Table 3.

Pragmatic study design examples applied to community interventions during COVID-19.

Study design	Implementation problem the design can address	Example	Outcome and comparisons
Two-stage randomized preference	Health departments would like to offer incentives to contacts to remain in quarantine for the full recommended duration but cannot offer all types of possible incentives so would like to determine which are more impactful.	CONTACT TRACING PROGRAM INCENTIVES TO QUARANTINEChoices about different incentives to stay in quarantine for the full recommended duration may impact uptake of strict adherence to health department recommendations for contacts identified in contact tracing programs. Understanding to what extent there are benefits of different preferences on outcomes for different types of incentives can help programs plan for the highest impact. A two-stage randomized preference trial can help answer these questions. Example Research Questions: •Does randomizing contacts to receive a cash stipend, a package of resources or a choice between the two result in greater proportion of contacts staying in quarantine? •Does randomizing contacts to receive a cash stipend, a package of resources or a choice between the two result in different COVID-positive test probabilities after the quarantine is over? •How different is the uptake of these two approaches among contacts randomized to choice? What about sub-groups of interest (e.g., by age group, ethnicity, or employment status)?	•Participation/engagement levels for those randomized to different options vs. randomization to preference. •Impact of randomization vs. choice on self-reported or test positivity outcomes.
SMART	Clinic systems may not be able to offer video-visits to all patients, and can benefit from determining whether less intensive formats (e.g., telephone calls; email communications) are sufficient for some patients, allowing the more intensive formats to be offered to those who struggle with other formats, or whose health needs do not align with less intensive formats.	CLINIC-BASED TELEMEDICINEClinics (or individual providers) are randomized to one of two telemedicine approaches for registered patients (e.g., telephone visit or video visit). Those who are not engaging with options after a specified period of time, are re-randomized at an intermediate point to either the other intervention or an augmented form of care, such as a health coach call. Example Research Questions: •Are patients (or providers) given (a) a single intervention A or B equally likely to complete follow-up recommendations (e.g., labs, medications refills) as those given (b) a sequenced combination of the two (A and then B, or B and then A) or (c) an augmented intervention, such as A plus an augmented form of care?	Comparisons of: (1) patient-level and provider-level engagement with different telemedicine options; (2) levels of satisfaction; (3) outcome metrics such as completion of referrals, labs, refills of patients in different groups/no-show rates at the clinic.
Stepped wedge design (modified)	Schools may want to re-open but prefer a staggered approach, in which all schools start with on-line learning, and then depending on outcomes of COVID-19 testing after the school starts, changes in restrictions are made, such as in-person attendance.	STAGGERED IMPLEMENTATION OF IN-PERSON SCHOOL WITH TEST-BASED DECISION-MAKING AT EACH STAGEBy using a combination of a stepped wedge design (with staggered roll-out) plus modifications to the intervention at pre-specified time points as in a SMART design, a staggered modifiable implementation of school sites for in-person classes can be evaluated. For this design, at the end of each set time period, the COVID-19 prevalence is estimated and decisions about how to either stay in the most restrictive mode or to advance to a less restrictive approach are evaluated. A staged approach to testing different educational environments a school district may allow for alterations to restrictions as each new phase is rolled out.	•Do the schools/classrooms meet the advancement criteria for moving to the next school reopening level? •Adjustments to the school environment, such as outdoor classrooms, time spent in class, ventilation, classroom student numbers, etc., can also be incorporated to see if there are additional impacts on prevalence, if these options are feasible.