National Institutes Of Health Tender

Contracting Office Address the Department Of Health And Human Services, National Institutes Of Health, National Cancer Institute, Office Of Acquisitions, 9609 Medical Center Drive, Rockville, Md 20850, United States 1.0 Description the Division Of Cancer Control And Population Sciences (dccps), Surveillance Research Program (srp), Plans To Procure, On A Sole Source Basis, From Jason P. Fine, 202 Main Street, Wakefield, Massachusetts 01880. the Response Close Date Of The Notice For This Requirement Is In Accordance With Far 5.203(b). This Acquisition Will Be Processed Under Far Part 12 - Acquisition For Commercial Items And Will Be Made Pursuant To The Authority In Far Part 13.106-1(b)(1)(i); And Is Exempt From The Requirements Of Far Part 6. The North American Industry Classification System Code Is 541715 And The Small Business Size Standard Is 1,000 Employees. it Has Been Determined There Are No Opportunities To Acquire Green Products Or Services For This Procurement. 2.0 Background the National Cancer Institute (nci), Surveillance Research Program (srp) Supports Cancer Surveillance Programs And Research, Including The Surveillance, Epidemiology, And End Results (seer) Program. Seer Consists Of Population-based Registries Covering Nearly 50% Of The Us Population. Srp Releases Annual Reports On Incidence, Survival, Treatment, And Mortality Trends Using Seer Databases To Highlight The Cancer Burden In The Population. These Reports Are Made Widely Available And Are Often Cited In The General Media And Medical Literature. in Many Of These Reports, Survival Information Has Been Presented As Net Survival, Which Measures The Likelihood Of Dying Of Cancer-related Causes In The Absence Of Non-cancer Death. Net Survival Has Utility As A Cancer Control Measure, But It Is Less Useful For Cancer Patients And Clinical Decision Makers As Cancer Patients Are At Risk Of Dying Of Non-cancer Causes. Survival Measures That Can Account For Competing Risks Such As Other-cause Death Are Of Great Value To Patients And Clinicians To Make Informed Treatment Decisions. For Example, Patients With Multiple Comorbidities May Have An Increased Risk Of Non-cancer Death That Would Justify Not Pursuing Aggressive Cancer Treatment. Statistics That Account For Competing Causes Of Death May Be Calculated As Cumulative Probabilities Of Cancer And Non-cancer Death Along With The Probability Of Survival. srp Has Prioritized The Development Of Models To Better Utilize Seer Data And Estimate Competing Risks Survival Models. The Goal Of Such Models Is To Develop Tools That Provide Clinically Relevant Survival Information To Physicians And Individual Patients. The Development Of Methods That Can Produce More Accurate Measures Of The Risks Of Cancer And Non-cancer Death Are Of Great Interest To Srp. standard Competing Risks Analysis Models Cause-specific Hazard Functions From The Time Of Diagnosis For Both Cancer And Non-cancer Mortality Via The Proportional Hazards Model In Continuous Time, As Described In Cheng, Fine And Wei (biometrics, 1999). This Approach Was Later Adapted By Lee Et Al (2017, Biostatistics) To Model Cancer Mortality On The Time Since Diagnosis Scale And Model Other Cause Mortality On The Age Time Scale, Which Is A More Natural Scale To Model The Risk Of Non-cancer Death. Later, Lee Et Al (2019, Statistics In Medicine) Extended The Two Time Scale, Continuous-time Method To Discrete Time And Proposed A Flexible And Computationally Convenient Framework. In The Prior Work By Lee Et Al (2017, Biostatistics; 2019, Statistics In Medicine), The Other-cause Death Model Accounted For Left Truncation In The Estimation As Patients Become At Risk Only After Diagnosis. In A Previous Contract, The Contractor Extended The Discrete-time Methods Of Lee Et Al (2019, Statistics In Medicine) To The Continuous Time Setting Using Parametric Models. Continuous Time Is Advantageous As It Offers A Computationally Simpler And More Straight-forward Framework Than Discrete Time. Furthermore, The Majority Of Risk Calculators And Clinical Nomograms Are Presented In Continuous Time. To Improve The Application And Utility Of The Continuous-time Competing Risks Models, Enhancements Are Needed. one Such Enhancement Is To Develop And Explore Measures Of Predictive Accuracy Or Prediction Error For Competing Risks Models. Predictive Accuracy Refers To The Ability Of The Model To Produce Individual Predictions That Are Close To The Observed Values. In The Competing Risks Setting, Prediction Can Be Quantified For Each Cause (e.g. Cancer, Other Causes) And Overall, Both At Specific Time Points And Averaged Across All Times. There Has Been Limited Work In This Area, With Most Prior Work Focusing On Estimating Prediction Error For A Single Cause At A Particular Time Point. Methods To Produce An Overall Estimate Of Prediction Error Accounting For Both Causes Have Not Been Widely Disseminated. Quantifying Prediction Error Is Important As It Provides A Measure Of “confidence” In Model Predictions. Srp Is Interested In Developing Risk Prediction Tools For The General Public, And Including Measures Of Prediction Error In Such Tools Is Necessary To Provide Users With An Idea Of How Accurate The Predictions Produced By The Tool Are. 2.1 Objective the Purpose Of This Contract Is To Develop Extensions To Methods For Risk Prediction For Continuous-time Competing Risks Survival Models On Two-time Scales. The Competing Risks Models May Be Parametric Or Semi-parametric. Specifically, The Contract Aims To Extend The Competing Risk Models By Exploring Measures To Summarize Prediction Error. The Innovation Of This Contract Is To Provide Srp With Measures To Assess How Well The Competing Risks Models Can Predict The Risk Of Cancer And Other-cause Death. Srp Expects That The Methods In This Contract Will Be Used In Reports Or Web-based Tools To Contextualize The Probabilities Of Dying Of Cancer, Dying Of Other Causes, And Surviving That Will Be Useful For Clinical Decision Making. 3.0 Scope this Contract Shall Extend Methods For Risk Prediction With A Continuous Time Model For Survival Under Competing Risks Using Two-time Scales: Age For Other Causes Of Death And Time Since Diagnosis For Cancer Death. For The Competing Risks Models, The Contractor Shall Implement Both Semiparametric And Parametric Models. With Parametric Models, The Gompertz Or Piecewise Gompertz Baseline Hazard Will Be Used By The Government And Contractor For Other-cause Mortality Modeling. For Quantifying Prediction Error, The Government And Contractor Will Explore Time-dependent And Overall Measures Based On A Simple Brier Score. In Addition, Both Parties Will Also Develop Cause-specific And Overall Predictive Accuracy Measures. The Deliverables From This Work Will Be An Outline Of The Methodology For This Approach, Along With Associated Software Developed To Implement This Approach. 4.0 Contract Requirements/ And Personnel Qualifications the Contractor Shall Accomplish The Following Objectives: 4.1 Develop The Methodology For Prediction Error For Competing Risks Models On Two Time Scales the Contractor Shall Develop The Methodology For Summarizing Cause-specific And Overall Prediction Error For A Competing Risks Survival Models Fit On Two-time Scales. Models Will Be Fit With Other-cause Mortality Modeled On The Age Scale And Cancer Mortality Modeled On The Time Since Diagnosis Scale. The Competing Risks Model May Be Parametric Or Semiparametric. 4.1.1 Develop Predictive Accuracy Measures For Competing Risks the First Objective (ob1) Is To Develop Measures Of Predictive Accuracy For Continuous Time Competing Risk Models On Two-time Scales. Task 1.1 Of Ob1 Is To Specify And Fit The Continuous Time Models On Two-time Scales. Task 1.2 For Ob1 Is To Define Cause-specific And Overall Prediction Error Based On The Continuous Time Models Fit In Task 1.1. Task 1.3 For Ob1 Is To Study The Numerical And Statistical Properties Of The Measures Developed In Task 1.2. The Numerical Comparisons Will Use Artificially Generated Data And Real Datasets. 4.2 Develop The Software Program For Calculating Prediction Error the Contractor Shall Develop A Reproducible Software Program To Calculate The Prediction Error Measures Developed In Objective 4.1. The Program Will Be Implemented By The Government And Contractor In Standard Statistical Software, Such As Sas, R, Or Stata. The Program Will Allow Future Users To Calculate Prediction Error For An Existing Competing Risks Model On Two Time Scales. The Program May Also Include Functions To Generate Visualizations Based On The Data Or Model Output. The Program May Depend On Standard Functions, Functions Available Through Libraries, Or Original Code. 4.2.1 Develop Reproducible Software Program In Statistical Software: the Second Objective (ob2) Is To Provide User Friendly Software Implementing The Prediction Error Calculations Developed In Ob1. Task 2.1 For Ob2 Is To Produce A Reproducible Software Program Or Function To Implement The Proposed Measures. Such Measures Will Be Implemented In Standard Statistical Software (r, Sas) And Will Allow Future Users To Fit Competing Risks Models On Two-time Scales, Obtain Risk Predictions, And Derive The Cause-specific And Overall Error In Prediction For A Given Dataset. It Will Include Functions To Generate Visualizations Of The Data And Model Output. It Will Use Standard Functions Available Through Libraries Where Possible, And Original Code Where Standard Functions Are Not Available. 4.3 Summarize Findings In A Technical Report the Contractor Shall Summarize The Methodologies And Results In A Technical Report That Should Be Suitable For Publication In A Statistical Or Medical Research Journal. The Contractor Should Outline How The Methods Developed In This Contract Are Novel And Useful For Quantifying The Predictive Accuracy Of Competing Risks Models. The Report Should Include Both Theoretical Results And An Application To A Cancer-related Data Set. 4.3.1 Generate Manuscript Summarizing Major Findings : the Third Objective (ob3) Is To Provide Reports On The Probabilities Of Dying Of Cancer, Dying Of Other Causes, And Surviving That Will Be Useful For Clinical Decision Making, Along With Associated Prediction Errors. Task 3.1 For Ob3 Is To Write A Technical Report Describing The Proposed Predictive Accuracy Measures For Continuous Time Risk Predictions And The Numerical Results, Including Probabilities Of Cancer And Other Cause Death From Real Datasets. Task 3.2 For Ob3 Will Be To Write A Technical Report Describing The Statistical Software Implemented From Ob2. These Technical Reports Will Be Disseminated In Peer Reviewed Journals. 5.0 Type Of Order this Is A Firm Fixed Price Purchase Order. 6.0 Period Of Performance The Period Of Performance Shall Be: Base Period: 12 Months From The Date Of Award. Option Period One: 12 Months Following The Base Period. 7.0 Place Of Performance the Work Shall Be Performed At The Contractor’s Facility. 8.0 Non-government Information/documents the Contractor May Develop The Prediction Error Measures In A Standard Statistical Software Program Such As R, Sas, Or Stata. Depending On Which Language The Contractor Chooses, They May Implement Their Methods As A Set Of Functions In A Singular Program, As A Separate Library Or Package, Or As A Macro Function. Regardless, If The Contractor Would Like To Submit The Programs Associated With The Contract To A Wider Audience, Such As Through An R Package, The Program Must Adhere To Standards Set Forth By Cran (https://cran.r-project.org/submit.html). 9.0 Report(s)/deliverables And Delivery Schedule all Written Deliverables Shall Be Sent Electronically To The Nci Technical Point Of Contact (tpoc), Tbd At Award, In Appropriate Formats (including But Not Limited Bam, Cram, Vcf, Excel, Doc, Csv, Pdf, Png, Etc.) Unless Approved By The Tpoc In Accordance With The Deliverable Schedule Below. The Tpoc Shall Review The Contents Of All Draft Deliverables. If No Comments Or Requests For Revisions Are Provided To The Contractor Within 30 Business Days, The Deliverables Shall Be Considered Acceptable. If Revisions Are Required, Nci Shall Respond To The Contractor Within Five (5) Business Days Of Receiving The Deliverable, Specifying The Required Changes/revisions. Final Copies Of Approved Drafts Shall Be Delivered To The Nci Tpoc Within Five (5) Business Days After Receipt Of The Government’s Comments. the Tpoc For This Order Is: (tbd) all Deliverables Shall Be Sent Electronically Per The Following Deliverable Schedule: deliverable Deliverable Description / Format Requirements Due Date base Period base Period objective 4.1 outline Of The Methodology For Quantifying Cause-specific And Overall Prediction Error For Competing Risks Model On Two-time Scales. 11 Months After Award base Period objective 4.2 software Program(s) For Implementation Of The Proposed Methodology. 11 Months After Award option Period One option Period One objective 4.2 software Program(s) For Implementation Of The Proposed Methodology. 11 Months After Award option Period One objective 4.3 technical Report On The Proposed Methodology That Includes An Application To A Cancer Data Set. 11 Months After Award 10.0 Unique Qualifications Of The Contractor jason Fine’s (“fine”) Unique Expertise In The Field Of Competing Risks And Prior Work With Nci Ensure That He Is The Only Contractor Who Can Accomplish The Aims Of The Proposed Contract To Develop And Apply Prediction Error Measures To Complex Competing Risks Data In A Timely And Efficient Manner. Dr. Fine Has A Deep Understanding Of Two Statistical Research Areas Required To Fulfill The Requirements Of This Order: 1) The Statistical Framework Of Competing Risks When Considering Cancer Mortality And Other-cause Mortality, And 2) The Statistical Aspects Of Developing Competing Risks Survival Models Incorporating Multiple Time Scales. dr. Fine Has Demonstrated Capability In Both Of The Required Subject-matter Areas And Has Published Numerous Methodological Papers In Competing Risks, Including A Seminal Paper In The Field (fine And Gray, 1999). Many Papers Published By Dr. Fine In Recent Years Have Focused On Competing Risks Models Using Two Time Scales And Have Been Applied To Seer Data (lee Et Al 2017, 2018, 2019). Dr. Fine Has Also Previously Developed The Methodology And Expertise To Produce Other-cause Mortality Estimates For The Surveillance Research Program That Has Been Published In High-impact Journals (https://seer.cancer.gov/survivalcalculator/). To Ensure The Consistency Of The Nci’s Methodology And Thus The Consistency Of The Results Of The Institute’s Research, Similar Methodologies Must Be Developed For This Project. information On Survival In Reports Based On Seer Data Has Been Net Survival, I.e., Likelihood Of Dying Of Cancer In Absence Of Other Causes Of Death. While Useful As A Cancer Control Measure, Net Survival Is Not As Relevant For Cancer Patients At Risk Of Dying Of Non-cancer Related Causes. in Addition To Considering Competing Risks, We Must Carefully Consider The Time Scale For Analysis. Cancer Mortality Can Be Modeled From The Time Since Diagnosis, But It Is More Natural To Model Non-cancer Mortality On The Age Scale. Other Considerations Include Whether To Model Time Continuously Or Discretely. while There Are Other Sources Working On Competing Risks Models, These Sources Lack The Understanding And Familiarity With Both The Theoretical Models And The Data Source Needed For This Contract, Namely Working With Competing Risks Models On Two Time Scales. This Is Evidenced By The Fact That The Most Recent Papers In This Specialized Area Of Competing Risks Analysis Have All Involved Dr. Fine As A Senior Author Responsible For Overseeing The Development Of The Methodology (lee Et Al 2017, 2018, 2019). procuring These Services From A Different Source Would Result In Delays To The Overall Project, As Any New Contractor Would Need To Develop The Requisite Expertise In Competing Risks And Become Familiar With The Previous Efforts From Previous Orders. Dr. Fine Has Developed The Methodology That Will Be Used In The Proposed Contract, And To Extend This Methodology, It Is Imperative To Employ Him As The Contractor. 11.0 Submission Instructions this Notice Is Not A Request For Competitive Quotations. However, If Any Interested Party Believes It Can Meet The Above Requirements, It May Submit A Proposal Or Quote For The Government To Consider. The Response And Any Other Information Furnished Must Be In Writing And Must Contain Material In Sufficient Detail To Allow Nci To Determine If The Party Can Perform The Requirement. All Responses And Questions Must Be Sent Via Email To The Contract Specialist, Dana Summons, At Dana.summons@nih.gov By No Later Than 3:00 Pm Est On July 29, 2024, (7/29/24). A Determination By The Government Not To Compete This Proposed Requirement Based Upon Responses To This Notice Is Solely Within The Discretion Of The Government. Information Received Will Be Considered Solely For The Purpose Of Determining Whether To Conduct A Competitive Procurement. In Order To Receive An Award, Contractors Must Be Registered And Have Valid Certification Through Sam.gov. Reference: 6957372 On All Correspondence.