FAQ.pdf

FAQ Questions Scientific analysis and mathematical modeling of clinical research data

Questions are grouped by the attachment they most closely relate to.

Organization Summary

• Combined Synopsis-Solicitation - 12 questions

• Evaluation Criteria - 2 questions

• Statement of Work - 33 questions

Combined Synopsis-Solicitation

Evaluation Method / Basis for Award

1. The Combined Synopsis/Solicitation references an evaluation "most advantageous to the

Government, price and other factors considered" under FAR 52.212-2, while the Evaluation Criteria attachment specifies a Lowest Price Technically Acceptable (LPTA) basis for award. Please confirm that the LPTA basis in the Evaluation Criteria attachment governs.

A. LPTA Governs

2. Could you confirm the basis for award? The synopsis references FAR 52.212-2 tradeoff language, while the evaluation addendum specifies LPTA. We want to make sure we respond correctly.

A. LPTA

3. The Combined Synopsis/Solicitation describes a best-value tradeoff under FAR 52.212-2, while the Evaluation Criteria attachment specifies Lowest Price Technically Acceptable (LPTA). Which evaluation method will control?

A. LPTA Governs

Set-Aside Status

4. Paragraph (iv) of the Combined Synopsis/Solicitation states that the acquisition is unrestricted, while SAM.gov and item 21 identify a total small-business set-aside. Please confirm the applicable set-aside status. This acquisition is a Total Small Business Set- Aside under FAR 19.5. The reference in paragraph (iv) of the Combined Synopsis/Solicitation stating that the solicitation is “released unrestricted” was included in error. Offerors should follow the set-aside status identified in SAM.gov and item 21 of the Combined Synopsis/Solicitation.

5. Set-aside status. Item (iv) states the solicitation is "released unrestricted," while item (21) lists "Set-aside Status: Small Business." Please confirm whether this is a total small-business set-aside or unrestricted. This acquisition is a Total Small Business Set-Aside under FAR 19.5. The reference in paragraph (iv) of the Combined Synopsis/Solicitation stating that the solicitation is “released unrestricted” was included in error. Offerors should follow the set-aside status identified in SAM.gov and item 21 of the Combined Synopsis/Solicitation.

Missing Addenda / Attachments

6. The synopsis references addenda to FAR 52.212-1 and 52.212-4 and an Attachment 5 containing applicable clauses. Only the Combined Synopsis/Solicitation, Evaluation Criteria, and Statement of Work currently appear on SAM.gov. Will the referenced addenda and Attachment 5 be posted? A. The referenced addenda and Attachment 5 was a mistake.

Please disregard.

Submission Method / Administrative Requirements

7. Please confirm the quotation submission method and required administrative documents, including whether quotations should be emailed directly to you and whether any specific subject line or quotation form is required.

A. Documents should be emailed directly to David Shen David.Shen@nih.gov ; include Solicitation ID number (75N98026Q00752) in the subject line. See the Evaluation Criteria attachment for submission instructions.

Incumbency / New Requirement

8. Is this a new requirement, or a follow-on to prior work?

9. Incumbency: Is this a new requirement, or a follow-on to prior work? If a follow-on, can the

Government share the prior contract number?

A. No previous PO or Contract. This is a new Requirement.

Budget / Cost Estimate

10. What is the independent government cost estimate or an anticipated budget range for this requirement?

A. The Government will not release the Independent Government Cost Estimate or anticipated budget range. Offerors are responsible for proposing pricing based on their technical approach, the Statement of Work, and the solicitation requirements. The Government will evaluate price for fairness and reasonableness in accordance with the solicitation.

11. Budget: Can the Government share the independent government cost estimate or an anticipated budget range for this requirement?

A. The Government will not release the Independent Government Cost Estimate or anticipated budget range. Offerors are responsible for proposing pricing based on their technical approach, the Statement of Work, and the solicitation requirements. The Government will evaluate price for fairness and reasonableness in accordance with the solicitation.

Evaluation Criteria

Page Limits / Format / Attachments

1. Page limits and format. Does the 8-page combined limit for Volumes I and II exclude a cover page, table of contents, and acronym list? Aside from the Minimum Performance Standards Checklist, pricing table, and resumes/CVs, are additional excluded attachments (e.g., a brief technical appendix or publications list) permitted.

A. The 8-page combined narrative limit includes any cover page, table of contents, acronym list, tables, and graphics submitted as part of Volumes I and II. Items excluded from the page limit are limited to those specifically identified in the posting.

mailto:David.Shen@nih.gov

Pricing / Invoicing / Payment

2. Should offerors assume invoicing and payment upon Government acceptance of D1 through D6? If so, will NIH provide an anticipated review period for deliverables other than D1?

A. The contract award type is firm-fixed price. Invoicing and payment is not dependent on acceptance of D1-D6.

Statement of Work

Government-Furnished Dataset Scope / Availability / Data Characteristics

1. What is the approximate size of the dataset? A. Dataset size will be approximately < 100

GB.

2. Are simultaneous SPG/PPG data available and when would it be available. A. Yes, SPG and PPG data were acquired simultaneously from each patient and will be made available starting at the planning stage and throughout period of performance.

3. How many participants are included and how many samples per modality are collected?

A. Main study recruited from n = 33 participants. Data was collected at 250 Hz (SPG/PPG) and 100 Hz (blood pressure). Duration varies between 15-35 minutes for each measurement session. Participants were asked to complete a series of physiological challenges to alter their pressure (i.e. cold pressor, Valsalva, mental arithmetic, etc) and had the option to opt out of challenges they did not want to perform.

4. Will Government-furnished SPG data be available during the period of performance, and if so, roughly when? We'd like to plan D3/D4 accordingly. A. Yes, SPG and PPG data will be made available starting at the planning stage and throughout period of performance.

5. Dataset. Is a paired SPG and PPG dataset available for this effort? If so, could you share how many subjects/records it includes, whether the SPG and PPG signals are recorded simultaneously from the same subjects, and when it would be provided relative to kickoff?

A. Yes, SPG and PPG data were acquired simultaneously from each patient and will be made available starting at the planning stage and throughout period of performance.

Main study recruited from n = 33 participants. Data was collected at 250 Hz (SPG/PPG) and 100 Hz (blood pressure). Duration varies between 15-35 minutes for each measurement session. Participants were asked to complete a series of physiological challenges to alter their pressure (i.e. cold pressor, Valsalva, mental arithmetic, etc) and had the option to opt out of challenges they did not want to perform.

6. What is the approximate total size of the Government-furnished SPG/PPG datasets? A rough order of magnitude (GB vs. TB) is fine. A. Dataset size will be approximately < 100

GB.

7. Approximately how many subjects and recordings are included, and roughly how many samples per modality (SPG, PPG, and any reference blood-pressure signal)? A. Main study recruited from n = 33 participants. Data was collected at 250 Hz (SPG/PPG) and 100 Hz (blood pressure). Duration varies between 15-35 minutes for each measurement session. Participants were asked to complete a series of physiological challenges to alter their pressure (i.e. cold pressor, Valsalva, mental arithmetic, etc) and had the option to opt out of challenges they did not want to perform.

8. Government-furnished datasets: Can the Government describe the scope of the de-identified datasets to be furnished — approximate number of subjects/recording sessions, SPG/PPG sampling rates, and recording durations? Will the datasets be available to the contractor at award, or on a later schedule? A. Main study recruited from n = 33 participants. Data was collected at 250 Hz (SPG/PPG) and 100 Hz (blood pressure).

Duration varies between 15-35 minutes for each measurement session. Participants were asked to complete a series of physiological challenges to alter their pressure (i.e. cold pressor, Valsalva, mental arithmetic, etc) and had the option to opt out of challenges they did not want to perform. Yes, SPG and PPG data will be made available starting at the planning stage and throughout period of performance.

9. What Government-furnished datasets are anticipated? If available, please identify the approximate number of participants or records, approximate data volume, file formats, sampling rates, and whether the data include raw SPG image/video data or derived SPG waveforms, synchronized PPG/ECG, reference blood-pressure measurements, event labels, and participant-level metadata. A. Yes, SPG and PPG data will be made available starting at the planning stage and throughout period of performance. Study recruited from n = 33 participants. Data was collected at 250 Hz (SPG/PPG) and 100 Hz (blood pressure). Duration varies between 15-35 minutes for each measurement session.

Dataset size will be approximately < 100 GB. Raw data is composed of text files housing the derived SPG and PPG waveforms as well as ECG data. Participants were asked to complete a series of physiological challenges to alter their pressure (i.e. cold pressor, Valsalva, mental arithmetic, etc) and had the option to opt out of challenges they did not want to perform. Event labels will be included to distinguish between different physiological challenges. Participant meta-data will include age and gender.

Blood Pressure Reference / Ground Truth / Calibration

10. What was the reference method used for the blood pressure values. Were they time-synced to the SPG/PPG waveforms. A. Reference blood pressure values were obtained with a continuous non-invasive blood pressure (NIBP) monitor. Blood pressure values were not precisely synced to the SPG/PPG waveforms.

11. What is the blood pressure range represented in the dataset. Is per-subject calibration anticipated? A. Blood pressure ranges acquired were approximately between 100/60 to 160/110 mmHg. Per-subject calibration is anticipated and can be discussed during planning stage.

12. Blood-pressure reference. For the reference blood pressure used to train and evaluate the models (D4), could you clarify the measurement method (e.g., cuff or arterial line), whether it is time-aligned to the waveforms, and the blood-pressure range represented across subjects? Confirmation of whether per-subject calibration is anticipated would also be helpful. A. Reference blood pressure values were obtained with a continuous non-invasive blood pressure (NIBP) monitor. Blood pressure values were not precisely synced to the SPG/PPG waveforms. Blood pressure ranges acquired were approximately between 100/60 to 160/110 mmHg. Per-subject calibration is anticipated and can be discussed during planning stage.

13. Ground-truth reference: What reference method was used to obtain the blood pressure values paired with the SPG/PPG recordings (e.g., oscillometric cuff, auscultatory, finger arterial pressure/continuous NIBP)? This bears directly on achievable validation against the example accuracy thresholds in D4. A. Reference blood pressure values were obtained with a continuous non-invasive blood pressure (NIBP) monitor. Blood pressure values were not precisely synced to the SPG/PPG waveforms. Blood pressure ranges acquired were approximately between 100/60 to 160/110 mmHg. Per-subject calibration is anticipated and can be discussed during planning stage.

14. Blood-pressure reference standard (D4): What BP reference method and dataset will be used to evaluate MAE/RMSE and Bland–Altman limits of agreement (e.g., an ISO 81060-2 reference), and will paired reference measurements be provided with the dataset? A.

Reference blood pressure values were obtained with a continuous non-invasive blood pressure (NIBP) monitor. Blood pressure values were not precisely synced to the SPG/PPG waveforms. Blood pressure ranges acquired were approximately between 100/60 to 160/110 mmHg. Per-subject calibration is anticipated and can be discussed during planning stage. Paired reference measurements will be provided

Storage / Computing Environment / Cloud / Local Processing

15. What is the total storage capacity anticipated. A. Estimated storage capacity for source data and any intermediate data generated would be approximately 250 GB at minimum.

16. What is the reference computing environment used for real-time demonstration? A. Initial specs would be targeted for a standard Microsoft surface pro 10 (Windows 11). More specific hardware constraints can be discussed during planning stage

17. Is there a reference hardware/environment spec for the D5 real-time demonstration (target sampling rate, latency, CPU/memory)? A. Initial specs would be targeted for a standard Microsoft surface pro 10 (Windows 11). More specific hardware constraints can be discussed during planning stage

18. What total storage capacity should the contractor plan to provision for the source data plus derived and intermediate data? An approximate figure is fine. A. Estimated storage capacity for source data and any intermediate data generated would be approximately 250 GB at minimum.

19. Reference computing environment: Section 8 of the SOW lists a reference computing environment specification as Government-furnished. Can the Government share that specification, including operating system, available compute/GPU, and any container or Python version constraints relevant to the D5 demonstration package? A. Initial specs would be targeted for a standard Microsoft surface pro 10 (Windows 11). More specific hardware constraints can be discussed during planning stage

20. For D5, what reference hardware, operating system, software environment, and input-stream interface should offerors assume? Is a containerized demonstration using a recorded or simulated sample stream acceptable? A. Initial specs would be targeted for a standard Microsoft surface pro 10 (Windows 11). More specific hardware constraints can be discussed during planning stage. Preference would be to test on a live demonstration but this can be negotiated during planning stages.

21. Does NIH anticipate providing access to an NIH-hosted analysis environment, or should offerors price a contractor-furnished environment? Please also confirm whether FedRAMP Moderate is required only if CUI is processed in a contractor-furnished cloud environment.

A. Our intent will be to provide de-identified data to the contractor to process on contractor-provided infrastructure. However, if alternate method would facilitate contractor access, this can be discussed during kickoff phase.

22. SOW section 10 identifies NIH cloud-based systems as a potential mechanism for accessing Government-furnished de-identified data, and requires any contractor cloud use for CUI to be FedRAMP Moderate (or higher). Could you please clarify: A. Our intent will be to provide de-identified data to the contractor to process on contractor-provided infrastructure. However, if alternate method would facilitate contractor access, this can be discussed during kickoff phase.

23. Will the Contractor be granted access to perform storage and computation directly within an NIH-provided cloud environment for this effort, or will de-identified data instead be furnished to the Contractor for processing in Contractor-provided infrastructure? A. Our intent will be to provide de-identified data to the contractor to process on contractor-provided infrastructure. However, if alternate method would facilitate contractor access, this can be discussed during kickoff phase.

24. If data is furnished to the Contractor, may it be downloaded to and processed locally on a Contractor-managed system outside an NIH cloud environment, and if so, what safeguarding requirements apply to that system? A. Yes, this data is not considered sensitive; The data is to be used only for the purposes of the deliverables in the SOW.

25. Reference environment and cloud. Will the Government provide the reference computing-environment specification (for the D5 <100 ms at 60 Hz target) at kickoff? Is performing the work in the ≥ contractor's own controlled environment on de-identified data acceptable, or is use of an NIH FedRAMP Moderate cloud required? A. Government will provide reference computing specifications at kickoff. Initial specs would be targeted for a standard Microsoft surface pro 10 (Windows 11). More specific hardware constraints can be discussed during planning stage.

Acceptance Criteria / Performance Metrics / Predefined Markers

26. Are the acceptance criteria (MAE ≤7 mmHg systolic / ≤5 mmHg diastolic, etc) in D5 binding or are they example values? A. These targets are example research targets, acceptance criteria will be discussed and agreed upon during planning stage.

27. For D4, are the accuracy figures (MAE ≤7 mmHg systolic / ≤5 mmHg diastolic) firm acceptance thresholds, or example research targets? A. These targets are example research targets, acceptance criteria will be discussed and agreed upon during planning stage.

28. Acceptance targets. The SOW lists example thresholds for some deliverables e.g., MAE ≤ 7/5 mmHg for D4, r ≥ 0.6 for D3 and refers to certain criteria as "predefined" or "agreed."

Could you confirm whether those example values are the binding acceptance targets, and how the remaining criteria - such as the D2 signal-quality targets and the specific D3 markers - would be finalized (for example, agreed during the D1 Project Plan)? A. These targets are example research targets, acceptance criteria will be discussed and agreed upon during planning stage.

29. For D2 and D4, will NIH define the artifact-suppression/SNR criteria and agreed datasets after award? Are the stated blood-pressure thresholds (7 mmHg systolic MAE, 5 mmHg diastolic MAE, and limits of agreement within +/-15 mmHg) illustrative research goals or mandatory acceptance thresholds? A. These targets are example research targets, acceptance criteria will be discussed and agreed upon during planning stage.

30. Predefined markers and metrics (D3/D4). Will the Government specify the "at least three predefined markers," the agreed metrics, and the datasets at kickoff, or should offerors propose them in their technical approach? A. Government will provide examples of the markers. Performance metrics and datasets will be provided during kickoff.

Software / Tools / Code / Test Materials

31. Are there any restrictions on software, programming languages, libraries, etc the contractor may use for security or compliance? A. While we do not anticipate any restrictions on software, any questions about compliance will be evaluated on a case by case basis and directed to NIH IT for confirmation during kickoff.

32. Are there any restrictions on the software, programming languages, libraries, or tools the contractor may use (e.g., limits on open-source components, license requirements, or prohibited tools) for security or compliance? A. While we do not anticipate any restrictions on software, any questions about compliance will be evaluated on a case by case basis and directed to NIH IT for confirmation during kickoff.

33. Will NIH furnish existing signal-generation or preprocessing code, data dictionaries, baseline algorithms, reference outputs, internal test protocols, or unit tests? A. NIH will supply preliminary versions of pre-processing code but the bulk of analysis code will be expected to be developed by offerors.