The healthcare industry is rapidly moving forward with the electronic collection and exchange of data among providers, payers and patients. Propelled by federal requirements, business needs and evolving technologies, developers are leveraging the flexibility and interoperability of HL7’s Fast Healthcare Interoperable Resources (FHIR) standard to address numerous challenges and use cases. Among the latest examples of value-added capability include questionnaires and calculators to better support improved health assessments, patient engagement, risk scoring and value-based care (VBC).  

These tools are taking on growing importance in the areas of patient engagement and metrics for value-based care. A rapidly emerging body of research indicates that when patients are engaged in their own healthcare, it results in better outcomes and lower costs. But how can those be measured? Questionnaires and calculators can help. For example, questionnaires can elicit--and yield--vital information about patients’ health practices and gaps in care. Calculators can help provide critical insights on the costs of prescription drugs and a patient’s health status, such as body mass and blood sugar management (A1C).  Questionnaires and calculators also can provide data that address metrics for value-based care models, including satisfaction and quality improvement.    

Why FHIR? 

FHIR is already a game-changing approach in the industry for interoperability. As a standard, it specifies a data model (synced with the USCDI – the US Core for Data Interoperability) and is central to the creation of open application programming interfaces (APIs), as specified by the federal government to implement part of the 21st Century Cures Act. FHIR also is based on a modern internet-based (RESTful) transport mechanism.  In addition, the normative release R4 in 2019 fueled industry adoption by creating a  “locked” and stable baseline, which organizations then felt comfortable utilizing in their business processes and data enterprises.  

FHIR’s inherent properties and flexibility can be leveraged to create two areas of value-added capability: questionnaires and calculators. Although standardized questionnaires and calculators have been in use for many years, combining their use with FHIR makes the approach even better. Reasons lie in FHIR’s functionalities. FHIR’s most common use is as an exchange bridge between two systems, either of which could be legacy or more modern. The data models of the two systems are mapped to the common FHIR data model, after which either system can read or write to the other system. A powerful aspect of FHIR is the ability to request just the data fields that are needed, instead of an array of what might be requested. These fields can be grouped into common questionnaires, and the field values can be used in calculators..  ( For additional information, see https://www.hl7.org/fhir/questionnaire.html and http://hl7.org/fhir/uv/sdc/2019May/behavior.html ) 

Let’s take a look at FHIR in action. 

Questionnaires 

Consider a forms-based questionnaire, let’s say for a health assessment.  A challenge with any forms-based approach is standardization – there may be many versions of the form in use within and between various organizations and systems. While mapping an organization’s specific forms to the FHIR data model would help, a potentially better approach would be to create a completely standardized FHIR-based questionnaire that allows for pre-population and extraction of data in fields, and advanced form rendering and behavior controls. This would solve version control issues and approach data gathering, whether for health assessments or electronic prior authorization, in a standards-based way which can result in better data quality.  

At the recent ONC Tech Forum, the National Library of Medicine (NLM) described a number of useful FHIR-based tools – including an open-source (GitHub) approach using FHIR structured data capture (SDC) functions and JavaScript to build dynamic questionnaires and web-based input forms. It supports nesting, skip logic, validation, pre-population of data from EHRs, etc. Questionnaires can be stored, edited, and retrieved. Examples of its use include reference in the FHIR Social Determinants of Health Clinical Care Implementation Guide, and for lab data in support of 2000 LOINC panels, which are automatically generated as FHIR Questionnaires. 

Calculators 

A similar situation applies to calculators – such as for dosing and risk score calculations. During the tech forum, National Institutes of Health (NIH) used an example for the PHQ-9 depression scoring, and use of their FHIRpath.js to compute the Framingham Risk score for cardiac events. By utilizing standardized calculators with FHIR, the specific fields required by the calculation are readily retrieved and the calculation can take place dynamically within the FHIR environment. Results are then passed on to the requesting system. This has advantages over the calculations being embedded deeper inside various back-end systems where they may be harder to access, update, or keep in synch. Use of tools such as NLM’s validator for UCUM (Unified Code for Units of Measure) also helps reduce errors and ensures that calculations and conversions are standardized.   

Old and New Ways to Support Forms and Calculations  

Expanding Opportunities for FHIR-Based Questionnaires and Calculators 

Among the stakeholders that can make use of questionnaires and calculators beyond clinicians and health systems include payers and researchers. Payers are increasingly focused on accessing clinician data for all sorts of purposes, including prior authorizations (retrieving the most pertinent data) and risk assessments (scoring). Researchers retrieving clinical data want to have confidence they are receiving the correct data fields in a consistent manner across health systems. The NIH has also developed a Research Data Finder which allows users to find and retrieve data based on name or constraint (e.g., values above a certain amount). They are working to make the giant dbGaP database (Genotype and Phenotype - 5 billion records) at NLM FHIR accessible. 

Looking Forward 

The use of more standardized forms, questionnaires, calculators, and retrieval tools makes for easier prototyping, development, and adoption. There is no doubt: it is a trend likely to continue.  Expanding the idea still further to more standardized workflows and common use cases is, of course, at the heart of the FHIR-based accelerator programs. These include Da Vinci for Provider-Payer, Gravity for Social Determinants of Health, Vulcan for researchers, and Codex for cancer care. If you would like to discuss more about these topics, please feel free to reach out to me at ken.kleinberg@pocp.com.