medCPU Radiology Case Study

The AcademyThe Health Management Academy

Clinical Decision SupportThe Academy Innovation Series:

Reducing Inappropriate CT Imaging in the ED

Clinical Decision Support The Health Management Academy © 2015 2

Authors

James (Jay) Flounlacker, M.B.A.Senior Vice PresidentThe Health Management Academy

Charles Watts, M.D.Executive-in-ResidenceThe Health Management Academy

The Academy Innovation Series: Clinical Decision SupportThe Health Management Academy, in partnership with medCPU, created the Center for Advanced Solutions in Healthcare (the Solutions Center). The Solutions Center will accelerate the adoption of innovative clinical decision support solutions designed to increase quality of care and improve clinical, financial, and operational performance at the point-of-care.

The Academy interviewed hospitals and health systems to document the dissemination of clinical decision support software tools in multiple clinical settings for various diagnoses. This case study reports on the medCPU Radiology Advisor tool and a methodology developed by Imaging Advantage to study non-traumatic headache diagnoses. It was utilized in the Emergency Department at West Suburban Hospital, located in Oak Park, IL, owned by Tenet Healthcare Corporation (Tenet).

The project was undertaken by Imaging Advantage in partnership with West Suburban Hospital with funding from the Center for Medicare and Medicaid Innovation (CMMI) located within the U.S. Center for Medicare and Medicaid Services (CMS).


Contents

Authors 2

The Academy Innovation Series: Clinical Decision Support 2

Key Findings 4

Study Introduction: An Opportunity for Innovative Improvement 4

Project Background and Evolution 4

Designing a Simple Solution 4

Efficient Solution Deployment 5

Identify Physician Champion 5

Develop Implementation Plan and Teams 5

Protocol Refinement 5

Server Installation & Data Feeds 5

Deployment 5

Ongoing Experience, Behavioral Change and Outcomes 6

Physician Behavioral Change 6

User Feedback 6

Acceptance and Outcomes 7

Expansion 7

Conclusions & Key Takeaways 8

Benefits 8

Success Factors 8

Lessons Learned 8

Footnotes 9

References 9

The Health Management Academy 10

medCPU 10


Key Findings �Clinical decision support tools reduced utilization of unnecessary and duplicative imaging services and identified cases where imaging services were needed but not ordered.

� Physician adoption and acceptance resulted from use of evidence-based protocols and integration of clinical decision support within the existing CPOE workflow

�Clinical decision support tools can be integrated into an EMR with relatively low IT time and resource investment.

� Increased revenues resulted from performance of appropriate imaging services and reduction in the number of claims denied due to incomplete coding.

Study Introduction: An Opportunity for Innovative ImprovementThe healthcare environment is adapting to increased pressure to improve performance and reduce costs as both consumers and providers are taking on greater risk for the costs of care. This case study explores how one hospital has deployed an innovative clinical decision support solution designed to reduce unnecessary utilization of CT imaging in the emergency department for patients with non-traumatic headache. This experience illustrated the roles that leadership, culture, and solution design play in successful deployment of clinical decision support that drives enduring change in physician behavior.

Project Background and EvolutionIn 2012, Mark Montoney, then CMO of Vanguard Health System (acquired by Tenet in 2013) began a collaborative project with Imaging Advantage to design a more efficient radiology model. The project was accepted by CMMI as a method to test new payment and delivery models. Imaging Advantage assumed the lead role on the project working with West Suburban Hospital.

For purposes of the study, the team narrowed the objective of the project to the performance of appropriate CT scans of the head for non-traumatic headache to improve quality and reduce the documented over-utilization of CT scans. Dr. Montoney highlighted that the project presented “a good opportunity to improve quality by reducing utilization that didn’t bring value, and could potentially

be harmful through unnecessary exposure to radiation”.1 Tenet’s Chicago market hospitals were identified as the “laboratory” for the initial study, which would build on the Imaging Advantage radiology platform combined with an innovative clinical decision support solution from medCPU.

Designing a Simple Solution The hospital required a solution that would efficiently integrate

with its existing EMR, read both structured and unstructured data (such as physician progress notes) in real time, fit into the existing clinical workflow, apply best-practice protocols, and provide information for clinicians to evaluate performance of the solution. A recent article published in the Annals of Internal Medicine identified two solution design factors that correlate with effectiveness of clinical decision support in improving appropriate diagnostic imaging: characteristics of the intervention, and the use of audit and feedback.2 Interventions were characterized into different levels based upon whether they ranged from passive informational only alerts that required no action to a “hard stop” requiring the clinician to gain a second opinion before proceeding.

The project utilized a point-of-care clinical decision support application that monitors both structured data and free text clinician notes in real time during encounters. This monitoring process runs in the background without distracting the clinician from patient care and assesses orders issued by the physician with evidenced-based clinical guidelines and protocols.

A CULTURE OF INNOVATION“We have a culture of Innovation that creates new solutions for the challenges and opportunities in the healthcare system”

– Tenet Healthcare, Core Values

QUALITY IMPROVEMENTThe project was “a good opportunity to improve quality by reducing utilization that didn’t bring value, and could potentially be harmful”

– Mark Montoney, M.D.

INTEGRATING WITH CLINICAL WORKFLOW“We inserted the automated decision support platform into the clinical workflow”

– Mark Montoney, M.D.


An unobtrusive but visible alert is triggered when a discrepancy is noted between the physician’s orders and the best-practice protocol, and the patient’s clinical history. The alert is simple to understand and indicates if a CT scan should or should not be ordered and provides relevant data from the patient’s clinical history and the best-practice evidence supporting the recommendation (Level B intervention).

Clinicians quickly scan the alert, access more information if needed, apply clinical judgment, and determine the appropriate course of action – either change their orders, update the order with additional documentation supporting their order or bypass the alert to continue with the current treatment plan. Reports are provided that show the action taken for each prompt (audit and feedback are included).

Efficient Solution DeploymentTenet’s West Suburban Medical Center (West Suburban) was identified as the initial site for deployment of the solution, which has now been implemented at three other Tenet hospitals in the Chicago market (MacNeal Hospital, Weiss Memorial Hospital, and Westlake Hospital). The process includes the following major activities:

Identify Physician ChampionKip Adrian, M.D., Medical Director – Department of Emergency Medicine at West Suburban, served as the clinical champion and overall leader for the implementation.

Develop Implementation Plan and TeamsWest Suburban partners in the project, medCPU and Imaging Advantage, collaborated to develop a straightforward implementation process that required minimal effort and could be completed in three to four months. The implementation required a small IT team that included an IT application lead, a desktop and infrastructure lead, and a clinical lead to assist with the integration to the hospital’s EMR application. The IT team and the medCPU and Imaging Advantage teams convened for an initial planning meeting and allocated responsibility for each of the major activities of the implementation. Protocol refinement and server installation & data feeds operate in parallel, with deployment following once development and testing has been completed.

Protocol RefinementBest-practice guidelines were developed following a thorough review of clinical best practices, as well as input from clinicians with relevant insight and experience. Dr. Adrian was able to review the 200+ pages of detailed protocols allocating a few hours each day over a two week period, editing them where West Suburban’s care practices varied from the initial protocol. The refined protocols were made available for the medical staff at West Suburban prior to deployment for their input on any changes they felt were necessary.

Server Installation & Data FeedsIn parallel, the IT staff were working to install the solution database server, create the environment and integration accounts, prepare the EMR user desktop installation package which communicates with the database server and to create the necessary data feeds for patient admission and clinical results to populate the solution database with patient medical histories. The level of IT effort for the project was minimal, with 4-6 weeks of part-time effort required for 1 coordinator and 2 analysts to work through data mapping and infrastructure activities.

DeploymentThe actual deployment was anti-climactic. Dr. Adrian noted that the medCPU Radiology Advisor was “one of the easiest things I’ve had to roll out”. Two weeks before go-live, Dr. Adrian introduced the new solution at a department meeting with a 10 minute tutorial. Once live, Dr. Adrian commented that “nothing seemed to be happening, as it was several hours before the first alert was triggered.” Dr. Adrian noted that the first alert was ironic, as it appropriately recommended a CT study where none had been ordered.

EASE OF DEPLOYMENT“One of the easiest things I’ve had to roll out”

– Kip Adrian, M.D.

EVIDENCE-BASED MEDICINE“I was impressed with their evidence-based algorithms. The program is powerful, and they are receptive and responsive to adjusting their clinical algorithms.”

– .Kevin Tao, M.D

INTERVENTION LEVELS2

A. Present information only

B. Pop-up window that the selected intervention does not meet current guidelines

C. Intervention that requires an active override to proceed

D. Intervention that requires peer consultation before proceeding


Ongoing Experience, Behavioral Change and Outcomes

Physician Behavioral ChangeDuring the first three months, Dr. Adrian reviewed daily reports identifying alerts generated by the advisor and the action taken by the physician. Dr. Adrian would discuss cases where the clinician bypassed the alert and referred to them as “teachable moments” discussed in an academic, non-punitive fashion, as he sought to understand whether an adjustment may be needed with the protocol, or whether the physician should consider a change in their clinical practice.

Refinements were regularly made to the protocol, but after several months, Dr. Adrian indicated that the number of alerts bypassed dropped substantially. Refinements to the protocol are now rare. Dr. Adrian currently reviews these reports on a monthly basis given the reduced frequency of alerts. Today when an alert is bypassed, in approximately 2/3 of the cases the clinicians should consider adopting the recommended best-practice, and in the other 1/3 of the cases, clinical documentation was not accurately recorded.

User FeedbackUser’s feedback on the solution has been positive. Clinicians have praised the data-driven approach that does not disrupt their clinical workflow. Dr. Kevin Tau, MacNeal Hospital Medical Director – Department of Emergency Medicine, commented “I was impressed with their evidence-based algorithms. The program is very powerful, and medCPU is receptive and responsive to adjusting the clinical algorithms.” Dr. Adrian noted that clinicians “want to do the right thing, and when provided with evidence they are willing to change”.

339 338 330

293 275 275

245 236

194 198

113 101 89 65 55 50

37 36 21 22

33%

30%

27%

22% 20%

18%

15% 15%

11% 11%

Months from Go-Live at Each Hospital

Total number of Alerts Total Alerts Bypassed Average rate of non-compliance

FIGURE 1. ADOPTION ACROSS CHICAGO HOSPITALS

BEHAVIORAL CHANGEClinicians “want to do the right thing, and when provided with evidence, they are willing to change.”

– Kip Adrian, M.D.


Acceptance and OutcomesWest Suburban’s approach to make the process a learning opportunity helped to drive acceptance and adoption. Today inappropriate and duplicate head CT scans have been reduced and compliance with the best-practice protocol has increased. As compliance and documentation have increased, fewer scans are denied resulting in increased revenue. Results began to emerge relatively soon after go-live, and Dr. Montoney noted “we began to see an impact within a couple of quarters, and the results have been sustained.”

ExpansionThe non-traumatic headache pilot implementation was expanded to other facilities as benefits were realized. The pilot phase was not fully integrated within the IT change management and training programs which did create communication issues as upgrades were made to the Radiology Advisor solution and environment. As deployment to other facilities progressed, the project was better integrated into these IT processes. Based upon the successful deployment for non-traumatic headache, work is in progress at the Chicago market hospitals to expand the program to cover additional indications, including CT imaging for all headache, chest and abdominal pain patients.

0%

10%

20%

30%

40%

50%

60%

Jun-14 Jul-14 Aug-14 Sep-14 Oct-14 Nov-14 Dec-14 Jan-15 Feb-15 Mar-15

MacNeal Weiss Westlake West Suburban

FIGURE 3. PERCENT DUPLICATE EXAMS

113

101

89

65

55 50

37 10.6% 11.9%

27.0%

15.4%

9.1%

26.0%

Rate of non-compliance Month-to-month reduction %

FIGURE 2. NON-COMPLIANCE: EVENTS & RATE OF REDUCTION


Conclusions & Key Takeaways

BenefitsWest Suburban realized benefits in its population of ED patients presenting with non-traumatic headache in three ways:

�The number of inappropriate CT scans was sustainably reduced to 5% or less by December 2014, with duplicate exams virtually eliminated by March 2015;

�Compliance with the CT scan protocol increased;

� Fewer inappropriate CT scans resulted in fewer denials providing an increase in revenue.

Success FactorsPositive results have led to expanded usage for additional indications in the ED, with implementation of Sepsis protocols planned in the ED and hospital-wide in the near future.

Tenet’s successful experience can be attributed to the following success factors:

�The solution was implemented within the context of Tenet’s learning culture;

� Protocols were adapted quickly to real-world clinical feedback and experience;

�There is a strong evidence-base supporting the protocols;

�The solution fits within existing clinical workflow;

�The solution is readily integrated with existing EMRs within 3-4 months.

Lessons LearnedWest Suburban Hospital experienced positive outcomes from the program, but also recognized improvement opportunities. These considerations have been addressed with the expanded use of the Radiology Advisor. Lessons learned:

�Allocate more time for protocol peer review and testing to reduce post go-live refinements;

� Integrate pilot programs better with the organization’s Information Technology change management and educational processes to minimize communication issues.

0%

5%

10%

15%

20%

25%

Mar-14 Apr-14 May-14 Jun-14 Jul-14 Aug-14 Sep-14 Oct-14 Nov-14 Dec-14 MacNeal Weiss Westlake West Suburban

FIGURE 4. INNAPROPRIATE HEAT CTs


Footnotes1 Callaghan BC, Kerber KA, Pace RJ, Skolarus LE, Burke JF. Headaches and Neuroimaging: High Utilization and Costs Despite Guidelines. JAMA Intern Med. 2014; 174(5):819-21.2 Goldzweig CL, Orshansky G, Paige NM, Miake-Lye IM, Beroes JM, Ewing BA, Shekelle PG. Electronic Health Record-Based Interventions for Improving Appropriate Diagnostic Imaging. Ann Intern Med. 2015; 162(8):557-65.

ReferencesBecker LA, Green LA, Beaufait D, Kirk J, Froom J, Freeman WL. Use of CT scans for the investigation of headache: a report from ASPN, Part 1. J Fam Pract. 1993; 37(2):129-34.

Brenner DJ, Hall EJ. Computed tomography – an increasing source of radiation exposure. N Engl J Med. 2007; 357(22):2277-84.

Callaghan BC, Kerber KA, Pace RJ, Skolarus LE, Burke JF. Headaches and Neuroimaging: High Utilization and Costs Despite Guidelines. JAMA Intern Med. 2014; 174(5):819-21.

Edlow JA, Panagos PD, Godwin SA, Thomas TL, Decker WW. Clinical Policy: Critical Issues in the Evaluation and Management of Adult Patients Presenting to the Emergency Department With Acute Headache. Ann Emerg Med. 2008; 52(4):407-21.

Goldzweig CL, Orshansky G, Paige NM, Miake-Lye IM, Beroes JM, Ewing BA, Shekelle PG. Electronic Health Record-Based Interventions for Improving Appropriate Diagnostic Imaging. Ann Intern Med. 2015; 162(8):557-65.

Miglioretti DL, Johnson E, Williams A, Greenlee RT, Weinmann S, Solberg LI, Feigelson HS, Roblin D, Flynn MJ, Vanneman N, Smith-Bindman R. The use of computed tomography in pediatrics and the associated radiation exposure and estimated cancer risk. JAMA Pediatr. 2013; 167(8):700-7.

Parizel PM, Voormolen M, Van Goethem JW, van den Hauew L. Headache: When is neuroimaging needed?. JBR-BTR. 2007; 90(4):268-71.

Schwartz DT. Counter-Point: Are we really ordering too many CT scans?. West J Emerg Med. 2008; 9(2):120-2.

UC Davis Health System. Reducing unnecessary and high-dose pediatric CT scans could cut associated cancers by 62 percent. Web. 10 Jun 2013. http://www.ucdmc.ucdavis.edu/publish/news/newsroom/7854. Accessed Jun 2015.


The Health Management AcademyThe Health Management Academy ( The Ac ademy) provides un ique, peer-learning, complemented by highly-targeted research and advisory services, to executives of Leading Health Systems. These s e rvices e n able h e alth s y stem a n d industry members to cultivate relationships, perspectives, and knowledge.

In 1998, The Academy created the first knowledge network exclusively focused on Leading Health Systems. This learning model, refined over 16 years of working side-by-side with members, combines peer learning (Executive Forums, Trustee Institute, Collaboratives), research (Health System, Consumer, Health Policy, Advisory), and leadership development (Leadership Programs and Fellowships).

medCPUmedCPU delivers accurate real-time enterprise decision support software and services through its proprietary Advisor technology. medCPU captures the complete clinical picture from clinicians’ free-text notes, dictations, discharge summaries and structured documentation entered into any Electronic Medical Record (EMR), and analyzes it against a growing library of best-practice content, generating real-time precise prompts for best care consideration. medCPU’s founding multi-disciplinary team has been pioneering new clinical decision support for nearly 20 years, delivering intelligent error reduction software systems to hospitals across the United States. medCPU’s applications include clinical and compliance support solutions.

Click here to learn more

medCPU Radiology Case Study

Documents