All Resources
User Guides

Preparing Data for Digital Measurement

Introduction into digital measures data preparation and readiness for non-technical users.

FHIR Supports Interoperability and Data Standardization

Organizations must ensure their data are clean, structured and accessible, in order to generate accurate insights and enhance their decision-making process. Preparing the data also mitigates errors, inefficiencies and compliance issues, and ultimately enables a more agile and responsive environment.

Fast Health Care Interoperability Resources (FHIR®) is a standard developed by Health Level Seven International (HL7®) for exchanging health care information electronically. FHIR is designed for sharing health care data across systems and platforms, facilitating interoperability. This makes it an essential tool for developers and health care providers. FHIR-enabled interoperability supports improved patient outcomes by ensuring that all providers have access to the same current information, thus reducing errors, avoiding redundant tests and enabling coordinated care.

The importance of FHIR in digital quality assessment and health care interoperability cannot be overstated. By standardizing the format and exchange protocols of health care data, FHIR ensures that critical health information is consistently available, accurate and actionable. This enhances the ability of health care providers to conduct comprehensive digital quality assessments because they can access and analyze data from diverse sources seamlessly.

Understanding FHIR Data Requirements

FHIR uses a modern web-based suite of Application Programming Interface (API) technology, which allows developers to build applications that can access data from EHRs and other health care systems. FHIR supports RESTful (representational state transfer) architectures, making it flexible, scalable and easier to implement than previous standards. This interoperable approach to health care data exchange addresses many limitations of traditional data formats, and paves the way for more efficient and effective health care information systems.

Traditional Data Formats vs. FHIR

TRADITIONAL FORMATSFHIR
FLEXIBILITY AND MODULARITYOlder health care data formats, such as HL7 v2 and CDA, tend to be rigid and complex, and often require significant customization to fit specific use cases.Modular design allows greater flexibility. Resources can be easily extended and adapted to meet specific needs without altering the core standard.
EASE OF IMPLEMENTATIONCan be challenging to implement due to their complexity and the need for specialized knowledge.Use of standard, modern web technologies, like RESTful APIs and JSON/XML, simplifies implementation.
INTEROPERABILITYOften require significant effort to map and translate data between different systems.Standardized resource definitions and use of common web protocols reduce the need for complex data transformation.
DATA EXCHANGEData exchange often relies on batch processing and point-to-point integrations, which can be slow and prone to errors.Supports real-time data exchange through APIs, enabling faster and more reliable communication between systems.

FHIR Data Structure and Format

FHIR’s core components provide a comprehensive framework for health care data interoperability, supporting a range of applications from basic data exchange to advanced clinical decision support and mobile health applications. Key concepts pertaining to FHIR data structure and format are outlined below.

Resources

Core building blocks of FHIR, representing specific types of health care information (e.g., patients, medications).

Standardized Elements

Compose resources, and include attributes and relationships.

Formats

FHIR supports multiple data formats for resource representation (primarily JSON and XML).

Terminology and Coding

FHIR supports standardized terminologies and coding systems such as SNOMED CT, LOINC and ICD.

Profiles and Extensions

Profiles are customized versions of standard resources to meet specific requirements of different use cases. Extensions are mechanisms to add additional attributes or resources without altering the core standard.

Value Sets and Code Systems

Define and manage sets of permissible codes for specific data elements, ensuring consistency and interoperability.

Security and Privacy

FHIR supports various security protocols, such as OAuth2. Audit trails log access and changes to data, ensuring accountability and compliance with regulations like HIPAA.

Meeting FHIR Specifications with Limited Claims Data

woman looking at a computer screen

Claims data can be expressed in FHIR using CARIN Blue Button® FHIR profiles, providing a standardized way to capture and share information related to health care services. Data offer a longitudinal summary of a patient’s health, chronicling encounters across different health care providers and settings over time. Data encompass an array of information, such as diagnoses, procedures, medications and financial transactions, giving a comprehensive picture of a patient’s interactions with the health care system. This information is invaluable for understanding long-term health trends, identifying patterns in health care utilization and monitoring the effectiveness of treatments over extended periods.

Data from EHRs and other clinical systems present a detailed and deeper view of patient health, and include clinical notes, lab results, imaging studies, vital signs and detailed patient histories recorded by health care professionals. EHR data offer granular insights into current health status, diagnostic processes and clinical decision-making and treatment plans.

When combined, claims and EHR data are valuable in measuring and assessing quality of care, and provide a powerful complementary perspective on patient health. The broader perspective of claims data helps identify long-term health trends and cost implications; the detailed insights from EHR data enhance understanding of a patient’s clinical needs and interventions.

Resources

Preparing Data for Digital Quality Transformation

Organizations must consider multiple factors before they embark on digital quality initiatives.

STAKEHOLDER ENGAGEMENTBring together relevant stakeholders across departments and teams, and establish channels for clear communication and collaboration.
SYSTEM READINESSEnsure the availability of all necessary IT infrastructure and technical expertise.
DATA SOURCES AND MAPPINGIdentify all relevant data sources that need to be integrated, and establish accurate mapping to FHIR resources and elements.
STANDARDIZATION AND INTEROPERABILITYIdentify data standards and appropriate coding systems (e.g., ICD, LOINC); ensure data formats are consistent and compatible with FHIR requirements.
GOVERNANCE AND POLICIESEstablish appropriate data governance policies for quality and compliance.
TESTING AND VALIDATIONEstablish processes for comprehensive validation of the initial implementation, as well as ongoing data quality monitoring and improvement.
COMPLIANCE AND SECURITYEnsure data security measures (encryption and access controls) are in place.
SCALABILITY AND FLEXIBILITYEnsure that systems can scale to handle increased data volumes and additional data sources, and are flexible enough to adapt to future changes
TRAINING AND EDUCATIONLeverage existing and/or develop new resources to train staff on related processes, standards and technologies.

Quick Guide to Data Preparation

The following are some high level steps in the process of preparing your data for digital quality measurement.

1. Data Assessment and Inventory

  • Identify all data sources, including EHRs, claims data, lab systems and others.
  • Assess data quality, identifying issues such as missing values, inconsistencies and errors.

2. Data Standardization

  • Map data to standardized coding systems like ICD, LOINC, SNOMED CT.

3. Data Integration

  • Integrate data from different sources into a unified system.
  • Transform data into FHIR-compliant formats.

4. Data Mapping

  • Map existing data elements to FHIR resources and their corresponding attributes.
  • Ensure that data accurately reflect standard structure and relationships.

5. Data Enrichment

  • Supplement data with additional information from other sources to fill gaps and enhance completeness.
  • Add metadata to provide context and improve data usability.

6. Testing and Validation

  • Conduct thorough testing to ensure that integrated data are accurate and complete.
  • Use validation tools to check that data conform to FHIR standards and specifications.

7. Security and Privacy

  • Encrypt data to for protection during storage and transmission.
  • Implement robust access controls to ensure only authorized personnel can access sensitive data.

8. Data Governance

  • Establish clear data governance policies and procedures.
  • Assign data stewards to oversee data quality, use, and security.

Common Challenges in Data Mapping

DATA QUALITY ISSUES
  • Variations in data formats, units and coding systems across sources.
  • Missing/incomplete data lead to gaps in accurate mapping and analysis.
  • Duplicate records can skew results and analytics.
COMPLEX DATA STRUCTURES
  • Difficulty handling complex, hierarchical data structures in legacy systems.
  • Challenges in processing and mapping unstructured data, such as free-text clinical notes.
SEMANTIC DISCREPANCIES
  • Inconsistent terminologies and codes used in different systems for the same clinical concepts.
  • Lack of clarity in data definitions and use.
DATA VOLUME AND SCALABILITY
  • Handling large volumes of data during transformation can be challenging.
  • Ensure data transformation processes are performant and do not bottleneck the system.
TECHNICAL AND COMPATIBILITY ISSUES
  • Legacy systems may not support modern data formats or standards.
  • Ensure data can be correctly interpreted and utilized by different systems, post-transformation.
regulatory iconREGULATORY COMPLIANCE
  • Ensure data transformation processes adhere to privacy regulations such as HIPAA.
  • Maintain traceability and audit trails of data transformations for compliance.

Data Mapping Best Practices

Data mapping is a critical process in data preparation, especially when transforming legacy data into a standardized format like FHIR. It involves translating data from the source format to a target format, ensuring that data align with the target system’s structure and requirements.

Perform a comprehensive analysis of source data

  • Identify and document all data sources; catalog all data elements in source systems that need to be mapped.
  • Conduct thorough data profiling to understand data distributions, identify anomalies and discover hidden relationships in the source data.
  • Work with domain experts, such as clinicians and data stewards, to ensure understanding of the source data context.

Gain full understanding of the target data structure

  • Understand the structure, elements and relationships of FHIR resources; get familiar with coding systems like ICD, LOINC, SNOMED CT.
  • Standardize data using common coding systems before mapping to ensure uniformity, and normalize the data to ensure conformity with expected formats and values in the target system.

Create detailed mapping specifications

  • Define mapping of each source data element to a FHIR resource and its fields.
  • Define clear and precise transformation rules to handle complex data conversions and mappings.
  • Implement robust error handling to manage exceptions and discrepancies during data transformation.

Automate and streamline with specialized tools

  • Use data mapping and ETL (Extract, Transform, Load) tools to automate and streamline the mapping process.
  • Develop automated scripts to validate the accuracy and completeness of mapped data.

Incorporate iterative testing and feedback cycles for validation and acceptance

  • Ensure that mapped data maintains consistency and accuracy; validate that data comply with FHIR standards and requirements.
  • Conduct initial testing with a subset of data to resolve issues early; implement a process of continuous stakeholder feedback to refine logic.

Maintain comprehensive documentation

  • Document the mapping process, including source-to-target field mappings and transformation rules, and decisions and assumptions.
  • Create a traceability matrix to link source data elements to their corresponding target elements, facilitating auditing and troubleshooting.

Mapping Non-FHIR Data to FHIR

There are several approaches to transform and integrate data from legacy systems, claims databases or other non-standardized sources into the FHIR framework. Here are some common approaches.

APPROACH 1:
Direct Mapping

Each data element in the source system is directly mapped to a corresponding element in a FHIR resource. Data type conversion is performed (if needed), and values are translated to conform to FHIR standards (e.g., map diagnosis codes to ICD-10).

Example:

  • Source: A patient's demographic information in a legacy system.
  • Target: Patient resource in FHIR with fields like name, birthdate, gender.

APPROACH 2:
Scripted Transformation

Utilizes scripts or transformation languages to automate the data mapping process, allowing more flexibility and repeatability. Specific rules are applied to transform and clean data; validation steps are implemented in the script to ensure data integrity.

Example:

  • Source: XML file containing patient records.
  • Target: JSON representation of FHIR patient resource, transformed using XSLT script.

APPROACH 3:
ETL Tools or APIs

Can be used to handle data retrieval from the source, transformation into FHIR format and loading or posting data into the target FHIR-compliant database or system.

Example:

  • Source: Relational database with health care data or API endpoint providing patient data in a proprietary format.
  • Target: FHIR server or direct creation of resources such as observation, medication, condition.

Evaluating Mapping Quality and Completion

CRITERIA
AccuracyData values should be correctly translated from the source to the target format, and consistent data representation maintained across mappings and transformations.
CompletenessAll relevant data elements from the source should be mapped to the target system and verified that there is no critical data have been lost in the transformation process.
TimelinessData transformation should be timely and able to meet required time frames without significant delays.
ScalabilityThe system should have the ability to handle increasing data volumes without degradation in performance.
Future-ProofingThe mapping process should accommodate future data growth and additional data sources.
InteroperabilityThe transformed data should conform to relevant standards (e.g., FHIR, HL7, ICD, SNOMED CT), and be able to be used seamlessly by other systems and applications.
Validation and VerificationThe mechanisms for error detection and handling should be robust, and validation rules should be applied to check data integrity and conformity.
TOOLS
DATA PROFILING TOOLSProfile, cleanse, transform data.
ETL TOOLSAutomate data flow and transformation processes with built-in validation capabilities.
VALIDATION SCRIPTSValidate data integrity, accuracy and automated testing frameworks for validating transformation workflows.
SCHEMA VALIDATION TOOLSCheck conformance to predefined schemas (e.g., validate FHIR resources against FHIR schema to ensure compliance).
DATA COMPARISON TOOLSIdentify differences and reconcile transformed data against source data to ensure completeness and accuracy.
METADATA MANAGEMENT TOOLSProvide governance and documentation of data mappings, ensuring traceability and compliance.

Addressing Specific Use Cases

Although there are general considerations for all types of digital quality initiatives, there is often a need for tailored data mapping strategies to address the unique requirements of health care plans, populations, programs and settings. Customization ensures that data mapping efforts effectively support the goals of each initiative—whether it is quality reporting, value-based care, population health management or specific care settings and requirements.

Unique Considerations for Plans and Populations

Medicare Advantage Plans:

  • Risk adjustment data
  • Coverage and benefits
  • Dual eligibility

Medicaid Managed Care Plans:

  • Medicaid eligibility criteria
  • Beneficiary demographics
  • State-specific requirements
  • Service authorizations

Commercial Health Insurance Plans:

  • Network provider data
  • Utilization management
  • Prior authorization
  • Employer group data

Value-Based Care Programs:

  • Outcome measures
  • Performance metrics
  • Care coordination data

Population Health Management:

  • Risk stratification and population segmentation
  • SDOH

Specialty Care and Subspecialties:

  • Specialty-specific data
  • Procedure-specific mapping

Planning for Digital Quality Transformation

1

Project Initiation and Planning

  • Define project scope, objectives, deliverables.
  • Identify stakeholders and establish communication channels.
  • Develop a project plan, with timelines and resource allocation.
2

Data Assessment and Analysis

  • Assess existing data sources, formats, quality.
  • Analyze data mapping requirements and transformation needs.
  • Document data dictionaries, schemas, metadata.
3

Design and Architecture

  • Design a unified data model and mapping strategy.
  • Determine the technology stack and infrastructure requirements.
  • Develop architecture diagrams and integration workflows.
4

Development and Implementation

  • Develop transformation scripts, ETL workflows, system interfaces.
  • Implement mapping logic and transformation rules.
  • Test data transformation processes and validate results.
5

Testing and Quality Assurance

  • Conduct unit testing, integration testing, user acceptance testing.
  • Validate data quality, accuracy, completeness.
  • Identify and address any issues or discrepancies.
6

Deploy and Go Live

  • Deploy tools into production environments.
  • Conduct final validations and readiness checks.
  • Transition to operational mode and monitor system performance.
7

Post-Implementation Optimization

  • Review project outcomes against initial objectives and success criteria.
  • Gather feedback from stakeholders and end-users.
  • Identify areas for optimization and improvement.

Who Can Help With Data Transformation?

Within Your Organization

  • IT department, data engineers and database administrators: Assist with data integration, transformation and mapping, and are skilled in managing databases.
  • Health informaticians: Knowledgeable about health care data standards and mapping processes.
  • Clinicians and medical staff: Provide domain expertise, validate clinical data mappings.
  • Data stewards: Oversee data quality, compliance, documentation initiatives.
  • Project managers and business analysts: Help coordinate the project, manage resources, monitor progress, gather and document mapping requirements from stakeholders.

Within the Community/Vendors

  • Health care IT vendors: Offer FHIR data mapping tools and services as part of product offerings, in addition to cloud-based integration solutions for data mapping and interoperability.
  • Health care IT consultants: Specialize in FHIR implementation, data mapping, ETL processes, interoperability solutions.
  • FHIR Implementation Community: On the HL7 FHIR website—valuable resources, forums, discussion groups.
  • Open-source FHIR libraries and tools: Available for FHIR data mapping and conversion.

FHIR® is the registered trademark of Health Level Seven International and use does not constitute endorsement by HL7.

Related Resources

NCQA Launches Coalition to Assess Bulk FHIR Data Quality

How Digital Quality Can Improve the Member Experience

When Exchanging (and Trusting) Data Grows Up

The Digital Transformation-Data Quality Puzzle With NCQA’s Brad Ryan

The New Rules of Data Sharing

Updated TEFCA SOPs for Health Care Operations and NCQA

Top Three Challenges in the Transition to Digital Quality

The Move to Digital Quality Measurement

View All Resources