Implementing SNOMEDCT Challenges and Solutions

1
Implementing SNOMED-CTChallenges and Solutions?

• Dr Nick Booth
• Chair, SNOMED International Primary Care Working
  Group

2
Presentation outline

• Acute Sector
• Legacy Systems
• Culture of the various sectors
• Impact of Payment by Results
• Working together in Teams

3
Acute Sector

• PAS type functionality with added clinicals now
  being installed
• SNOMED implementations beginning to appear
• Systems not designed around SNOMED
• Implemented SNOMED functionality is less than
  complete
• Flat files
• No or limited post co-ordination
• Problems with term length
• Need for a wide range of subsets in a national
  dimension
• Poor arrangements for large-scale creation of
  subsets and other NHS CRS metadata

4
Legacy Systems

• No native capability of utilising the full
  functionality of SCT to the extent envisaged by
  NPfIT
• Challenging to implement as idiosyncracies vary
  from system to system

5
Shoe-horning SNOMED in to legacy

• Inability to display terms more than 40
  characters
• Inability to store adequate number of concepts in
  data tables
• Inability to display or navigate hierarchy
• Inability to deal with synonyms (more than one
  description)
• Absent or limited ability to post-coordinate
• Limited ability to append text
• Uncertainty over standard logical NHS Clinical
  Information Model

6
Repercussions of perpetuating this legacy

• Poor professional perception of SNOMED
• New legacy of poor data quality
• Problems in choosing the right term
• What to do if cannot find the appropriate term?

7
NHS culture in the Acute Sector

• Compare and contrast GP and Hospital experience
  of health IT
• Previous focus on business support
• Future dependence on ICD/OPCS/HRG coding for
  hospital income/survival
• Existing IT establishment vs. clinically focused
  IT
• Do they GET IT?
• Timing of PbR couldnt be worse

8
Further Reading

• Why general practitioners use computers and
  hospital doctors do not
• Tim Benson - BMJ, 2002 - bmj.bmjjournals.com

9
Payment by Results

• Revolution in hospital funding
• National tariff for reimbursement of work
  actually done
• Vulnerability of hospitals specialising in
  complex cases
• Major dependency on accuracy and optimisation of
  coding
• Focus on statistical coding in OPCS and ICD10
• SNOMED a distraction in the short term and an
  advantage in the long term

10
More on PbR

• Policy clarity on the coding process from SNOMED
  directly to HRGs and PbR is not evident at the
  front-line
• Resultant uncertainty in clinical coding teams
• Mixed messages from the clinical drivers for CRS
  and finance departments

11
Working in Teams

• A way forward!
• NHS CRS is being implemented now
• Early systems are not mature SNOMED
  implementations
• Pace of programme is rapid
• Political
• Technical
• Contractual-commercial
• Socio-technical

12
The problem of NHS CRS metadata

• CRS applications come empty
• Process of tailoring
• Site specific e.g. wards/clinics etc.
• New problem of NHS specific clinical information
  rather than site specific information
• Terminology Subsets
• Rules for Decision support
• Business logic
• Common representational form
• Assessments and assessment forms
• All of this is NHS CRS metadata

13
Previous approaches

• Collect clinical input in best practice groups /
  clinical reference groups
• Typically CfH and users
• Observed by LSP and supplier
• Documentation prepared with some technical rigour
• Not related closely enough to actual system
  functionality

14
Perceived problem

• Users from next wave of real implementations not
  necessarily represented
• No clear ownership
• Best practice advice too far from current system
  capability
• User requirement communicated to application
  developer via intermediaries

15
Capturing requirements
LSP
S
U
CfH
16
T2
LSP
In this initial phase there is a steep learning
curve for all participants Who work together in a
team to produce basic requirements for main areas
of Clinical use e.g. AE/eye casualty. A small
trust clinical analyst team is formed LSP and
CfH Subject matter specialists begin cascade
training programme
In this accelerated phase there is a need to
apply lessons learned from initial terminology
configuration and to begin to specialise roles
the rudiments of a federated group of cluster
clinical analysts are formed at supra-trust level
In this maturation phase, trusts have developed
clinical analyst teams who are increasingly
independent from CfH and LSP subject matter
specialists. Skills such as subsetting and
decision support rule definition are present in
more mature trusts and there is an increasingly
well developed national set of subsets and
ruesets in the DOAS library
CfH
S
T3
LSP
T2
LSP
T2
LSP
T2
LSP
CfH
S
Tn
LSP
CfH
S
T1
LSP
CfH
S
CfH
S
T4
LSP
CfH
S
CfH
S
n
CfH
S
definition
systematisation
production
17
Early adopters

• Get NHS users involved more closely with building
  the relevant phase of the system
• E.g. subset development
• From the trusts imminently taking the system
• Clinician analysts
• Clinical advocates
• Get them to work together between trusts
• CfH LSP and supplier expertise to catalyse the
  process
• Develop NHS clinician governance
• Encourage links with professional organisations

18
Watch out for SSeRP

• SNOMED in Structured electronic Records Programme
• SNOMED Early Adopters Project
• Education and Training
• Co-ordination of subset development