Hepatitis A Vaccination Coverage Among People With Chronic Liver Disease in England (HEALD): Protocol for a Retrospective Cohort Study

BACKGROUND

Hepatitis A outbreaks in the United Kingdom are uncommon. Most people develop mild to moderate symptoms that resolve, without sequelae, within months. However, in high-risk groups, including those with underlying chronic liver disease (CLD), hepatitis A infection can be severe, with a higher risk of mortality and morbidity. The Health Security Agency and the National Institute of Health and Care Excellence recommend preexposure hepatitis A vaccination given in 2 doses to people with CLD, regardless of its cause. There are currently no published reports of vaccination coverage for people with CLD in England or internationally.

OBJECTIVE

This study aims to describe hepatitis A vaccination coverage in adults with CLD in a UK primary care setting and compare liver disease etiology, sociodemographic characteristics, and comorbidities in people who are and are not exposed to the hepatitis A vaccine.

METHODS

We will conduct a retrospective cohort study with data from the Primary Care Sentinel Cohort of the Oxford-Royal College of General Practitioners Clinical Informatics Digital Hub database, which is nationally representative of the English population. We will include people aged 18 years and older who have been registered in general practices in the Research and Surveillance Centre network and have a record of CLD between January 1, 2012, and December 31, 2022, including those with alcohol-related liver disease, chronic hepatitis B, chronic hepatitis C, nonalcohol fatty liver disease, Wilson disease, hemochromatosis, and autoimmune hepatitis. We will carefully curate variables using the Systematized Nomenclature of Medicine Clinical Terms. We will report the sociodemographic characteristics of those who are vaccinated. These include age, gender, ethnicity, population density, region, socioeconomic status (measured using the index of multiple deprivation), obesity, alcohol consumption, and smoking. Hepatitis A vaccination coverage for 1 and 2 doses will be calculated using an estimate of the CLD population as the denominator. We will analyze the baseline characteristics using descriptive statistics, including measures of dispersion. Pairwise comparisons of case-mix characteristics, comorbidities, and complications will be reported according to vaccination status. A multistate survival model will be fitted to estimate the transition probabilities among four states: (1) diagnosed with CLD, (2) first dose of hepatitis A vaccination, (3) second dose of hepatitis A vaccination, and (4) death. This will identify any potential disparities in how people with CLD get vaccinated.

RESULTS

The Research and Surveillance Centre population comprises over 8 million people. The reported incidence of CLD is 20.7 cases per 100,000. International estimates of hepatitis A vaccine coverage vary between 10% and 50% in this group.

CONCLUSIONS

This study will describe the uptake of the hepatitis A vaccine in people with CLD and report any disparities or differences in the characteristics of the vaccinated population.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)

PRR1-10.2196/51861.

Methodological Issues in Using a Common Data Model of COVID-19 Vaccine Uptake and Important Adverse Events of Interest: Feasibility Study of Data and Connectivity COVID-19 Vaccines Pharmacovigilance in the United Kingdom

BACKGROUND

The Data and Connectivity COVID-19 Vaccines Pharmacovigilance (DaC-VaP) UK-wide collaboration was created to monitor vaccine uptake and effectiveness and provide pharmacovigilance using routine clinical and administrative data. To monitor these, pooled analyses may be needed. However, variation in terminologies present a barrier as England uses the Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT), while the rest of the United Kingdom uses the Read v2 terminology in primary care. The availability of data sources is not uniform across the United Kingdom.

OBJECTIVE

This study aims to use the concept mappings in the Observational Medical Outcomes Partnership (OMOP) common data model (CDM) to identify common concepts recorded and to report these in a repeated cross-sectional study. We planned to do this for vaccine coverage and 2 adverse events of interest (AEIs), cerebral venous sinus thrombosis (CVST) and anaphylaxis. We identified concept mappings to SNOMED CT, Read v2, the World Health Organization's International Classification of Disease Tenth Revision (ICD-10) terminology, and the UK Dictionary of Medicines and Devices (dm+d).

METHODS

Exposures and outcomes of interest to DaC-VaP for pharmacovigilance studies were selected. Mappings of these variables to different terminologies used across the United Kingdom's devolved nations' health services were identified from the Observational Health Data Sciences and Informatics (OHDSI) Automated Terminology Harmonization, Extraction, and Normalization for Analytics (ATHENA) online browser. Lead analysts from each nation then confirmed or added to the mappings identified. These mappings were then used to report AEIs in a common format. We reported rates for windows of 0-2 and 3-28 days postvaccine every 28 days.

RESULTS

We listed the mappings between Read v2, SNOMED CT, ICD-10, and dm+d. For vaccine exposure, we found clear mapping from OMOP to our clinical terminologies, though dm+d had codes not listed by OMOP at the time of searching. We found a list of CVST and anaphylaxis codes. For CVST, we had to use a broader cerebral venous thrombosis conceptual approach to include Read v2. We identified 56 SNOMED CT codes, of which we selected 47 (84%), and 15 Read v2 codes. For anaphylaxis, our refined search identified 60 SNOMED CT codes and 9 Read v2 codes, of which we selected 10 (17%) and 4 (44%), respectively, to include in our repeated cross-sectional studies.

CONCLUSIONS

This approach enables the use of mappings to different terminologies within the OMOP CDM without the need to catalogue an entire database. However, Read v2 has less granular concepts than some terminologies, such as SNOMED CT. Additionally, the OMOP CDM cannot compensate for limitations in the clinical coding system. Neither Read v2 nor ICD-10 is sufficiently granular to enable CVST to be specifically flagged. Hence, any pooled analysis will have to be at the less specific level of cerebrovascular venous thrombosis. Overall, the mappings within this CDM are useful, and our method could be used for rapid collaborations where there are only a limited number of concepts to pool.

Differences in Clinical Presentation With Long COVID After Community and Hospital Infection and Associations With All-Cause Mortality: English Sentinel Network Database Study

BACKGROUND

Most studies of long COVID (symptoms of COVID-19 infection beyond 4 weeks) have focused on people hospitalized in their initial illness. Long COVID is thought to be underrecorded in UK primary care electronic records.

OBJECTIVE

We sought to determine which symptoms people present to primary care after COVID-19 infection and whether presentation differs in people who were not hospitalized, as well as post-long COVID mortality rates.

METHODS

We used routine data from the nationally representative primary care sentinel cohort of the Oxford-Royal College of General Practitioners Research and Surveillance Centre (N=7,396,702), applying a predefined long COVID phenotype and grouped by whether the index infection occurred in hospital or in the community. We included COVID-19 infection cases from March 1, 2020, to April 1, 2021. We conducted a before-and-after analysis of long COVID symptoms prespecified by the Office of National Statistics, comparing symptoms presented between 1 and 6 months after the index infection matched with the same months 1 year previously. We conducted logistic regression analysis, quoting odds ratios (ORs) with 95% CIs.

RESULTS

In total, 5.63% (416,505/7,396,702) and 1.83% (7623/416,505) of the patients had received a coded diagnosis of COVID-19 infection and diagnosis of, or referral for, long COVID, respectively. People with diagnosis or referral of long COVID had higher odds of presenting the prespecified symptoms after versus before COVID-19 infection (OR 2.66, 95% CI 2.46-2.88, for those with index community infection and OR 2.42, 95% CI 2.03-2.89, for those hospitalized). After an index community infection, patients were more likely to present with nonspecific symptoms (OR 3.44, 95% CI 3.00-3.95; P<.001) compared with after a hospital admission (OR 2.09, 95% CI 1.56-2.80; P<.001). Mental health sequelae were more strongly associated with index hospital infections (OR 2.21, 95% CI 1.64-2.96) than with index community infections (OR 1.36, 95% CI 1.21-1.53; P<.001). People presenting to primary care after hospital infection were more likely to be men (OR 1.43, 95% CI 1.25-1.64; P<.001), more socioeconomically deprived (OR 1.42, 95% CI 1.24-1.63; P<.001), and with higher multimorbidity scores (OR 1.41, 95% CI 1.26-1.57; P<.001) than those presenting after an index community infection. All-cause mortality in people with long COVID was associated with increasing age, male sex (OR 3.32, 95% CI 1.34-9.24; P=.01), and higher multimorbidity score (OR 2.11, 95% CI 1.34-3.29; P<.001). Vaccination was associated with reduced odds of mortality (OR 0.10, 95% CI 0.03-0.35; P<.001).

CONCLUSIONS

The low percentage of people recorded as having long COVID after COVID-19 infection reflects either low prevalence or underrecording. The characteristics and comorbidities of those presenting with long COVID after a community infection are different from those hospitalized. This study provides insights into the presentation of long COVID in primary care and implications for workload.

Developing a Long COVID Phenotype for Postacute COVID-19 in a National Primary Care Sentinel Cohort: Observational Retrospective Database Analysis

BACKGROUND

Following COVID-19, up to 40% of people have ongoing health problems, referred to as postacute COVID-19 or long COVID (LC). LC varies from a single persisting symptom to a complex multisystem disease. Research has flagged that this condition is underrecorded in primary care records, and seeks to better define its clinical characteristics and management. Phenotypes provide a standard method for case definition and identification from routine data and are usually machine-processable. An LC phenotype can underpin research into this condition.

OBJECTIVE

This study aims to develop a phenotype for LC to inform the epidemiology and future research into this condition. We compared clinical symptoms in people with LC before and after their index infection, recorded from March 1, 2020, to April 1, 2021. We also compared people recorded as having acute infection with those with LC who were hospitalized and those who were not.

METHODS

We used data from the Primary Care Sentinel Cohort (PCSC) of the Oxford Royal College of General Practitioners (RCGP) Research and Surveillance Centre (RSC) database. This network was recruited to be nationally representative of the English population. We developed an LC phenotype using our established 3-step ontological method: (1) ontological step (defining the reasoning process underpinning the phenotype, (2) coding step (exploring what clinical terms are available, and (3) logical extract model (testing performance). We created a version of this phenotype using Protégé in the ontology web language for BioPortal and using PhenoFlow. Next, we used the phenotype to compare people with LC (1) with regard to their symptoms in the year prior to acquiring COVID-19 and (2) with people with acute COVID-19. We also compared hospitalized people with LC with those not hospitalized. We compared sociodemographic details, comorbidities, and Office of National Statistics-defined LC symptoms between groups. We used descriptive statistics and logistic regression.

RESULTS

The long-COVID phenotype differentiated people hospitalized with LC from people who were not and where no index infection was identified. The PCSC (N=7.4 million) includes 428,479 patients with acute COVID-19 diagnosis confirmed by a laboratory test and 10,772 patients with clinically diagnosed COVID-19. A total of 7471 (1.74%, 95% CI 1.70-1.78) people were coded as having LC, 1009 (13.5%, 95% CI 12.7-14.3) had a hospital admission related to acute COVID-19, and 6462 (86.5%, 95% CI 85.7-87.3) were not hospitalized, of whom 2728 (42.2%) had no COVID-19 index date recorded. In addition, 1009 (13.5%, 95% CI 12.73-14.28) people with LC were hospitalized compared to 17,993 (4.5%, 95% CI 4.48-4.61; P<.001) with uncomplicated COVID-19.

CONCLUSIONS

Our LC phenotype enables the identification of individuals with the condition in routine data sets, facilitating their comparison with unaffected people through retrospective research. This phenotype and study protocol to explore its face validity contributes to a better understanding of LC.

Effectiveness of the Use of Standardized Vocabularies on Epilepsy Patient Cohort Generation

Objectives

This study investigated the effectiveness of using standardized vocabularies to generate epilepsy patient cohorts with local medical codes, SNOMED Clinical Terms (SNOMED CT), and International Classification of Diseases tenth revision (ICD-10)/Korean Classification of Diseases-7 (KCD-7).

Methods

We compared the granularity between SNOMED CT and ICD-10 for epilepsy by counting the number of SNOMED CT concepts mapped to one ICD-10 code. Next, we created epilepsy patient cohorts by selecting all patients who had at least one code included in the concept sets defined using each vocabulary. We set patient cohorts generated by local codes as the reference to evaluate the patient cohorts generated using SNOMED CT and ICD-10/KCD-7. We compared the number of patients, the prevalence of epilepsy, and the age distribution between patient cohorts by year.

Results

In terms of the cohort size, the match rate with the reference cohort was approximately 99.2% for SNOMED CT and 94.0% for ICD-10/KDC7. From 2010 to 2019, the mean prevalence of epilepsy defined using the local codes, SNOMED CT, and ICD-10/KCD-7 was 0.889%, 0.891% and 0.923%, respectively. The age distribution of epilepsy patients showed no significant difference between the cohorts defined using local codes or SNOMED CT, but the ICD-9/KCD-7-generated cohort showed a substantial gap in the age distribution of patients with epilepsy compared to the cohort generated using the local codes.

Conclusions

The number and age distribution of patients were substantially different from the reference when we used ICD-10/KCD-7 codes, but not when we used SNOMED CT concepts. Therefore, SNOMED CT is more suitable for representing clinical ideas and conducting clinical studies than ICD-10/KCD-7.

Harmonization of Measurement Codes for Concept-Oriented Lab Data Retrieval

Measurement concepts are essential to observational healthcare research; however, a lack of concept harmonization limits the quality of research that can be done on multisite research networks. We developed five methods that used a combination of automated, semi-automated and manual approaches for generating measurement concept sets. We validated our concept sets by calculating their frequencies in cohorts from the Columbia University Irving Medical Center (CUIMC) database. For heart transplant patients, the preoperative frequencies of basic metabolic panel concept sets, which we generated by a semi-automated approach, were greater than 99%. We also made concept sets for lumbar puncture and coagulation panels, by automated and manual methods respectively.

Mapping Korean National Health Insurance Reimbursement Claim Codes for Therapeutic and Surgical Procedures to SNOMED-CT to Facilitate Data Reuse

South Korea has a public and single-payer system for healthcare services based on fee-for-service payments. The National Health Insurance (NHI) reimbursement claim codes are used by all healthcare providers for reimbursement. This study mapped NHI reimbursement claim codes for therapeutic and surgical procedures to the Systematized Nomenclature of Medicine Clinical Terms (SNOMED-CT) to facilitate semantic interoperability and data reuse for research. The Source codes for mapping were 2,500 reimbursement claim codes for therapeutic and surgical procedures such as surgery, endoscopic procedures, and interventional radiology. The target terminology for mapping was the 'Procedure' hierarchy of the international edition of SNOMED-CT released in July 2019. We translated Korean terms into English, clarified their meaning, extracted characteristics of the source codes, and mapped them to pre-coordinated concepts. If a source concept was not mapped to a pre-coordinated concept, we mapped it to a post-coordinated expression. The mapping results were validated internally using dual independent mapping and group discussion by trained terminologists, and by two physicians with experience of SNOMED-CT mapping. Out of 2,500 source codes, 1,298 (51.9%) codes were mapped to pre-coordinated concepts, and 1,202 (48.1%) codes were mapped to post-coordinated expressions. The mapping of the NHI reimbursement claim codes for therapeutic and surgical procedures to SNOMED-CT is expected to support clinical research by facilitating the utilization of health insurance claim data.

The U.S. National Library of Medicine and Standards for Electronic Health Records: One Thing Led to Another

When Donald A.B. Lindberg M.D. became Director in 1984, the U.S. National Library of Medicine (NLM) was a leader in the development and use of information standards for published literature but had no involvement with standards for clinical data. When Dr. Lindberg retired in 2015, NLM was the Central Coordinating Body for Clinical Terminology Standards within the U.S. Department of Health and Human Services, a major funder of ongoing maintenance and free dissemination of clinical terminology standards required for use in U.S. electronic health records (EHRs), and the provider of many services and tools to support the use of terminology standards in health care, public health, and research. This chapter describes key factors in the transformation of NLM into a significant player in the establishment of U.S. terminology standards for electronic health records.

Strategies for Adopting and Implementing SNOMED CT in Korea

Objectives

The objective of this study was to introduce the Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT), to describe use cases of SNOMED CT with the barriers and facilitators, and finally, to propose strategies for adopting and implementing SNOMED CT in Korea as a member of SNOMED International.

Methods

We reviewed a collection of SNOMED CT documents, such as introductory materials, practical guides, technical specifications, and reference materials provided by SNOMED International and the literature on SNOMED CT published by researchers to identify use cases of SNOMED CT with barriers and facilitators. We also surveyed the attendees of SNOMED CT education and training series offered by the Korea Human Resource Development Institute for Health and Welfare to identify perceived barriers to adopting SNOMED CT in Korea.

Results

We identified the barriers and facilitators to adopt SNOMED CT experienced by international terminology experts and prospective Korean users. They were related to governance and infrastructure, support services for use, education and training programs, use cases, and vendor capability to implement SNOMED CT. Based on these findings, we identified strategies for adopting and implementing SNOMED CT in Korea. They included the establishment of SNOMED CT management infrastructure, the development of SNOMED CT education and training programs for various user groups, the provision of support services for SNOMED CT use, and the development of SNOMED CT use cases.

Conclusions

These strategies for the adoption and implementation of SNOMED CT need to be executed step by step.

Consultation analysis: use of free text versus coded text

General practice in the United Kingdom has been using electronic health records for over two decades, but coding clinical information remains poor. Lack of interest and training are considerable barriers preventing code use levels improvement. Tailored training could be the way forward, to break barriers in the uptake of coding; to do so it is paramount to understand coding use of the particular clinicians, to recognise their needs. It should be possible to easily assess text quantity and quality in medical consultations. A tool to measure these parameters, which could be used to tailor training needs and assess change, is demonstrated. The tool is presented and a preliminary study using a randomised sample of five recent consultations from thirteen different clinicians is used as an example. The tool, based on using a word processor and a spread-sheet, allowed quantitative analysis among clinicians while word clouds permitted a qualitative comparison between coded and free text. The average amount of free text per consultation was 68.2 words, (ranging from 25.4 and 130.2 among clinicians); an average of 6% of the text was coded (ranging from 0 to 13%). Patterns among clinicians could be identified. Using Word cloud, a different text use was demonstrated depending on its purpose. Some free text could be turned into code but nomenclature probably prevented some of the codings, like the expression of time. This proof of concept demonstrated that it is possible to calculate what percentage of consultations are coded and what codes are used. This allowed understanding clinicians’ preferences; training needs and gaps in nomenclature.

Implementation of a Terminology Server with SNOMED CT in Graph Databases

This paper described the implementation of a terminology server performed at a health institution in Uruguay, whose architecture is based on SNOMED CT using graph databases (NoSQL). The aim of this development was to create an intuitive terminological service, making the most of SNOMED CT's ontology, and which can be used from a clinical, statistical, management, decision support and research point of view, among others, with good performance.

Using SNOMED-CT to Help the Transition from Microbiological Data to ICD-10 Sepsis Codes

Assigning ICD-10 code of sepsis in regard of a pathogenic bacterium found in an haemoculture requires knowledge of microbiology because of the difference of granularity. The aim of this paper is to automate this coding thanks to the use of SNOMED-CT. A dichotomous classification of bacteria causing sepsis has been generated in respect of ICD-10. Our algorithm follows this and explores SNOMED-CT to assign the right ICD-10 code of the sepsis. Applied to a list of 164 bacteria, the system has an error rate of 1.22 %.

Building an Experimental German User Interface Terminology Linked to SNOMED CT

We describe the process of creating a User Interface Terminology (UIT) with the goal to generate a maximum of German language interface terms that are mapped to the reference terminology SNOMED CT. The purpose is to offer a high coverage of medical jargon in order to optimise semantic annotations of clinical documents by text mining systems. The first step consisted in the creation of an n-gram table to which words and short phrases from the English SNOMED CT description table were automatically extracted and entered. The second step was to fill up the n-gram table with human and machine translations, manually enriched by POS tags. Top-down and bottom-up methods for manual terminology population were used. Grammar rules were formulated and embedded into a term generator, which then created one-to-many German variants per SNOMED CT description. Currently, the German user interface terminology contains 4,425,948 entries, created out of 111,605 German n-grams, assigned to 95,298 English n-grams. With 341,105 active concepts and 542,462 (non FSN) descriptions, it corresponds to an average of 13 interface terms per concept and 8.2 per description. An analysis of the current quality of this resource by blinded human assessment terminology states equivalence regarding term understandability compared to a fully automated Web-based translator, which, however does not yield any synonyms, so that there are good reasons to further develop this semi-automated terminology engineering method and recommend it for other language pairs.

SNOMED CT Coding and Analytics of in vitro Diagnostics Observations

This work investigates the capability of SNOMED CT to encode microbiology laboratory data with the goal of fully describing multidrug resistance and breakpoint assignment by specimen.

Mapping Korean EDI Medical Procedure Code to SNOMED CT

The Electronic Data Interchange (EDI) medical procedure code is the code used for health insurance claims in Korea. We mapped Korean EDI codes to SNOMED CT to explore the global interoperability of health insurance claims data. We developed rules for mapping based on the mapping guideline provided by SNOMED CT International. The first and second authors mapped 726 EDI codes used to claim reimbursement in five specialty areas to SNOMED CT. Eight subject matter experts reviewed the mapping results. Out of 726 procedure codes, 82.5% were exactly or partially mapped to SNOMED CT. An EDI code was mapped to an average of 2.04 SNOMED CT concepts. Twenty-one attributes were identified in the EDI codes mapped to SNOMED CT concepts. We identified strategies to improve the EDI code in this study. They include introducing hierarchical structures, adding inclusion and exclusion criteria for procedure codes, and improving EDI code labels.

Expanding Evolutionary Terminology Auditing with Historic Formal and Linguistic Intensions: A Case Study in SNOMED CT

A method is described to use SNOMED CT's history mechanism as a means to compute how the formal and linguistic intensions of its concepts change over versions. As a result of this, it is demonstrated that the intended principle of concept permanence is not always adhered to. It is shown that the evolution of formal intensions can be monitored fully automatically and that the proposed procedure includes a method to suggest missing subsumers in a concept's transitive closure set by identifying mistakes that have been made in the past. Changes in linguistic intensions were found to be much more labor-intensive to identify. It is suggested that this could be improved if the history mechanism would come with more detailed motivations for change than the current and insufficiently used annotation to the effect that a fully specified name 'fails to comply with the current editorial guidance'.

SNOMEDtxt: Natural Language Generation from SNOMED Ontology

SNOMED Clinical Terms (SNOMED CT) defines over 70,000 diseases, including many rare ones. Meanwhile, descriptions of rare conditions are missing from online educational resources. SNOMEDtxt converts ontological concept definitions and relations contained in SNOMED CT into narrative disease descriptions using Natural Language Generation techniques. Generated text is evaluated using both computational methods and clinician and lay user feedback. User evaluations indicate that lay people prefer generated text to the original SNOMED content, find it more informative, and understand it significantly better. This method promises to improve access to clinical knowledge for patients and the medical community and to assist in ontology auditing through natural language descriptions.

The emergence of modern muscle names: the contribution to the foundation of systematic terminology of Vesalius, Sylvius, and Bauhin

Although its history is complicated, today's anatomical nomenclature, including muscle terminology, has acquired a system of naming using epithets. The objective of this literary research paper was to ascertain the founder of modern muscle terminology. The texts of four anatomists, Galen, Andreas Vesalius, Jacobus Sylvius, and Gaspard Bauhin, who have all been identified as being influential in the establishment of early modern anatomy and its nomenclature, were analyzed. Particular emphasis was given to the naming method, and to the consistency of that method. The analysis shows that each of these four anatomists had a different conception of muscle naming, and that three early modern anatomists, Vesalius, Sylvius, and Bauhin, contributed to the development of modern muscle terminology. This investigation revealed the types of contributions they made: Vesalius was an originator of rule-governed muscle terminology with a univocal naming method, Sylvius was an inventor of epithet naming, and Bauhin applied Sylvius's epithet naming method to Vesalius's concept of rule-governed terminology with a univocal naming method.

Pheochromocytoma in Denmark during 1977-2016: validating diagnosis codes and creating a national cohort using patterns of health registrations

Background

Pheochromocytoma and catecholamine-secreting paraganglioma (PPGL) are rare but potentially life-threatening tumors. We aimed to validate diagnosis codes for PPGL in the Danish National Patient Registry, the Danish National Pathology Registry, and the Danish Registry of Causes of Death and to create a national cohort of incident PPGL patients by linking these three registries.

Patients and methods

We obtained data from the three abovementioned registries for all individuals registered with pheochromocytoma or catecholamine hypersecretion in Denmark during 1977–2016 (average population 5.30 million). We then reviewed health records for all individuals living in the North Denmark Region and Central Denmark Region (average population 1.75 million) to validate the diagnosis of PPGL. We tested a number of algorithms for accurately identifying true cases of PPGL to maximize positive predictive values (PPVs) and completeness. The best algorithm was subsequently validated in an external sample.

Results

We identified 2626 individuals with a PPGL diagnosis code in Denmark, including 787 (30.0%) in the North Denmark Region and Central Denmark Region. In this subsample, we retrieved the health records of 771/787 (98.0%) individuals and confirmed 198 incident PPGL patients (25.3%). The PPV of PPGL diagnosis codes was 21.7% in the Danish National Patient Registry, 50.0% in the Danish Registry of Causes of Death, and 79.5% in the Danish National Pathology Registry. By combining patterns of registrations in the three registries, we could increase the PPV to 93.1% (95% confidence interval [CI]: 88.5–96.3) and completeness to 88.9% (95% CI: 83.7–92.9), thus creating a national PPGL cohort of 588 patients. PPV for the optimal algorithm was 95.3% (95% CI: 88.5–98.7) in the external validation sample.

Conclusion

Diagnosis codes for pheochromocytoma had low PPV in several individual health registries. However, with a combination of registries we were able to identify a near-complete national cohort of PPGL patients in Denmark, as a valuable source for epidemiological research.

Evaluation of SNOMED CT Content Coverage: A Systematic Literature Review

BACKGROUND

One of the most important features studied for adoption of terminologies is content coverage. The content coverage of SNOMED CT as a large scale terminology system has been evaluated in different domains by various methods.

OBJECTIVES

This study provided an overview of studies evaluating SNOMED CT content coverage.

METHODS

This systematic literature review covered Scopus, Embase, PubMed and Web of Science. It included studies in English language with accessible full-text from the beginning of 2002 to November 2017.

RESULTS

Reviewing 62 studies revealed that 76 percent of studies were carried out in the US and other countries started to study in this regard from 2007. Most of the studies focused on the comparison of SNOMED CT with disease classifications in the domain of "diagnosis and problem list".

CONCLUSION

Studying the trend of studies in different countries shows that SNOMED CT content coverage is not limited to the early stages of SNOMED CT adoption. However, evaluation methods are likely different due to the stage of SNOMED CT implementation. Therefore, it is recommended to identify and compare evaluation methods of SNOMED CT content coverage in future studies.

Attitudes toward and beliefs about the use of a dental diagnostic terminology: A survey of dental care providers in a dental practice

BACKGROUND

Attitudes and views are critical to the adoption of innovation. Although there have been broadening calls for a standardized dental diagnostic terminology, little is known about the views of private practice dental team members regarding the adoption of such a terminology.

METHODS

The authors developed a survey by using validated questions identified through literature review. Domain experts' input allowed for further modifications. The authors administered the final survey electronically to 814 team members at a multioffice practice based in the US Pacific Northwest.

RESULTS

Response proportion was 92%. The survey had excellent reliability (Cronbach α coefficient = 0.87). Results suggested that participants showed, in general, positive attitudes and beliefs about using a standardized diagnostic terminology in their practices. Additional written comments by participants highlighted the potential for improved communication with use of the terminology.

CONCLUSIONS

Dental care providers and staff in 1 multioffice practice showed positive attitudes about the use of a diagnostic terminology; specifically, they believed it would improve communication between the dentist and patient, as well as among providers, while expressing some concerns about whether using standardized dental diagnostic terms helps clinicians to deliver better dental care.

PRACTICAL IMPLICATIONS

As the dental profession is advancing toward the use of standardized diagnostic terminology, successful implementation will require that dental team leaders prepare their teams by gauging their attitude about the use of such a terminology.

Challenges in Archetypes Terminology Binding Using SNOMED-CT Compositional Grammar: The Norwegian Patient Summary Case

In order to cover the requirements for interoperability in the Norwegian context, we studied the terminology binding of archetypes to terminology expressions created with the SNOMED-CT compositional grammar. As a result we identified important challenges categorized as technical, expressivity, human, and models mismatch.

Interoperability of Disease Concepts in Clinical and Research Ontologies: Contrasting Coverage and Structure in the Disease Ontology and SNOMED CT

OBJECTIVES

To contrast the coverage of diseases between the Disease Ontology (DO) and SNOMED CT, and to compare the hierarchical structure of the two ontologies.

METHODS

We establish a reference list of mappings. We characterize unmapped concepts in DO semantically and structurally. Finally, we compare the hierarchical structure between the two ontologies.

RESULTS

Overall, 4478 (65%) the 6931 DO concepts are mapped to SNOMED CT. The cancer and neoplasm subtrees of DO account for many of the unmapped concepts. The most frequent differentiae in unmapped concepts include morphology (for cancers and neoplasms), specific subtypes (for rare genetic disorders), and anatomical subtypes. Unmapped concepts usually form subtrees, and less often correspond to isolated leaves or intermediary concepts.

CONCLUSION

This detailed analysis of the gaps in coverage and structural differences between DO and SNOMED CT contributes to the interoperability between these two ontologies and will guide further validation of the mapping.

LOINC in prehospital emergency medicine in Germany - experience of the `DIRK´-project

BACKGROUND

Treatment of patients picked up by emergency services can be improved by data transfer ahead of arrival. Care given to emergency patients can be assessed and improved through data analysis. Both goals require electronic data transfer from the emergency medical services (EMS) to the hospital information system. Therefore a generic semantic standard is needed.

OBJECTIVES

Objective of this paper is to test the suitability of the international nomenclature Logical Observation Identifiers Names and Codes (LOINC) to encode the core data-sets for rescue service protocols (MIND 2 and MIND 3). Encoding diagnosis and medication categories using ICD-10 and ATC were also assessed.

METHODS

Protocols were broken down into concepts, assigned to categories, translated and manually mapped to LOINC codes. Each protocol was independently encoded by two healthcare professionals and in case of discrepancies a third expert was consulted to reach a consensus.

RESULTS

Currently 39% of parameters could be mapped to LOINC. Additional use of other coding systems such as International Statistical Classification of Diseases and Related Health Problems (ICD-10) for diagnoses and Anatomical Therapeutic Chemical Classification System (ATC) for medications increases the rate of 'mappable' parameters to 56%.

CONCLUSIONS

Although the coverage is low, mapping has shown that LOINC is suitable to encode concepts of the rescue services. In order to create a generic semantic model to be applied in the field our next step is to request new LOINC codes for the missing concepts.

An ontology-based similarity measure for biomedical data-application to radiology reports

BACKGROUND

Determining similarity between two individual concepts or two sets of concepts extracted from a free text document is important for various aspects of biomedicine, for instance, to find prior clinical reports for a patient that are relevant to the current clinical context. Using simple concept matching techniques, such as lexicon based comparisons, is typically not sufficient to determine an accurate measure of similarity.

METHODS

In this study, we tested an enhancement to the standard document vector cosine similarity model in which ontological parent-child (is-a) relationships are exploited. For a given concept, we define a semantic vector consisting of all parent concepts and their corresponding weights as determined by the shortest distance between the concept and parent after accounting for all possible paths. Similarity between the two concepts is then determined by taking the cosine angle between the two corresponding vectors. To test the improvement over the non-semantic document vector cosine similarity model, we measured the similarity between groups of reports arising from similar clinical contexts, including anatomy and imaging procedure. We further applied the similarity metrics within a k-nearest-neighbor (k-NN) algorithm to classify reports based on their anatomical and procedure based groups. 2150 production CT radiology reports (952 abdomen reports and 1128 neuro reports) were used in testing with SNOMED CT, restricted to Body structure, Clinical finding and Procedure branches, as the reference ontology.

RESULTS

The semantic algorithm preferentially increased the intra-class similarity over the inter-class similarity, with a 0.07 and 0.08 mean increase in the neuro-neuro and abdomen-abdomen pairs versus a 0.04 mean increase in the neuro-abdomen pairs. Using leave-one-out cross-validation in which each document was iteratively used as a test sample while excluding it from the training data, the k-NN based classification accuracy was shown in all cases to be consistently higher with the semantics based measure compared with the non-semantic case. Moreover, the accuracy remained steady even as k value was increased - for the two anatomy related classes accuracy for k=41 was 93.1% with semantics compared to 86.7% without semantics. Similarly, for the eight imaging procedures related classes, accuracy (for k=41) with semantics was 63.8% compared to 60.2% without semantics. At the same k, accuracy improved significantly to 82.8% and 77.4% respectively when procedures were logically grouped together into four classes (such as ignoring contrast information in the imaging procedure description). Similar results were seen at other k-values.

CONCLUSIONS

The addition of semantic context into the document vector space model improves the ability of the cosine similarity to differentiate between radiology reports of different anatomical and image procedure-based classes. This effect can be leveraged for document classification tasks, which suggests its potential applicability for biomedical information retrieval.

Comparison of Knowledge Levels Required for SNOMED CT Coding of Diagnosis and Operation Names in Clinical Records

Objectives

Coding Systematized Nomenclature of Medicine, Clinical Terms (SNOMED CT) with complex and polysemy clinical terms may ask coder to have a high level of knowledge of clinical domains, but with simpler clinical terms, coding may require only simpler knowledge. However, there are few studies quantitatively showing the relation between domain knowledge and coding ability. So, we tried to show the relationship between those two areas.

Methods

We extracted diagnosis and operation names from electronic medical records of a university hospital for 500 ophthalmology and 500 neurosurgery patients. The coding process involved one ophthalmologist, one neurosurgeon, and one medical record technician who had no experience of SNOMED coding, without limitation to accessing of data for coding. The coding results and domain knowledge were compared.

Results

705 and 576 diagnoses, and 500 and 629 operation names from ophthalmology and neurosurgery, were enrolled, respectively. The physicians showed higher performance in coding than in MRT for all domains; all specialist physicians showed the highest performance in domains of their own departments. All three coders showed statistically better coding rates in diagnosis than in operation names (p < 0.001).

Conclusions

Performance of SNOMED coding with clinical terms is strongly related to the knowledge level of the domain and the complexity of the clinical terms. Physicians who generate clinical data can be the best potential candidates as excellent coders from the aspect of coding performance.

Exploring the Possibility of Information Sharing between the Medical and Nursing Domains by Mapping Medical Records to SNOMED CT and ICNP

Objectives

The purpose of this study is to explore possibility of information sharing between the medical and nursing domains.

Methods

Narrative medical records of 281 hospitalization days of 36 gastrectomy patients were decomposed into single-meaning statements. These single-meaning statements were combined into unique statements by removing semantically redundant statements. Concepts from the statements describing patients' problem and medical procedures were mapped to Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT) and International Classification for Nursing Practice (ICNP) concepts.

Results

A total 4,717 single-meaning statements were collected and these single-meaning statements were combined into 858 unique statements. Out of 677 unique statements describing patients' problems and medical procedures, about 85.5% statements were fully mapped to SNOMED CT. The remaining statements were partially mapped. In the mapping to the ICNP concepts, 17.4% of unique statements were fully mapped, 62.8% were partially mapped, and 19.8% were not mapped. About 32.3% of 705 concepts extracted from the statements were mapped to both SNOMED CT and ICNP concepts.

Conclusions

These mapping results suggest that physicians' narrative medical records can be structured and can be used for electronic medical record system, and also it is possible for medicine and nursing to share patient care information.

Proposed Algorithm with Standard Terminologies (SNOMED and CPT) for Automated Generation of Medical Bills for Laboratory Tests

OBJECTIVES

In this study, we proposed an algorithm for mapping standard terminologies for the automated generation of medical bills. As the Korean and American structures of health insurance claim codes for laboratory tests are similar, we used Current Procedural Terminology (CPT) instead of the Korean health insurance code set due to the advantages of mapping in the English language.

METHODS

1,149 CPT codes for laboratory tests were chosen for study. Each CPT code was divided into two parts, a Logical Observation Identifi ers Names and Codes (LOINC) matched part (matching part) and an unmatched part (unmatched part). The matching parts were assigned to LOINC axes. An ontology set was designed to express the unmatched parts, and a mapping strategy with Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT) was also proposed. Through the proceeding analysis, an algorithm for mapping CPT with SNOMED CT arranged by LOINC was developed.

RESULTS

75% of the 1,149 CPT codes could be assigned to LOINC codes. Two hundred and twenty-five CPT codes had only one component part of LOINC, whereas others had more than two parts of LOINC. The system of LOINC axes was found in 309 CPT codes, scale 555, property 9, method 42, and time aspect 4. From the unmatched parts, three classes, 'types', 'objects', and 'subjects', were determined. By determining the relationship between the classes with several properties, all unmatched parts could be described. Since the 'subject to' class was strongly connected to the six axes of LOINC, links between the matching parts and unmatched parts were made.

CONCLUSIONS

The proposed method may be useful for translating CPT into concept-oriented terminology, facilitating the automated generation of medical bills, and could be adapted for the Korean health insurance claim code set.