SOLICITATION NOTICE
A -- Complex Models
- Notice Date
- 8/16/2006
- Notice Type
- Solicitation Notice
- NAICS
- 541690
— Other Scientific and Technical Consulting Services
- Contracting Office
- Department of Health and Human Services, Center for Disease Control and Prevention, Procurement and Grants Office (Atlanta), 2920 Brandywine Road, Room 3000, Atlanta, GA, 30341-4146
- ZIP Code
- 30341-4146
- Solicitation Number
- 2006-Q-09080
- Response Due
- 8/25/2006
- Archive Date
- 8/25/2006
- Description
- This is a combined synopsis/solicitation for services prepared in accordance with the format in the Federal Acquisition Regulation (FAR), Part 12, Acquisition of Commercial Items, as supplemented with additional information included in this notice. The Centers for Disease Control and Prevention intends to award a Firmed Fixed Price purchase order to provide the tasks as listed in the attached Statement of Work. Minimum vendor qualifications are listed in the Statement of Work. This procurement?s destination and delivery will be to Centers for Disease Control, National Center for Health Statistics, Hyattsville, Maryland. The contractor shall deliver all items as outlined in the Statement of Work after receipt of order. The North American Industry Classification System Code is 541690. This notice is a request for competitive proposals. The offeror shall provide pricing for each of the tasks to be performed. Offeror shall also provide proof of credentials as required in the Statement of Work. Firms that believe they can provide the services need to submit written proposals and a statement of capabilities to the Contracting Specialist within 10 days from the publication date of this notice. The Government reserves the right to award any, all, or none of the items above. A determination by the Government not to compete this proposed action based on responses to this notice is solely within the discretion of the Government. Inquires may be sent to Bertha Russell email @ egy2@cdc.gov or by fax at 770-488-2671. Statement of work Complex Models to Estimate Mortalities in Small Domains with 30+ Years? of NCHS Data Requisition I. BACKGROUND The National Center for Health Statistics (NCHS) publishes complete sex and race specific national life tables annually and national and State specific complete life tables decennially. In addition, decennial ?causes of death? life tables are also published with updated ICD codes. NCHS is interested in improving methodologies for generating decennial tables. Methodologies for generating annual life tables are usually revised and updated respecting decennial life tables and then kept consistent for about 10 years until the next decennial life tables? methodology is updated. For decennial life tables 1999-2001, a major attempt is being made to overcome difficulties in estimating mortality in small domains (small states, some causes of death), so that life tables can be published with reliability. To fulfill this goal, research has been focused on methods for smoothing observed death data in small subpopulations by ?borrowing? strengths from other data sources and sophisticated statistical models. A previous contract study successfully established a semiparametric time series model for estimating age-specific mortality by utilizing NCHS 30 years? mortality data. Another recent study showed that the absence of deaths in some small domains are due to extremely low mortality rates and small populations, and therefore should be treated as ?zero? rather than ?missing? data. These ?zero? values should contribute to the calculation of an average of neighborhood (time and/or age) mortality. The concept of ?missing? for insufficient death observations used in previous applications leads to overestimation of mortality rates. To correct this problem, it is proposed to establish some proper probability model for estimating extremely low mortality rates with time series analysis. It is also necessary to develop the corresponding algorithm in concept of maximum likelihood for estimating practices. To optimize estimating results, several sophisticated methods, including Bayesian method, two-dimensional Kriging smoothing and semi-likelihood estimation, will be applied and compared. The result from this research is expected to significantly improve reliability and utility of NCHS decennial life tables. II. PURPOSE 1) Establish an estimating model for mortality rates for domains with ?zero? death observations based on the probability theory and neighborhood observations in a 2-dimensional window space (year x age); 2) Compare the classical maximum likelihood algorithm for mortality estimates with other alternatives or modifications, such as trans-Gaussian kriging, Bayesian transformed Gaussian, semi-likelihood estimation, etc.; 3) Evaluate proposed estimation methodology with NCHS mortality data from small domains such as ?small states? and ?rarer causes of death?; 4) Assist in preparing specifications for use in applications at NCHS. III . SCOPE OF WORK The vendor will provide all labor, materials and equipment required to complete the tasks outlined below. Vendor will deliver to Project Officer quarterly summaries of progress and problems. The vendor maybe required to travel an estimate of 2 trips to or communications with NCHS, (Hyattsville, MD) to obtain background information on the project. All documentation and software will be developed for IBM PC and PC compatible computers operating on a Windows XP Professional platform. Source and executable code and technical and user documentation will be provided to NCHS. Vendor will make interim software developed during the research available to NCHS in preliminary form to continue the research with sufficient instruction in its use to enable its use by NCHS and federal colleagues. Vendor will not have access to confidential NCHS data. Research will not involve testing human subjects. Research directions and goals will be set by project officer. IV. TASKS TO BE PERFORMED 1) Vendor will conduct a literature review to solidify the study base, report literature review on estimation models and corresponding algorithms related to the subject, and provide a summary and comments for reviewed methodologies; 2) Vendor will develop probability models and corresponding algorithms for estimating extremely low mortalities in small domains; 3) Vendor will adapt several alternative estimating methods to different types of life tables (e.g. complete vs. abridged life tables, decennial vs. annual tables); 4) Vendor will assess and compare proposed models and algorithms using NCHS mortality data; 5) Vendor will attendan estimate of 8 -10 work group meetings periodically to update study progresses; 6) Vendor will prepare quarterly and final reports; 7) Vendor will assist senior NCHS program personnel to prepareapproximatley 1-2 research papers suitable for publication in peer reviewed journals. V. GOVERNMENT FURNISHED MATERIALS N/A VI. PERIOD OF PERFORMANCE The period of performance of this contract is from the effective date of contract (EDOC) to 12 months later. VII. DELIVERABLES The vendor shall provide the following deliverables to the Project Officer by the dates reflected below. All electronic text files should be delivered in MS Word format and provided via e-mail. Deliverables and Date Required 1. Description and plan of research by the 15th day after EDOC 2. Report on literature review by the 30th day after EDOC 3. Provide details of proposed research models and alternatives for comparisons by the 40th day after EDOC 4. First interim report on model research by the 120th day after EDOC 5. Evaluation report of properties by the 180th day after EDOC 6. First draft of final report by the 240th day after EDOC 7. Final draft of final report (inclusive of papers for publication) by the 300th day after EDOC 8. Response to peer reviews by the 360th day after EDOC 9. Quarterly progress reports VIII. MINIMUM VENDOR QUALIFICATIONS This study will require a Ph.D, training and expertise in math statistics, especially in time series data modeling and small area estimating. The study will require knowledge of missing data imputation in time series prediction and complex probability estimation. Data analysis will require computing skill and familiarity with issues in population mortality data. The study also will require expertise in research development and consulting practice. The vendor should understand the background of this research and the key issues from the related studies prior to this project. The vendor shall be able to travel several times to Hyattsville, MD to fulfill the requirements of consultation with CDC.
- Place of Performance
- Address: Acqisition & Assistance Field Branch-Hy, 3311 Toledo Road, Hyattsville, MD
- Zip Code: 20782
- Country: UNITED STATES
- Zip Code: 20782
- Record
- SN01116484-W 20060818/060816220531 (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 |