NLP
Machine Learning
Decision Support
Technology
Grab Bag
100
Data needed to calculate patient care performance measures are often stored in this form in electronic medial records (EHRs) making them hard to extract automatically.
What is narrative text? (Will also accept "natural language" or "free text")
100
In this type machine learning, the correct classification of training set is used to train the system.
What is supervised learning?
100
This automated decision support message may be produced when a user enters a dangerous medication order into an EHR.
What is an alert? (Would also accept "What is a reminder?")
100
This ubiquitous device may be an ideal platform for mobile health applications.
What is a smart phone?
100
In information retrieval, sensitivity is to recall as specificity is to this.
What is precision?
200
This tool from the National Library of Medicine's UMLS project can identify concept names in text.
What is MetaMap?
200
MLI, developed by Sparks and colleagues, uses this technique to improve machine learning throughput.
What is parallelization? (Will also accept "what is distributed processing?"
200
Unlike Iliad and Meditel, these two diagnostic systems use non-Bayesian algorithms.
What are QMR and DXplain?
200
This technique incorporates elements of game design into computer applications that are otherwise boring and repetitive to increase user engagement.
What is gamification?
200
This is the process identifying patients who belong to a particular cohort, based on physical findings.
What is phenotyping?
300
MedLEE (Columbia University's Medical Language Extraction and Encoding system) uses this to represent semantic and syntactic components of its knowledge.
What is a grammar?
300
Dligach and colleagues studied these two types of disease modifiers in clinical texts.
What is are "body site" and "severity"?
300
Kassier gave computerized diagnosis systems this grade in response to a study by Berner and colleagues.
What is a "C"?
300
This approach to computerizing interactive role-playing games can yield engaging, ecologically valid and user-centric content.
What is virtual reality?
300
PubMed is to literature citations as this is to genetic sequence data.
What is GenBank?
400
Evans and colleagues used this NLP method to process dictated patient reports in order to identify patients with heart failure.
What is keyword/term detection?
400
In a paper by Osborn and colleagues, identification of cancer cases was best achieved by combining these two forms of coded data.
What are CUIs and ICD-9 Codes? (Also accept Concept Unique Identifiers and ICD-9-CM or International Classification of Diseases, 9th Edition, with Clinical Modifications)
400
In addition to noting the inclusion of the correct diagnosis in a diagnosis list, Berner and colleagues used this metric when scoring decision support systems for having the right answer.
What is rank?
400
Mobile phone ownership is at or above this percentage in the US, Africa and China.
What is 90%.
400
This measure, often symbolized by a single letter, estimates classification accuracy based on precision and recall.
What is the F measure?
500
Each concept in the UMLS Metathesaurus is assigned one or more of these to provide coded information about the concept's meaning.
What is a semantic type?
500
The performance of Tahsin and colleagues' rule-based extraction system was extremely sensitive to this.
What are errors in GenBank data?
500
In a study by of decision support tools by Moss and Berner, the use of this in the design of medication alerts was cited by participants as increasing their awareness of potentially important information.
What is color?
500
In an HIV gamification study by Nevin and colleagues, players' answers to reintroduced questions were more frequently correct by this percentage.
What is 11.9%? (Will accept 11% or 12%)
500
Known by the acronym ROC curves, these graphs depict the relationship between sensitivity and specificity.
What are Receiver Operating Characteristic curves?
M
e
n
u