Loren Data's SAM Daily™

FBO DAILY ISSUE OF AUGUST 17, 2008 FBO #2456
SOURCES SOUGHT

D -- Comprehensive Data Mining

Notice Date

8/15/2008
 

Notice Type

Sources Sought
 

NAICS

541511 — Custom Computer Programming Services
 

Contracting Office

Department of Health and Human Services, Food and Drug Administration, Office of Acquisitions and Grants Services, 5630 Fishers Lane, Room 2129, Rockville, Maryland, 20857-0001
 

ZIP Code

20857-0001
 

Solicitation Number

1042160
 

Response Due

8/20/2008
 

Point of Contact

Tawana Parker,, Phone: 301 827-7152, Ted L Weitzman,, Phone: 301-827-7178
 

E-Mail Address

tawana.parker@fda.hhs.gov, ted.weitzman@fda.hhs.gov
 

Small Business Set-Aside

N/A
 

Description

This is a SOURCES SOUGHT NOTICE to determine the availability and capability of qualified small businesses. The U.S. Department of Food & Drug Administration (FDA) proposed to apply WebVDME (web-based Visual Data Mining Environment), a data mining tool developed under a (Food and Drug Administration) FDA Cooperative Research and Development Agreement (CRADA) with Lincoln Technologies, Inc. to the Manufacturer and User Facility Device Experience (MAUDE) data at the Center for Devices and Radiological Health (CDRH). WebVDME supports the detection and quantification of safety signals in post-marketing data. The software utilizes advanced data mining techniques, including: Multi-Gamma Poisson Shrinker (MGPS), Proportional Reporting Ratio (PRR), Reporting Odds Ratio (ROR) and Logistic Regression. WebVDME is currently being used to detect signals of adverse events in the Adverse Event Reporting System (AERS) database and has been applied to the Vaccine Adverse Event Reporting System (VAERS) data. It is also being piloted at the FDA to analyze medical device adverse events reported into the MAUDE database. The contractor shall customize the software, data preparation and provide refreshing processes, training, and consulting support. An important component of this project is cleaning and preparing the MAUDE data which will provide FDA with an improved capability to extract actionable intelligence from its data. Device information from safety reports is captured in the MAUDE database as a verbatim representation of the original text in the report. As a result, the same device may be referred to by many different names, for example with or without the manufacturer name, type number or other specifics, as well as a variety of possible misspellings. In order to be able to consistently identify a specific device or device class for data mining purposes, this data will need to be cleaned up. The proposed approach for cleaning up the MAUDE data is based on proven techniques that are currently being applied on a monthly basis to internally released data from the AERS database, and which include mapping-based automatic spelling correction, synonym mapping and other data transformations. After such an approach has been developed and implemented, it can be re-applied on a regular basis with ever-increasing efficiency due to the gradual increase in size of mapping and synonym tables, covering an ever greater proportion of possible misspellings and synonyms. Once the cleanup has been completed it will be possible to apply quantitative signal detection against the cleaned MAUDE database at various aggregation levels dependent on the desired granularity of the result. During the data mining pilot the MAUDE database has been found to contain a significant number of duplicate reports which could negatively influence the reliability of the data mining results. Therefore the cleaning process will also include identification of duplicate reports in order to be able to exclude such reports from data mining. The contractor shall provide the following tasks: The Contractor shall assist FDA in developing and operating an advanced safety data mining application for use with the FDA Manufacturer and User Facility Device Experience (MAUDE) database which represents reports of adverse events involving medical devices. 1) In consultation with FDA staff, develop data cleaning procedures as necessary to prepare MAUDE for use in data mining, addressing issues such as manufacturer and device brand name normalization, use of hierarchical classification information to support data aggregation at multiple levels, and detection of duplicate reports. The work for coronary stents and oxygenators has been completed. The project will be extended to up to 15 device categories. The following device types will be included: a) Glucose meters and glucose test strips b) Implantable cardiac defibrillators (ICDs) and pacemakers (including leads and generators) c) Fluoroscopes and related accessories d) Intraocular lenses e) Electro-surgical units and accessories Furthermore, manufacturer name normalization will be extended for all device types. 2) MAUDE data for coronary stents and oxygenators has previously been prepared for use in data mining. This data will be refreshed monthly.3) MAUDE data for infusion pumps has been previously prepared for use in data mining. The contractor shall address issues such as device brand name normalization, use of hierarchical classification information to support data aggregation at multiple levels, and detection of duplicate reports. This data will be refreshed monthly. 4) Make up to four snapshots of the MAUDE data available over the course of year 1 for use with WebVDME, including transforming the MAUDE data to conform to the WebVDME data model, creating the WebVDME configuration (metadata) needed for Query and Reporting, and identifying MAUDE data cleaning needs. 5) The contractor shall work with FDA staff to explore alternative data mining techniques (including PRR, MGPS, and logistic regression) to determine which approaches are most productive for use in the generation of safety signals on MAUDE data. 6) Collaborate with FDA staff in the initial investigation of safety issues of interest, including for example safety issues related to ICD lead fractures. 7) Provide updates of the WebVDME software as required. 8) Support an initial cadre of FDA users (up to 20 users) by providing: system administration, training and assistance in the use of the system through telephone support, and up to 4 visits or WebEx remote sessions for support and training. 9) The contractor shall work with FDA OIT/SS staff to set up an FDA-provided dedicated computer server for production use with WebVDME, including installation and configuration of foundation software such as operating system, Oracle database, backup utilities, etc. 10) Install WebVDME safety data mining software on the server. 11) Develop a validated production process for the automated periodic refreshing and cleaning of the MAUDE data and place this process into routine operation (e.g., once a month). 12) Work with OIT/SS staff to prepare the necessary documentation, review and approval steps prior to production rollout. 13) Roll the system out to the full community of MAUDE users, including development and delivery of CDRH-specific on-site training at FDA. 14) Provide regular (e.g., first 6 months quarterly, then monthly) refreshes of the MAUDE data. 15) Provide project management resources to coordinate the above activities. The contractor must have experience in data mining software and the MAUDE adverse event reporting system. This includes prior experience in the development of webVDME and cleaning of MAUDE data to prepare for data mining. The contractor must have experience in the development and use of Bayesian data mining algorithms, including Multi-Gamma Poisson Shrinker (MGPS). The NAICS Code is 541511. The small business size standard is 23M. Interested small business must provide a capability statement demonstrating their experience, skills and capability to fulfill the Government requirement that is stated above. However, be advised that generic capability statements are not sufficient to evaluate your capability statement effectively. Information submitted must be in sufficient detail to permit the government to determine whether your organization meets the government requirement. This notice is for planning purposes only, and does not constitute a Request for Quotation. This request is not to be construed as a commitment on the part of the Government to award a purchase order or contract, nor does the Government intend to pay for any information submitted as a result of this request. The Government will not reimburse respondents for any cost associated with submission of the information being requested or reimburse expenses incurred to interested parties for responses to this announcement. Interested small business organizations that believe they possess the capabilities necessary to undertake this project should submit via email copies of their capability statement, addressing the areas above. Capability statements must identify the business status of the organization. NO COLLECT CALLS WILL BE ACCEPTED. RESPONDENTS MUST SUBMIT CAPABILITY STATEMENT VIA E-MAIL to Tawana Parker at tawana.parker@fda.hhs.gov, no later than August 20, 2008 at 2:00 PM Eastern Time for consideration.
 

Web Link

FedBizOpps Complete View
(https://www.fbo.gov/?s=opportunity&mode=form&id=b6d6f823eda29f4998cb6b00735b1dfd&tab=core&_cview=1)
 

Place of Performance

Address: White Oak Campus, Silver Spring, Maryland, 20993-0002, United States

Zip Code: 20993-0002
 

Record

SN01642115-W 20080817/080815224953-b6d6f823eda29f4998cb6b00735b1dfd (fbodaily.com)
 

Source

FedBizOpps Link to This Notice
(may not be valid after Archive Date)

---

| FSG Index  |  This Issue's Index  |  Today's FBO Daily Index Page |