Upload
terence-bridges
View
214
Download
0
Embed Size (px)
Citation preview
About SNOMED CT
• 40 year old medical terminology– 322,544 concepts (and growing)
• Attempting an ‘in situ’ migration to EL+– And ‘seamless’ deployment into an industry based
on enumerated classifications– 18-country international effort, $9.3M annually
Growth of SNOMED
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Concepts
SNOMED II
SNOMED International
SNOMED RT (Beta)
(V 3.0)
(V 3.1)(V 3.2)
(V 3.3)
(V 3.4)(V 3.5)
357,135 (Jan 2004)
368,590 (Jan 2006)
150,000
100,000
50,000
200,000
250,000
300,000
350,000
400,000
388,289(Jul 2009)
(Apr 2012)International Core + UK national extensions+ DM+D(+ US drug extensions)
Total: 709,742 concepts (of which 504,084 active)1,199,678 ‘en’ descriptions
READ2
37,502 (Jan 1991)
292,524(Oct 2009)
CTV3237,557 Oct 1999
249,081Mar 2002
262,231Oct 2004
81,885 Sep 1999
83,076 Oct 2001
90,956Apr 2007
SNOMED CT R1325,856 (Jan 2002)
SNOMED
93,717(Oct 2009)
2011 2012
270,600Apr 2006
282,253 Apr 2008
85,647Oct 2003
86,865Jul 2004
95,832(Apr 2012)
308,032(Apr 2012)
378,111 (Jan 2008)
395,346(Apr 2012)
SNOMED CT in UK
• Many secondary care sites– Some primary care
• 13,353,775 Summary Care Records• 33M Choose & Book referrals• Electronic TFR of prescriptions• Soon: Radiology & Pathology messaging
Issues
• Change management– Migration from/integration with legacy systems– Changes in SNOMED CT itself– Death by 1,000 mutual dependencies
• Implementation skills• User interfaces (or, data repair)• Tools
– Time to load & classify– Content refactoring– ‘Linkage’ to external resources
• Business case
SNOMED CT39 months @ a busy UK A&E Department
• One ‘reason for encounter’ code per completed visit
– 408,823 coded episodes • 39 months (Oct 2008 – Dec 2011)• 12,323 distinct codes selected at least once• 8,387 not coded (or uncodable?) = 1 in 50 episodes• 20-50% miscoding rate
Relative code use
12 537 1062 1587 2112 2637 3162 3687 4212 4737 5262 5787 6312 6837 7362 7887 8412 8937 9462 9987 105121103711562120870
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
1000 distinct SNOMED CT codes ac-count for 84% of the data
Top 15 tunes…(22.8% of all episodes)18332 282026002|Soft tissue injury (disorder)13995 21522001|Abdominal pain (finding)10588 82271004|Injury of head (disorder)9466 29857009|Chest pain (finding)6861 213257006|Generally unwell (finding)6113 367391008|Malaise (finding)4656 161891005|Backache (finding)4435 44465007|Sprain of ankle (disorder)4309 34014006|Viral disease (disorder)3777 399221001|Bleeding from vagina (disorder)2871 35933005|Laceration (morphologic abnormality)2690 281794004|Viral upper respiratory tract infection (disorder)2622 68566005|Urinary tract infectious disease (disorder)2404 19130008|Traumatic abnormality (morphologic abnormality)2280 281245003|Musculoskeletal chest pain (finding)
7.8% ‘Ontology-driven’ miscoding…(disorder) 234693 (morphologic abnormality) 17024(finding) 123430 (qualifier value) 4544(procedure) 6583 (body structure) 3517(situation) 6379 (substance) 2340(event) 4592 (attribute) 1293(regime/therapy) 1344 (observable entity) 1139
(physical object) 746(product) 233(cell) 226(navigational concept) 206(organism) 199(physical force) 152(record artifact) 97(ethnic group) 17(environment) 15(assessment scale) 14(person) 10(specimen) 9(administrative concept) 6(tumor staging) 5(social concept) 4(cell structure) 3(occupation) 2(inactive concept) 1
TOTALS 377021 31802
Miscoding examples
1097 Temperature 246508008|Temperature (attribute)|17 Drug used 246488008|Drug used (attribute)|373 ETOH - Alcohol intake 160573003|Alcohol intake (observable entity)|136 Nasogastric tube 17102003|Nasogastric tube, device (physical object)|82 Catheter 19923001|Catheter, device (physical object)|78 Dressing 37898001|Dressing, device (physical object)|
110 53570002|Removal of foreign body from eye (procedure)| 83 172828005|Removal of foreign body from nose (procedure)|43 172278002|Removal of foreign body from eyelid (procedure)|
293 82576008|Retained foreign body in eye (disorder)| 166 74699008|Foreign body in nose (disorder)7 25012008|Retained foreign body of eyelid (disorder)|
Variable Data Quality
• 23% of 74 abdominal aortic aneurysms miscoded as a Drug Trade Family (9192101000001100 AAA (product)); AAA make sore throat spray(and not much else)
• 25% of 939 stabbing victims miscoded as a qualifier value (‘stabbing sensation quality’, as in heart attack)
• 33% of 3771 patients with some form of high temperature miscoded as either an attribute, or a physical force
• 38% of 1101 failed consultations (patient left the department, or did not attend an appointment) miscoded as either a laterality (left) or as deoxyribonucleic acid (DNA = Did Not Attend)
• 44% of 575 patient attending with a fish bone stuck in their throat miscoded as the bone itself (7661006|Fish bone (substance)|)
• 49% of 5,062 alcohol-related attendances miscoded as either the substance (alcohol, ethyl alcohol) or just feeling elated/intoxicated but not necessarily involving alcohol intake at all
Not all bad news:Admissions for sickle cell
• Clinical impression– ‘They stop coming once they get older; most of the
attendances will be in the 15-20 age group’
• Clinical lore– Cold weather triggers attacks
‘People with sickle cell disease should try to avoid any potential triggers for a sickle cell crisis as much a possible. For example: try to keep warm in cold weather, try to avoid becoming dehydrated and take precautions if you undergo extreme exercise’(patient.co.uk)
• Clinical Data…???
Sunday, June 1, 2008 Monday, June 1, 2009 Tuesday, June 1, 2010 Wednesday, June 1, 2011 Thursday, May 31, 20120
10
20
30
40
50
60
70
80
AGE at presentation vs DATE of presentation 60 day average, rolling count of attendances in last 14 days
1336 A&E attendances, 410 patientsCompleted episodes October 2008 thru December 2011
Reason for attendance: sickle cell
WIN
TE
R 2
011-
12
WIN
TE
R 2
008-
9But:
We could probably have got this particular result using ICD.
Overall, is the ontology (as implemented) helping or hindering primary data capture and secondary data analysis?