The Potential of SNOMED CT to Improve Patient Care stephen.spencer@nhs.net stephen.spencer@nhs.net Dec 2015
What is SNOMED CT SNOMED CT: Is the most comprehensive, multilingual clinical healthcare terminology in the world. Is a resource with comprehensive, scientifically validated clinical content. Enables consistent, process-able representation of clinical content in electronic health records. Is mapped to other international standards. Is already used in more than fifty countries. International Health Terminology Standards Development Organisation (IHTSDO)
A brief history of the origins of SNOMED CT 1965 Systemized Nomenclature of Pathology (SNOP) 1975 Systemized Nomenclature of Medical Terms (SNOMED) 1979 SNOMED II was published 2000 Logic based version developed SNOMED RT 1980 Read terms developed in the UK (James Read) Read terms - Clinical Terms Version 3 (CTV3) 1998 3 yr project to merge SNOMED RT & CTV3 (CAP) 2002 SNOMED CT published 2007 IHTSDO acquired intellectual property rights for SNOMED SNOMED CT is now a trade name not an Acronym
Why is SNOMED CT so important It is necessary for electronic patient records It is intended to be used by clinicians to record clinical information in a consistent way at the patient bedside Any clinical concept can be incorporated It has a logical structure It supports data analysis
National Information Board Prospectus As we move progressively towards real-time digital record-keeping, we will achieve the objective of collecting all the information required to support direct clinical care and the analytic needs of health and care, once and once only at the point of care. Personalised Health and Care 2020: A Framework for Action
National Pilot Project to Develop SNOMED-CT subsets for outpatient specialties HSCIC UK Terminology Centre Clinical Data Standards AoMRC Speciality Groups
Score Lack of clinical validation Inappropriate coding rules e.g. Requirement to code a diagnosis when recorded as a clinical impression. The extraction of data from unstructured notes No differentiation between diagnosis present on admission and those developing in hospital No link between primary and secondary care records No record of outpatient diagnosis Inability to differentiate between active and inactive co-morbidities No record of adverse or critical incidents Inability to attribute care to professionals other than consultants No record of multi-professional team work No record of consultant team work All diagnostic information recorded only using ICD10 No consistent record of outpatient procedures No record of presenting complaint No record of anaesthetist No record of surgical operator All procedure information recorded only using OPCS 0 200 400 600 800 1000 1200 1400 1600 1800 Score
Clinical Terms Project 10 Outpatient Specialities Develop ~ 100 SNOMED terms each Aim to cover ~ 80% Primary diagnoses Mapping to ICD10 Testing Volunteer Trusts Clinicians collecting data in real time Conversion to ICD10 Submission to SUS Back end analysis
Volunteer Specialties General Paediatrics Community Child Health Paediatric Neurodisability Orthopaedics Foot & Ankle Rheumatology Ophthalmology Urology Respiratory Renal Gastro-enterology
The Outcomes Specialities were successful in developing subsets. Professional discussions were necessary to reach a consensus on the use of common terminology (One specialty produced a glossary) Different specialties took different approaches Some changes to SNOMED CT were requested Use of SNOMED CT resulted in tangible benefits especially in paediatric neurodisability
Changes to SNOMED CT 123 New Concepts 11 Concept Changes 10 Description Changes 5 Descriptions Added 6 Other Content Changes Approx 94% of content was available for top 100 When looked at a speciality area in detail, 70-80% of content available (example is rheumatology)
SNOMED CT subsets online http://www.diseasesdatabase.com/snomed/snomed_ subsets.asp
Neuro-disability Subset
Ian Gaywood 1, Ira Pande 1, Edward Cheetham 2 1 Department of Rheumatology, Nottingham University Hospitals NHS Trust, Nottingham 2 Information Standards Delivery, Health and Social Care Information Centre (HSCIC), Leeds Introduction The global move towards electronic health records (EHRs) and the need to make data available for wider purposes including research, registries, case mix analyses, epidemiological and population studies, commissioning and other secondary uses brings with it a requirement to standardize the way we record data in routine clinical practice. Methods Using inflammatory arthritis and connective tissue diseases (IA and CTD) as our prototype domain we (IP and IG) drew up a list of diagnoses which was necessary and sufficient to allow all relevant concepts to be captured. Having agreed this set we examined the available clinical terms to describe these concepts. Currently the most comprehensive systematically organised set of clinical terms is SNOMED CT. SNOMED CT is owned and maintained by the International Health Terminology Standards Organisation (IHTSDO). SNOMED Clinical Terms are arranged in domains including disorders, symptoms and findings and are widely used internationally. For each concept there is a preferred term and, usually, one or more permitted synonyms. Examination of existing SNOMED CT terms allowed us to divide our candidate set of diagnoses into three groups: Group 1: Diagnoses for which there is an existing SNOMED CT term which is correct and familiar to clinicians. Group 2: Diagnoses for which no appropriate SNOMED CT term could be found. New terms were agreed and submitted for SNOMED CT inclusion. Group 3: Diagnoses where an apparently appropriate SNOMED CT term was incorrect by virtue of hierarchical errors or where a group of disorders (e.g. psoriatic arthritis) was incompletely or incorrectly represented. Some members of this group remain under discussion with HSCIC. Throughout this stage, these sets were reviewed iteratively with an expert terminologist, (EC). Following this, the sets were submitted to the BSR Informatics Group for discussion which established a general preference for lumping into broad diagnostic concepts rather than splitting to capture phenotypic detail within a diagnostic term. After extensive discussion, consensus was reached on almost all of the members of group one. New candidate SNOMED CT terms were agreed for group two. Both groups were then submitted to HSCIC for SNOMED CT editorial review. Further consultation will take place on the more contentious group 3. The resulting set of clinical terms will be identified within SNOMED CT by a subset heading. Membership of that subset will grow as new domain sets are added. Eventually, it is anticipated that the SNOMED CT subset for musculoskeletal disorders will provide all necessary concepts and codes for capturing standardised data in an electronically processable form. The next domain (soft tissue disorders) is under construction. Examples: Group 1: Clinical term SNOMED Clinical term (code) Synonyms Rheumatoid arthritis - seropositive Rheumatoid arthritis - seronegative Group 3: Clinical term Rheumatoid arthritis of unspecified serology Psoriatic arthritis Results: This process has generated the first part of a standard set of terms for musculoskeletal disorders, ratified by BSR with full representation in SNOMED CT. Necessary changes to SNOMED CT (modifications to content and the publication of a subset file) are available for review and testing in the UK Edition published in April 2014. Conclusions: Seropositive rheumatoid arthritis (239791005) Seronegative rheumatoid arthritis (239792003) Problem for discussion The identification of a standard set of terms for musculoskeletal disorders is an important step in improving our ability to share data nationally and internationally and a necessary step in the transition to EHRs. None None Group 2: Clinical term New SNOMED Clinical Term (code) Comments Acute pyrophosphate crystal arthropathy Milwaukee shoulder Acute pyrophosphate crystal arthropathy (905311000000101) Apatite-associated destructive arthritis of shoulder (699942000) Further discussion required on synonyms Previous entry for Milwaukee shoulder incorrect. There is a parental SNOMED concept of rheumatoid arthritis but no way of differentiating between all rheumatoid arthritis and rheumatoid arthritis of unspecified serology. Currently has a SNOMED CT preferred term of psoriasis with arthropathy which presupposes the presence of skin disease. Remains under editorial discussion with HSCIC.
Rheumatology Poster Publication The identification of a standard set of terms for musculoskeletal disorders is an important step in improving our ability to share data nationally and internationally and a necessary step in the transition to EHRs.
What is now needed Development of ECRGs to work on maximising the benefits of professional engagement with SNOMED CT Gather evidence from sites and specialties that are using SNOMED CT SNOMED CT to be submitted into the national data collection Analysis of clinical data using SNOMED CT to be available to clinicians for audit, quality improvement etc. Introduction of SNOMED CT training to Undergraduate and Post Graduate Curricula
Work required of ECRG Agree on terminology to be used to describe common conditions ( The biggest issue is imprecise use of clinical language). Develop a Glossary where required to ensure consistent use of terms. Request changes to SNOMED CT content where these are required to reflect these agreements. Assist in the developments of subsets which will facilitate easy access to preferred terms.
UK Terminology Centre forward Plan Focus on migration of whole NHS estate to SNOMED CT in England by 2020 Withdrawal of Read codes from use in Primary Care Dedicated programme run out of GP Systems of Choice (GPSoC) Renewed impetus in supporting SNOMED CT Establishment of dedicated implementation support team
Wales SNOMED? Welsh Government Commitment The CANISC team (secondary and tertiary cancer system) moving from Read to SNOMED Joining the UK Primary Care programme to move to SNOMED CT Existing use with PatientCare system in Specialist Neurology Service across Wales Development progressing in Aneurin Bevan University Health Board toward general use in established EPR
>18,000 patients >120,000 clinical encounters Wide variety of patient outcome measures Ubiquitous use of SNOMED-CT Data entry / pick-lists from subsets and sub-hierarchies of SNOMED- CT (e.g. problem lists, diagnostic, medication DM&D, occupations etc). Simplifies development. Provides semantic understanding: e.g. if a patient has a type of epilepsy (whatever it might be) then additional functionality exposed. Derive SNOMED-CT codes from outcome scales (e.g. score less than 26/30 on MOCA = Impaired cognition [386806002] PatientCare https://patientcare.wales.nhs.uk Highly configurable very fast user interface allowing fixed curated lists of terms or selection of terms from a subset/subsets or hierarchy/hierarchies. Open data interfaces (REST) for interoperability
Conclusion Electronic Patient Records are the Future Consistent language is essential if data is to be used for quality improvement SNOMED CT is a key resource to achieve this Professional involvement at every stage of development is essential