WHO Family Planning Fast Healthcare Interoperability Resources Implementation Guides

Standardizing family planning and sexually transmitted infections clinical guidelines data in digital health tools to bridge the gap between human-readable clinical guidelines and machine-readable fast healthcare interoperability resources implementation guides.

Read On
Contact

Natschja Ratanaprayul
Digital Health Implementation Research Consultant
Email

Implementation Partners

World Health Organization
Regenstrief Institute
Global Health Informatics
Dynamic Content Group

Funder

UNFPA

Implementation Dates

November 2019 -May, 2020

Geographic Scope

Global

Target Users

Health Care Provider, Data Services Provider

Enabling Environment Building Blocks

Services and Applications, Standards and Interoperability

Family Planning Program Classification

Policy and Enabling Environment

Introduction

Clinical practice guidelines (CPGs) are used worldwide to inform clinical decision-making through the implementation of evidence-based clinical and public health practices. The use of digital technology, such as electronic health records (EHRs), continues to increase as countries work to facilitate public health interventions, to improve care delivery with decision support, and to ensure accountability at all levels of the health system. However, the translation of CPGs into digital systems often results in a subjective interpretation by implementers and software developers due to ambiguities during translation into an electronic format. These difficulties can lead to divergences in electronic CPG implementation, reducing usefulness of collected data outside of that implementation setting (Biondich et al., 2006; Gillois et al., 2001; Shiffman et al., 2004; and Tierney et al., 1995).

To resolve these challenges, the World Health Organization (WHO) created HL7 Fast Healthcare Interoperability Resources (FHIR), which reflect WHO recommendations in standards-based digital format. This is to ensure WHO’s evidence-based guideline content is implemented in digital systems with fidelity, using interoperability standards. 

Project/Digital Health Solution Overview

WHO created FHIR Implementation Guides (IGs) for the areas of family planning (FP) and sexually transmitted infections (STI). These IGs can be used by implementers and software developers to standardize FP and STI clinical guidelines in digital health solutions supporting these service areas. The FHIR IG for Family Planning was created based on the following WHO guidelines and guidance documents: 

 The FHIR IG contains the minimum dataset to be collected for service delivery and indicator reporting, according to these normative guidance documents, in a standards-based format. To support this, HL7 FHIR and standard semantic terminologies (see text box), including LOINC, SNOMED CT, ICD10, and RxNorm, provided structure and codes to take an additional step toward the goal of computable guidelines, critical for high-quality patient care (IHE Wiki, 2019). The structure and codes provided using FHIR and standard semantic terminologies offer clarification on the way a new healthcare application or system should be structured and what questions should be asked, which answer options should be available, and more. The Unified Medical Language System (UMLS) is one example of a tool that was used for terminology mapping across multiple standards.

To demonstrate that the creation of a FHIR IG for FP and STI is feasible and to support the processes and resourcing required—that is, a set of rules about how the FHIR resources can be applied—the project established the following objectives:

  • Create a mapping process and establish a collaborative environment where terminology and FHIR resource mapping can occur.
  • Map ~1,000 terms from FP and STI to FHIR resources and standard terminologies.
  • Create tooling that can accept mapped terms as an input and generate FHIR profiles and IGs.

Existing FHIR IGs, including the FHIR Clinical Guidelines IG Template, were reviewed to understand their goals and structures, and mapping recommendations were reviewed for the selected standard terminologies. Data dictionary terms were consolidated into a master data dictionary spreadsheet to serve as a collaborative environment where FHIR mapping, terminology coding, and progress evaluation occurred. A mapping process standard operation procedure (SOP) served as guidance while the FHIR and terminology mapping took place. This SOP also assisted in the development of tooling to generate the IGs. In the case that FHIR resources did not adequately address the mapping need, elements were added on to existing FHIR resources to address the need. Additionally, data modeling based on a data dictionary and minimum dataset provided by WHO could be altered in the case that a better modeling scheme was identified by clinical informatics personnel. Other data dictionary alterations to suit mapping were performed in collaboration with relevant subject matter experts.

Evaluation and Results Data

Agile methods were used to iteratively review the process and fine-tune the data dictionary, mapping processes, terminology coding, FHIR modeling, and IG tooling. The outputted FHIR IGs from the mapped terms allowed for iterative review and subsequent issue resolution. Of the 947 terms to be mapped, 907 have been assigned a FHIR mapping, and 805 have been given a related terminology code. These terms represent both FP and STI content, along with basic underlying information for a healthcare application or system. 

Because of the need for clarification in a term name or definition or of use of the data, not all terms have been mapped to terminology. Furthermore, due to the limitations of the standard terminology code sets themselves, not all data elements could be mapped. For example, the list of contraceptive methods needed to be mapped across multiple terminology code sets in order to be reflective of all the recommended contraceptive methods available. 

Some 801 terms have been mapped to both terminology and FHIR Resources, with 31 terms requiring custom attributes to be added to FHIR Resources, that is, additional attributes outside of base FHIR. IGs have been successfully generated from the data dictionary with the terminology mappings and FHIR resources assigned. This project continues to progress as mapping continues, the IG tooling is refined, and IGs are regenerated.

Lessons Learned

  1. Guidelines and guidance documents are ambiguous and require a certain level of subjective interpretation by FP subject matter experts, resulting in data dictionaries that maybe unclear to a health information or a software engineer. The creation of a data dictionary should include a basic information model that facilitate mapping to standards. The team recommends that future work in data dictionary creation engages a resource person skilled in the creation of information models, who works alongside subject matter experts, to provide guidance on the transition of the information displayed on the data entry form to its ideal storage location. This enables accurate and harmonized mapping to semantic terminology and FHIR Resources. Without clear naming and definitions, the mapping to FHIR and semantic terminology are at risk of providing inconsistent care and analysis needs.
  2. Clinical semantic terminology mapping is not an exact science, though CPGs require as much accuracy as possible to ensure clinical quality. Mapping existing forms/instruments to standards inevitably uncovers dissonance or a lack of clarity between concepts used in the data entry forms used in daily healthcare practices and existing international standards. This also causes challenges in estimating the labor-intensive resources required to align and map to appropriate terminology.
  3. Mapping data elements to FHIR resources is not an exact science; the resources overlap in real world scenarios. For example, some devices contain several drugs, all of which have a dedicated FHIR resource. In addition, there is an overlap between FHIR’s Observation and Condition resources, and the distinctions between real-world observations and conditions can often be unclear. 
  4. There need to be management processes that support additions and updates to the data dictionary need to be in place. For example, if there are changes to the recommendations, there will need to be a process to subsequently update the related digital tools. Changes to data dictionaries, including additions and updates, must be tracked so that semantic terminology and FHIR Resource mappings can be reviewed to ensure that any definitional or terminology changes are well supported in the resulting FHIR Implementation Guide.

Conclusion

While the project is still underway, FHIR profiles and IGs have been successfully generated from the data dictionary with the semantic terminology mappings and FHIR resources assigned. Future directions for this work will involve the refinement of this process and adaptation to other health domains. The establishment of a standardized mapping process, such as a terminology management system, is recommended to ensure consistency and usefulness. Finally, the products of this work require testing in various health care settings worldwide, allowing for further refinement.

References

  • Biondich PG et al. Collaboration between the medical informatics community and guideline authors: fostering HIT standard development that matters. AMIA Annu Symp Proc. 2006;2006:36–40. 
  • Integrating the Healthcare Enterprise (IHE) Wiki. Computable Care Guidelines. IHE Wiki website. https://wiki.ihe.net/index.php/Computable_Care_Guidelines . Published April 2, 2019. Accessed March 4, 2020. 
  • Gillois P. et al. From paper-based to electronic guidelines: application to French guidelines. Stud Health Technol Inform. 2001;84(Pt 1):196-200. 
  • Shiffman RN et al. Bridging the guideline implementation gap: a systematic, document-centered approach to guideline implementation. J Am Med Inform Assoc. 2004;11(5):418-26.
  • Tierney WM et al. Computerizing guidelines to improve care and patient outcomes: the example of heart failure. J Am Med Inform Assoc. 1995;2(5):316-22. 
  • Unified Medical Language System (UMLS). National Institutes of Health (NIH) National Library of Medicine website. www.nlm.nih.gov/research/umls/index.html . Published September 10, 2003. Accessed March 3, 2020.
  • Health Level Seven International. Welcome to FHIR. HL7FHIR Release 4 website. www.hl7.org/fhir/index.html. Published May 14, 2015. Accessed March 3, 2020. 
  • World Health Organization (WHO). WHO digital accelerator kits. WHO website. www.who.int/reproductivehealth/publications/digital-accelerator-kits/en/. Forthcoming. Accessed March 3, 2020.