Forty years of SNOMED A literature review

1
Forty years of SNOMED A literature review

• Ronald Cornet, Nicolette de Keizer
• Dept. of Medical Informatics
• Academic Medical Center Universiteit van
  Amsterdam

2
Research Background

• 10 years of experience in design, implementation
  and maintenance of a terminological system for
  Reasons for Admission in Intensive Care (DICE)
• Frame-based, experiments using Description Logic
• Compositional
• 2500 concepts, 1500 reasons for admission

3
Motivation Use of Patient Data

• Documentation in the EPR/EHR
• Decision support
• Clinical audit
• Reporting
• Summaries
• Administrative management information
• Epidemiology
• Billing
• Resource management

National Health Services United Kingdom
4
Motivation Clinical Terminology

• Cost reduction
• Terminology use benefits entire health system
• Save as much as 5 of total healthcare costs
• up to 100 Billion per year in US
• Source - Walker J et al., Market Watch
  200519th January10-18

Slide from Kevin Donnelly SNOMED international
5
Motivation SNOMED
Where is the science base for our work?
6
The science base

• Is there scientific evidence that patient care
  improves using SNOMED?

7
Method

• Literature review in Pubmed by 2 reviewers
• Retrieve all citations containing SNOMED or
  SNOP
• Exclude papers without English abstract
• Exclude papers that do not deal with SNOP/SNOMED,
  e.g. authored by Snop
• Analyze remaining abstracts

8
Characteristics

• SNOP/SNOMED Version
• Domain (e.g., cancer, cardiology)
• Usage (as primary/secondary study object)
• Evaluation phase (does it work, does it help,
  what are consequences)

9
(Cumulative) number of publications
10
Sources of publication

• 71 SCAMC / AMIA proceedings
• 41 pathology journals (22)
• 16 Medinfo proceedings
• 14 Methods of Information in Medicine
• 13 JAMIA
• 10 MIE proceedings
• 9 Med Inform (Lond)
• 9 Int J Medical Informatics
• 6 J Biomed Informatics
• 4 Biomed Sci Instrum
• 3 Int J Biomed Comput
• 3 Comput Inform Nurs
• 3 Comput Methods Programs Biomed
• lt 2 various (43 titles)

11
SNOMED versions
12
SNOMED versions
13
Is SNOMED CT really multi-professional?
Distribution of medical domains, when mentioned
in the paper
14
Is SNOMED CT really multi-professional?
15
Usage
16
Evaluation of SNOMED

• 129 papers describe evaluation
• 104 (81) Does it work?
• technical evaluation
• content coverage
• 25 (19) Does it help?
• retrieval of patients
• aggregation of patient groups
• No papers describe consequences

17
Discussion

• Limited to scientific publications
• No comparison to other trends / systems
• SNOMED is part of the solution, not the solution
  itself, e.g., for Kaiser Permanente

18
Conclusions

• The two largest groups of papers
• comparison of SNOMED to other TSs
• theory (automatic coding, natural language
  processing and description logics) illustrated by
  (case studies with) SNOMED
• Few studies are available on the usage of SNOMED
  in clinical practice

19
How can you improve care?

• SNOMED CT contributes to
• structured documentation
• decision support
• clinical audit
• The effect of using SNOMED CT on the care process
  and outcome needs to be studied