Sage Journals: Discover world-class research

Abstract

Background: Uniform interpretation of digital health messages is important to achieve semantic interoperability of electronic health information systems (eHIS). Whereas international digital health terminologies such as ICD, LOINC and SNOMED-CT exist, their design considerations regarding health processes, data collected, and technologies, among others do not necessarily match Uganda’s eHIS contextual needs. Objective: This research aimed to design a process that could be used to contextualise international digital health terminologies for Uganda’s eHIS. Methods: The Design Science approach was used in designing the contextualisation process while utilising a foundation contextualisation approach for mapping terminologies. Results: The contextualisation process constitutes six major phases; assessing the national digital health information system context, extracting data elements in the national digital health information system, mapping existing national data elements to international terminologies, identifying and coding unmatched data elements, validating contextualised terminologies and digitising the validated terminologies. The terminology standards contextualisation process was validated using the Delphi technique and the HIV Information Management Services use case. The validation results showed that the contextualisation process was relevant, usable, adaptable and interoperable to Uganda’s eHIS. Conclusion: Accordingly, this study demonstrated how international digital health terminologies could be contextualised for Uganda’s health information systems. The contextualisation process could also be applied to other disease information management services in Uganda.

Keywords

contextualisation digital health information systems semantic interoperability terminologies

Introduction

In a world with distributed health services, interoperable health information systems (HIS) between health facilities, patients and providers are paramount. The interoperability of national HIS facilitates the exchange and use of vital information across government systems.¹ The ability of systems to interpret exchanged linguistic information in meaningful and consistent ways is referred to as semantic interoperability.² Semantic interoperability enables information exchange without loss of meaning and addresses the different needs of users in terms of health data granularity.^3,4 However, semantic interoperability is a challenge because of the complexities in interpreting medical terminologies among health information users and information systems. Terminologies are presented in standard ways as medical codes and controlled vocabularies.⁵ These are key to achieving semantic interoperability because they provide a standardised way of presenting health information and contribute to solving the interpretation challenge by decoding the semantic differences prevailing in interoperable structures.⁶ Standardised ways are statistical classifications and nomenclatures or terminologies⁷ such as the International Classification of Diseases (ICD) of the World Health Organisation,⁸ Logical Observation Identifiers Names and Codes (LOINC) from Regenstrief Institute,⁹ Systematised Nomenclature of Medicine from the International Health Terminology Standards Development Organisation (SNOMED),¹⁰ Structured Product Labelling from Health Level Seven International,¹¹ Current Procedural Terminology from the American Medical Association,¹² and RxNorm for clinical drugs maintained by the United States National Library of Medicine.¹³ This study focused on ICD, SNOMED, and LOINC terminologies since these are already in use in Uganda’s health system,¹⁴ are open-source and accessible^8,15 and have been contextualised in other settings like Thailand, USA and Canada.^16–18

The term contextualisation in this context refers to the process of adapting a relevant terminology to a country setting.¹⁹ During the contextualisation process, it is important to identify and mitigate any design-actuality issues,²⁰ and also appreciate the local setting.²¹ Moreover, terminology guidelines ought to be contextualised to a given setting and not merely adopted as a standard.²² The process of contextualisation reduces information overload by filtering applicable information, highlighting the relevance of important information and guiding its understanding.²³

Several high-income and upper-middle-income countries including Thailand, USA and Canada have contextualised the ICD standard to their national contexts.²⁴ Japan has the HEaLth Information and Communication Standards Organisation (HELICS) which deals with standardisation and contextualisation of country-specific terminologies.²⁵ Germany has contextualised the ICD-10 standard.^26,27 The United States has contextualised ICD for getting the combined comorbidity score.²⁸ The Canadian public health system has implemented SNOMED-CT for managing clinical information^29,30 and contextualised ICD for her country’s information needs.³¹ France has contextualised both SNOMED-CT and LOINC.^15,32 Estonia contextualised LOINC and it became the standard coding terminology for the Health Information System.¹⁵ Much as there is evidence of contextualisation of the international terminologies by high-income and upper-middle-income countries, the researchers found neither documented processes of terminology contextualisation by high-income and upper-middle-income countries nor evidence of terminology contextualisation for low- and middle-income countries. As such, Uganda has the benefit of using learnings from high-income and upper-middle-income countries to contextualise her terminologies.

Currently, the government of Uganda is strengthening its health services through standardising digital health.³³ The HIS is undergoing a digital transformation in health programs such as HIV/AIDS, Reproductive and Child Health, Malaria, Tuberculosis, Immunisation and Nutrition, among others. The HIV programme was used as a use case because of the significant investment of donors in digitalisation to improve HIV data management. There are similar programs to support the digitalisation of HIV data management in countries supported by the U.S. President’s Emergency Plan for AIDS Relief³⁴ as documented in the PEPFAR’s 5-year strategy document of ending the HIV/AIDS Pandemic by 2030.³⁴

Standardisation improves digital health systems.^35–38 One of the core components of digital health standardisation is terminologies. The need for terminologies is well-documented in the Uganda Health Information and Digital Health Strategic Plan 2020/21-2024/25³⁹; It is one of the Ministry of Health (MoH) priorities to improve data interoperability of systems.⁴⁰ Currently, the implementation of international digital health terminologies in Uganda’s eHIS does not follow any national guidance; although some are implemented in the existing open-source systems like DHIS2 and OpenMRS.¹⁴ Even then, implementing multiple terminologies in a specific health context leads to complexity and incompatibility of health data standards.⁴¹ Worse still, the design considerations of the international digital health terminologies in terms of health processes, data collected, technologies, health workforce capabilities and management structures do not necessarily match the contextual needs of Uganda’s eHIS.⁴²

Thus, this research aimed to design a process that would guide the contextualisation of international digital health terminologies to support electronic patient data exchange for Uganda’s health services.^43–46 The design of the contextualisation process is a result of earlier research regarding the functional and non-functional requirements for contextualising digital health terminologies in Uganda’s Electronic Medical Records-based health information.⁴⁷

Methods

Study design

The study adopted a descriptive cross-sectional design. Descriptive cross-sectional designs represent a subset of the population at a specific time point.⁴⁸

Study setting and population

The study was carried out in Uganda considering the national and sub-national levels of the health system. The population constituted stakeholders initially involved in the requirements gathering and validation.⁴⁷ The stakeholders included policymakers (representatives from the MoH – Division of Health Information (DHI) and Uganda National Bureau of Standards), digital health development partners, academia and terminologies’ end-users. They were either technical officers or specialists. The stakeholder groups were purposively sampled because they were involved in the terminology management processes and had been involved in the requirements gathering. Twelve participants from the stakeholder groups were involved in the study. No study participants refused to participate in the study; however, some participants were unavailable on some workshop days due to engagements in other competing interests.

Design method and data collection techniques

The design science methods guided the designing of the terminology contextualisation process.^49,50 Design science was selected because it seeks to understand and improve artefacts and their development processes.^50,51 Besides design science methods are intrinsically problem-solving approaches to practical problems in natural settings.^51,52 Whereas action research would have been a suitable alternate method, it focuses more on social and organizational change, not artefact development.⁵³ The researchers particularly used the generic design science steps of Achampong and Dzidonu’s approach,⁵⁴ combined with the contextualisation approach of Nogueira et al.,⁵⁵ to design the contextualisation process, since the latter had also been successfully applied in another study on the foundation contextualisation approach for mapping terminologies.⁵⁵

Figure 1 illustrates how the design science approach⁵⁴ was combined with the foundation contextualisation approach⁵⁵ to generate the contextualisation process. Each design stage had a corresponding research outcome. At the planning stage, the researchers validated the problem and requirements for designing the contextualisation process with the stakeholders.^14,47,56 The outcome at this stage was a well-defined problem with validated requirements.

Figure 1.

Design science methodological approach (adapted from Achampong and Dzidonu)⁵⁴.

The researchers adapted the four foundation contextualisation steps⁵⁵ during the design stage. The data collection techniques used at this stage were literature review and focus group discussions. The study participants were sent invitation emails and requested to participate in the focus groups in person at conference rooms away from their workplaces to avoid workplace distractions.

The focus groups in the four workshops constituted a maximum of 12 members for each focus group meeting, taking about 90 minutes. AK, a male researcher trained in health information systems and qualitative research methods, facilitated the focus group meetings. AK established a good rapport with the study participants during the requirements-gathering process. He took notes during the focus group discussions. Although the researchers had no prior biases, they presented the four foundation contextualisation steps⁵⁵ to guide the discussions. The interviewer ensured data saturation before ending the focus group discussion. Themes were derived during the workshop discussions.

Additionally, the HIV Information Management Services use case was adopted following the Delphi technique^57–59 to validate the applicability/practicality of the contextualisation process. The outcome of this stage was a validated structured and logical contextualisation process.

Inclusion and exclusion criteria

Inclusion amongst respondents in the study considered stakeholders involved in the requirements gathering and validation process. The respondents knew the digital health landscape of Uganda due to their job roles and involvement in digital health projects. Stakeholders who were not involved in the requirements gathering and validation were excluded.

Ethical considerations

The ethical approval was obtained from a public Higher Institution of Learning’s Higher Degrees, Research and Ethics Committee. The MoH also permitted this research to be conducted. Additionally, written informed consent was obtained from the study participants before involving them in the research process. The study participants were informed that this study is part of a project leading to the facilitator’s award of a doctorate and the development of the Health Enterprise Architecture of Uganda.

Results

Designing of the digital health terminologies contextualisation process

Discussions from the focus group discussions yielded the themes which later constituted the phases and moderating factors. The terminology contextualisation process is a six-phase process with three moderating factors. The moderating factors determine the success of the contextualisation process and may account for the difference in outcomes for the process if used in different settings or health services^60,61; for example, may affect the acceptance of the contextualisation process. These factors include capacity building, public-private partnerships, and terminology governance. Figure 2 illustrates the contextualisation process phases and outputs at the different phases.

Figure 2.

The digital health terminologies contextualisation process.

Preliminary phase: need for terminologies

The national health system is assessed for opportunities and complexities leading to the need for contextualising international terminologies. Alignment of the need with national strategic plans facilitates gaining stakeholder buy-in from the government through the MoH. The need can be due to changes in healthcare practices, stakeholder feedback, new guidelines or regulations, standardisation efforts, legal requirements or advances in medical knowledge. Alternatively, a review may be caused by developing new guidelines or regulations. These drivers trigger either a need for new terminologies or a need to review existing terminologies. The MoH’s DHI should manage the review process.

Phase 1: Assess national HIS context

Assessing the national HIS context focuses on understanding the data collection tools and processes to identify the terminology systems used. This phase involves analysing the national digital health landscape to ascertain barriers and enabling factors for contextualising international terminologies. The assessment of the national HIS context includes health services and processes, capacity skillset, resources, non-homologated coding systems used, data collection tools, digital applications in use, and terminology opportunities and challenges. This phase posits the benchmark for assessing the success of the contextualisation process. Identifying the national terminology codes and systems that reference the international standards used in the national setting follows. This phase culminates in the current state of terminology implementation.

Phase 2: Extract data elements in the national HIS

All data fields in the HIS tools are extracted and duplicates in the forms are noted so that they are assigned the same terminology. The output of this phase is a comprehensive digital list of health service data elements.

Phase 3: Map existing national data elements to standard terminologies

Identify an international terminology that best matches the extracted data element. The matching data element is sought from the search boxes of the international terminologies. Nonetheless, manual review by subject matter experts should review the mappings to select the most relevant mappings to increase their accuracy. This phase produces a list of matched and unmatched health conditions represented in the data elements.

Phase 4: Assign codes to unmatched data elements

During this phase, it is important to identify national disease conditions and procedures unrepresented and those that cannot be exactly matched with the international digital health terminologies. These are called “unmatched data elements” in this study context. Coding unmatched data elements follows a four-step process as seen in Figure 3. First, identify the appropriate international terminology(ies) that addresses national disease conditions and procedures. Second, search for a similar code in one of the closely matching terminologies. Third, place a five-letter prefix at the beginning of the code to identify the terminology with a specific health service. Lastly, add a four-numbered suffix at the end of the code to serially identify the terminology and differentiate any other contextualised diseases or diagnoses added later. Stakeholders should suggest the selection of prefixes and suffixes. The respondents proposed the prefix “UG” signifying a national code plus three letters identifying the health service. The suffix is added serially as per each approved contextualised terminology.

Figure 3.

Assigning codes to unmatched data elements.

Phase 5: Validate contextualised terminologies

Contextualised standards are validated to ensure that the developed terminologies meet stakeholder requirements. The phase involves ensuring that the contextualised terminologies match the disease conditions or medical processes. All stakeholders should be involved in the validation and contextualisation process. Individuals can be involved in the validation process through communities of practices or by providing expert opinions. Stakeholders to validate contextualised standards include national ministries, departments and agencies, academia, technology standards organisations, health development partners, and software developing companies.

Phase 6: Digitise validated terminologies

Terminology services offer an opportunity to share and understand the multitude of recorded data in the health sector that’s usually stored in different formats. The terminology services translate the data into machine-readable codes that can be shared with other systems. The digitisation process ensures that all validated terminologies are easily accessed by application developers, healthcare workers and stakeholders through agreed-upon protocols. The terminology services and interface vocabularies should contain or reference national registries and contextualised terminologies. An interface vocabulary bridges the gap between terminologies and how we write and speak naturally, incorporating different jargon, acronyms and abbreviations. Security of the terminology services is paramount to avoid breaches. Digitally signed end-user agreements reduce the security liabilities of programmers and national bodies for implementing terminology services.

Contextualisation process design iterations

Designing the contextualisation process underwent three iterations conducted in four stakeholder workshops. In the first iteration, the researchers with the stakeholder groups brainstormed on how the final product of the contextualisation process should be. The researchers guided the discussions using the initial/foundation contextualisation process and the initial four steps before these were expanded to six. The initial four steps are extraction of terms, identification of previously published open EHR archetypes, assessment of the adequacy of those open EHR archetypes and development of new open EHR archetypes when required when modified to suit the contextualised process. Additionally, the stakeholders recommended an assessment phase to understand the health information system context and a validation phase to check the accuracy of the contextualised codes.

In the second iteration, the stakeholders recommended the inclusion of a preliminary phase to cater for any drivers that needed to be triggered before the contextualisation process started. The identified drivers are changes in healthcare practices, feedback from stakeholders, new guidelines or regulations, advances in medical technologies and legal requirements. Besides, the stakeholders recommended a contextualisation champion.

In the third workshop, the stakeholders recommended that the third to fifth phases permit a review of the mapped terminologies; that is, looping between map existing national data elements to validate contextualised terminologies. The discussion extended to the enabling and constraining factors for the contextualisation process in Uganda, and requirements for implementing terminologies that had been identified in prior research. Through the discussions, the identified conditions were coined as “moderating factors”. Feedback from each of the three iterations was reviewed to ensure the researchers had improved the contextualisation process as per the recommendations in the previous workshops. There was no further input from the stakeholders in the fourth workshop.

Moderating factors

Terminology governance

The development of guidelines by the national governing body to ensure all stakeholders access the terminologies is fundamental for the success of the terminologies contextualisation process. For instance, digital health application developers need an Application Programming Interface whereas healthcare workers need the terminologies to complete data collection forms. Governance also ensures terminologies cover all health services and there is a framework to monitor compliance with the terminology regulations.

Capacity building

Training should focus on both pre-service and in-service healthcare workers. Terminology concepts can be incorporated into health sciences training institutions’ curricula. Short courses for terminologies can be tailored for healthcare workers to fast-track their capacities in terminologies.

Public-private partnerships

Platforms for public and private digital health stakeholders’ collaboration and communication stimulate discussions and engagement on the future of digital health terminologies. Terminologies do not operate in a vacuum and technologies cannot exist independently of individuals.

Validation of the terminology contextualisation process

Validation involves establishing if an artefact is appropriate for its purpose and will perform as expected.⁶² As such, the Delphi method guided the validation process.^57,63,64 Twelve experts with graduate (Master’s) education levels participated in the Terminology Contextualisation Process validation workshop. The experts were considered to have specialised and adequate knowledge and training in digital health. The experts were from MoH (4), academia (1), Uganda National Bureau of Standards (1), digital health development partners (3) and terminologies’ end users (3). The validation workshop addressed all the issues that the stakeholders and national experts had. The validation considered the four requirements categories; mapping healthcare business processes, developing terminology services, governance structures and capacity building.⁴⁷ The adopted validation parameters⁵⁰ included; relevance (the degree to which the designed process addresses the lack of contextualised standards), usability (the extent to which the process is easy to use), adaptability (the ability of the contextualisation process to be used for other international terminologies and disease information management services), acceptance (degree to which the process is adopted and used by stakeholders) and satisfaction (stakeholders’ contentment of the process). Validation forms were also distributed to members of the MoH’s Health Information, Innovation and Research Technical Working Group (HIIRE TWG). The forms were pilot-tested before they were used. The TWG has representation from all digital health stakeholders in Uganda’s health sector. Table 1 shows the results of the validation metrics. The metric scores were ranked on a Likert scale of 1 to 5 where 1 = strongly disagree, 2 = disagree, 3 = neutral, 4 = agree and 5 = strongly agree.

Table 1.

Validation metric scores.

Validation statement	Average score
The terminology contextualisation process can easily be used by multiple stakeholders in the healthcare industry	5
The terminology contextualisation process is relevant for Uganda’s health services and can effectively contextualise international terminologies to suit the Ugandan digital ecosystem to address semantic interoperability challenges	5
The terminology contextualisation process can be adapted for other health services in Uganda like malaria	5
The terminology contextualisation process contextualises standards that are interoperable with other international standards such as HL7 and FHIR	5
The terminology contextualisation process satisfactorily addresses the semantic interoperability⁶⁵ challenges of the health sector	5
The terminology contextualisation process can be used as is by healthcare stakeholders	5

Applicability of the terminology contextualisation process using HIV information management services use case

Preliminary phase

The need for terminologies was triggered by new national guidelines that required standardisation of digital health in Uganda. The MoH championed the HIV Information Management Services contextualisation process.

Phase 1

Information management in Uganda’s healthcare system is guided by the national Health Management Information System (HMIS), which is a paper-based system monitored by MoH DHI. The paper-based HMIS comprises identifying cards, forms, registers and charts as summarised in Figure 4.

Figure 4.

Uganda HIV service categories and corresponding data collection tools.

Phase 2

Data elements from the cards, forms and registers were extracted from the national HMIS data tools.

Figures 5 and 6 show extracts from the HMIS ACP (AIDS Control Programme) 001: Viral Load Non-Suppressed Register and the HMIS ACP 002: laboratory request form for HIV viral load analysis/HIV drug resistance testing, respectively. For this paper, the researchers restricted the proof-of-concept for terminologies’ contextualisation to only the mentioned forms.

Figure 5.

HMIS ACP 001 - viral load non-suppressed register.

Figure 6.

Hmis ACP 002 - laboratory request form for HIV viral load analysis/HIV drug resistance testing.

Phase 3

The extracted data elements were mapped to three international digital health terminologies; SNOMED, LOINC and ICD because they are open-source and accessible. For CD4 results, the researchers were informed that the commonly used test is the rapid diagnostic test that measures antigens, this enabled them to type CD4 in the SNOMED CT search browser https://browser.ihtsdotools.org/?. There were over 93 matches representing a high level of granularity. The researchers zeroed down to Cell positive for CD4 antigen (cell) after reviewing the concept details. An example of matched indicators of the HMIS ACP 001 register is shown in Table 2.

Table 2.

HIV services matched to terminologies.

National indicator	International terminology	Terminology code	Terminology code identifier
CD4 result	SNOMED	116745007	Cell positive for CD4 antigen (cell)
Crag results	LOINC	31791-7	Cryptococcus sp Ag - units/volume in serum
Current WHO stage	ICD11	1C62.0	HIV disease without mention of tuberculosis or malaria
Current WHO stage	ICD11	1C62.1	HIV disease clinical stage 2 without mention of tuberculosis or malaria
Current WHO stage	ICD11	1C62.2	HIV disease clinical stage 3 without mention of tuberculosis or malaria
Current WHO stage	ICD11	1C62.3	HIV disease clinical stage 4 without mention of tuberculosis or malaria

Phase 4

Non-matched data elements were assigned codes using the four-step process described in Figure 3. Using an example of eMTCT (Elimination of Mother-to-Child Transmission) in HMIS ACP 001: Viral Load Non-Suppressed Register column 6, the researchers noted that the data element is highly granular and the SNOMED closely aligns with the data element. Second, the most similar terminology in SNOMED was human-to-human transmission with a code 409699001. A five-letter prefix “UGHIV” was assigned to signify that it is a Uganda terminology in the HIV services. A four-digit suffix “0001” was added, to signify that it’s the first terminology to be contextualised; resulting in eMTCT “UGHIV.409699001.0001”.

Phase 5

The contextualised terminologies for the HIV services were compiled as a list in an MS Excel document, and redundancies like demographic data obtained from different forms were removed to have a minimum set of terminologies for HIV services. HIV information management services terminologies were validated in a workshop with members of MoH’s HIIRE TWG. The discussions rotated around the mappings of the HMIS datasets with the international standards.

Phase 6

The contextualised terminologies were not digitised because the MoH had not yet developed the terminology services at the time this study was taken.

Discussion

This study designed a process for contextualising international digital health terminology standards to improve the semantic interoperability of health information systems. Grounded in design science methods, this study has developed an artefact “contextualisation process” that has the practical utility of addressing an existing problem “health terminologies”.^66,67 Notwithstanding its capabilities, the contextualisation process partly addresses the challenge of semantic interoperability from the terminologies’ perspective.

The contextualisation process

This study has extended four methodological mapping steps to design a six-phase process for contextualising international digital health terminology.⁵⁵ The preliminary phase of the contextualisation process focuses on understanding the complexities and opportunities of the national HIS,⁵⁶ setting the environment for the next phases of contextualisation and ensuring the right buy-in from relevant stakeholders.^68,69

While assessing the national HIS context, the focus is on data elements of a specific health service programme or the entire national HIS targeting both individual and health service-based records.⁷⁰ The contextualisation scope decision is agreed upon by stakeholders. Importantly, data elements of health services that can be an opportunity for contextualisation can be procedures, disease conditions or drugs used in the national HIS.⁷¹ The outcome of the assessment should be a landscape analysis showing the extent of terminology implementation.¹⁴

Phase 2 involves extracting data elements in the national HIS. Given the national HIS is mostly paper-based in many developing countries in Uganda, this calls for manual means of extraction. In this study, the researchers used an MS Excel spreadsheet to develop a list of the HIV services data elements with their corresponding cards, forms and registers. Terminology servers are an opportunity for contextualisation in a digitalised environment.⁷²

Phases 3 and 4 focus on mapping existing national data elements to standard terminologies and assigning codes to unmatched data elements. International digital health terminologies^73–76 are identified through desktop reviews and exploratory assessments.⁷⁷ Identification and extraction of international terminologies considers categorisation used in the terminology nomenclature, that is, titles of the terminology or classifications tables are used as keywords.⁵⁵ Literature shows that semi-automated approaches like Regenstrief LOINC Mapping Assistant (RELMA) for LOINC are used to improve the mapping quality and reduce the time spent during mapping.⁷⁸ Digital clinical documentation is automatically mapped to SNOMED^79,80 and automated processes are also used for ICD.^81–83 Shi et al., showed that ICD can be automatically coded using neural language models.⁸³ However, Uganda’s eHIS data elements are still collected manually; thus, the automated mapping processes cannot be applied. In this study, the researchers used the exact phrases or words to maintain the same meanings and preserve semantic interoperability to map the national HIS data elements to the international digital health terminologies.

It is important to ensure the exact meanings and the terminology semantic structure are maintained through validation and post-mapping coordination.⁸⁴ Validation in this study was done using stakeholders that were involved in the requirements gathering and designing of the process since they had an understanding of what they expected to solve the problem of semantic interoperability from the terminology perspective. Having validated the contextualised standards, it is paramount to digitise the contextualised terminologies into a data dictionary and vocabulary interface to facilitate the development of digital health information systems⁷¹ and terminology services.^85–87 Vocabularies interfaces consider localism of jargon and spelling errors.⁸⁸ A vocabulary interface is usually online and includes all domains facilitating easy retrieval of terminologies although this comes at the disadvantage of having invalid terms and spelling errors.⁷³ Vocabulary interfaces help healthcare professionals who may not be technical in terminology speciality to use the terminologies in a more context-specific language.⁸⁹ This points to further research on nomenclatures and ontologies of medical information⁹⁰ to understand the naming and classifying systems for health procedures and conditions, as well as the structured vocabulary, concepts and their relationships.

Moderating factors

Terminology governance structures, capacity building and collaboration through private-public partnerships are crucial for the successful contextualisation and implementation of terminologies. Successful terminology governance depends on a collaborative structure that includes a working committee, an approval committee and a policy committee.⁹¹ Separation of roles ensures that the necessary checks are done to maintain the proper functioning of the terminology management system. Uganda’s MoH has similar digital health structures that the governance of terminologies aligns with.

Capacity building is critical for the successful contextualisation and implementation of digital health terminologies. Clinical terminology training for health workers reduces the number of errors in clinical coding.¹⁶ Online resources for consultation about the terminologies increase acceptance of the available terminologies and those developed by users.⁷²

Accelerated adoption of terminologies requires public-private partnerships and sufficient incentives, rather than relying on technical innovations only.⁶⁸ Collaborative efforts by different public health and clinical stakeholders lead to better population health data and outcomes.⁹² Collaboration at the initial steps of the contextualisation process improves stakeholder acceptance of all processes and increases acceptance.

Study contribution to research and practice

This study contributes to information systems research a process for contextualising international digital terminology standards to support semantic interoperability of digital health information systems. To practice, this study demonstrates the “how” to contextualise and or adapt international digital health terminologies to achieve semantic interoperability of health information systems.

Study limitation

As a limitation of the study findings, the validation form had not been used in prior studies. The authors mitigated the limitation by pilot-testing the tool before it was used to validate the contextualisation process. Additionally, the composition of the stakeholders in the four stakeholder workshops during the design iterations was not the same. The authors ensured that improvements recommended in the previous meetings were re-echoed in subsequent meetings.

Conclusion

This study designed and validated a process for contextualising international digital health terminologies for Uganda’s eHIS. This work contributes to solving the semantic interoperability challenges of Uganda’s health information systems. The HIV information management services use case demonstrated how the contextualisation process could be applied, and proved to be useful, usable, adaptable and satisfactory. Accordingly, the terminology contextualisation process is generalisable and scalable to other health disease information management services not only in Uganda because the terminologies used in this study, ICD, SNOMED and LOINC, are internationally used in different disease domains beyond the HIV case study. The next research points to developing a data dictionary for Uganda’s health services and implementing terminology services to further improve solving the semantic health information system interoperability challenges in Uganda’s health system.

Supplemental Material

Supplemental Material - A process for contextualising digital health terminology standards for Uganda’s health information systems: A use case of HIV information management services

Supplemental Material for A process for contextualising digital health terminology standards for Uganda’s health information systems: A use case of HIV information management services by Achilles Kiwanuka and Josephine Nabukenya in Health Informatics Journal.

Footnotes

Acknowledgements

The authors acknowledge the contribution of the Makerere University Health Informatics Research Group towards the realisation of this manuscript.

Author contributions

AK: conceptualisation, methodology, data collection, data analysis, drafting the manuscript, reviewing the manuscript. JN: conceptualisation, methodology, reviewing the manuscript.

Declaration of conflicting interest

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Ethical statement

Author credentials

AK: BSc, MSc, PhD candidate (Makerere University), JN: BSc, MSc, PhD.

ORCID iD

Josephine Nabukenya

Supplemental Material

Supplemental material for this article is available online.

References

Mandl

Kohane

. Escaping the EHR trap—the future of health IT. N Engl J Med 2012; 366(24): 2240–2242.

Ide

Pustejovsky

(eds) What does interoperability mean, anyway? Toward an operational definition of interoperability for language technology. In: Proceedings of the second international conference on global interoperability for language resources, Hong Kong China, 18–20 January 2010.

de Mello

Rigo

da Costa

, et al. Semantic interoperability in health records standards: a systematic literature review. Health Technol 2022; 12(2): 255–272.

Arvanitis

. Semantic interoperability in healthcare. Integrating information technology and management for quality of care. Athens, Greece: ICIMTH IOS Press, 2014, 202, pp. 5–8.

Lopez

Blobel

. A development framework for semantically interoperable health information systems. Int J Med Inf 2009; 78(2): 83–103.

Bhartiya

Mehrotra

Girdhar

. Issues in achieving complete interoperability while sharing electronic health records. Procedia Comput Sci 2016; 78: 192–198.

Madden

Sykes

Ustun

. World Health Organization family of international classifications: definition, scope and purpose. Geneva, Switzerland: World Health Organization, 2007.

World Health Organisation . International statistical classification of diseases and related health problems (ICD) 2022. https://www.who.int/standards/classifications/classification-of-diseases

Regenstrief Institute . LOINC release notes, august 2023 2023. Version 2.75. https://loinc.org/kb/release-notes-archive/2.75/

10.

Bhattacharyya

. Overview of SNOMED CT. Introduction to SNOMED CT. Singapore: Springer, 2016, pp. 1–2.

11.

U.S. Food & Drug Administration . Structured product labeling resources 2023. https://www.fda.gov/industry/fda-data-standards-advisory-board/structured-product-labeling-resources

12.

Hirsch

Leslie-Mazwi

Nicola

, et al. Current procedural terminology; a primer. J Neurointerventional Surg 2015; 7(4): 309–312.

13.

National Library of Medicine . RxNorm technical documentation 2023. https://www.nlm.nih.gov/research/umls/rxnorm/docs/techdoc.html.

14.

Kiwanuka

Bagyendera

Wamema

, et al. Establishing the state of practice about data standards in monitoring healthcare interventions for HIV in Uganda’s EMR-based health information systems. Biomedical Engineering Systems and Technologies. 2021; 5: 200–211. doi:10.5220/0010264302000211.

15.

Bodenreider

Cornet

Vreeman

. Recent developments in clinical terminologies - SNOMED CT, LOINC, and RxNorm. Yearb Med Inform 2018; 27(1): 129–139.

16.

Paoin

Yuenyongsuwan

Yokobori

, et al. Development of the ICD-10 simplified version and field test. Health Inf Manag 2018; 47(2): 77–84.

17.

Aggrey

Steven

Daniel

, et al. Developing a national-level concept dictionary for EHR implementations in Kenya. In: Medinfo 2015: eHealth-enabled Health. Amsterdam, Netherlands: IOS Press, 2015, pp. 780–784.

18.

Meke

Singh

. A process analysis of the Namibian Health System: an exploratory case study. Ethiop J Health Dev. 2018; 32(4): 200–209.

19.

Abdusamadovich

. Global standards and local applications: case of implementing ICD-10 standard in HMIS Tajikistan. Journal of Health Informatics in Developing Countries. 2013; 7(1): 20–33.

20.

Heeks

. Information systems and developing countries: failure, success, and local improvisations. Inf Soc 2002; 18(2): 101–112.

21.

Rosemann

Recker

Flender

. Contextualisation of business processes. Int J Bus Process Integrat Manag 2008; 3(1): 47–60.

22.

Walsh

. Ethical standards in research: contextualising guidelines to fit ethnographic needs. Ecclectica 2003; 1: 2003.

23.

Klemke

Nick

(eds) Case studies in developing contextualising information systems. international and interdisciplinary conference on modeling and using context. Salmon Tower Building New York City: Springer, 2001.

24.

Alharbi

Isouard

Tolchard

. The development of ICD adaptations and modifications as background to a potential Saudi Arabia’s National version. Global J Health Sci. 2019; 11(11): 158–167.

25.

Kimura

Ueno

. Trends in health information and communication standards in Japan. Journal of the National Institute of Public Health 2020; 69(1): 52–62.

26.

de Sordi

Kappen

Otto‐Sobotka

, et al. Validity of hospital ICD‐10‐GM codes to identify anaphylaxis. Pharmacoepidemiol Drug Saf 2021; 30(12): 1643–1652.

27.

Stausberg

Pollex-Krüger

Semler

, et al. Field tests for the beta version of the ICD-11-MMS in Germany: background and methods. Bundesgesundheitsblatt - Gesundheitsforsch - Gesundheitsschutz 2018; 61: 836–844.

28.

Sun

Rogers

Her

, et al. Adaptation and validation of the combined comorbidity score for ICD-10-CM. Med Care 2017; 55(12): 1046–1051.

29.

Block

Wong

Handfield

, et al. Comparison of terminology mapping methods for nursing wound care knowledge representation. Int J Med Inf 2021; 153: 104539.

30.

Zhao

Alghamdi

Schmidt

, et al. (eds) Tracking logical difference in large-scale ontologies: a forgetting-based approach. In: Proceedings of the AAAI conference on artificial intelligence, Honolulu, Hawaii, USA, 27 January 2019–1 February 2019.

31.

Eastwood

Southern

Khair

, et al. Field testing a new ICD coding system: methods and early experiences with ICD-11 Beta Version 2018. BMC Res Notes 2022; 15(1): 343.

32.

Merabti

Soualmia

Grosjean

, et al. Assisting the translation of SNOMED CT into French. MEDINFO 2013. Amsterdam, Netherlands: IOS Press; 2013. p. 47–51.

33.

Ministry of Health . Strengthening Uganda’s health system through standardizing digital health. Requirements for digital health standards and Enterprise architecture framework. Republic of Uganda: Ministry of Health, 2021.

34.

PEPFAR . PEPFAR’s five-year strategy: fulfilling America’s promise to end the HIV/AIDS pandemic by 2030 2022.

35.

Shanks

Gloet

Asadi Someh

, et al. Achieving benefits with enterprise architecture. J Strat Inf Syst 2018; 27(2): 139–156.

36.

Wears

. Standardisation and its discontents. Cognit Technol Work 2015; 17: 89–94.

37.

Schulz

Stegwee

Chronaki

. Standards in healthcare data. Fundamentals of clinical data science. 2019; 19–36.

38.

Eslami Andargoli

Scheepers

Rajendran

, et al. Health information systems evaluation frameworks: a systematic review. Int J Med Inf 2017; 97: 195–209.

39.

Ministry of Health . The Uganda health information and digital health strategic plan 2020/21-2024/25. Kampala, Uganda: Ministry of Health, 2021.

40.

Ministry of Health . The Uganda health information and digital health strategic plan 2020/21-2024/25. Kampala, Uganda: Ministry of Health, 2021.

41.

Alkraiji

Jackson

Murray

. Factors impacting the adoption decision of health data standards in tertiary healthcare organisations in Saudi Arabia. J Enterprise Inf Manag 2016; 29(5): 650–676.

42.

Heeks

. Health information systems: failure, success and improvisation. Int J Med Inf 2006; 75(2): 125–137.

43.

Wamema

Alunyu

Amiyo

, et al. Enterprise architecture requirements for standardising digital health in Uganda's health system. Health Policy and Technology 2023; 12: 100805.

44.

World Health Organisation . Global strategy on digital health 2020-2025 2021.

45.

Alunyu

Amiyo

Nabukenya

. State of digital health communication infrastructure in LMICs: theory, standards and factors affecting their implementation. In: Current and future trends in health and medical Informatics. Cham, Switzerland: Springer, 2023, pp. 109–134.

46.

Abima

Nakakawa

Kitui

. Service-oriented framework for developing interoperable e-health systems in a low-income country. The African Journal of Information Systems 2023; 15(3): 141–174.

47.

Kiwanuka

Nabukenya

(eds) Requirements for implementing digital health terminology standards in Uganda’s electronic medical records-based health information systems. In: Gehin

et al. Biomedical engineering systems and technologies. BIOSTEC 2021. Communications in computer and information science. Vol 1710. Springer, Cham. https://doi.org/10.1007/978-3-031-20664-1_11

48.

Maier

Thatcher

Grover

, et al. Cross-sectional research: a critical perspective, use cases, and recommendations for IS research. Amsterdam, Netherlands: Elsevier, 2023, 102625.

49.

Vom Brocke

Hevner

Maedche

. Introduction to design science research. Design science research Cases. 2020; 1–13.

50.

Hevner

Chatterjee

. Design science research in information systems. Design research in information systems: theory and practice. New York: Springer, 2010, pp. 9–22.

51.

Hevner

. Design science research. Computing handbook. 3rd ed. Boca Raton, USA: Taylor & Francis Group, 2014.

52.

Johanesson

Perjons

. Design science. In: An introduction to design science. Cham, Switzerland: Springer, 2014, vol 10, pp. 978.

53.

Iivari

Venable

. Action research and design science research-seemingly similar but decisively dissimilar. ECIS 2009 Proceedings. 2009; (53).

54.

Achampong

Dzidonu

. Methodological framework for artefact design and development in design science research. Journal of Advances in Science and Technology Research (JASTER). 2017; 4(1): 1–8.

55.

Nogueira

Cook

Cavalini

. Mapping a nursing terminology subset to openEHR archetypes. A case study of the international classification for nursing practice. Methods Inf Med 2015; 54(03): 271–275.

56.

Kiwanuka

Amiyo

Nabukenya

. Constraints to and enablers for contextualizing digital health terminology standards in Uganda’s health services: a qualitative case study. Oxford Open Digital Health 2023; 1: oqad003.

57.

Alarabiat

Ramos

. The delphi method in information systems research (2004-2017). Electron J Bus Res Methods 2019; 17(2): 86–99.

58.

Laick

. Using Delphi methodology in information system research. Int J Managmnt Cases 2012; 14(4): 261–268.

59.

Ogbeifun

Agwa-Ejon

Mbohwa

, et al. The Delphi technique: a credible research methodology. Proceedings of the 2016 International Conference on Industrial Engineering and Operations Management. 2016.

60.

Sun

Zhang

. The role of moderating factors in user technology acceptance. Int J Hum Comput Stud 2006; 64(2): 53–78.

61.

Keane

Cormican

Flynn

. User acceptance of mandated technology: an analysis of the role of moderating factors. Procedia Comput Sci 2019; 164: 72–78.

62.

Marwedel

. Evaluation and validation 2011. https://www.researchgate.net/publication/241027419_Evaluation_and_Validation/citations

63.

Yousuf

. Using experts opinions through Delphi technique. Practical Assess Res Eval 2019; 12(1): 4.

64.

Habibi

Sarafrazi

Izadyar

. Delphi technique theoretical framework in qualitative research. Int J Eng Sci 2014; 3(4): 8–13.

65.

International Health Terminology Standards Development Organisation . IHTSDO launches global health terminology workbench - tools to benefit member nations and other SNOMED CT users 2008 2020. https://www.nlm.nih.gov/research/umls/Snomed/IHTSDOLaunchesGlobalHealthTerminologyWorkbenchJanuary2009.pdf

66.

Hevner

. A three cycle view of design science research. Scand J Inf Syst 2007; 19(2): 4.

67.

Hevner

March

Park

, et al. Design science in information systems research. MIS Q 2004; 28: 75–105.

68.

Aspden

Corrigan

Wolcott

, et al. Components of a national health information infrastructure. Patient safety: achieving a new standard for care. Washington, DC: National Academies Press (US), 2004.

69.

Dixon

Holmgren

Adler-Milstein

, et al. Health information exchange and interoperability. Clinical Informatics study guide: text and review. Cham, Switzerland: Springer, 2022, pp. 203–219.

70.

World Health Organisation . Framework and standards for country health information systems. Geneva, Switzerland: World Health Organization, 2008.

71.

Zhao

Yang

, et al. Primary health information standard system based on semantic interoperability. BMC Med Inf Decis Making 2018; 18(5): 112–127.

72.

Navas

Osornio

Baum

, et al. Creation and evaluation of a terminology server for the interactive coding of discharge summaries. Stud Health Technol Inf 2007; 129(1): 650–654.

73.

Pan American Health Organisation . Introduction to semantic interoperability. Digital transformation toolkit knowledge tools. 2021. https://iris.pahp.org/bitstream/handle/10665.2/55417/PAHOEIHIS21023_eng.pdf

74.

Awaysheh

Wilcke

Elvinger

, et al. A review of medical terminology standards and structured reporting. J Vet Diagn Invest 2018; 30(1): 17–25.

75.

Benson

Grieve

. Principles of health interoperability. Cham, Switzerland: Springer International, 2021, pp. 21–40.

76.

Van Laere

Verbeke

Questier

, et al. Mapping of terminology standards - a way for interoperability. In Proceedings of the Fourth International Conference on Telecommunications and Remote Sensing (ICTRS 2015). 2015; 76–80.

77.

Saunders

Lewis

Thornhill

. Research methods for business students. 6th ed. London: Pearson, 2014.

78.

Zunner

Bürkle

Prokosch

H-U

, et al. Mapping local laboratory interface terms to LOINC at a German university hospital using RELMA V. 5: a semi-automated approach. J Am Med Inf Assoc 2013; 20(2): 293–297.

79.

Adlassnig

(ed) Automatic mapping of clinical documentation to SNOMED CT. In: Medical informatics in a united and healthy Europe: proceedings of MIE 2009, the XXII international congress of the European federation for medical informatics. Amsterdam, Netherlands: IOS Press, 2009.

80.

Allones

Martinez

Taboada

. Automated mapping of clinical terms into SNOMED-CT. An application to codify procedures in pathology. J Med Syst 2014; 38(10): 134.

81.

Lam

Pang

, et al. (eds) Multimodal machine learning for automated ICD coding. In: Machine learning for healthcare conference, Ann Arbor, MI, 8–10 August 2019, PMLR.

82.

Fung

Richesson

Smerek

, et al. Preparing for the ICD-10-CM transition: automated methods for translating ICD codes in clinical phenotype definitions. eGEMs 2016; 4(1): 1211.

83.

Shi

Xie

, et al. Towards automated ICD coding using deep learning. arXiv preprint. arXiv:171104075. 2017.

84.

Karlsson

Nyström

Cornet

. Does SNOMED CT post-coordination scale? e-Health–for continuity of care. Amsterdam, Netherlands: IOS Press, 2014, pp. 1048–1052.

85.

Metke-Jimenez

Steel

Hansen

, et al. Ontoserver: a syndicated terminology server. J Biomed Semant 2018; 9(1): 24.

86.

Lab

. Open Concept Lab is an open-source terminology management system 2022. https://openconceptlab.org/

87.

HL7 . Welcome to FHIR. Michigan, USA: HL7, 2019. https://www.hl7.org/fhir/terminology-module.html.

88.

Bakhshi-Raiez

Cornet

de Keizer

. Development and application of a framework for maintenance of medical terminological systems. J Am Med Inf Assoc 2008; 15(5): 687–700.

89.

McCormick

. Business use-case modeling. 2013.

90.

Lehne

Sass

Essenwanger

, et al. Why digital medicine depends on interoperability. NPJ Digit Med 2019; 2(1): 79.

91.

Theeraroungchaisri

Phuksaritanon

Kijsanayotin

. A medicines terminology governance model for Thailand. Thai Journal of Pharmaceutical Sciences 2017; 41(2): 82–92.

92.

Dixon

Vreeman

Grannis

. The long road to semantic interoperability in support of public health: experiences from two states. J Biomed Inf 2014; 49: 3–8.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.40 MB