東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Automated Problem Summarization from...

Van Vleck, Tielman Trevor.

Linked to FindBook

Google Book

Amazon

博客來

Automated Problem Summarization from Clinical Notes.

Record Type:	Language materials, printed : Monograph/item
Title/Author:	Automated Problem Summarization from Clinical Notes./
Author:	Van Vleck, Tielman Trevor.
Description:	213 p.
Notes:	Source: Dissertation Abstracts International, Volume: 72-06, Section: B, page: .
Contained By:	Dissertation Abstracts International72-06B.
Subject:	Biology, Bioinformatics. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3451809
ISBN:	9781124572550

Automated Problem Summarization from Clinical Notes.
Van Vleck, Tielman Trevor.

Automated Problem Summarization from Clinical Notes. - 213 p.

Source: Dissertation Abstracts International, Volume: 72-06, Section: B, page: .

Thesis (Ph.D.)--Columbia University, 2011.

Most modern industries approach core decision-making tasks in a hierarchical manner, in which crucial decisions are made by experts based on several levels of analyst aggregate and organize relevant data for the decision maker. Clinical practice relies on the physician to serve not only as the primary decision maker, but also as the primary information gatherer and analyst as well. Some amount of clinical data is collected in forms and nursing notes, but this is generally supplemental information for the physician in addition to information found in clinical databases and preceding clinical notes. Medical Informatics and Health Information Technology (HIT) have facilitated the collection and aggregation of structured clinical data for physicians to use at the time of patient care. In practice much patient information is documented in free-text notes that cannot be represented as coded data, and thus remains unintelligible for computation.

ISBN: 9781124572550Subjects--Topical Terms:

1018415
Biology, Bioinformatics.

Automated Problem Summarization from Clinical Notes.
LDR:04791nam 2200313 4500 001 1403041
005 20111108080359.5
008 130515s2011 ||||||||||||||||| ||eng d
020 $a 9781124572550
035 $a (UMI)AAI3451809
035 $a AAI3451809
040 $a UMI $c UMI
100 1 $a Van Vleck, Tielman Trevor. $3 1682276
245 1 0 $a Automated Problem Summarization from Clinical Notes.
300 $a 213 p.
500 $a Source: Dissertation Abstracts International, Volume: 72-06, Section: B, page: .
502 $a Thesis (Ph.D.)--Columbia University, 2011.
520 $a Most modern industries approach core decision-making tasks in a hierarchical manner, in which crucial decisions are made by experts based on several levels of analyst aggregate and organize relevant data for the decision maker. Clinical practice relies on the physician to serve not only as the primary decision maker, but also as the primary information gatherer and analyst as well. Some amount of clinical data is collected in forms and nursing notes, but this is generally supplemental information for the physician in addition to information found in clinical databases and preceding clinical notes. Medical Informatics and Health Information Technology (HIT) have facilitated the collection and aggregation of structured clinical data for physicians to use at the time of patient care. In practice much patient information is documented in free-text notes that cannot be represented as coded data, and thus remains unintelligible for computation.
520 $a Modern patient care is generally overseen by a primary care doctor, but an increasing percent of patient care is provided by a cadre of specialists often with little or no coordination. To support this decentralized model of collaborative care, clinical documentation has become increasingly important for sharing patient information between physicians. This has resulted in an information management challenge, with important patient data spread between many doctors in different location.
520 $a While systemic need for information management is high, the time and cognitive demands on the individual physician are great. Decades of advances in Biomedical Informatics have facilitated the digital capture of vast quantities of clinical data, and as such the volume of patient data available is rapidly increasing. Additionally, health information exchange (HIE) efforts to share patient information between institutions are multiplying the information available at the point of care.
520 $a With the convergence of physicians' increasing need for accurate patient information and rising volumes of disjoint data, the problem of physician information overload becomes increasingly significant. We have proposed the generation of high-level summaries of patient medical history as a potential solution to mitigate this problem.
520 $a While structured data such as discharge diagnoses, medications, allergies and lab results are crucial for a physician to be aware of, in their present form these do not convey to a physician the same holistic, longitudinal perspective that may be described in clinical narrative. Only limited research has investigated the summarization of information embedded in clinical notes. Here we present generic methods for automatic identification of relevant information from a corpus of clinical notes and their presentation in the form of a clinical abstract referred to as a Brief. For purposes restricting research to a manageable task, information assessed is limited to clinical problems.
520 $a Analysis of expert summaries collected on four patients in a separate study demonstrated that together, physicians classified about 22% of problems in patient notes to be relevant for a summary. In this study, we assess a number of ways to use algorithms, specifically machine learning classifiers, to mimic experts' selection of which problems were relevant for a summary and which were not. Under the safest parameters identified, we created a classifier identifying relevant problems with F-measures as high as 0.91.
520 $a This research builds on existing work in clinical NLP and machine learning to explore methods for the identification of patient information in previously unused data sources that could be crucial for helping a physician to understand key aspects of patient history at the time of care. Findings could be incorporated into future Electronic Health Records (EHRs) or other physician portals to provide a starting point from which all other information could be assessed.
590 $a School code: 0054.
650 4 $a Biology, Bioinformatics. $3 1018415
690 $a 0715
710 2 $a Columbia University. $3 571054
773 0 $t Dissertation Abstracts International $g 72-06B.
790 $a 0054
791 $a Ph.D.
792 $a 2011
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3451809