NOTE: Instructions for completion are shown in green italics and are included where additional clarification may be helpful to the user. These instructions should be deleted or hidden before sharing with decision makers.

Table S1. Documentation of SPIFD2 Steps 1 - 5

STEP 1: STATE RESEARCH AIM, QUESTION, AND OBJECTIVES

STEP 1a: Overarching research aim

STEP 1b: Research question

STEP 1c: Primary objective(s)

Primary Objective:

DESIGN ELEMENTS

STEP 2:
DESCRIBE HYPOTHETICAL TARGET TRIAL

STEP 3:
DESCRIBE RWD STUDY EMULATION OF HYPOTHETICAL TARGET TRIAL

STEP 4:
IDENTIFY FIT-FOR-PURPOSE RWD SOURCE

STEP 5:
DOCUMENT FINAL RW OPERATIONALIZATION, RATIONALE, VALIDITY CONCERNS
AND APPROACHES TO ADDRESS THESE CONCERNS

2a. Conceptual definition

2b. Trial operationalization

3a. Best-case operationalization under routine clinical care

4a. Minimal criteria for valid operationalization in RWD source

4b. RWD minimal criteria ranking with regard to uniqueness and importance

4c. Detailed data source feasibility assessment findings and summary heat map

5a. Final RW operationalization in STaRT-RWE

5b. Rationale for final operationalization of RW study design element or variable

5c. Validity concerns related to RW operationalization, selected RW data source and/or study design

5d. How these validity concerns are/will be addressed

Data Source 1 <replace column header with name of data source 1 assessed>

... <replace column header with name(s) of data source(s) assessed>

Data Source N <replace column header with name of data source N assessed>

INSTRUCTIONS

GENERAL

Estimator [1]

NA

NA

NA

NA

NA

NA

NA

NA

See STaRT-RWE Table 3G

Treatment group "assignment"

<Randomization. List assignment ratio of treated-to-untreated. Note assumption that patients comply with treatment assignment during entire length of follow-up>

NA

NA

NA

NA

NA

NA

NA

NA

See STaRT-RWE Table 4

Time zero (start of follow-up)

<Generally, date of randomization>

NA

NA

NA

NA

NA

NA

NA

NA

See STaRT-RWE Table 4

Length and frequency of follow-up

5

5

5

See STaRT-RWE Table 3H

Sample size needed

5

5

5

See STaRT-RWE Table 7

VARIABLE-RELATED

INSTRUCTIONS

Study Population

Inclusion Criterion 1

5

5

5

See STaRT-RWE Tables 3C/3D

...

5

5

5

Inclusion Criterion N

5

5

5

Exclusion Criterion 1

5

5

5

...

5

5

5

Exclusion Criterion N

5

5

5

Treatment Group

5

5

5

See STaRT-RWE Table 3B

Comparator Group

5

5

5

Primary Outcome
(Definition & Ascertainment)

5

5

5

See STaRT-RWE Table 3G

Key Secondary Outcome 1
(Definition & Ascertainment)

5

5

5

...

5

5

5

Key Secondary Outcome N
(Definition & Ascertainment)

5

5

5

Subgroup Variable 1

5

5

5

See STaRT-RWE Table 4

...

5

5

5

Subgroup Variable N

5

5

5

Confounding Variable 1

[2]

5

5

5

See STaRT-RWE Tables 3E/3F

...

[2]

5

5

5

Confounding Variable N

[2]

5

5

5

4d. Practical considerations

LOGISTICAL ASPECT

Describe Details of Practical Considerations

Data Source 1 <replace column header with name of data source 1 assessed>

... <replace column header with name(s) of data source(s) assessed>

Data Source N <replace column header with name of data source N assessed>

Budgetary Considerations

$

$

$

Timeline Considerations

Fast

Fast

Fast

Other Practical Considerations

5

5

5

Data source previously used for peer-reviewed published research?

YES

YES

YES

Data source previously used for research to support a regulatory decision?

YES

YES

YES

Final Data Source Selected?

Footnotes:

[1] ICH E9 Addendum - statistical principles for clinical trials. https://database.ich.org/sites/default/files/E9-R1_Step4_Guideline_2019_1203.pdf

[2] Prior to completing these "Confounding Variable" cells, ensure the identification of each variable as a potential confounder is justified. Use Table A1 to document the rationale for selection of variables for confounding control in the RWD study emulation. Minimally, variables that should be measured at baseline in the hypothetical target trial should be considered for selection. Once the list of variables is finalized, copy the conceptual definition and paste here to complete Step 2a, and complete Steps 2b and 3a

[3] Schneeweiss et al. Graphical Depiction of Longitudinal Study Designs in Health Care Databases. Ann. Intern. Med. 170:398-406 (2019). doi:10.7326/M18-3079.

[4] We generally recommend including only "must have" criteria in Step 4a, and then among these must-have criteria, numerically ranking them 1 - N by uniqueness and/or how easy they are to evaluate. The goal is to efficiently / quickly narrow the candidate data sources. Often only the top 10 must have criteria are needed to narrow the data sources to included in the detailed feasibility assessment to <5.

[5] After this column is complete, go to Appendix Figure A1 to narrow / identify data sources for detailed feasibility assessment (we suggest narrowing to <5)

Keys for Ranking

Scoring for Data Sources by Data Elements

Scoring Data Sources by Budget Constraints

Scoring Data Sources by Timeline Constraints

Scoring

Description
<Modify description if needed>

Scoring

Description
<Modify description if needed>

Scoring

Description
<Modify description if needed>

1

Data Requirements are not met

$

Well Below Budget

Slow

May cause delays in timeline

2

$$

Within Budget

Medium

Expected to meet timeline constraints

3

Some data requirements are met

$$$

Well Above Budget

Fast

Expected to be delivered well before timeline constraints (expedited).

4

5

All or nearly all data requirements are met