality Performance Measurement Sets by Specialty Groups. Utilized in the Docket project, it helped develop pay-for-performance. Dr. Hale pointed out this was a broad topic ranging from drug-drug to PEER. She said the biggest issue was frequency of alerts; how much they could be turned off and still be useful. The challenge was deciding what to push and what to make available. Pilot projects were underway. Ms. Narcisi said sometimes e-prescribing took longer than writing a paper script and following that process. Ms. Schumann said ANA was working on protocols and decision support models to determine the significance of signs and symptoms and how to coach patients/families. Much clinical practice work had to feed into this.
Asked where to look for minimum data elements, Dr. Hale said there was agreement on some things in the EHI table. Dr. Kibbe suggested the XML tagging for CCR included some of this. Ms. Narcisi suggested the state law might need to be the floor rather than ceiling.
Mr. Blair commented on e-prescribing networks' and vendors' progress. Standards gaps filled by proprietary solutions provided pharmacy connectivity, PBM formularies, and contact with drug KB vendors. Asked if providing information to physicians for their decision making was the major gap, Dr. Hale agreed in general, but emphasized that multiple different proprietary dictionaries were an issue only tested in “perfect, focused” settings. Mr. Blair said he had the impression that gaps in terms of messages could be filled by vendors turning over some proprietary standards to SDOs. NLM was working on RxNorm and mapping to drug KB vendors. He asked what was suitable for a demonstration project in 2006 that would increase interoperability and enable physician involvement in patient safety. Dr. Hale asked whether the goal for low hanging fruit was to “pick perfect pieces across the country” or capture the best diverse practices. Ultimately they had to deal with the small, rural type. Ms. Schumann noted that 46 percent of hospitals were under 100 beds, largely rural, and did not care for 46 percent of patients. Dr. Kibbe agreed that looking at low-hanging fruit was appropriate. Noting everyone contributing to the EHI initiative was cooperative, Dr. Hale said she was optimistic.
Dr. Huff noted structured-coded SIGs were done in hospitals; standardization was needed. They heard code sets had to accommodate alternative care and devices, Second signatures were needed. Ms. Schumann suggested starting with coding and second signature. The multiplicity of settings, was more an implementation issue. Pointing out a disconnect on the PBM and payer side, Ms. Narcisi said prior authorization had to be accommodated.
Mr. Reynolds asked if the practitioner groups could supply service-level agreements giving a clear delineation of success for the pilots. Ms. Narcisi suggested gathering a sense of the gaps from applications already in place. Dr. Hale said the first criterion was would physicians use it for all patients all the time? Dr. Kibbe suggested considering solutions for quality and safety/efficiency enhancement in the context of the Future of Family Medicine report. Ms. Schumann said the recipe for success might not necessarily be speed and cost but in the balance between convenience and payback to patients.
Ms. Narcisi said vendors had to be at the table in order to be able to supply the requisite systems. Dr. Hale said a guideline should clarify that an unnecessary number of steps should not reflect bias. Ms. Cronin said that criterion could be part of the product certification process.
Open Microphone
Mr. Martin from Pfizer emphasized SDOs' understanding providers who often do not have the technical depth to get involved. He noted ways to structure RxNorm to capture differentiating features: codify them all into what a product was or use a pick list. He noted standardization of the list was missing. Manufacturers could assert to a standardized list. RxNorm function could “lump together” all manufactured products that fulfilled specific picked items, representing a middle ground. Dr. Martin suggested that a hierarchy of needs for RxNorm needed to be articulated in a structured fashion. Dr. Cohn noted NLM needed to be heard from soon.
Testimony from ExpressScripts suggested measuring success by lower cost and higher quality. In the managed benefit world, adoption and use was sporadic. It was important to plan now in order to gather pre-data to measure after implementation. The formulary component became secondary as components were added. Noting a primary barrier to adoption was accessing member data in the physician practice, the participant proposed standardized extracting methodology for e-prescribing to get at PMS data. Looking at e-prescribing from a Medicare cost savings perspective, the testifier cautioned about restricting system capabilities so much that possibilities for cost savings were lost. The testifier emphasized that driving from one drug to another was beneficial to patients. In Medicare, people not getting the most cost-effective choice often would not get anything. Mr. Blair agreed, suggesting that some of the anxiety heard had to do with the transparency needed in presentation of formulary. The participant noted getting the actual price was difficult and the drive had always been toward relative cost. The perception was that something that did not exist was being hidden.
Dr. Bizzaro noted confusion among ultimate end-users about the responsibility of content, application and network vendors and the standards. He agreed that issues had to be addressed through each entity, but urged physician end-users to look at current systems, emphasizing that the marketplace drove the best systems to the forefront.
While banner ads were disappearing, Mr. Martin said vendors still wanted to “talk” to physicians but were subtler and effective. No commercial messaging was a principle at Pfizer. Clinical information provided at POC for prescribing should be created and managed by someone no financial interest. Editorial independence was critical throughout the chain. Mr. Martin also emphasized that the total cost of care had to be considered, especially if more expensive drug might keep patients out of the hospital and reduce the cost to the system.
Dr. Steindel suggested asking pharmacists about how to handle off-label use. Dr. Cobuzzi from VA's Pharmacy Benefits Management said e-prescribing represented 95 percent of their e-prescribing. He advised e-prescribing standards be broad as possible and said efficiencies of process allowed doing more in the same time. He noted that functionality of collaborative practice agreements between pharmacies and physicians worked well in some states.
Mr. Robinson noted standards would not fix the problem of data stored in disparate locations. Nowhere was there a proposal for a centralized patient record. He emphasized CDS was only as good as the record and lack of a patient-centric and holistic, disease management approach. He suggested a CMS initiative for case management fees might be a way for someone to act as agent for the patient. Ms. Cronin said FDA was considering pay-for-performance with e-prescribing. Details will be discussed at the NHII conference.
Wrap Up/Next Steps
Dr. Cohn noted the Subcommittee will compile thoughts on this hearing at the break-out session during the June 16-17 full Committee meeting. He noted the June 24-25 Quality Workgroup hearings will review changes to billing and data collection practices related to HIPAA administrative and financial transactions; if enough Subcommittee members are available this might be a joint hearing. Members will hear from PBMs, state boards of pharmacies and medicine, and state Medicaid agencies at the July 29-30 hearings. Hearings are also scheduled for August 17-18. If July and August hearings are consolidated, private-sector and government payors will also be heard and the meeting might begin July 28. Notes from the May and previous hearings will be compiled into a ready reference of issues, themes (gaps) and questions/answers to help frame future conversations that, hopefully, members will review at the June breakout. Members said they would not make decisions before hearing from all testifiers as specified in MMA but noted many testified of gaps the SDOs should work on. Dr. Huff suggested applying resources for addressing issues in NCPDP SCRIPT and RxNorm and supporting HL7/SCRIPT collaboration for the structured SIG. Noting they also needed SDO input on their ability to respond, Mr. Blair suggested an accelerated schedule to September rather than mid-June. Mr. Reynolds said they had heard a lot that needed to be discussed at the break out session. Dr. Huff's straw structure would allow them to listen and ask questions differently. Remarking that, hopefully, what was learned in these three days would help everyone gain a better understanding, Dr. Cohn adjourned the meeting.
I hereby certify that, to the best of my knowledge, the foregoing summary of minutes is accurate and complete.
___/s/______________________________________________
Chairman, Date