SNOMED Research Articles

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.

Map-Assisted Generation of Procedure and Intervention Encoding (Magpie): An Innovative Approach for ICD-10-PCS Coding.

ICD-10-PCS coding is challenging because of the large number of codes, non-intuitive terms and paucity of the ICD-10-PCS index. We previously repurposed the richer ICD-9-CM procedure index for ICD-10-PCS coding. We have developed the MAGPIE tool based on the repurposed ICD-9-CM index with other lexical and mapping resources. MAGPIE helps the user to identify SNOMED CT and ICD-10-PCS codes for medical procedures. MAGPIE uses three innovative search approaches: cascading search (SNOMED CT to ICD-9-CM to ICD-10-PCS), hybrid lexical and map-assisted matching, and semantic filtering of ICD-10-PCS codes. Our evaluation showed that MAGPIE found the correct SNOMED CT code and ICD-10-PCS table in 70% and 85% of cases respectively, without any user intervention. MAGPIE is available online from the NLM website: magpie.nlm.nih.gov.

Aggregation and Visualization of Laboratory Data by Using Ontological Tools Based on LOINC and SNOMED CT.

With the proliferation of digital communication in healthcare, the reuse of laboratory test data entails valuable insights into clinical and scientific issues, basically enabled by semantic standardization using the LOINC coding system. In order to extend the currently limited potential for analysis, which is mainly caused by structural peculiarities of LOINC, an algorithmic transformation of relevant content into an OWL ontology was performed, which includes LOINC Terms, Parts and Hierarchies. For extending analysis capabilities, the comprehensive SNOMED CT ontology is added by transferring its contents and the recently published LOINC-related mapping data into OWL ontologies. These formalizations offer rich, computer-processable content and allow to infer additional structures and relationships, especially when used together. Consequently, various reutilizations are facilitated; an application demonstrating the dynamic visualization of fractional hierarchy structures for user-supplied laboratory data was already implemented. By providing element-wise aggregation via superclasses, an adaptable, graph representation is obtained for studying categorizations.

Comparing Two Standardized Value Sets of Infectious Agents: Implications for Semantic Interoperability.

Infections are a global public health concern. For coordinated actions against infectious diseases, semantic interoperability between infection control systems is crucial. This requires a consistent use of standard terminologies such as SNOMED CT. Here, we compare two value sets of infectious agents annotated with SNOMED CT (WHONET 2018 vs. pathogens reported under the German Protection against Infection Act). Our comparison revealed several inconsistencies, highlighting the importance of the consistent and coordinated use of standard terminologies.

Validating Auto-Suggested Changes for SNOMED CT in Non-Lattice Subgraphs Using Relational Machine Learning.

An attractive feature of non-lattice-based ontology auditing methods is its ability to not only identify potential quality issues, but also automatically generate the corresponding fixes. However, exhaustive manual evaluation of the validity of suggested changes remains a challenge shared with virtually all auditing methods. To address this challenge, we explore machine learning techniques as an aid to systematically evaluate the strength of auto-suggested relational changes in the context of existing knowledge embedded in an ontology. We introduce a hybrid convolutional neural network and multilayer perception (CNN-MLP) classifier using a combination of graph, concept features and concept embeddings. We use lattice subgraphs to generate a curated, loosely-coupled training set of positive and negative instances to train the classifier. Our result shows that machine learning techniques have the potential to alleviate the manual effort required for validating and confirming changes generated by non-lattice-based auditing methods for SNOMED CT.

Evaluating the Performance of a Terminology Search Engine Using Historical Data.

Clinical terms are noisy descriptions typed by healthcare professionals in Spanish language in the electronic health record system (EHR). Thus, an evaluation of terminology search engine that extends SNOMED CT and an approach that uses historical data of clinical terms is described. We show how to measure precision and recall using historical search data, and we show how the performance of the search engine can be improved significantly using the technology available in the search engine.

Role of Nursing Informatics in Implementation of SNOMED-CT in India.

SNOMED-CT project under the Ministry of Health and Family Welfare is operational at AIIMS since July 2016. A team of nurses were recruited under SNOMED project who actively works for integrating existing EHR with SNOMED-CT, monitoring, training of users auditing the data, resets creation and development of National Drug Database. This paper emphasizes role of Nursing Informatics in implementation of SNOMED-CT project in India as well as in any other country.

Rapid-Cycle Implementation of a Multi-Organization Registry for Heart Failure with Preserved Ejection Fraction Using Health Information Exchange Standards.

Constructing multi-site specialty registries typically proves time-consuming. Electronic health record (EHR) data collected during clinical care affords a pragmatic approach to accelerating registry implementation. Heart failure with preserved ejection fraction (HFpEF) is an increasingly common and morbid condition. Building a multi-site registry for HFpEF proved feasible using EHR data coded in standard terminologies (SNOMED CT, LOINC) and shared via Health Information Exchanges.

Can Wikipedia Be Used to Derive an Open Clinical Terminology?

Clinical terminologies play an essential role in enabling semantic interoperability between medical records. However, existing terminologies have several issues that impact data quality, such as content gaps and slow updates. In this study we explore the suitability of existing, community-driven resources, specifically Wikipedia, as a potential source to bootstrap an open clinical terminology, in terms of content coverage. In order to establish the extent of the coverage, a team of expert clinical terminologists manually mapped a clinically-relevant subset of SNOMED CT to Wikipedia articles. The results show that approximately 80% of the concepts are covered by Wikipedia. Most concepts that do not have a direct match in Wikipedia are composable from multiple articles. These findings are encouraging and suggest that it should be possible to bootstrap an open clinical terminology from Wikipedia.

Validity of first-time diagnoses of congenital epidermolysis bullosa in the Danish National Patient Registry and the Danish Pathology Registry.

PurposeCongenital epidermolysis bullosa (CEB) is a group of rare monogenic genodermatoses. Phenotypically, the diseases vary in both severity and dissemination, which complicates studies of their epidemiology. To investigate the potential of using the Danish National Patient Registry (DNPR) for epidemiological research on CEB, we examined the positive predictive value (PPV) of a first-time diagnosis of CEB.MethodsWe identified patients with a record of CEB in DNPR and the Danish Pathology Registry (DPR) during January 1, 1977, until December 31, 2015. We restricted diagnoses from two dermatological departments and one regional hospital. Diagnoses in the DNPR are coded by the eighth and tenth revisions of the ICD (ICD-8 and ICD-10) and in the DPR by the Systematized Nomenclature of Medicine (SNOMED). We used clinical description in medical records, family history, histological findings, and molecular genetic investigations to validate diagnoses and classified them as rejected and confirmed. We estimated PPVs for any diagnosis, according to coding systems used, and for additional subdivisions of ICD-10 codes.ResultsWe identified 116 cases from the hospital departments investigated and evaluated 96 medical records for validity. The overall PPV for probable CEB was 62.5% (95% CI: 52.5–71.5). For ICD-8, ICD-10, and SNOMED codes, the PPVs were 30.8% (95% CI: 11.4–57.7), 76.7% (95% CI: 65.8–84.9), and 0.0% (95% CI: 0.0–21.7), respectively. For the ICD-10 codes, we found the highest PPVs for diagnoses arising from the dermatological departments. For subdivisions of ICD-10 codes, PPVs were high for epidermolysis bullosa simplex and dystrophica.ConclusionThe PPVs for first-time diagnoses of CEB registered in the two Danish nationwide registries investigated, DNPR and DPR, ranged from low to average. We therefore recommend that these data be used with caution and restricted to ICD-10 diagnoses from specialized dermatological departments.

Development of Clinical Terminology System for Human Body : Convergence Research of SNOMED CT

Development of Clinical Terminology System for Human Body : Convergence Research of SNOMED CT

Risk of Venous Thrombosis in Patients with Sickle Cell Trait after Orthopedic Surgery

IntroductionSickle cell trait (SCT) is considered to confer a hypercoagulable state. Historically, venous thromboembolism (VTE, deep vein thrombosis and PE) rates for untreated patients after major orthopedic surgery (hip or knee replacement or hip fracture surgery) have been close to 4.3%, however with the introduction of post-op anticoagulation, the rates have been as low as 1.15%. Although guidelines exist regarding anticoagulation for up to 35 days after major orthopedic surgery, there are no specific recommendations for patients with SCT. The purpose of this retrospective study is to examine the rates of VTE after major orthopedic surgery in a cohort of patients with SCT. We hypothesize that rates of VTE would be higher in patients with SCT and the risk of VTE would persist beyond 35 days.MethodsA commercial database (Explorys Inc, Cleveland, OH, USA), an aggregate of electronic health record data from 26 major integrated US healthcare systems representing a sixth of the US population, was queried for data, using Systematized Nomenclature of Medicine (SNOMED) clinical terms or codes. Cases were defined as patients with SCT who underwent major knee or hip surgery. Since a majority of the US population with SCT are African American (AA) patients, controls were defined as AA patients without SCT undergoing major orthopedic surgery. For the primary end point of VTE, only adult patients (≥18 years) were selected. Those with previous history of VTE, thrombophilia, malignant disease, antiphospholipid antibody syndrome and other hemoglobinopathies such as sickle cell disease were excluded. 30 and 90-day rates of VTE were recorded for both groups. Logistic regression models were used to adjust of confounding variables (defined a priori as age > 65 or< 65, smoking, gender and presence or absence of body mass index > 30). Of note, SCT is likely under-estimated due to incomplete diagnosis. Rates or proportions were compared using Chi-squared testusing Medcalc software (2018). Logistic regression analysis was done using Statistical Package for Social Sciences (SPSS, version 21, IBM Corp, Armonk, NY). P< 0.05 was considered statistically significant.ResultsA total of 1360 major orthopedic surgeries in patients with SCT and 74040 surgeries in non-SCT patients were identified. 30 and 90-day VTE for SCT patients undergoing major orthopedic surgery was 9.7% each. 30 and 90 day VTE for non-SCT patients undergoing major hip and knee surgery were 5.9 % and 6.4 % respectively. The difference in 30-day and 90-day VTE rates between the SCT and non-SCT group was statistically significant (30 day VTE difference 3.1%, 95% CI 1.6650-4.7569, p < 0.001; 90 day VTE difference=3.6%; 95% CI 2.1658-5.2562, p= <0.001). The rates of anticoagulant dispensation (oral Xa inhibitors, enoxaparin or warfarin) after surgery were 56% and 46% in SCT and non-SCT group respectively (difference = 10%, 95% CI 7.32-12.64, p <0.001). Despite the higher proportion of patients prescribed for anticoagulants in the SCT population, there was still a higher 30 and 90-day VTE rate in that group. Compliance to anticoagulation and mortality from VTE could not be assessed in this study.Logistic regression of risk factors associated with risk of VTE revealed age over 65 years of age, female gender, active smoking status, obesity (BMI >30), and presence of sickle cell trait were all significantly associated with increased risk of both 30 and 90 day VTE post major orthopedic surgery. Please see Table 1 and 2 for further details.ConclusionOur study represents real life data outside of a clinical trial. We found that patients with SCT who underwent major hip and knee surgery had an increased 30 and 90-day VTE rates compared to non-SCT patients undergoing the same procedures. Overall, this cohort of AA patients had VTE rates higher than that were described in literature. Of note, AA patients overall are at a higher risk of VTE than are their Caucasian counterparts. The results from the study seem to suggest a role for extended prophylaxis in people with SCT who are undergoing orthopedic procedures, and warrants further study. [Display omitted] DisclosuresLittle:Doris Duke Charitable Foundations: Research Funding; NHLBI: Research Funding; PCORI: Research Funding; Hemex: Patents & Royalties: Patent, no honoraria.

Safety of Apixaban in Patients with CKD Stage V and ESRD with Venous Thromboembolism

IntroductionHistorically, warfarin (Coumadin) was the only choice of anticoagulation in patients with chronic renal disease (CKD) stage V and end stage renal disease on dialysis (ESRD). Although apixaban (Eliquis) is approved for use in patients with CKD for venous thrombosis (VTE), its role in patients with CKD 5 and on dialysis is less defined. The purpose of this retrospective study is to compare the rates of bleeding, particularly gastrointestinal (GI) and intracranial bleeding in patients with CKD stage V or dialysis who are anticoagulated with warfarin or apixaban for indication of VTE.MethodsData for this study was queried from a commercial database (Explorys Inc, Cleveland, OH, USA), an aggregate of electronic health record data from 26 major integrated US healthcare systems representing a sixth of the US population. Cases and controls were identified using Systematized Nomenclature of Medicine (SNOMED) clinical terms or codes. Cases were defined as patients with CKD stage 5 or ESRD and were on apixaban for VTE. Controls were defined as those with CKD stage 5 or ESRD and were on warfarin for VTE. For the primary end point of bleeding, only patients above the age of 15 were selected. Those with previous history of GI or intracranial bleeding, atrial fibrillation, traumatic injury, cancer and cirrhosis were excluded. In the coumadin group, in addition to above, patients who were on for indication of artificial valves, arterial thrombosis and PVD were also excluded. 30 and 90-day rates of GI and intracranial bleeding were recorded for both groups. Logistic regression models were used to adjust of confounding variables (defined a priori as age > 65 or< 65, use of antiplatelet agents, gender, race and presence or absence of hypertension greater than 160/100 mm of hg). Rates or proportions were compared using Chi-squared test using Medcalc software (2018).Logistic regression analysis was done using Statistical Package for Social Sciences (SPSS, version 21, IBM Corp, Armonk, NY). P< 0.05 was considered statistically significant.ResultsA total of 990 cases and 8110 controls were identified. 30 and 90-day pooled GI and intracranial bleeding rates for apixaban users were 40% and 35% respectively. 30 and 90 day pooled GI and intracranial bleeding rates for warfarin users were 58% and 54% respectively. Difference in 30 day bleeding between cases and controls was 18% (95% CI 14.7-21.2; p< 0.0001) and 90 day bleeding was 19% (95% CI 15.7-22.1; p <0.001)Due to the limitations of the study, both major and minor GI bleeds were included. Given this study was retrospective, efficacy data in preventing recurrent VTE could not be assessed.Logistic regression of risk factors associated with risk of bleeding revealed that age over 65 years, presence hypertension> 160/100 mm of Hg, use of antiplatelet agents, African american race, and use of warfarin as opposed to apixaban were all significantly associated with increased risk of both 30 and 90 day bleeding. Please see Table 1 and 2 for further details.DiscussionOur study has shown that pooled rates of GI and intracranial bleeds were lower in the apixaban users even when used in patients with CKD stage V or ESRD patients, showing that apixaban is safe even when used for the indication of VTE prevention in this cohort. The results of this study are in concert with a recently published meta-analysis, which showed similar lower risks of bleeding when used in the advanced CKD cohort when used for the indication of atrial fibrillation. However, more studies are needed to prove its efficacy in this cohort. [Display omitted] DisclosuresNo relevant conflicts of interest to declare.

How well can familial hypercholesterolemia be identified in an electronic health record database?

BackgroundFamilial hypercholesterolemia (FH) is a condition characterized by high cholesterol levels and increased risk for coronary heart disease (CHD) that often goes undiagnosed. The Dutch Lipid Network Criteria (DLNC) are used to identify FH in clinical settings via physical examination, personal and family history of CHD, in addition to the presence of deleterious mutations of the LDLR, ApoB, and PCSK9 genes. Agreement between clinical and genetic diagnosis of FH varies. While an ICD diagnosis code was not available for coding FH until 2016, Systematized Nomenclature of Medicine (SNOMED) clinical concept codes, including genetic diagnoses, for FH have been utilized in electronic health records (EHRs).ObjectiveTo evaluate the concordance of identifying FH via SNOMED and ICD-10 CM codes vs the DLNC in an EHR database.MethodsUsing the Practice Fusion EHR database, the sensitivity, specificity, positive predictive value (PPV), and negative predictive value were calculated comparing an FH cohort identified via SNOMED and ICD-10 CM codes to one identified via the DLNC.ResultsAmong 907,616 patients with hypercholesterolemia, 2,180 were identified as FH via SNOMED code (zero were identified via ICD-10 CM), 259 had a DLNC score 6–8 (probable FH), and 45 had a DLNC score >8 (definite FH). Compared to DLNC score >8, the sensitivity, specificity, and PPV of the FH SNOMED code were 84.4%, 99.4%, and 6.4%, respectively. Compared to DLNC score ≥6, the sensitivity was 36.8% and the specificity was 99.5% with a PPV of 18.7%.ConclusionCompared to the clinical criteria for FH, identification of FH patients via SNOMED diagnosis codes had high sensitivity and specificity, but low PPV. The discordance of these two techniques in identifying FH patients speaks to the challenges in identifying FH patients in large electronic databases such as administrative claims and EHR.

Non-Conformity in the Anatomical Description of the Duodenum and the Impact on Clinical Communication

Background: Communication errors are frequently identified sources of sentinel events and improving clinician communication is now a focus of patient safety organizations. Medical schools also have an opportunity and interest in improving clinical communication. One area for potential improvement is the miscommunication that can stem from lack of conformity in terminology student doctors learn from academic anatomists and terminology used by practicing physicians who serve as preceptors during third and fourth clinical years. While transitioning from medical school to a clinical setting, student doctors must relearn some clinical anatomical terms .The lack of uniformity in clinical terminology could lead to a miscommunication in the health care field and become detrimental to patient outcomes. Our goal is to determine and implement a universal standardized terminology for the duodenum. Methods: We used nineteen references from six surgical, seven radiological, and six anatomical texts. We compared the resources’ terminology in describing the parts of the duodenum for similarities and differences. Results: The medical resources used five different anatomical descriptions of the duodenum among the three categories of anatomical texts. Additionally, individual sources within the same discipline of each category used different descriptions of the duodenum. Conclusion: Student doctors and teaching clinicians would benefit fromthe use of easily identifiable landmarks instead of vague, conflicting anatomical terms for the duodenum to help solve communication errors. We recommend adoption of the American Association for the Surgery of Trauma duodenum nomenclature system, which divides the duodenum into four parts utilizing easily identifiable landmarks. Keywords: DUODENUM; INTERDISCIPLINARY COMMUNICATION; HEALTH COMMUNICATION; SYSTEMATIZED NOMENCLATURE OF MEDICINE

An Empirical Approach to Enhancing Terminology Binding - An HL7 FHIR SNOMED CT Example.

Information exchange at the level of semantic interoperability requires that information models and clinical terminologies work well together. In HL7 FHIR resources, terminology binding to standard terminologies such as SNOMED CT are suggested, and even though most are suggestions rather than rules, they still must reflect the clinical domain accurately. In this study, we suggest a method for empirically evaluating whether a terminology binding represents the value sets used in practice. We evaluated the terminology binding associated with the MedicationRequest.reasonCode using the Danish national indication value set which we mapped to SNOMED CT. We found two problems with the terminology binding, namely, that the reason for prophylactic treatment and that medication given as part of a procedure, but not related to the patients' problems per se could not be expressed within the boundary of HL7 FHIR's example terminology binding. Future work will include showing how more complex terminology binding issues could be informed by looking at value sets in use.

Comparison of Three English-to-Dutch Machine Translations of SNOMED CT Procedures.

Dutch interface terminologies are needed to use SNOMED CT in the Netherlands. Machine translation may support in their creation. The aim of our study is to compare different machine translations of procedures in SNOMED CT. Procedures were translated using Google Translate, Matecat, and Thot. Google Translate and Matecat are tools with large but general translation memories. The translation memory of Thot was trained and tuned with various configurations of a Dutch translation of parts of SNOMED CT, a medical dictionary and parts of the UMLS Metathesaurus. The configuration with the highest BLEU score, representing closeness to human translation, was selected. Similarity was determined between Thot translations and those by Google and Matecat. The validity of translations was assessed through random samples. Google and Matecat translated similarly in 85.4% of the cases and generally better than Thot. Whereas the quality of translations was considered acceptable, machine translations alone are yet insufficient.

Interface Terminologies, Reference Terminologies and Aggregation Terminologies: A Strategy for Better Integration.

The time has come to end unproductive competitions among different types of biomedical terminology artefacts. Tools and strategies to create the foundation of a seamless environment covering clinical jargon, clinical terminologies, and classifications are necessary. Whereas language processing relies on human interface terminologies, which represent clinical jargon, their link to reference terminologies such as SNOMED CT is essential to guarantee semantic interoperability. There is also a need for interoperation between reference and aggregation terminologies. Simple mappings between nodes are not enough, because the three kinds of terminology systems represent different things: reference terminologies focus on context-free descriptions of classes of entities of a domain; aggregation terminologies contain rules that enforce the principle of single hierarchies and disjoint classes; interface terminologies represent the language used in a domain. We propose a model that aims at providing a better flow of standardized information, addressing multiple use cases in health care including clinical research, epidemiology, care management, and reimbursement.

Simplifying the use of pharmacogenomics in clinical practice: Building the genomic prescribing system

BackgroundA barrier to the use of genomic information during prescribing is the limited number of software solutions that combine a user-friendly interface with complex medical data. We built and designed an online, secure, electronic custom interface termed the Genomic Prescribing System (GPS). MethodsActionable pharmacogenomic (PGx) information was reviewed, collected, and stored in the back-end of GPS to enable creation of customized drug- and variant-specific clinical decision support (CDS) summaries. The database architecture utilized the star schema to store information. Patient raw genomic data underwent transformation via custom-designed algorithms to enable gene and phenotype-level associations. Multiple external data sets (PubMed, The Systematized Nomenclature of Medicine (SNOMED), National Drug File – Reference Terminology (ND-FRT), and a publically-available PGx knowledgebase) were integrated to facilitate the delivery of patient, drug, disease, and genomic information. Institutional security infrastructure was leveraged to securely store patient genomic and clinical data on a HIPAA-compliant server farm. ResultsAs of May 17, 2016, the GPS back-end housed 257 CDS encompassing 112 genetic variants, 42 genes, and 46 PGx-actionable drugs. The GPS user interface presented patient-specific CDS alongside a recognizable traffic light symbol (green/yellow/red), denoting PGx risk for each genomic result. The number of traffic lights per visit increased with the corresponding increase in the number of available PGx-annotated drugs over time. An integrated drug and disease search functionality, links to primary literature sources, and potential alternative PGx drugs were indicated. The system, which was initially used as stand-alone CDS software within our clinical environment, was then integrated with the institutional electronic medical record for enhanced usability. There have been nearly 2000 logins in 43months since inception, with usage exceeding 56 logins per month and system up-times of 99.99%. For all patient-provider visits encompassing >3years of implementation, unique alert click-through rates corresponded to genomic risk: red lights clicked 100%, yellow lights 79%, green lights 43%. ConclusionsSuccessful deployment of GPS by combining complex data and recognizable iconography led to a tool that enabled point-of-care genomic delivery with high usability. Continued scalability and incorporation of additional clinical elements to be considered alongside PGx information could expand future impact.

Large-scale adverse effects related to treatment evidence standardization (LAERTES): an open scalable system for linking pharmacovigilance evidence sources with clinical data

BackgroundIntegrating multiple sources of pharmacovigilance evidence has the potential to advance the science of safety signal detection and evaluation. In this regard, there is a need for more research on how to integrate multiple disparate evidence sources while making the evidence computable from a knowledge representation perspective (i.e., semantic enrichment). Existing frameworks suggest well-promising outcomes for such integration but employ a rather limited number of sources. In particular, none have been specifically designed to support both regulatory and clinical use cases, nor have any been designed to add new resources and use cases through an open architecture. This paper discusses the architecture and functionality of a system called Large-scale Adverse Effects Related to Treatment Evidence Standardization (LAERTES) that aims to address these shortcomings.ResultsLAERTES provides a standardized, open, and scalable architecture for linking evidence sources relevant to the association of drugs with health outcomes of interest (HOIs). Standard terminologies are used to represent different entities. For example, drugs and HOIs are represented in RxNorm and Systematized Nomenclature of Medicine -- Clinical Terms respectively. At the time of this writing, six evidence sources have been loaded into the LAERTES evidence base and are accessible through prototype evidence exploration user interface and a set of Web application programming interface services. This system operates within a larger software stack provided by the Observational Health Data Sciences and Informatics clinical research framework, including the relational Common Data Model for observational patient data created by the Observational Medical Outcomes Partnership. Elements of the Linked Data paradigm facilitate the systematic and scalable integration of relevant evidence sources.ConclusionsThe prototype LAERTES system provides useful functionality while creating opportunities for further research. Future work will involve improving the method for normalizing drug and HOI concepts across the integrated sources, aggregated evidence at different levels of a hierarchy of HOI concepts, and developing more advanced user interface for drug-HOI investigations.