Supratherapeutic INR Clinical Decision Support

Transcription

1 Supratherapeutic INR Clinical Decision Support 11/29/2009 Northwestern University MMI Barbara Antuna, MD Susan Bersheer Mary Moore

2 Contents Overview/Knowledge... 3 The Model Selection of CCDS Interventions and Logic... 6 ERROR PREVENTION AND HOW CCDS CAN HELP... 6 CARE PROCESS AND HOW CCDS CAN HELP... 7 THE MODEL SUPRATHERPUTIC INR RESULTS... 7 Available Information Systems Meaningful use: Qualifying EHRs Demonstrate Meaningful Use goals: goals: goals: Hospital Incentives $2 Million Base, Adjustment by Discharges Physician Medicare Incentives HITECH Act Evaluation Preliminary Validation Final Validation References:... 20

3 Overview/Knowledge Many disease states utilize antithrombotic therapy. As medical care becomes more complex and the population ages, this list will continue to grow. A partial list is included in table 1. Table 1 Disease states utilizing antithrombotic therapy Venous and pulmonary thrombosis Atrial fibrillation Ischemic stroke Prosthetic heart valves Coronary artery disease Heart failure Post operative lower extremity surgery Antiphospholipid syndrome Protein S deficiency Antithrombotic therapy can be accomplished through the use of parenteral or oral medications. Most commonly, oral therapy occurs with warfarin, also known by its trade name Coumadin. Of patients enrolled into a National Trauma DataBank, warfarin use was estimated at 2.3% in 2002 and increased to 4.0% in In patients greater than 65 years of age, this prevalence went from 7.3% in 2002 to 12.8% in This increase in utilization in warfarin parallels both the increasing complexity of healthcare and the aging of the population. Unfortunately, the major adverse effect of warfarin is bleeding. One study reported bleeding rates that included an incidence of fatal bleeding ranging from 0-4.8%. (see Table 2) 2 Bleeding risk increases with age. In addition, bleeding risk increases exponentially with anticoagulant effect therefore there will be a substantial increase in this risk with over anticoagulation. Efficacy of warfarin can be measured in the serum of patients taking it. The Prothrombin Time (PT) is a measure of extrinsic pathway of coagulation. Historically, this has been the most reliable test to measure this effect. Due to problems with a wide variety of Table 2 Annual Bleeding Rates Fatal 0-4.8% Major % Minor about 15% thromboplastin reagents the protime undergoes a mathematical correction and is reported as an International Normalized Ratio (INR). An INR greater than 3 is considered supratheraputic reflecting that the patient is over anticoagulated. Patients who have been over anticoagulated can be managed with treatments ranging from simply withholding warfarin to infusions of vitamin K, fresh frozen plasma (FFP) and Prothrombin complex Concentrate (PCC). Vitamin K overrides the warfarin induced antagonism on the coagulation pathway within the liver. FFP contains all the coagulation factors while PCC Fitzmaurice D, Blann A, Lip G. Bleeding Risks of antithrombotic therapy. BMJ 2002 Oct. 325(7368):

4 contains factors II, VII, IX, X. This latter treatment is emerging as highly beneficial but costly addition. Clinical guidelines exist for the treatment of supratheraputic INR values. The American College of Chest Physicians published evidence based guidelines for management of nontheraputicinrs in their eighth edition in Strength of evidence ratings in this guideline consist of four 1C, one 2A and one 1B ratings. See Table 3. 3 Table 3 Strength of Evidence Ratings 1C INR <5 w/o bleed Withhold/lower dose Monitor more frequently 1C INR>5 but <9 w/o bleed Withhold/lower dose Monitor more frequently 1C Serious bleed, any INR Withhold IV vit K FFP/PCC, may repeat 1C Life threatening bleed, any INR Withhold FFP/PCC IV vit K, may repeat 1B INR>9 w/o bleed Withhold If bleed risk-higher dose PO vit K Monitor more frequently 2A INR>5 but <9 w/o bleed Withhold If bleed risk-low dose PO vit K Monitor more frequently There are many safety issues surrounding the anticoagulation treatment. In a study conducted by the National Coordinating Council for Medication Error Reporting and Prevention (NCC MERR) for , heparin and warfarin accounted for 49% of high alert medication errors that resulted in harm. (see Fig 1) 3 Hirsch J, Guyatt G et al. Pharmacology and Management of Vitamin K Antagonists. Chest. 2008; 133:160S-198S.

5 FIG 1 As a result of these safety issues, several national organizations have created safety goals surrounding anticoagulation treatment. These organizations include the Institute for Safe Medication Practices (ISMP), the Agency for Health Care Research and Quality (AHRQ), the Leapfrog Group, the Joint Commission and the Institute for Healthcare Improvement. PROBLEM STATEMENT Patients on warfarin are at risk for serious bleeding episodes. This risk correlates to the INR result. Decision support can be incorporated at the lab alert for elevated INR resulting in better adherence to clinical guidelines and documentation of a treatment plan for these patients thus improving safety overall. Computerized Clinical Decision Support (CCDS) exists for antithrombotic therapy but the majority of these tools support oral anticoagulation dosing. One study reviewed effectiveness of computer support for determining optimum drug dose. The conclusions were that this decision support reduced time to achieve therapeutic benefits with fewer unwanted effects and helped practioners tailor drug doses more closely to the needs of individual patients. 5 Another study compared computer decision support within a prescription application that provided a nomogram or made specific dosage adjustment recommendations. 6 A study from Vanderbilt University implemented a script that included decision support for treatment of DVT/PE in adults. 7 One study evaluated the use of CCDS for high INR values and blood product transfusion. Their conclusions were that education and computerized CS both decreased the percentage of inappropriate transfusions. 8 At the present time, no specific CCDS exists in the literature for treatment of patients with supratheraputic INR values. In this evaluation, the clinical objective is to improve treatment of and appropriately utilize resources for patients with supratheraputic INR values. The desired action is to provide specific Walton R, Dovey S et al. Computer support for determining drug dose: systematic review and meta-analysis BMJ, 1999 Apr 10;318(7188) Colombet I, Bura-Riviere A et al. Personalized versus non-personalized computerized decision support system to increase therapeutic quality control of oral anticoagulant therapy: an alternating time series analysis. BMC Health Ser Res, 2004;4: Rothschild J, McGurk S et al. Assessment of education and computerized decision support interventions for improving transfusion practice. Transfusion, 2006; 47:

6 computerized clinical decision support at the lab alert embedded within an EMR. The desired outcome is 100% compliance in documentation of a treatment plan for all patients with supratheraputic INR values. The Model Selection of CCDS Interventions and Logic The process of noting and treating supratheraputic INR values can break down in many areas. This can result in errors, poor decision making and poor medical care. To help prevent errors, errors of commission and omission need to be addressed. Decision making can be optimized by improving compliance with simple care guidelines and improving the overall appropriateness of the workup and development of a treatment plan. Improvements in documentation will improve with the care process. ERROR PREVENTION AND HOW CCDS CAN HELP Errors of commission generally occur when a key piece of data has not been taken into account during the process. CCDS can help by providing alerts. Alerts can occur at several workflow points including ordering, order handling, therapies and when results arrive. In dealing with supratheraputic INR values, the alert would be most beneficial when high INR results arrive. See Fig 2. Fig 2 Lab order results ALERT CCDS can also address errors of commission by providing relevant data for ordering. This occurs when results arrive and in the ordering steps of workflow. This helps optimize decision making by ensuring all pertinent data have been considered and organized to provide an overall view and highlight needed actions. In this scenario, this process again occurs when high INR results arrive. See Fig 3. Fig 3 Lab order results alert Provide Clinical guideline Errors of omission occur when a needed therapy is not prescribed. CCDS can help by providing pertinent order sets and clinician documentation forms. Order sets involve the ordering workflow step. Placed here, these sets help promote adherence to standards of care. Specific INR value driven order sets conforming to the guideline are inserted into the workflow. Clinician documentation forms involve the documentation workflow piece. This CDSS step provides documentation for quality and continuity of care as well as helping with reimbursement and legal issues. In treatment of high INR values, this documentation occurs when the corresponding order set if selected within the EMR. See Fig 4.

7 Fig 4 Lab order results alert DECISION MAKING AND HOW CCDS CAN HELP Provide Clinical guideline ORDER SETS Decision making can be optimized by improving compliance with care guidelines as well as improving the appropriateness of the overall workup and development of a treatment plan. To improve guideline compliance, CCDS can provide order sets with the secondary result being the documentation of how each high INR result was treated. Appropriate workup and treatment plans benefit from CCDS order sets as well. CARE PROCESS AND HOW CCDS CAN HELP Care processes are improved with improved documentation which is accomplished when INR value driven order sets are selected. THE MODEL SUPRATHERPUTIC INR RESULTS In this case, protime orders are placed via CPOE within the EMR for patients of warfarin. Labs are resulted with those greater than 3 triggering an alert. If the clinician feels the result may be in error, the test can be repeated and the result will either fall out of the CCDS arm or remain in. If the INR remains greater than 3, the clinical guideline in the form of order sets is provided. See Fig 5. Fig 5 Protime order Labs resulted Low or Normal INR CDS For low INR HIGH INR CCDS For high INR?lab error Repeat protime Clinical guideline

8 The next decision arm requires documentation whether the patient if bleeding by choosing the bleeding order set with any INR>3. If bleeding does not exist, the next decision will flow according to specific INR values. The provider will be directed to those specific sets. See Table 4 and 5. Table 4 INR RESULT Document if Pt BLEEDING IF NO Y/N Order set for INR Y/N Order set for INR Y/N Order set for INR >20 Y/N Order set for INR > 20 IF YES Bleeding Order set Bleeding Order set Bleeding Order set Bleeding Order Set

9 Table 5 Order set For INR For INR 5-9 For INR For INR >20 Order set BLEEDING This Hold coumadin Hold coumadin Hold coumadin Hold coumadin Hold coumadin CCDS for New Dose order New Dose order New Dose order Bleed risk? Go to Vit K order Vit K order FFP/PCC Order Vit K order FFP/PCC Order VitK order Repeat INR 2 d Repeat INR 24 hr Repeat INR hr Repeat INR 6-12 hr Repeat INR After FFP/PCC

10 supratheraputic INR results is an attractive model due to the ease of development, acceptability and impact. The alert activity is likely to have high acceptability and impact because of high specificity for the alert. Almost every high INR value needs some change in management. The reactive alert is easily incorporated within the EMR structure as lab alerts and CPOE with the ability to develop order sets already exist. Performance measures will include reviewing utilization of orders for withholding warfarin, vitamin K, fresh frozen plasma and prothrombin complex concentrate. Additionally, providers use of the CCDS will be monitored via audit tools. Lastly, quality of care and costs of service will be followed. Available Information Systems Through a 1994 AHRQ grant, our Health Information Exchange had its beginnings. The need for sharing data (labs, transcription and diagnostic images) will prove to reduce healthcare costs by eliminating repetition of expensive tests. The MedicityProAccess application is not dependent upon certain EMR software but shares information through HL 7 codes.medicityproaccess gathers all patient information into a central repository, where the patient information can be shared and accessed by all participating healthcare providers.currently the hospital EMR application is Siemens Soarian. We have utilized Zynx Health evidence based order sets to support our CPOE and Clinical Decision Support rules. Currently ARRA funding is available to qualifying healthcare providers in an effort to encourage the adoption and use of Health Information Technology (HIT) systems and provide the infrastructure required to exchange medical information. Payment incentives are offered for early adopters, and will be paid out over four years beginning in Hospitals are eligible for from $5 $10 million (adjusted from a base of $2 million) dispersed over a maximum of four payments, dependant on when they qualify for the first payment. Annual metrics (discharges, charity care and inpatient days) will be used to determine award.eligible professionals can receive up to $44,000 total per individual provider, with up to five payments made to individuals based on when they qualify for the first payment and the physician s Medicare patient volume. 9 Timeline: payments start in Hospitals have until 2015 to demonstrate meaningful use of EHRs to begin receiving incentives. Following is an explanation of meaningful use and incentives to healthcare providers to encourage early adoption of EMRs and patient information sharing. Meaningful use: Improve safety, quality, efficiency; reduce health disparities Engage patients/ familiesensure privacy and security protections for health information Improve care coordination Improve population and public health 9 Eclipsys, (2009). Leveraging the American Recovery and Reinvestment Act Critical Success Factors to Maximize ARRA HITECH Incentives,

11 Qualifying EHRs Provide clinical decision support Supporting physician order entry Capture and query information relevant to health care quality Exchange electronic health information from other sources Demonstrate Meaningful Use Electronic exchange of information to improve quality and care coordination including e- prescribing Reporting on quality measures goals: Use computerized physician order entry (CPOE) with 10% of all orders directly entered by authorizing provider Implement one clinical decision rule relevant to high clinical priority Include lab tests in EHRs and results shared electronically with public health agencies Generate patient problem lists of conditions for quality improvement and outreach Provide patients and family with timely access to key clinical information 2013 goals: Generate and transmit prescriptions electronically Manage chronic conditions using patient lists and decision support tools Bar coding for medication administration Patient access to personal health record (PHR) populated with real-time patient data 2015 goals: Minimal levels of performance on quality, safety and efficiency measures Comprehensive patient data from all available sources Automated real-time surveillance of adverse events, disease outbreaks and bio-terrorism Incorporation of clinical dashboards into the EHR TempDev, (2009). Maximizing Stimulus Incentives for Your Organization,

12 Hospital Incentives $2 Million Base, Adjustment by Discharges Year of Adoption Payment for adopting before % 75% 50% 25% If first adopting in % 75% 50% 25% If first adopting in % 75% 50% 25% If first adopting in % 50% 25% If first adopting in % 25% Penalties if not adopting by % 66.66% 100% Physician Medicare Incentives HITECH Act Year of Adoption Adopted EHR in 2011 $18,000 $0 $0 $0 $0 Adopted EHR in 2012 $12,000 $18,000 $0 $0 $0 Adopted EHR in 2013 $8,000 $12,000 $15,000 $0 $0 Adopted EHR in 2014 $4,000 $8,000 $12,000 $15,000 $0 Adopted EHR in 2015 $2,000 $4,000 $8,000 $12,000 $0 Adopted EHR in 2016 $0 $2,000 $4,000 $8,000 $0 Total $44,000 $44,000 $39,000 $35,000 $0 10% additional payment for health $4,400 $4,400 $3,900 $3,500 $0 professional shortage areas Total $48,400 $48,400 $42,900 $38,500 $0 13 The Hospital Clinical information systems are designed to manage the administrative, financial and clinical aspects of a hospital. The software components of these information systems we are concerned with are those of Lab and CPOE. Using CPOE physician orders are transmitted over the hospital network to the lab. After the laboratory receives the physician order, the lab information management system or LIM, automates lab sample management. When gathering patient samples a bar-code is printed that is patient specific, ensuring correct patient Eclipsys, (2009). Leveraging the American Recovery and Reinvestment Act Critical Success Factors to Maximize ARRA HITECH Incentives, 12 Beacon Partners, (2009). ARRA Position Paper, 13 Athena Health, (2009). A Summary of the HITECH Act,

13 identification. The bar-codes are scanned at the beginning and end of analysis, where the results are validated in the LIM system. Once the abnormal lab is registered, the LIM in turn reports those abnormal results to the CDS engine, which in turn pushes an alert out to the ProAccess application and will be triggered when the physician views that patient s labs. Outside Physician Practices will accessthe system externally from a secure connection. For our purposes, we have our outside practices connect via Citrix since it is encrypted and provides a secure exchange of the patient information. The ProAccess application will form the interface with the hospital EMR, Lab and physicians. The Lab interfaces with the ProAccess database through the HL7 architecture that we will discuss next.health Level 7 Architecture focuses on the application layer protocols for health care. The 7 refers to the 7 th layer in the OSI model that is the application layer providing messaging, file transfers, and network management. The 7 th layer is where a user interfaces with all the electronic services on the information system. 14 HL7 allows the interoperability between many information systems but again, our focus will be limited to Electronic Medical Records, Laboratory Information Systems, and Pharmacy systems. Hospital Information Systems / Intranet Data Sources Internet Interface System Information Health Access Layer Privacy Security Rules Based CDS Engine Confidentiality Audit Trail Validation Health Access Layer Patient Encounters Healthcare Providers Internal / External Data Sources Order Results Patient Consent Terminology CDS Rules Access / Audit Logs HL 7 Architecture Clinical Guideline Documentation Tech-Faq, (2009). What is the OSI Model? The Seven Layers of the OSI Model, 15 CMS, Medicaid Information Technology Architecture, HL7%20MITA%20Project%20and%20the%20MITA%20IA.pd

14 Hospital Information Systems accessed by intranet users will access the EMR or ProAccess at their discretion since both are tied to the clinical decision support engine. Internet interface will tie healthcare providers outside the organization into the shared patient information housed in the ProAccess database. You see above that the Health Access Layer is interfaced with Patient encounters and healthcare providers through internal and external data sources. Within the health access layer, are the information protection rules that will be applied to patient information. 1. Privacy HIPAA, and the HITECH act have tightened the rules around patient privacy and information breaches. Information systems in use must secure the shared information as well as allow its access to authorized users. 2. Access that is needed to fulfill your job responsibilities is just one way that permissions can be restricted and a sound security plan implemented. 3. The Rules based Clinical Decision Support Engine houses the knowledge database that triggers the physician alerts when a supratheraputic INR is transmitted from the lab. The medical logic module in the CDS engine identifies the needed data elements from the lab values and then uses computer logic to analyze those data elements. The CDS engine then generates an alert and sends it to the ProAccess database. 4. Confidentiality is insured in the same ways that information is held private and secure. 5. Audit trails are an important way to monitor access to patient information. A break glass feature is in place to force the user to say why a patient s information is being viewed if the users job normally does not allow that access. 6. Validation of the HL7 communication must be done before the CDS alert can be sent to the live EMR environment. The messages must present the data in the proper format and include all the information that is to be presented to the end user. 16 The system architecture that exchanges patient information is a network connected by VPN or physical optical fiber or copper connections. Clinical decision support is at the heart of the system, with all physician orders routing through it. The physician from outside the hospital is allowed through secure access to see patient lab results and be alerted when those results are out of the normal range. Once an alert has been generated, it is passed back through the application interfaces into the patient EMR. All of the components that go into this information exchange include the terminologies used by all clinical applications. LOINC used in lab, RX Norm used in pharmacy, and SNOMED vocabulary are some of the terminologies that must be correctly interpreted by the CDS rules engine. A further complication occurs if a patient refuses to consent to share their information. In a case such as this, the ProAccess application must have rules that do not allow that patient s information to be viewed. This brings up the necessity to audit access logs. Not only do we monitor the possibility of a breach of information, we must also audit the information that is accessed, and that it is only accessed by authorized users. With all the security precautions in place we can turn our attention to the clinical guideline documentation. 16 Wright, Knobloch, Pecher, Mejicano, Hall, (2007). Clinical Decision Support Systems Use in Wisconsin,

15 Proposed Process Internet Interface CDS Rules Engine Laboratory CPOE 5 Hospitals 1 ProAccess Database Walk-in Clinics Hospitals 1. Physician lab order 2. Lab retrieves order: results transmitted to ProAccess DB 3. If INR is HIGH evaluated by CDS rules engine, transmits to ProAccess DB 4. MD receives CDS results alert / info displayed 5. MD confirms message / documents action MD Practices 6. CDS engine evaluates results, transmitting to ProAccess DB 1. Physician lab order will be transmitted either internally or externally. 2. The Lab will retrieve the order: results transmitted to ProAccess database 3. If INR is HIGH The ProAccess database sends the lab value to the evaluated by CDS rules engine, which interprets the result and transmits the alert back to the ProAccess database where it waits for physician retrieval. 4. MD receives the alert information displayed for that patient. 5. MD confirms the alert message, and then documents the clinical action to be taken. 6. CDS engine evaluates the response results, transmitting them back to the ProAccess database. Once a physician is in the system, they are required to respond to any alert that they encounter by documenting the action they plan to take when triggered by the abnormal alert. So from the ProAccess clinical inbox, the physician clicks on a patient who has an abnormal lab. When the physician discovers it is a supratherapeutic INR, the alert generated by the CDS engine will notify the physician that immediate attention is required for this patient s lab value. In this way the physician can electronically document their course of action, and this documentation will in turn become a legal part of the patient s medical record the EMR.

16 Evaluation Evaluation is a means to assess the quality, value, effects and impact of information technology and applications in the health care environment, to improve health applications and to enable the emergence of an evidence-base health informatics profession and practice 17 Our designed decision support tool for Identification of Supratherapeutic INR Results incorporates calculated INR results and integrates the decision tool for the management of supratherapeutic INR value guidelines, recommended by Baylor University Medical Center in These guidelines explicitly address guidance and assessment for elevated INR results. Our aim of this tool is that physicians receive immediate feedback of clinically important information such as elevated INR with the aim to correctly treat each specific elevated IRN result utilizing our clinical decision support tool thereby decreasing variations of practice. Our tool accounts for management of supratherapeutic INR values 3.1 to greater than 20.0 and addresses the resulting treatment and assessment necessary. The decision support tool is not designed at the individual level only to address treatment of each level of an INR result, not taking into account individualized patient data such as age, gender, diagnosis or other risk factors. The tool using the Baylor guidelines automatically incorporated the resulted INR with the accompanying decision analysis for each patient. The efficiency of resulting treatment and assessment guidelines at the time of the high INR alert uses the integration of the system and results in better quality of care then non-integrated system. The hospital HCIS collects and results the patients INR, our clinical decision support tool is used by doctors and nurses in order to receive an alert to the elevated INR result, visualize the result, select the correct procedure for treatment to a specific INR result, and documentation of decision. The documentation within the HCIS is integrated with the EMR application. Quality assurance of the system is aimed at improving the quality of care by correctly addressing the accurate treatment for each elevated INR level. Our data elements were identified on the basis what was important to the involved stakeholders, and consensus was reached on the management by using previously similar published literature regarding thrombolytic treatment therapy. Data elements utilized to query patients included gender, age, race, medical record number, ICD-9 diagnosis codes, laboratory results, allergies, drug-drug interactions and provider. These data element drive the rules for our clinical decision support tool. Preliminary Validation Validation of data output is accomplished by collection of past patient charts presenting with similar disease process and treatment of diagnosis such as atrial-fibrillation, stroke, or surgery, with the resulting manual decision management made by responsible providers. These same cases are then put through our decision support tool and the resulting recommended treatment is 17 Ammenwerth E, Brender J, et al.(2004) HIS-EVAL Workshop. Visions and strategies to improve evaluation of health information systems. Innsbruch. Int J Med Inform Jun 30;73(6):

17 review by our stakeholders who give their opinion on appropriate management and treatment therapy. Adjustment to our tool is according to the advice of our stakeholders in order to achieve consensus on treatment therapy. Final Validation Finally a trial is ran where the tool gives advice on several patients first and then is compared once more to the advise of stakeholders. Subsequently all possible inputs into our tool is analyzed and responses are developed for each piece of input and in turn stakeholders again provide responses for input into the tool this type of completeness will assess how comprehensive our tool is in terms of it ability to assess INR results related to our data input. This feedback loop mechanism will establish an alignment with published guidelines and must continually be updated for operational management. The benefit was based on treatment alternatives in order to improve patient outcomes, alert physicians to results, compliance with guidelines, documentation and cost effectiveness of treatment. A CDSS for anticoagulant study was demonstrated at Whittington Hospital, a teaching hospital in London. The study and subsequent building of the CDSS was for the purpose of oral warfarin control in patients in the area of cardiovascular medicine. The study audits demonstrated trainee doctors having poor knowledge of therapeutic range even when treatment guidelines were printed on pathology forms. Most anticoagulant studies are for the poor control of heparin therapy in patients. 18 Discussion Components important to our CDSS tool such as patient demographics, allergies, medical history, current or prior procedures, care provider and current medication list must be available in the system in order for the CDSS tool to function properly and must be entered by reliable staff. The tool must be used appropriately for the intended use, which is the physician receives the alert to treat only supratherapeutic INR results. Patient safety is a high priority the tool improves the efficiency of health care delivery to reflect this goal with the improved treatment of high INR s. Implementation This current model is ready for implementation in an organization using an integrated system with CPOE. The implementation will impact care providers and require a cultural change in practice (the tool does not give the ability to by-pass or change treatment). Success is dependent on physician and staff buy-in and adoption of the new treatment along with compliance in documentation of the treatment plan. Intense staff education, training and support are necessary to alleviate fears, concerns and questions. The intense education of staff is two-fold, one aspect 18 Wess, M., Schauer, D. et al..(2007). Application of a Decision Support Tool for Anticoagulation in Patients with Non-valvular Atrial Fibrillation. PubMedCentral April;23(4); Retrieved November 22, 2009 from

18 is to the benefits of this change in practice and to prevent false expectations; and the second aspect is to educate the patient to promote the hospitals effort of increasing patient safety. Evaluation and monitoring is crucial for the continued success, metrics such as physician adoption evidenced by no overrides to the system tool, tracking audits tools to monitor performance, turn-around-time from alert to treatment, and ADRs are utilized. Feedback mechanisms include , suggestion box, support teams, hospital intranet, surveys, observations and direct communication. Evaluation of usage, software and hardware effectiveness is monitored and adjusted according to the end-user requirements along with system upgrades and issue patches as necessary. The strengths of the model is the reduction of errors in ordering and alerts to the physician which in turn allows for quicker turn-around in time of patient care. The weakness is the change in process and issues with any downtime due to computer maintenance or unscheduled downtime. Key Assumptions Laboratory values are correct including instrumentation effect on the International Sensitivity Index (ISI) and with no erroneous reporting by the thromboplastin manufacture. 19 Physician buy-in Hardware Secure access from off site In-Scope Analysis and requirements Design Business Process mapping Construction Subject matter experts (physicians) Consensus on evidence-based context and therapy by Physician Advisory Group Testing Care providers competency Implementation End-user training Evaluation Streamline workflow as needed Development of policy and procedures Development of reports specific to physician management Development of metrics for the evaluation of tool Out of Scope 19 National Guideline Clearinghouse, Antithrombotic therapy Supplement.(2007) Aug. 64p. Retrieved November 24, 2009 from &string=elevated+and+inr

19 Anticoagulation therapy prone to poor control Any functional outside the realm of the system Outpatients Hardware procurement Future Extensions The model directs the care provider to treatment for specific INR levels however cannot guarantee that treatment such as Fresh Frozen Plasma is given in the correct dosage or if enough FFP is administered. 20 Improvement would include the inclusions list of possible causes of the elevated INR this might preclude administration of Vitamin K or FFP. 20 Rodgers,M.(2006). The Use of Vitamin K. Retrieved November 24, 2009 from

20 References: guideline 14045/antithrombotic_ther apy_supplement guideline_.html ND+INR title=mgmt%20supratherapeutic%20inr guideline 14045/antithrombotic_ther apy_supplement guideline_.html