Replicating medication trend studies using ad hoc information extraction in a clinical data warehouse.
Data warehouse
Information extraction
Medication extraction
Journal
BMC medical informatics and decision making
ISSN: 1472-6947
Titre abrégé: BMC Med Inform Decis Mak
Pays: England
ID NLM: 101088682
Informations de publication
Date de publication:
18 01 2019
18 01 2019
Historique:
received:
27
07
2018
accepted:
21
12
2018
entrez:
20
1
2019
pubmed:
20
1
2019
medline:
16
7
2019
Statut:
epublish
Résumé
Medication trend studies show the changes of medication over the years and may be replicated using a clinical Data Warehouse (CDW). Even nowadays, a lot of the patient information, like medication data, in the EHR is stored in the format of free text. As the conventional approach of information extraction (IE) demands a high developmental effort, we used ad hoc IE instead. This technique queries information and extracts it on the fly from texts contained in the CDW. We present a generalizable approach of ad hoc IE for pharmacotherapy (medications and their daily dosage) presented in hospital discharge letters. We added import and query features to the CDW system, like error tolerant queries to deal with misspellings and proximity search for the extraction of the daily dosage. During the data integration process in the CDW, negated, historical and non-patient context data are filtered. For the replication studies, we used a drug list grouped by ATC (Anatomical Therapeutic Chemical Classification System) codes as input for queries to the CDW. We achieve an F1 score of 0.983 (precision 0.997, recall 0.970) for extracting medication from discharge letters and an F1 score of 0.974 (precision 0.977, recall 0.972) for extracting the dosage. We replicated three published medical trend studies for hypertension, atrial fibrillation and chronic kidney disease. Overall, 93% of the main findings could be replicated, 68% of sub-findings, and 75% of all findings. One study could be completely replicated with all main and sub-findings. A novel approach for ad hoc IE is presented. It is very suitable for basic medical texts like discharge letters and finding reports. Ad hoc IE is by definition more limited than conventional IE and does not claim to replace it, but it substantially exceeds the search capabilities of many CDWs and it is convenient to conduct replication studies fast and with high quality.
Sections du résumé
BACKGROUND
Medication trend studies show the changes of medication over the years and may be replicated using a clinical Data Warehouse (CDW). Even nowadays, a lot of the patient information, like medication data, in the EHR is stored in the format of free text. As the conventional approach of information extraction (IE) demands a high developmental effort, we used ad hoc IE instead. This technique queries information and extracts it on the fly from texts contained in the CDW.
METHODS
We present a generalizable approach of ad hoc IE for pharmacotherapy (medications and their daily dosage) presented in hospital discharge letters. We added import and query features to the CDW system, like error tolerant queries to deal with misspellings and proximity search for the extraction of the daily dosage. During the data integration process in the CDW, negated, historical and non-patient context data are filtered. For the replication studies, we used a drug list grouped by ATC (Anatomical Therapeutic Chemical Classification System) codes as input for queries to the CDW.
RESULTS
We achieve an F1 score of 0.983 (precision 0.997, recall 0.970) for extracting medication from discharge letters and an F1 score of 0.974 (precision 0.977, recall 0.972) for extracting the dosage. We replicated three published medical trend studies for hypertension, atrial fibrillation and chronic kidney disease. Overall, 93% of the main findings could be replicated, 68% of sub-findings, and 75% of all findings. One study could be completely replicated with all main and sub-findings.
CONCLUSION
A novel approach for ad hoc IE is presented. It is very suitable for basic medical texts like discharge letters and finding reports. Ad hoc IE is by definition more limited than conventional IE and does not claim to replace it, but it substantially exceeds the search capabilities of many CDWs and it is convenient to conduct replication studies fast and with high quality.
Identifiants
pubmed: 30658633
doi: 10.1186/s12911-018-0729-0
pii: 10.1186/s12911-018-0729-0
pmc: PMC6339317
doi:
Types de publication
Journal Article
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
15Références
J Biomed Inform. 2009 Apr;42(2):377-81
pubmed: 18929686
Nature. 2012 Mar 28;483(7391):531-3
pubmed: 22460880
Sci Rep. 2017 Apr 07;7:46226
pubmed: 28387314
Eur J Prev Cardiol. 2012 Apr;19(2):213-20
pubmed: 21450611
Lancet. 2014 May 31;383(9932):1912-9
pubmed: 24881995
J Am Med Inform Assoc. 2017 May 1;24(3):607-613
pubmed: 28339516
J Am Med Inform Assoc. 2011 Jul-Aug;18(4):387-91
pubmed: 21672908
Stud Health Technol Inform. 2011;169:584-8
pubmed: 21893816
AMIA Annu Symp Proc. 2009 Nov 14;2009:391-5
pubmed: 20351886
Summit Transl Bioinform. 2010 Mar 01;2010:71-5
pubmed: 21347153
Curr Hypertens Rep. 2013 Jun;15(3):134-6
pubmed: 23536128
J Biomed Inform. 2018 Jan;77:34-49
pubmed: 29162496
J Am Med Inform Assoc. 2010 Jan-Feb;17(1):19-24
pubmed: 20064797
Acta Psychiatr Scand. 2011 May;123(5):360-7
pubmed: 20860726
J Biomed Inform. 2009 Oct;42(5):839-51
pubmed: 19435614
J Biomed Inform. 2001 Oct;34(5):301-10
pubmed: 12123149
Proc AMIA Symp. 2001;:105-9
pubmed: 11825163
J Clin Hypertens (Greenwich). 2018 Jan;20(1):106-114
pubmed: 29220556
Circulation. 2012 Oct 23;126(17):2105-14
pubmed: 23091084
Int J Chronic Dis. 2018 Feb 25;2018:1382705
pubmed: 29682516
Methods Inf Med. 2018 May;57(1):e22-e29
pubmed: 29801178
Nature. 2016 May 25;533(7604):452-4
pubmed: 27225100
Eur Heart J. 2017 Mar 21;38(12):899-906
pubmed: 28110293
Am J Hypertens. 2017 Oct 1;30(10):1008-1014
pubmed: 28531239
Am J Hypertens. 2016 Jan;29(1):104-13
pubmed: 25968124
Biomed Inform Insights. 2013 Jun 24;6(Suppl 1):7-16
pubmed: 23847423
J Am Med Inform Assoc. 2011 Dec;18 Suppl 1:i144-9
pubmed: 21946242
Clin Rheumatol. 2015 May;34(5):949-56
pubmed: 24420724
Arch Intern Med. 2004 Jan 12;164(1):55-60
pubmed: 14718322
Sci Rep. 2016 Aug 11;6:31477
pubmed: 27510920
BMJ Open Diabetes Res Care. 2016 Apr 11;4(1):e000154
pubmed: 27110365
J Am Med Inform Assoc. 2010 Sep-Oct;17(5):532-5
pubmed: 20819858