Data Analytics Flashcards
Which type of analytics is used in many health care settings to determine what might happen next?
Descriptive
Diagnostic
Predictive
Prescriptive
Answer: Predictive analytics
Descriptive Analytics = based on live data. What is happening in real time?
Diagnostic Analytics = Why are things happening? Might involve root cause analysis.
Predictive Analytics = What is likely to happen next? Predictive Analytics are likely the most used/developed in health care. Example: analyzing EHR data to identify patients at risk for a complication from a disease
Just because we can predict something doesn’t mean we can prevent it.
Predictive models aim to identify associations between “predictor” variables (also called “features”) and an “outcome” variable.
Prescriptive Analytics = what to do next? Decide future actions based on advanced algorithms.
The process of original research being submitted for peer review publication and possibly secondary publications is best described as:
Knowledge management
Knowledge generation
Knowledge acquisition
Knowledge representation
Answer: Knowledge generation
Knowledge management (KM) includes:
- Knowledge generation (publishing research)
- Knowledge acquisition (info retrieval from literature)
- Knowledge representation for persistence and sharing (using rules/IfThen, algorithms, Bayesian, scoring systems; sharable forms such as Arden Syntax which codes clinical decision support rules)
Which of the following is not true of public health reporting?
A. Notifiable diseases must be reported to the CDC.
B. Reportable diseases must be reported to states, and these vary by state.
C. HITECH required public health reporting (ex. Vaccine info) in meaningful use measures.
D. Syndromic surveillance monitors the prevalence of births with genetic abnormalities to screen for public health issues.
D is not true
Syndromic surveillance seeks to identify illness clusters, respond quickly, and reduce morbidity and mortality.
Notifiable diseases are reported to CDC
Reportable diseases are reported to states
HITECH did include public health reporting for meaningful use.
A transfer summary is prepared by the transferring hospital and sent to the accepting facility. This is an example of what type of health information exchange?
A. Directed
B. Query-based
C. Consumer mediated
D. Private
Answer: A. Directed
Types of HIE:
Directed = direct sending and receiving of info to support planned care. “Push” Examples include referrals and transfers. Is analogous to secure email.
Query-based = seeking out and finding info to support unplanned care. Examples include emergency care and using a record locator service. “Pull”
Consumer-mediated = consumers aggregating and using their own info. Possible example includes patient registry developed by patient advocacy group
All can be classified as public or private
Which of the following best describes the process of data migration?
Bonus: what do the other items describe (ie name them)?
A. Checking for accuracy of data in the source and target system
B. Setting up definitions to map data from the source system to target system.
C. Converting data into a different format through conversion
D. Transferring data from one storage system to another, typically on a large scale.
E. Using mathematical methods to perform high level accuracy check of large amounts of data that have been moved
Answer: D
Data migration is defined as the process of transferring data from one storage system or computing environment to another, typically on a large scale. Note that this overlaps with the definition of Data Transfer except that data transfer is not necessarily on a large scale.
A. Data transfer validation is the process of checking the accuracy of data in the source and target systems. Examples include checking for truncated or missing data, such as can happen with the ~ character in HL7 functions.
B. Data mapping is the process of Setting up definitions to map data from the source system to target system.
C. Data transformation is the process of Converting data into a different format or value through conversion or calculation. Never plug and play. Errors can result in loss of granularity or loss of context. When done during data transfer, must balance needs for transformation with potential for confusion because the source data will no longer visually match the data in the target system.
E. Data reconciliation is the process of Using mathematical methods to perform high level accuracy check of large amounts of data that have been moved/ migrated. Looks for gross errors but does not tell you the source of the error. Examples include master data reconciliation (ex. Check basic patient demographics), accuracy of activity (are transactions valid), transactional data reconciliation (computations/revenue), and automated data reconciliation (uses meta tables and algorithms). Should be followed up by data validation on smaller subsets of data, which is the process of validating that the data from the source system has crossed to the target system “accurately AND completely AND with the same context/interpretation.” Data validation can be internal (done by internal healthcare employees) or external (by external 3rd party service).
DRAFT
Assuming a large scale migration must occur, put these data management processes in the correct order
Data transformation
Data reconciliation
Data generation
Data migration
Data mapping
Data validation
Work in progress…
