Solor Analysis Normal Form (ANF)
Summary
Why Standardized Clinical Statements?
Standardized clinical statements provide a structure that allows clinical terminology data to be used to its fullest potential. Consider medical data as its own language that requires both terminology and a structure to effectively communicate a patient’s health story. In language, words each have their own inherent meaning, but a word on its own communicates very little. However, providing structure and order to the words allows for the creation of sentences, resulting in higher levels of communication and understanding.
Just as with other languages, medical data is able to communicate much more when structured into an ordered statement. Standardized clinical terminology is essential, but it cannot capture much of the context that is important to clinical care, such as purpose, prerequisites, causality, etc. Incorporating structure through effective order and rules can allow terminology to produce a standardized clinical statement. In this way, clinical terminology data can capture information and improve the analysis of data that is currently lost in numerous rows of unassociated and non-normalized data.
As a data structure, ANF is designed to prevent terminology misrepresentations, reduce data variation, increase data integrity, and improve clinical decision support (CDS) to ensure patients receive proper care. In fact, this data structuring methodology can be applied to a wide variety of use cases across clinical domains, health IT technologies, and EHR systems, including:
|
|
What are ANF Statements?
ANF statements are a type of standardized clinical statement that follow the principle of separation of concerns. Simply put, ANF statements are designed to allow the clinical terminology (i.e., SNOMED, LOINC, etc.) and the data structure (i.e., clinical statement framework) to each operate as intended; the clinical terminology is responsible for defining clinical terms, such as, types of procedures, disorders, findings, and so on while the data structure is responsible for considering how clinicians and health information systems will form and use the data.
From analyzing the clinical use of data, the most important considerations for patient care are (1) the necessary procedure; (2) the purpose of the procedure; (3) the frequency with which it is to be completed; and (4) the outcome of the procedure. Clinicians can represent any procedure or request for a procedure using these four simple concepts. Finally, the data can be stored in a highly reliable and easily reproducible structure, which simplifies clinical decision support by providing a uniform representation of data to improve analysis.
ANF and its components enable users to transform ambiguous and disparate clinical data into a representation that adheres to the Understandable, Reproducible, and Useful (URU) model. The URU model serves as a guiding standard when building exact and meaningful clinical statements of a patient's data (e.g., FHIR). More specifically, a URU model ensures that clinical statements are highly reliable by:
Ensuring ANF statements can be understood without referencing private or inaccessible information
Simplifying data representation and creation of equivalent statements across both users and sources
Mandating ANF statements are fit-for-purpose to data analysis, supporting CDS, research, and extensive clinical data analysis initiatives (e.g., Population Health, Pandemic Preparedness Assessments)
ANF and High Reliability
ANF is a foundational standard for improved data representation, thereby enhancing data association and querying, improving analytical and searching capabilities, and promoting usability and shareability. Utilizing ANF can also facilitate healthcare organizations with becoming high reliability organizations (HROs)—characterized as having high safety levels under inherently risky, demanding, and technologically complex conditions. In most cases, the greatest barrier to measuring and improving outcomes is the quality of data that is aggregated from various systems. ANF aims to minimize these data quality challenges and provides a common format with semantic clarity, which allows for normalized data sets and meaningful secondary use of clinical data for patient care.
Other types of clinical statements, such as FHIR, emphasize data entry and exchange by providing multiple ways for the same data to be entered. However, this lack of a standard way of representing clinical data increases the chances of unforeseen issues when it comes to querying and using the data. For example, FHIR Observation resources allow for data elements to store values in multiple attributes (e.g., laterality can be stored in interpretation, bodySite, or method), which results in incompatible clinical statements that were originally meant to represent the same thing. By design, these resources are not reproducible, and if systems are not configured to expect certain structural combinations of terms, they are not understandable or useful to the ingesting system.
ANF, on the other hand, produces highly reliable and URU-based clinical statements by establishing order and rules for how clinical terminology data is structured, expanding the capabilities of the data for future use and analysis. This approach allows HROs to gain insight into their clinical data and reach high levels of safety, thereby producing the blueprint to replicate HROs across the industry.
Creating an ANF Clinical Statement
In order to successfully build out reproducible, standardized ANF statements, one must understand the pieces that fit together to form the ANF clinical statement. The flow chart below highlights the three main steps of taking a clinical narrative and building out a standardized ANF statement. Additional detail for each step is provided in the following sections:
Impact on Strategies
Â
References
Â