CDS Connect: CMS’s Million Hearts® Model Longitudinal ASCVD Risk Assessment Tool for Baseline 10-Year ASCVD Risk

CMS’s Million Hearts® Model Longitudinal ASCVD Risk Assessment Tool for Baseline 10-Year ASCVD Risk

Description

This artifact provides the ability to calculate a baseline 10-Year ASCVD risk score to support primary prevention of ASCVD. It utilizes the 2013 ACC/AHA pooled cohort equation to calculate the risk of developing a first time "hard" ASCVD event, defined as: nonfatal myocardial infarction (MI), coronary heart disease (CHD) death, nonfatal stroke, or fatal stroke.

It addresses the first of 3 clinical scenarios where CMS's Million Hearts® Model Longitudinal ASCVD Risk Assessment Tool might be used:
1. Calculation of a baseline 10-Year ASCVD risk assessment score
2. Prospective estimations of ASCVD risk in support of shared decision-making while considering the benefits of therapies, alone or in combination
3. Calculation of updated risk after preventive therapies have been initiated

CDS guidance that facilitates shared decision-making based on prospective risk estimates, along with updated risk calculation after the initiation of therapy will be included in separate CDS artifacts.

Artifact Type
Creation Date
Version
0.1.1
Unique Identifier
CDS 003
Status
Experimental
False

Artifact Creation and Usage

Contributors

This artifact was developed by MITRE software engineers and clinical informaticists, in collaboration with clinical subject matter experts and leaders from CMS and the Million Hearts® initiative.

Additional information about MITRE's health expertise is available here.

If you would like further information, would like to give us feedback, or have any questions about this artifact, please contact us at ClinicalDecisionSupport@ahrq.hhs.gov

License
IP Attestation The author asserts that this artifact has been developed in compliance with the intellectual property rights attributed to the source material.
Implementation Details
Engineering Details
This CDS logic is expressed using Clinical Quality Language (CQL) and the FHIR Draft Standard for Trial Use 2 (DSTU2) data model. All value sets referenced in the logic are published on the Value Set Authority Center (VSAC). Additional details about these resources can be accessed via the following URLs:
     CQL: https://ecqi.healthit.gov/cql-clinical-quality-language
     FHIR DSTU2: https://www.hl7.org/fhir/DSTU2/index.html       
     VSAC: https://vsac.nlm.nih.gov/
Repository Information
Approval Date
Publication Date
Last Review Date
Knowledge Level

Structured code that is interpretable by a computer (includes data elements, value sets, logic)

Purpose and Usage
Purpose

Provides the ability to calculate and display a patient's baseline 10-Year ASCVD risk score to inform a plan of care

Intended Population

"The 10-year ASCVD Risk Tool is intended for use in a broad population aged 40-79 years and eligible for primary prevention of ASCVD. The baseline 10-year ASCVD risk estimate is calculated using the ACC/AHA 2013 Pooled Cohort Equations (PCE), which provide sex- and race-specific 10-year estimates of ASCVD risk, have been validated in a broadly representative sample of U.S. whites and African-Americans, and are well calibrated for the Medicare population." Source: Final Technical Report: Estimating Benefits in Risk Reduction From Cardiovascular Preventive Therapies in Medicare Patients: Development of the Longitudinal ASCVD Risk Estimator.

Usage

This artifact is intended for use by providers while delivering care in an outpatient setting.

Cautions

Online calculators and the Million Hearts® Model Longitudinal ASCVD Risk Assessment Tool appear to use more precise values than those published in the 2013 Report on the Assessment of Cardiovascular Risk: Full Work Group Report Supplement. The CQL code in this artifact aligns with the precision used in the Million Hearts® Model Longitudinal ASCVD Risk Assessment Tool and online calculators.

ASCVD risk scores calculated by the CQL code in this artifact consistently align with scores calculated by online calculators using identical input data. Discrepancies were identified in the test data included in the 2013 Report on the Assessment of Cardiovascular Risk: Full Work Group Report Supplement (e.g., copy/paste errors and incorrect rounding, which can lead to results that are off by as much as 0.001). Ensure robust integration testing if this artifact is implemented in an EHR and validate scores against more than one source.

Since baseline calculation of 10-year ASCVD risk is not reliant on a patient's Low-Density Lipoprotein (LDL) value, or potential treatment with Statin Therapy or Aspirin Therapy, these concepts are not expressed in this artifact. If implementers choose to track these concepts at a "baseline" level prior to initiating or changing therapy, these concepts can be added to the CQL code at that time.

"At present, the risk equation applies most accurately to non-Hispanic Whites and African Americans.  For non-White and non-African American ethnic groups, the equations for Whites of the same sex were used, which may provide overestimation of risk for some groups (e.g., East Asian Americans) and underestimation in others (e.g., South Asian Americans)." Source: 2013 Report on the Assessment of Cardiovascular Risk: Full Work Group Report Supplement

Use of the tool is not indicated for individuals who have ASCVD.

"Patients with end-stage renal disease were not included in the derivation sample for the PCEs; such patients require highly individualized care with respect to use of aspirin and blood-pressure-lowering therapies, and data on use of statin medications in ESRD patients do not indicate overall benefit. For some symptomatic and for advanced heart failure patients, similar considerations and highly individualized decision-making may be necessary. However, recent data reinforce the importance of ASCVD risk-reducing therapies even among patients with heart failure. Use of this tool is not indicated for individuals that have ASCVD. For older (and younger) individuals, guidelines recommend individualized care decisions." Source: Final Technical Report: Estimating Longitudinal Risks and Benefits from Cardiovascular Preventive Therapies Among Medicare Patients: The Million Hearts Longitudinal ASCVD Risk Assessment Tool.

Risk scores calculated on patients with a history of familial hypercholesterolemia may under-represent the patient's true ASCVD risk. "Patients with LDL-cholesterol of at least 190 mg/dL should be evaluated and considered for statin therapy regardless of age and estimated 10-year ASCVD risk." Source: Final Technical Report: Estimating Benefits in Risk Reduction From Cardiovascular Preventive Therapies in Medicare Patients: Development of the Longitudinal ASCVD Risk Estimator.

In 2018 the ACC/AHA released updated guidelines on cardiovascular risk assessment. Per Arps et al. in New Aspects of the Risk Assessment Guidelines: Practical Highlights, Scientific Evidence and Future Goals, "the basic themes of the 2013 guidelines are largely preserved with an adjusted framework accentuating the value of non-traditional risk factors, enhancing the role for coronary artery calcium (CAC) scoring and introducing low-density lipoprotein cholesterol (LDL-C) thresholds for consideration of intensifying lipid lowering therapy." This ASCVD risk calculator does not incorporate the adjusted framework outlined in this article. Source: Arps, K., Blumenthal, R., Martin, S. (2018) New Aspects of the Risk Assessment Guidelines: Practical Highlights, Scientific Evidence and Future Goals. November 18, 2018. Accessed on August 30, 2020 at: https://www.acc.org/latest-in-cardiology/articles/2018/11/14/07/10/new-aspects-of-the-risk-assessment-guidelines.

Supporting Evidence
References

DeFelippis AP, Young R. (2016) Estimating Longitudinal Risks and Benefits from Cardiovascular Preventive Therapies Among Medicare Patients: The Million Hearts Longitudinal ASCVD Risk Assessment Tool. Nov 4, 2016. Accessed on Feb 2, 2017 at: http://www.acc.org/latest-in-cardiology/articles/2016/11/04/08/56/estimating-longitudinal-risks-and-benefits-from-cv-preventive-therapies.

Lloyd-Jones DM, Ning H, Huffman MD, Karmali K,  Berendsen M, Goff D. (2016) Final Technical Report: Estimating Benefits in Risk Reduction From Cardiovascular Preventive Therapies in Medicare Patients: Development of the Longitudinal ASCVD Risk Estimator. 

Goff DC Jr, Lloyd-Jones DM, Bennett G, Coady S, D’Agostino RB Sr, et al.  (2013) 2013 ACC/AHA Guideline on the Assessment of Cardiovascular Risk: a Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation. 2013;00:000–000.

Goff DC Jr, Lloyd-Jones DM, D’Agostino RB Sr, Gibbons R, Greenland P, et al. (2013) 2013 Report on the Assessment of Cardiovascular Risk: Full Work Group Report Supplement. National Heart, Lung, and Blood Institute.

Recommendation

The race- and sex-specific Pooled Cohort Equations* to predict 10-year risk of a first hard ASCVD event should be used in non-Hispanic African Americans and non-Hispanic whites, 40–79 years of age. (ACC/AHA Class of Recommendation I - Procedure/Treatment SHOULD be performed/administered) Source: 2013 ACC/AHA Guideline on the Assessment of Cardiovascular Risk: a Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation. 2013;00:000–000.

Strength of Recommendation

Recorded above immediately following the Recommendation Statement.

The recommendation utilizes the ACC/AHA grading system for strength of recommendation is available on page S8 of ACC/AHA 2013 Guideline on the Treatment of Blood Cholesterol to Reduce Atherosclerotic Cardiovascular Risk in Adults accessible at: http://circ.ahajournals.org/content/early/2013/11/11/01.cir.0000437738.63853.7a

Quality of Evidence

Level of Evidence: AHA/ACC Level B (Limited populations evaluated)

Complete descriptions of the ACC/AHA level of evidence grading system is available on page 2938 of the ACC/AHA Guideline on the Assessment of Cardiovascular Risk.

Artifact Decision Notes

Decision notes are listed in the attached implementation guide.

 

Artifact Representation
Triggers

N/A

Inclusions
Age >=40 and <=79 years
Exclusions

None

Interventions and Actions
Intervention DISPLAY segment of the CMS's Million Hearts® Model Longitudinal ASCVD Risk Tool that provides the ability to input the following data concepts:
1) Age
2) Gender (M/F)
3) Race (White/African American/Other)
4) Total Cholesterol - MOST RECENT within past 6 years (measured as mg/dL). Allowable result range: 130-320 mg/dL
6) HDL Cholesterol - MOST RECENT within past 6 years (measured as mg/dL). Allowable result range: 20-100 mg/dL
8) Systolic Blood Pressure - MOST RECENT within past 6 years (measured as mmHg). Allowable result range: 90-200 mmHg
9) Treated for high blood pressure - (Y/N) determined by a diagnosis of Hypertension AND an active anti-hypertensive medication
10) Diabetes (Y/N)
12) Current Smoker - (Y/N) - MOST RECENT within the past year
  Note: The risk score should be calculated whenever possible to facilitate workflow and provider decision making. To this end, if a patient's lab (i.e., Total Cholesterol and HDL Cholesterol) or systolic blood pressure (SBP) results are outside the range specified by the calculator, then the result should be replaced with the nearest value that will be accepted by the tool. For example, the tool allows a SBP value of 90-200 mmHg. If the most recent patient value is "212", it would be replaced with "200" to enable calculation of the score.
Intervention POPULATE fields in CMS's Million Hearts® Model ASCVD Longitudinal Risk Tool
Intervention CALCULATE risk score
Intervention DISPLAY risk score
Intervention DISPLAY the following ERRORS, when indicated:
  • ERROR: This CDS is not applicable, as patients must be age 40-79 years (inclusive).
  • ERROR: Inadequate data to process CDS: birthdate is missing.
  • ERROR: Inadequate data to process CDS: race is missing.
  • ERROR: Inadequate data to process CDS: gender is missing or is other than male / female.
  • ERROR: Inadequate data to process CDS: total cholesterol result is missing or more than six years old.
  • ERROR: Inadequate data to process CDS: HDL result is missing or more than six years old.
  • ERROR: Inadequate data to process CDS: systolic blood pressure result is missing or more than six years old.
  • ERROR: Inadequate data to process CDS: smoking status is missing or more than one year old
Intervention DISPLAY the following WARNINGS, when indicated:
  • WARNING: For non-White and non-African American ethnic groups, the equations for Whites of the same sex were used, which may provide overestimation of risk for some groups (e.g., East Asian Americans) and underestimation in others (e.g., South Asian Americans).
  • WARNING: Total cholesterol <actual> mg/dL is not in the allowable range from 130 to 320 mg/dL. The closest boundary was used for the calculation.
  • WARNING: HDL <actual> mg/dL is not in the allowable range from 20 to 100 mg/dL. The closest boundary was used for the calculation.
  • WARNING: Systolic blood pressure <actual> mmHg is not in the allowable range from 90 to 200 mmHg. The closest boundary was used for the calculation.
Intervention DISPLAY link to relevant guidelines related to the tool (e.g., http://www.acc.org/latest-in-cardiology/articles/2016/11/04/08/56/estimating-longitudinal-risks-and-benefits-from-cv-preventive-therapies)
Action DOCUMENT risk score
Testing Experience
Pilot Experience

This artifact was tested by b.well® Connected Health over an 8-week period from June 2019 - August 2019, along with the following 4 artifacts (also published on the CDS Connect Repository):

  • Healthful Diet and Physical Activity for CVD Prevention in Adults with Cardiovascular Risk Factors
  • Abnormal Blood Glucose and Type 2 Diabetes Mellitus: Part 1, Screening
  • Abnormal Blood Glucose and Type 2 Diabetes Mellitus: Part 2, Counseling
  • Statin Use for the Primary Prevention of CVD in Adults: Patient-Facing CDS Intervention

b.well offers a platform with personalized health management resources targeted to consumers to help them self-manage the entire healthcare process. AHRQ and MITRE would like to thank b.well for their partnership on the pilot. The collaborative effort provided a valuable opportunity to test the CQL CDS expression and learn from the implementation process and end user experiences. High-level details are outlined below. Detailed information on the pilot implementation can be found in the CDS Connect Pilot Report: Preventive Health CDS Interventions document posted within this artifact.

1. What went well?
·         The CDS was successfully integrated into the pilot site’s platform.
·         The pilot partner was extremely collaborative and responsive, meeting all deadlines and exceeding expectations despite a very tight timeframe.
·         The use of several communication tools helped streamline the sharing of documentation and software and enhanced the collaboration between the CDS Connect and b.well technical teams (i.e., Accellion kiteworks and Slack).
·         Leveraging the pilot partner’s expertise in patient/consumer engagement and health management resources and tools enhanced the pilot intervention process and contributed greatly to a successful pilot.
2. What did not go well?
·         The pilot technical team did not have prior experience with the standards used (e.g., Fast Healthcare Interoperability Resources [FHIR], CDS Hooks, and Clinical Quality Language [CQL]), which extended the technical integration time and efforts.
·         Data mapping consumed almost 25 percent of the pilot partner’s technical resource hours (80 hours) and an additional 20 hours of clinical and data/analytics resource time.
·         Some required data was simply not available. Given more time, the pilot partner could have enhanced existing capabilities to obtain the required data directly from the end users.
3. What would you do differently?
·         Initiate the pilot partnership no later than March to allow more time for integration and testing efforts, as well as end user engagement.
·         If more time was available, consider implementing  the CDS using the FHIR Release 4 (R4) in addition to FHIR Draft Standard for Trial Use 2 (DSTU2).
   4. What enhancements were suggested by the pilot partner?
·         Allow the calculation of an “inferred” diagnosis or potential “at risk” status for end users without a confirmed diagnosis, based on other available data. This would enable reaching a larger population of end users and providing them with education that may be valuable. The CDS Connect team opted not to implement this enhancement since expansion of the target population would diverge from the USPSTF recommendation.
·         If time had allowed, it would have been helpful to obtain end user input on the educational materials prior to implementing them. Instead, the pilot partner used health literacy and health communication best practices when developing the educational content.
·         It would have been helpful to survey a larger number of end users, and to provide the survey to the user directly after the intervention. This would allow more conclusive analysis of the survey results; however, the project team was required to follow Paperwork Reduction Act requirements which limited the number of survey recipients to 9 or less
·         Provision of more detailed information from CDS Hooks, to provide information such as the specific user data that qualified that user for the CDS intervention. This information could be used for both testing purposes as well as personalization of the intervention text.
   5. What were the key takeaways/lessons learned from the pilot experience?
·         Aggregating data from multiple sources provides a rich source of clinical information, but also presents challenges when using the data to satisfy CDS logic requirements. This is primarily due to the lack of standards employed by each data source (e.g., claims, electronic health systems, pharmacy benefits management systems, reference laboratories, and patient-generated health data). Much of the data were not linked to a standard clinical terminology code, and many of the required FHIR attributes were missing (e.g., “status,” “verificationStatus,” “onsetDateTime,” “valueQuantity”).
·         Data mapping to account for the issues listed above is a resource-intensive process that also requires knowledgeable clinical informatics expertise. The mapping efforts can impact the integration timeline.
·         The experience of the pilot technical team with FHIR and other evolving standards should be considered during the technical evaluation and planning stages.
·         Patient-facing CDS pilots should consider a pilot partner with consumer/patient-facing experience and expertise and should ensure a pilot timeframe that allows end user personalization and engagement opportunities to be realized.