PMID- 29275493
OWN - NLM
STAT- MEDLINE
DCOM- 20180827
LR  - 20181113
IS  - 1573-689X (Electronic)
IS  - 0148-5598 (Linking)
VI  - 42
IP  - 2
DP  - 2017 Dec 23
TI  - Automatic Decision Support for Clinical Diagnostic Literature Using Link Analysis 
      in a Weighted Keyword Network.
PG  - 27
LID - 10.1007/s10916-017-0876-3 [doi]
AB  - We present a novel approach to recommending articles from the medical literature 
      that support clinical diagnostic decision-making, giving detailed descriptions of 
      the associated ideas and principles. The specific goal is to retrieve biomedical 
      articles that help answer questions of a specified type about a particular case. 
      Based on the filtered keywords, MeSH(Medical Subject Headings) lexicon and the 
      automatically extracted acronyms, the relationship between keywords and articles 
      was built. The paper gives a detailed description of the process of by which 
      keywords were measured and relevant articles identified based on link analysis in 
      a weighted keywords network. Some important challenges identified in this study 
      include the extraction of diagnosis-related keywords and a collection of valid 
      sentences based on the keyword co-occurrence analysis and existing descriptions 
      of symptoms. All data were taken from medical articles provided in the TREC (Text 
      Retrieval Conference) clinical decision support track 2015. Ten standard topics 
      and one demonstration topic were tested. In each case, a maximum of five articles 
      with the highest relevance were returned. The total user satisfaction of 3.98 was 
      33% higher than average. The results also suggested that the smaller the number 
      of results, the higher the average satisfaction. However, a few shortcomings were 
      also revealed since medical literature recommendation for clinical diagnostic 
      decision support is so complex a topic that it cannot be fully addressed through 
      the semantic information carried solely by keywords in existing descriptions of 
      symptoms. Nevertheless, the fact that these articles are actually relevant will 
      no doubt inspire future research.
FAU - Li, Shuqing
AU  - Li S
AUID- ORCID: 0000-0001-9814-5766
AD  - College of Information Engineering, Nanjing University of Finance & Economics, 
      Nanjing, 210023, China. leeshuqing@gmail.com.
FAU - Sun, Ying
AU  - Sun Y
AD  - Department of Library and Information Studies, Graduate School of Education, 
      University at Buffalo, New York State University, Buffalo, NY, 14260, USA.
FAU - Soergel, Dagobert
AU  - Soergel D
AD  - Department of Library and Information Studies, Graduate School of Education, 
      University at Buffalo, New York State University, Buffalo, NY, 14260, USA.
LA  - eng
GR  - 16BTQ030/Chinese National Social Science Foundation/
PT  - Journal Article
DEP - 20171223
PL  - United States
TA  - J Med Syst
JT  - Journal of medical systems
JID - 7806056
SB  - IM
MH  - Algorithms
MH  - Clinical Decision-Making/*methods
MH  - Consumer Behavior
MH  - Data Mining/methods
MH  - *Diagnostic Techniques and Procedures
MH  - Humans
MH  - Information Storage and Retrieval/*methods
MH  - Medical Subject Headings
MH  - *Periodicals as Topic
MH  - Semantics
OTO - NOTNLM
OT  - Clinical decision support
OT  - Keyword co-occurrence analysis
OT  - Link analysis
OT  - Literature recommendation service
EDAT- 2017/12/25 06:00
MHDA- 2018/08/28 06:00
CRDT- 2017/12/25 06:00
PHST- 2016/01/10 00:00 [received]
PHST- 2017/12/13 00:00 [accepted]
PHST- 2017/12/25 06:00 [entrez]
PHST- 2017/12/25 06:00 [pubmed]
PHST- 2018/08/28 06:00 [medline]
AID - 10.1007/s10916-017-0876-3 [pii]
AID - 10.1007/s10916-017-0876-3 [doi]
PST - epublish
SO  - J Med Syst. 2017 Dec 23;42(2):27. doi: 10.1007/s10916-017-0876-3.