Supporting Clinical Shared Care in the UK and globally

Size: px

Start display at page:

Download "Supporting Clinical Shared Care in the UK and globally"

Anthony Gibbs
6 years ago
Views:

1 Supporting Clinical Shared Care in the UK and globally Sharing Clinical Documents and Integrating Workflow IHE Conference, 9th April 2008, Oxford Dr Dipak Kalra Centre for Health Informatics and Multiprofessional Education (CHIME) University College London

2 Jennings, Miller, Materna 1997

3 Clinical drivers for the EHR Manage increasingly complex clinical care Connect multiple locations of care delivery Support team-based care Deliver evidence-based health care Improve safety reduce errors and inequalities reduce duplication and delay Improve cost effectiveness of health services Underpin population health and research Empower and involve citizens Protect patient privacy

4 The status quo is not an option The US Institute of Medicine report "To Err is Human" has estimated that 100,000 US citizens die each year through medical errors Medical errors may rank as the eighth leading cause of death in the US, and contribute 4% ($37.6 billion) to the cost of US healthcare Other studies have shown that: 20% of outpatients had missed diagnoses, failures of follow up or missed screening tests 10% of elderly people attending an outpatient clinic had out of date (erroneous) medication records The lack of access to a complete health record for a patient is an important contributor to clinical error

5 High risk areas of communication breakdown Shared care communications survey (Kalra, D. London, 1997) Percent of GPs citing this as a high risk situation

6 The NHS communications challenge 54 million patients 1.2 million staff 8,800 GP practices and 28,000 GPs 302 Primary Care Trusts 273 acute, community, mental health Trusts ~ The third largest employer in the world Per annum 730 million patient interactions 300 million consultations in primary care 13m outpatient consultations 25m home visits Over 5.3m people admitted to hospital 617m prescription items issued ~ 40 million shared care communications

7 Support of unplanned care Telephone Repositories of clinical information Paper health records supporting shared care through: letters, summaries, results, reports Support of continuity of care Post, Fax Organisation-specific systems NHS Care Record Summary Structured medications and allergies Word processed letters and reports Post, Fax Secure NHS Care Record Service Extended summary details Electronic letters, summaries, results, reports XDS with notification Regional sharing Patient sharing XDS pull Detailed Care Records? HealthSpace? Niche shared care web applications XDS with enhanced semantics and access control Full EHR interoperability e.g. EN and openehr archetypes Patient centric EHR systems Full EHR interoperability e.g. EN and openehr archetypes

8 A physically or logically centralised document repository, for an Affinity Domain GP systems GP systems Ambulatory & secondary care systems Whittington Hospital Date: Healthcare Record John Smith DoB: Ambulatory & secondary care systems Tertiary care systems Tertiary care systems

9 Clinical value of IHE XDS provides a means of sharing clinical documents between care teams and sites ensures that a cumulative collection of documents is accessible across a document-sharing domain could be implemented rapidly with todayʼs systems via an export/import interface is readily usable because many useful clinical reference materials already exist in document form letters, discharge summaries, pathology and radiology reports, medical images) Many clinicians would be happy just to be able to access rapidly and read recent reports and images on a patient: e.g. date ordered PDFs

10 XDS semantic tags (metadata) Title classcode (i.e. a coarse grained document type) Type (finer grained) authorspeciality healthcarefacilitytypecode practicesetting eventcodelist (a list of the main Acts in the document) plus date, author, organisation, etc.

11 Semantic limitations of XDS A date ordered list of tagged documents will only be useful some of the time, and will not scale A document centric approach only meets some of the clinical requirements

12 Semantic limitations of XDS A date ordered view will only be useful some of the time clinicians often want to locate specific information rather than browse (except for the most recent contacts) they may wish to check that things are not present i.e. for the absence of specific conditions, findings or interventions and will not scale we do not yet know the kinds of indices and search criteria that will prove the most useful for searching and filtering but the present set of metadata tags are unlikely to suffice for searching across lifetime patient records

13 Example oncology outpatient letters 85 letters, for one patient, over 6 years How will a clinician find the right letter(s) to answer his/her question?

14 Number of clinical documents within the EHR at specialist cancer hospital Average cancer care documents per patient (n=21,000) Cytology reports Histopathology reports Death certificate Primary diagnosis Operation notes Radiotherapy reports Radiology reports Clinic letters Haematology results Chemistry results Drug prescriptions

15 Classifying documents A set of document type codes must be defined precisely and used consistently but will only offer a coarse grained aid to navigation

16 LOINC document type codes (used in HL7 CDA release 2) Anesthesia Hospital General Pre-op Medicine Note Outpatient Outpatient Consultation Confirmatory Note Consultation Note Attending Outpatient History Supervisory and Physical Note Note Outpatient Initial Evaluation Note Autopsy Note Home Health Educational Outpatient Visit Note Note Cardiology Attending Home Outpatient Health Initial Supervisory Evaluation Outpatient Note Note Operative Note Cardiology Consultation Home Health Note Progress Note Outpatient Progress Note Cardiology Hospital Hospital Admission Admission NoteH & Pathology P Note Note Cardiology Outpatient Hospital Progress Consultation Note Note Pharmacist Outpatient Progress Note Case Manager Home Hospital Health Discharge Care Progress Summary Physician Note Note Hospital Discharge Summary Note Clinical Encounter Hospital Note Group Counseling Physician NoteNursing Home H & P Note Comprehensive H&P Hospital NoteProgress NoteProgress Note Conference NoteInpatient Note Psychiatry Hospital Consultation Note Consultation Note Interventional Procedure Pulmonary Note Consultation Note Critical Care Consultation Medical Student Note Hospital Summary H & P Note Critical Care Progress Nurse Note Telephone Encounter Targeted NoteH & P Note Dental Hygienist Nursing Outpatient Home Progress Comprehensive Note Transfer H Referral & P Note Note Dentist Outpatient Nursing Progress Home Note Conference Transfer NoteSummary Note Diabetology Outpatient Nursing Note Home Initial Evaluation Over the Note Counter Drug Label Emergency Department Nursing Note Home Note Prescription Drug Label Gastroenterology Operative Attending Note Outpatient Supervisory Reference Document Note

17 Classifying documents The eventcodelist is not well defined, and it is unclear how it might be populated automated means will remain unreliable for many years the author could populate this list, rather like publication keywords, but will not have time to do it for every document their content needs to be relevant to the kinds of searches that clinicians need to perform, which are not yet well understood

18 Striving towards automated information extraction A hand-crafted example: doing this automatically is very very hard

19 Clinical documents are full of tricky phrases Original clinical phrase no evidence of disseminated disease widespread metastatic disease to bone his chest X ray shows stable disease now Alternative interpretations? Normal No evidence of disease No evidence only of disease that is disseminated Bone disease Bone cancer Metastases located in bone (i.e. not a bone cancer) Chest X ray is normal Chest X ray shows cancer Chest X ray is unchanged from last time

20 Semantic limitations of XDS A document centric approach only meets some of the clinical requirements the longer term requirement is for the sharing of semantically interoperable detailed health records to support the use of care pathways, alerts, decision support, and to underpin education, public health and research

21 Purposes requiring a semantically computable solution Manage increasingly complex clinical care Connect multiple locations of care delivery Support team-based care Deliver evidence-based health care Improve safety reduce errors and inequalities reduce duplication and delay Improve cost effectiveness of health services Underpin population health and research Empower and involve citizens Protect patient privacy

22 Alerting systems have been shown to save lives drug interactions, contraindications and adverse effects reminders for overdue screening and monitoring events alert algorithms for laboratory values pager-based alerting system for physicians looking after intensive care patients in a tertiary care hospital image interpretation of mammograms patient reminders for medication dosing, disease monitoring test results These require a semantically computable solution

23 Clinical data life-cycle Point of care delivery Citizen in the community Teaching Research Clinical trials explicit consent Continuing care (within the institution) Education Research Epidemiology Data mining de-identified +/- consent implied consent Public health Health care management Clinical audit Long-term shared care (regional national, global) Governance requirements Faithfulness Completeness Medico-legal integrity Standards conformance Consistent semantics Privacy management Holistic requirements Empowering and respecting Educating Supporting diverse cultures and professions Capable of evolution

24 The virtual EHR Clinical trials, functional genomics, public health databases EHR repositories Decision support, knowledge management and analysis components Date: Shared logical EHR Whittington Hospital Healthcare Record Personnel registers, security services John Smith Shared domain DoB: knowledge Mobile devices Clinical devices, instruments Clinical applications

25 The role of EHR interoperability standards Clinical trials, functional genomics, public health databases EHR repositories Decision support, knowledge management and analysis components ISO/EN Date: Whittington Hospital Personnel registers, security services Healthcare Record openehr.org John Smith DoB: EHR archetypes Mobile devices Clinical devices, instruments Clinical applications

26 The ISO/EN standard for EHR Communications defines a logical model for the core EHR supporting interoperability between heterogeneous systems providing a common view across message paradigms meets published EHR requirements draws on 15 years of significant R&D including multi-national implementation experience is being translated into an HL7 v3 RMIM includes mappings to the XDS metadata incorporates archetypes for sharing semantic structures offers a framework for sharing disclosure consent and access control information

27 mapping to XDS meta-data

28 mapping to XDS meta-data

The openehr Foundation An international community A set of specifications for: a highly stable reference model for a health computing platform a language and model for

29 The openehr Foundation An international community A set of specifications for: a highly stable reference model for a health computing platform a language and model for expressing archetypes and templates General purpose clinical archetypes Open source software implementations An online archetype library and governance framework

30 What is an openehr/13606 Archetype? A formal sharable model of a clinical domain concept e.g. blood pressure, discharge summary, fundoscopy Allows data quality constraints to be placed on the organisation and content of health record entries Can be published and shared within a clinical community, or globally May be mapped to the specific information in each clinical (EHR) system Defines a systematic EHR target for search queries, and for alerts and decision support

31 openehr / Archetypes: a shared library of clinical data structures

32 openehr / Archetypes: a shared library of clinical data structures

33 NHS investigation of archetypes: findings (published 2007) openehr / EN archetypes provide a useful formalism for reconciling and re-using detailed clinical data specifications across different use cases They are generally suitable for human review and for use in software development easier for humans than reviewing equivalent semantics within an HL7 V3 messaging model Archetypes can provide a useful mechanism for establishing and expressing national guidance for the use of SNOMED CT The NHS should use archetypes for user interface, messaging and dictionary content requirements

34 Over 400 NHS Archetypes: emergency care, maternity, and the beginnings of mental health

35 Policies for EHR access need to specify: Which institutions should normally have access If certain teams or specialities should have privileged access If particular users should be excluded, or given wide access If consent has been granted for teaching, specific research or generic research EHR communication needs to ensure that such consent specifications are represented and communicated in a standardised way

36 : Potential users of EHR data Clinical governance Clinical team Teaching team Healthcare organisation the patient Research team Date: Health service Whittington Hospital Healthcare Record John Smith DoB: Research funder Publisher, the press family & carers Public health Commercial organisation The public Professional organisation

37 : Communicating EHR access consent ACCESS_POLICY policy_id [1]: II author [1]: II date_committed [1]: TS previous_version [0..1]: II effective_start [0..1]: TS effective_end [0..1]: TS policy_attestation EHR_target 1..* ATTESTATION time [1]: TS performer [1]: II proof [0..1]: ED function [0..1]: CV 0..* TARGET access_rules request_characteristics rc_ids [0..1]: SET<II> archetype_ids [0..1]: SET<II> time_period [0..1]: IVL<TS> other_criteria [0..1] SET<String> 1 MAX_SENSITIVITY_CONSTRAINTS access [1]: INT write [0..1]: INT modify [0..1]: INT communicate [0..1]: INT version_history [1]: BL other_constraints [0..1]: SET<String> NOTE: The INT value corresponds to one of the values of CS_SENSITIVITY and matches the values of the sensitivity attribute of RECORD_COMPONENT 0..* NOTE: If no target criteria are specified this policy applies to the whole EHR_EXTRACT REQUEST functional_roles [0..1]: SET<CS_FUNC_ROLE> structural_roles [0..1]: SET<CV> functional_responsibilities [0..1]: SET<CV> clinical_settings [0..1]: SET<CS_SETTING> specialities [0..1]: SET<CV> parties [0..1]: SET<II> other_characteristics [0..1]: SET<String> NOTE: If no requestor characteristics are specified this policy applies to all requests

38 : Who has been looking at my record?

39 Conclusions: IHE XDS Focuses on shared access to whole documents this ensures that the basic interpretation contexts are kept intact meets the pragmatic need expressed by many clinicians: simply to be able to read the notes of colleagues sharing care for the patient on different sites Relies upon meta-data to help navigate to the right sub-set of documents but the creation of RELEVANT meta-data is an unproven challenge this will be increasingly important if XDS is a success! Confidentiality management needs to be demonstrably trustworthy: research still needed

40 Conclusions: the EHR Medical summaries are presently in fashion but the drivers for these are not predominantly clinical A consistently structured and context-safe EHR is now a realistic goal: via openehr and architectures for good EHR systems EHR interoperability access policy interoperability archetypes: systematising clinical data structures help to index and retrieve relevant documents ehealth needs to work towards this goal XDS is a practical starting point, meeting many immediate needs

I n t e ro p e r a b i l i t y o f content and systems. using CEN/tc251 European Standards. Paradigm shifts

I n t e ro p e r a b i l i t y o f content and systems using CEN/tc251 European Standards Paradigm shifts Gerard Freriks v2 Brussel Ride Workshop 8-12-2006 1 Slides from Dipak Kalra, GESI and Francois