260 likes | 372 Views
VA Drug Terminology Projects Update: NDF-RT, the New Drug Transaction, And Beyond. Presented to HL7 Vocabulary TC Meeting. VA: Steve Brown, Mike Lincoln Apelon: Mark Tuttle, John Carter, Mark Erlbaum October 3, 2002. Overview. NDF-RT: Current state and plans
E N D
VA Drug Terminology Projects Update: NDF-RT, the New Drug Transaction, And Beyond Presented to HL7 Vocabulary TC Meeting VA: Steve Brown, Mike Lincoln Apelon: Mark Tuttle, John Carter, Mark Erlbaum October 3, 2002
Overview • NDF-RT: Current state and plans • New Drug Transaction: Introduction • Interagency cooperation update HL7 Baltimore
Background:VA National Drug File (NDF) • Centrally maintained and distributed, locally modified and deployed. Uses include: • VistA POE Decision Support • Mail out pharmacy (57 million in 2001) • Single-inheritance hierarchy including • VA Drug Classes (anti-hypertensives, beta-blockers) • Products (Ampicillin 250 mg tab) • Ingredients (carbidopa, codeine phosphate) • NDCs (acetaminophen 325 mg tab, bottle of 100) HL7 Baltimore
NDF Reference Terminology (NDF-RT) • Explicit, multi-hierarchical model • Centered on drug ingredients for function, maintenance and economies of scale • Semi-algorithmic initialization followed by human review • Authoritative, collaborative content (FDA, NLM, HL7, others) • VA’s ERT strategy • Subject Matter Experts use COTS tools to develop, modify, & maintain enterprise standards • Abstracted terminology services HL7 Baltimore
Structural ID VA Drug Classes Database Links MeSH, CUIs, … Chemical Structure Mechanism of Action Active Ingredients Physiologic Effect Strength Dose Form Route of Adm. (HL-7) Pharmacokinetics Clinical Drugs Therapeutic Intent National Drug Codes (NDCs) Packaged Drugs Simplified NDF-RT Model HL7 Baltimore
NDF-RT Content • Drug Hierarchy • 3,977 Active Ingredients (incl. salts & esters) • 11,345 Orderable Drugs (= VA Products) • 87,210 Packaged Drugs (NDCs) • Initialized from … • VA National Drug File (Sept 2001) • NLM RxNorm Drug Names (Dec 2001) HL7 Baltimore
NDF-RT Content • Reference Hierarchies • 3,994 Diseases & Manifestations (“Intended Therapeutic Use” hierarchy) • 489 Chemical Structure categories • 402 Mechanism of Action & Physiologic Effect categories • 154 HL7 Dose forms • 58 Clinical Kinetics categories • Initialized from MeSH & HL7 HL7 Baltimore
Modeling Active Ingredients • 4000 active ingredients & 87,000 NDC level products • Descriptive aspects inherited down the hierarchy to orderable drugs & packaged NDC drug products which contain the active ingredients modeled • Increases modeling efficiency & productivity via “inheritance” HL7 Baltimore
Covering Drugs Ordered in One VAMC (June 2000) • Extracted 488 active NDF-RT ingredient names from sample of ordered drugs • Modeling 488 (12.3%) of all 3977 active ingredients in NDF-RT would cover at least one ingredient in: • 63.7% (7228/11345) of all NDF-RT orderable drugs • 78.2% (68194/87210) of all NDF-RT packaged drugs (NDCs) HL7 Baltimore
Est. Time to Review NDF-RT • Time & motion study • Averaged 12-13 minutes per concept, range 5-30 minutes • Est. throughput = 4 concepts per hour • Approx. 1000 person hours for ~4000 active ingredient concepts in NDF-RT HL7 Baltimore
Upcoming Work • VA Contract awarded 09/2002, to be completed Spring 2003 • NIGMS • Includes support for reference hierarchy development and some human review • Focuses on meeting needs of pharmacogenomic research network HL7 Baltimore
Background: VANew Drug Transaction • Thousands of NDC changes per year • Hundreds of clinical drug changes per year • Tens of New Molecular Entities per year • How to maintain? HL7 Baltimore
VA Formulary Updates: Current Situation • Manual system for review and entry into national formulary • Daily entry, quarterly publication • Multiple sources of data • Commercial KB’s, FDA, trade pubs, requests from field • Local hospitals also can add updates, these are then transmitted monthly to central office for review (workload dependent) HL7 Baltimore
New Drug Transaction Plan • Create an XML message based on publicly available data sources to carry drug information • Develop a semi-automated method and software tool to review and accept new drugs into the formulary HL7 Baltimore
A Little Plan Detail • Multiple “levels”: NDC, Clinical drug, NME • 3 “flavors”: new, change, retire • A single new NDC may cause an upward chain of updates HL7 Baltimore
New Dose Form Example: Copaxone • Generic Name: Glatiramer Acetate • Treats: Multiple Sclerosis • Currently, the only GA packaged drug in NDF-RT is: • GLATIRAMER ACETATE 20MG/ML SYR [NDC: 66914-1009- 1] • Injectable solution • Add: • GLATIRAMER ACETATE 20MG/VIL [NDC: 0480-0126-84] • Powder, lyophilized for solution HL7 Baltimore
Necessary Transaction Steps • “Clinical Drug” is the level of the dose form information • “Dose form” is different (???) because existing NDC is already in solution • Potential new concepts: • Clinical drug concept (???) • NDC-level concept HL7 Baltimore
Government Drug Terminology “Partners” Drug Company FDA NCI DEA VA CMS DoD NLM IHS In Place CDC NIGMS PTO HL7 Baltimore Drafted
FDA Labeling • Being created as a Structured Document using HL7 Clinical Document Architecture • Label Warehouse System prototype in development • Selected data to be shared with NLM to help populate RxNorm HL7 Baltimore
NIGMS • Helping develop reference hierarchies, especially pharmacokinetics • Tied to PharmGKB project at Stanford • Thanks to Chris Chute! HL7 Baltimore
NCI Experimental Agents • Approximately 300 agents, ranging from the common to the very experimental • A separate drug class hierarchy, focused on mechanisms of action and specialized anti-neoplastic categories HL7 Baltimore
Vision:Interagency Information Flow for the New Drug Transaction (Off-Label Indications, VA MoA, other VA DSS aggregations, and other usage information) (Approved Drug Information) (FDA+NLM Information) FDA NLM VA NDT Electronic Label HealthePeople& Other Federal Systems (Additional structure information) Non-Federal Users These Flows to be HL7 candidate Standards HL7 Baltimore
Challenges • Delivering drug definitions and reference hierarchies with “face validity” • Managing multi-agency, multi-institution collaboration • Varying needs & expectations • Information model, terminology model, knowledge base interaction HL7 Baltimore