The Art of Requesting Data from IT
1 like288 views
Vanderbilt University Medical Center has annual operating expenses of $2.3 billion, an annual sponsored research budget of $471.6 million, and annual unrecovered costs of charity care, community benefits, and other costs of $843.6 million. The document then discusses challenges in accessing and analyzing healthcare data from their databases due to issues such as lack of integration, improper structuring of the data, and cultural barriers between operations and IT. Strategies provided to help address these challenges include establishing standard data requests, designating cross-functional leads, and developing relationships with different types of "data people".
The Art of Requesting Data from IT
- 2. ABOUT VANDERBILT UNIVERSITY MEDICAL CENTER • $2.3 Billion Annual Healthcare Operating Expenses (excludes academics and research) • $471.6 Million Annual Sponsored Research Budget • $843.6 Million Annual Charity Care, Community Benefits, and other Unrecovered Costs
- 5. Freakonomics Radio - June 26, 2014
- 6. I never would have imagined that it is an IT problem that you simply cannot get the data you want. We’ll spend three or six person months just put together a data set to do a basic analysis.
- 7. How can we view the new ICD-10 codes in the revenue reports? ja' chel De' yotlh 'ej 'oH Hoch items cha' choH.
- 8. Table Field Data Type Primary Key Record Normalization Foreign Key EDW OLAP OLTP Cubes Aggregation Index SQL View
- 9. SLIDES AND HELPFUL TERMS ARE POSTED AT TERMINALIQ.COM/BLOG/ ALSO POSTED TO LINKEDIN
- 10. IT’S 1 AM, DO YOU KNOW WHERE YOUR DATA IS?
- 11. OLTP • Online Tranaction Processing • Designed for adding, deleting, and updating individual pieces of data. • Applications used by operations will use OLTP. OLAP • Online Analytical Processing • Designed for selecting data. • Reporting systems and data warehouses will use OLAP.
- 12. IT’S 2 AM, DO YOU WHO YOUR DATA IS WITH?
- 13. THE CURMUDGEON • Advantages: They know the system, data, and the history of how things came to be. • Disadvantages: Won’t let others access their data, let alone each anyone how to do what they do because they are afraid they will become unnecessary to the organization. • Strategies: Convince them you don’t want to upset the status quo but that they are your only hope. Give them something more to do then pull the data like validate it for reasonableness. Praise their contributions and helpfulness in front of their 1-ups.
- 14. THE NEWBIE • Advantages: Wants to help and do a good job. • Disadvantages: Doesn’t know very much and might act like they know more than they do which can lead to bad data. • Strategies: Partner with the Newbie, teach them what the data means to operations. Invite them to your meetings or give them an in-depth tour of your area.
- 15. THE GURU • Advantages: Knows the data and operations. Can you get the data you want the first time. • Disadvantages: No time to get to every request that comes their way. • Strategies: Your goal is to become her favorite end-user. Clearly define the data you need, going back for multiple iterations can be your death with the Guru. If there is anything you can do for them, do it.
- 16. THE IDIOT • Advantages: None • Disadvantages: They are slow, do a half-assed job, and you have to double-check their work. • Strategies: Avoid the Idiot if you can. If you can’t, figure out a way to at least eliminate the slowness disadvantage. Candy bars or cokes are often good bribes.
- 18. MORE STRATEGIES • Request forms for standard(ish) data needs. • For ad-hoc requests, have IT include a copy of the query with the data so you can easily request it again later. • For frequent data needs, ask if reports can be scheduled to run overnight. • Designate leads/experts from both operations, finance, IT, etc. to work together on reporting needs.
- 20. The Art of Requesting Data from IT Presented by Brad Adams Helpful Terms Term Definition Aggregation Any time data is grouped or summarized. Many user-friendly reporting systems will automatically group and summarize data automatically. Cubes Pre-summarized datasets that increase response time of OLAP databases. When reporting from the general ledger, performance would be greatly improved by pre-summarizing monthly transitions by account, month/year, and department. Data Type Defines the acceptable inputs to a field. Common data types include text, number, date, and time. Field Similar to a column in a spreadsheet, it contains information of a particular type (e.g., First Name). Foreign Key A field that is a primary key in another table used to create a relationship between tables. Index Indexes are added to tables on one or more fields to speed up random lookups and improve overall performance and efficiency of the database. Normalization Database design methodology to minimize data redundancy by splitting data into it’s smallest useful parts. Rather than have a single field for name, a normalized database will have separate fields for first, middle, last, and sometimes preferred and maiden names. OLTP databases will typically be normalized while OLAP not be fully normalized. OLAP Online Analytical Processing is a database architecture specializing in selecting information. Reporting systems and data warehouses will typically use an OLAP architecture. OLTP Online Transaction Processing is a database architecture specializing in adding, updating, and deleting records but is not efficient in selecting records (not good for reporting). Applications used by operations (e.g., registration) will typically use an OLTP architecture. Primary Key A field or combination of fields that uniquely identify a record in a table. Record Similar to a row in a spreadsheet, in contains a data set representing a single item. terminaliq.com!1 of !2
- 21. The Art of Requesting Data from IT Presented by Brad Adams Term Definition SQL Structured Query Language (pronounced Sea-Quill) is a common program language used to manipulate databases. Many commercial database platforms will have their own flavor of SQL such as pl/sql for Oracle and T-SQL for Microsoft SQL Server. Table A collection of fields and records. Think the rows and columns of a spreadsheet. View A saved query that can be accessed similiar to a table. EDW Enterprise Data Warehouse is where all of your data should reside and should make it easy to get and create useful, timely, reports. Recommended Reading I’ve purchased these two books for new team members multiple times and highly recommend them to anyone who needs to better understand databases. If you use the links below I’ll earn about $1 on each book. • Database Design for Mere Mortals (3rd Edition) - Michael Hernandez http://amzn.to/1CWMm18 • SQL Queries for Mere Mortals (3rd Edition) - Michael Hernandez http://amzn.to/1rgXRLr terminaliq.com!2 of !2