DA002 - Sampling for NDB Enrollment

DA002 Sampling for NDB Enrollment
Review Committee: Planning Start Date: 10/1/2012
Attachments: None Last Revised Date: 8/15/23
Forms: Last Reviewed Date: 8/15/23

Introduction:

The TBIMS uses established criteria to determine eligibility for enrollment and participation in the TBIMS National Database (NDB). To date, it has been the expectation that TBIMS centers approach every eligible subject for enrollment in the NDB. However, beginning on January 1, 2020, it is expected that through randomization of eligible cases a center will have a target of 35 cases per year. This SOP outlines the procedures that will take place for all centers, providing them with a sampling strategy that maintains the representativeness of their sample of subjects.

Purpose:

To institute standard procedural steps for enrolling subjects in the NDB and to address the issue of sampling for NDB enrollment.

Scope:

Current TBIMS centers enrolling subjects into the NDB.

Responsibilities:

TBIMS center staff responsible for enrolling subjects into the NDB (e.g., TBI researchers or clinicians, research assistants, study coordinators).

Procedural steps:

  1. Centers are expected to enter into a screening database all TBI cases at or close to the time of admission.

  2. Each subject will be screened for eligibility and if not eligible a reason for ineligibility will be marked (note: not all of the eligibility criteria need to be identified, just the first one noted in the record).

  3. Once a subject is determined eligible, or if the case meets over-enrollment criteria (disorder of consciousness criteria) the researcher will click a randomization button which is linked to an algorithm that will choose whether or not that subject should be approached for consent based on the target number (35) and historical information such as number of eligible cases seen in a year and, of those eligible, percent enrolled.

  4. All sampling strategies and randomization procedures will be centralized at the NDSC and will be available for audit.

  5. Subjects will be randomized after eligibility is determined but before being approached for consent.

  6. The sampling methodology will be designed to enroll approximately 35 subjects per center in a 12-month period. The percentage of subjects randomly selected for consent will be based on a center’s most recent year’s statistics for eligible subjects and percent successfully enrolled.

  7. It is expected that with small variations in performance, on any given year the actual number enrolled may fall short of or exceed the target of 35. The NDSC will track quarterly the enrollment trends of centers and if the projections differ by +/- 5 subjects then a recalculation of the randomization cut point may be considered.

  8. There may be instances when a Project Director wishes to override the algorithm and change their center’s target to exceed 35. Based on their understanding of admission rates, enrollment rates, and so forth, the Project Director of a center may use their professional judgment to choose a different target number in excess of 35. This new target number will apply only to that center and only for that year. The process by which a Project Director can override the algorithm is as follows:

    1. The Project Director will notify their Project Officer that the center will be increasing their target enrollment to a specified number in excess of the 35 for the current year, as afforded through this SOP (#DA002).

    2. The Project Director will email the NDSC to alert them of their new target number for the current year.

    3. The NDSC will alter the algorithm, only for that specific center and only for the current year, to calculate a new randomization cut point.

Algorithm:

\[Randomization\ cut\ point = \ \frac{\mathbf{Target}\ \mathbf{Number}}{Eligible\ last\ year*Percent\ enrolled\ last\ year}\]

Example:

If a center has 58 eligible persons in the past year and successful consented 85% of them, their randomization cut point will be 71.

Randomization generator:

The database will generate a random number between 1 and 100 using the formula

@Result = 1 + ABS(CHECKSUM(NEWID())) % 100; Comparing this number to the randomization cut point will decide whether an eligible subject should be approached. If the randomization number is equal to or less than the randomization cut point, then the researcher will be directed to approach patient for consent; conversely, if the number is greater than the cut point then they will be directed to not approach the patient for inclusion into the study.

Training requirements:

None

Compliance:

All TBIMS centers who are enrolling subjects into the NDB and the NDSC will comply with this procedure.

References:

None

History:

Date Action
6/17/2011 SOP Created and Approved by Project Directors
10/1/2012 Start date of SOP
6/12/2015 Updated NIDRR to NIDILRR
1/27/2016 Added example calculations
10/1/2019 The SOP changed from randomizing only centers who had large enrollments (over 70) to randomizing all centers to target 35 enrollments per year
10/1/2019 Removed NIDILRR from Compliance section
8/15/23 Added paragraphs allowing centers to override the algorithm

Review schedule:

At least every 5 years.