Implementation Package_Data Repository_v0.6

advertisement
Implementation Package
Project Name:
Business Owner:
Data Repository
Max Erhmann
Project ID:
Product Owner:
Zac Wyner
Project Manager:
Zac Wyner
Please see the accompanying presentation, Implementation Package Development Training for help in completing this template. The
training slides break down each section of this document by its intent, guidelines for use, and examples.
Note: Mandatory sections of the Initiation Package have “(M)” after their title. These sections must be completed in order for the project to
be considered in the Change Control Board (CCB) process.
1.
Background / Rationale: (M)
Intent: Provide sufficient information to outline and define the specific business opportunities or problems that will be addressed by the project
and the reasons which support the development of the new system.
Format: Text, yet can also be augmented by combining diagrams and/or illustrations of current systems.
BACKGROUND:
Currently, the Sentinel QA group produces and collects information about data completeness, characterization,
and quality of data across the 18 Sentinel data partners. This involves several processes and tools as outlined
below:
QA Processes
Process
Output
Output Content
Output Sample
SAS datasets
“Output files” with summary
statistics MSCDM variables
K:\TIDE Projects\FDA_Sentinel\03.
Data Activities\03. Data Checking\04.
Output from Data Partners\01. Data
Quality and
Reports\Y5R4\HMGHC\MSOC
ETL_Compare
.sas
SAS datasets (not
saved)  Excel file
(refresh_tables.xls)
Differences between current and
previous ETL
K:\Sentinel\requests\qa\kpnc\packag
es\to16_upd_qar_wp002\to16_upd_
qar_wp002_kpnc_v01\msoc
QA runs DaRT to do
error checks
(Imports refresh_tables
into DaRT)
“DaRT”
Excel with
VBA
Excel
(MS Core Data QA
Review
for[dp].xlsx)
Error codes and their
descriptions for MSCDM and
table and variable,
communication between DP and
MSOC about issues
K:\Sentinel\requests\qa\kpnc\packag
es\to16_upd_qar_wp002\to16_upd_
qar_wp002_kpnc_v01\reports
QA Report sent to DP
SFTP
MSCDM ETL Data
Partner Survey
PDF form
sent via PMN
file transfer
PDF form
A mixture of categorical and freetext responses from each DP.
Some are table specific; some are
table/variable specific; some are
more general.
K:\TIDE Projects\FDA_Sentinel\03.
Data Activities\03. Data Checking\04.
Output from Data Partners\07. Data
Partner Questionnaire\Clinical Data
Elements Questionnaire
DP correspondence
Email
Emails that
characterize or
explain data
Data partner notifies
HPHC that new ETL is
ready
QA sends package to DP
DP runs QA package
DP returns QA package
QA compares new
version of QA output
tables to previous
version
File Name: Document1
Tool
Email
PMN
SAS
PMN DMC
Page 1 of 4
Note: All of the above processes need to be unified in a single resource in which an investigator or analyst at MSOC is
able to query the repository to obtain all information available on a DP, variable, or a component of a variable (a specific
NDC or PX code in the MSDD).
RATIONALE:
The proposed QA repository would benefit the MSOC and its stakeholders in the following ways:
1. Reduce time spent on answering specific data related questions
2. Increase the capacity of the QA group to perform QA processes
3. ETLs approved and moved to production faster
4. Increase the speed at which QF can complete data requests that are waiting for refreshed data
5. Allow for more complicated questions to be asked of QA than they currently have the capacity to
handle
3.
Driving Requirements: (M) (system function)
Intent: Provide a high-level, executive overview of what the system must do. The requirements should not specify solutions—only the key
capabilities necessary to solve the business problems.
4.
DR Key
Driving Requirement
DR 1
The system must pool data from the first 3 outputs listed above into a query-able relational database
structure
DR 2
The system must make available all QA information about a chosen data element from the MSDD
DR 3
The system must be able to compare all data / metadata about a current ETL to a previous ETL
DR 4
The system must present all information about any data element in the MSDD for a single data partner or
across any combination of data partners
DR 5
The system should be accessible via a web-based interface
DR 6
The system should be integrated into the PopMedNet system
Supported Scenarios: (business process)
Intent: Document the list of business scenarios that the overall system will need to address and support.
SS Key
Supported Scenario
SS 1
Get counts of any continuous data elements in the system stratified by values
SS 2
Get proportions for any values of categorical data elements in the system
SS 3
Get completeness/missingness counts on any data elements in the system
SS 4
Get summary statistics on any data elements in the system
SS 5
Compare current ETL information about any data element in the system to information from previous ETL
versions
SS 6
Get all error codes for to a single data element across all data partners
SS 7
Get all error codes relating to a one or more data partners
File Name: Document1
Page 2 of 4
5.
Context Diagrams: (M)
QA Process Diagram - Complete
Proposed Context Diagram
Steps 5 - 11, 13, and 14, in the QA Process diagram above, produce metadata that needs to be captured by the
Repository in phase 1 as diagramed below. (Note: Data from step 4 will be incorporated in a future phase).
Repository Phase 1
File Name: Document1
Page 3 of 4
7.
Description of the Expected Business Outcomes: (M)
Intent: Acknowledge and communicate the expected outcome for the business as a result of the project. Expected business outcomes
measure the business success of the initiative. Also note when the outcome is expected, how that outcome will be measured and who will be
responsible for that measurement.
Expected Business Outcome
Due
Overall time spent responding to queries
about QA data is reduced by 1 month per
year
The QA group’s average time to complete
a data partner ETL is reduced by at least
5%
The complexity of questions asked by
requesters (MSOC, FDA, and occasionally
Data Partners) increases
1 month post
implementation
DMQA
1 year post
implementation
DMQA
1 year post
implementation
DMQA
File Name: Document1
Page 4 of 4
Measurement Methodology
Who
Download