Rapid-cycle designs to adapt interventions for COVID-19 in safety-net healthcare systems

Abstract

Racial/ethnic minority, low socioeconomic status, and rural populations are disproportionately affected by COVID-19. Developing and evaluating interventions to address COVID-19 testing and vaccination among these populations are crucial to improving health inequities. The purpose of this paper is to describe the application of a rapid-cycle design and adaptation process from an ongoing trial to address COVID-19 among safety-net healthcare system patients. The rapid-cycle design and adaptation process included: (a) assessing context and determining relevant models/frameworks; (b) determining core and modifiable components of interventions; and (c) conducting iterative adaptations using Plan-Do-Study-Act (PDSA) cycles. PDSA cycles included: Plan. Gather information from potential adopters/implementers (e.g., Community Health Center [CHC] staff/patients) and design initial interventions; Do. Implement interventions in single CHC or patient cohort; Study. Examine process, outcome, and context data (e.g., infection rates); and, Act. If necessary, refine interventions based on process and outcome data, then disseminate interventions to other CHCs and patient cohorts. Seven CHC systems with 26 clinics participated in the trial. Rapid-cycle, PDSA-based adaptations were made to adapt to evolving COVID-19-related needs. Near real-time data used for adaptation included data on infection hot spots, CHC capacity, stakeholder priorities, local/national policies, and testing/vaccine availability. Adaptations included those to study design, intervention content, and intervention cohorts. Decision-making included multiple stakeholders (e.g., State Department of Health, Primary Care Association, CHCs, patients, researchers). Rapid-cycle designs may improve the relevance and timeliness of interventions for CHCs and other settings that provide care to populations experiencing health inequities, and for rapidly evolving healthcare challenges such as COVID-19.

SCALE-UP Utah used real-time information on changes in COVID-19 policy (e.g., vaccination authorization), local case rates, and the capacity of safety-net healthcare systems to iteratively change interventions to be relevant and timely for patients.

Implications.

Practice: Population health management approaches can be rapidly adapted to reach patients of safety-net healthcare systems to address COVID-19 and to reduce health inequities.

Policy: Policy makers should consider how to support safety-net healthcare systems in the development and implementation of population health management systems that can be rapidly adapted to address COVID-19 and other health issues.

Research: Future research is needed to examine the effectiveness of rapid-cycle designs to address COVID-19 and other health conditions, and how rapid-cycle designs can be used to reduce health inequities.

INTRODUCTION

Over 1 million deaths in the USA have been caused by the SARS-CoV-2 virus (hereafter referred to as COVID-19), with a disproportionate burden of COVID-19 in, racial/ethnic minority [1–4], rural populations [5], and low socioeconomic status populations [6]. After accounting for age, the rate of COVID-19 cases, hospitalizations, and deaths are greater among African American/Black (1.1x, 2.8x, 2.0x), Hispanic/Latino (1.9x, 2.8x, 2.3x), and American Indian or Alaska Native individuals (1.7x, 3.4x, 2.4x) compared with Non-Hispanic Whites [4]. Spared early in the pandemic, nonmetropolitan counties currently have a higher death rate from COVID-19 compared with their metropolitan counterparts (183 deaths per 100,000 for metropolitan vs. 211 deaths per 100,000 for nonmetropolitan counties) [7]. Vaccination rates are markedly lower among rural compared with urban areas [8]. Lower socioeconomic status is associated with higher COVID-19 incidence and mortality [6] and lower vaccination rates [9]. Consequently, developing and evaluating interventions to address COVID-19 among these populations are crucial to controlling the pandemic and to reducing health inequities [10].

The Centers for Disease Control and Prevention (CDC)’s strategic plan to address inequities of COVID-19 includes expanding programs and policies for testing, contact tracing, and vaccination to reach populations at increased COVID-19 risk [10]. The plan emphasizes establishing collaborations and partnerships between populations experiencing a disproportionate burden of COVID-19 and federal, state, local, and tribal agencies [10]. Federally Qualified Health Centers and Community Health Centers (hereafter referred to as CHCs) provide comprehensive primary care to diverse, low socioeconomic status, and rural populations. In 2020, CHCs in the USA served over 28.5 million patients—68% of patients were <100% of the federal poverty level, ~37% were Hispanic or Latino/a, and 24% were best served in a language other than English [11]. CHCs reach populations at increased risk for COVID-19 and approximately 80% of individuals in the USA, including low socioeconomic status individuals, see a primary care provider at least annually [12]. Consequently, partnering with CHCs has the potential to reach at-risk populations with interventions to address COVID-19.

The dynamic nature of the COVID-19 pandemic has resulted in interventions that need to be rapidly “developed, adapted, employed, and abandoned” to keep pace with scientific and policy advances, as well as the status of the pandemic (e.g., changes in testing technology, approval of vaccines, changes in public health guidance, local infection rates) [13]. Rapid, relevant, and responsive research designs have been promoted to both increase the pace of traditional, linear research development, and to provide relevant evidence on rapidly evolving questions such as those related to COVID-19 [14–16]. While still allowing rigorous evaluation, rapid-cycle designs enable iterative adaptation of interventions by incorporating data from multiple sources, including policy (e.g., vaccination approvals, eligibility), setting capacity (e.g., clinic testing capacity, vaccine availability), and stakeholder priorities (e.g., infection “hot spots” for testing/vaccination).

Numerous frameworks have been developed to guide researchers and practitioners in adaptation of interventions, and largely cover the following broad elements: assess community; understand the intervention; select intervention; consult with experts and stakeholders; decide what needs adaptation; adapt the original intervention; test adaptation; and implement and evaluate adaptation [17]. Applying these elements in a rapid-cycle design has the potential to improve intervention relevance and responsiveness according to rapid changes in the context of COVID-19. Furthermore, by engaging potential adopters, implementers, and end-users in the intervention development and implementation process, tacit community knowledge can be incorporated throughout the design and implementation of interventions [18]. These perspectives can provide expertise on the local capacity, available resources, and preferences of clinic staff and patients to ensure that intervention adaptations are relevant to the targeted populations and settings [19].

The objective of this paper is to describe the rapid-cycle design process and exemplars of adaptions for an ongoing, multilevel, pragmatic clinical trial to address COVID-19 among underserved populations (SCALE-UP Utah). The goal of SCALE-UP Utah is to increase the reach, uptake, and sustainability of COVID-19 screening, testing, and vaccination among underserved populations using clinic- and patient-level interventions. The study was funded as part of the National Institutes of Health Rapid Acceleration of Diagnostics-Underserved Populations (RADx-UP) initiative that aims to “ensure that all Americans have access to COVID-19 testing, with a focus on communities most affected by the pandemic [20].” This initiative is focused on using community-engaged research to “rapidly increase reach, access, acceptance, uptake, and sustainment of Food and Drug Administration (FDA)-authorized/approved diagnostics among vulnerable populations in geographic locations that are underserved [20].” The SCALE-UP Utah randomized clinical trial is ongoing, and the primary and secondary outcomes will be presented upon completion of the trial. Therefore, the primary focus of this paper is to describe the rapid-cycle design and adaptation process.

METHODS

The protocol for this ongoing study was approved by the Institutional Review Board of the study authors’ university.

SCALE-UP Utah overview

SCALE-UP Utah is a pragmatic, randomized clinical trial to evaluate clinic- and patient-level interventions to increase COVID-19 screening, testing, and vaccination among CHC patients. Additional information on the study design, measurement, outcomes, and evaluation plan can be found at clinicaltrials.gov (NCT04939532).

Clinic-level intervention

Ask-Advise-Connect (AAC) is a clinic-level health information technology intervention that focuses on enhanced system supports at the point of care using the electronic health record (EHR). AAC is a CDC “best practice” for tobacco control [21], and similar “e-referral” programs are effective for a wide range of conditions [22–24]. In this intervention, we proposed to use AAC to increase COVID-19 screening and testing uptake. AAC is a proactive strategy, in which the EHR includes supports for the clinical staff to screen for COVID-19 testing eligibility with every patient as a part of every visit intake, advise eligible patients to obtain testing, and directly connect patients with a test site.

Patient-level population health management interventions

Population health management (PHM) is a proactive and coordinated approach to healthcare, where patients at risk or likely to benefit from intervention are identified and receive repeated offers to engage in interventions [25]. SCALE-UP Utah is evaluating the effectiveness of two PHM interventions: text messaging (TM) and text messaging plus patient navigation (TM + PN) on reach and uptake of screening, testing, and vaccination for COVID-19. Text messages in both conditions include a brief message regarding risk, encouragement for COVID-19 testing/vaccination, and provision of information for scheduling testing/vaccination. Patients who accept testing/vaccination via text message receive additional text messages with instructions for how to schedule either a test or vaccination (e.g., locations, hours, contact information to schedule, online scheduler, or other COVID-19-related information specific to the patient’s clinic). Patients who decline testing/vaccination receive a text message with the clinic phone number and a note to call if anything changes.

Patients in the TM + PN condition who accepted testing/vaccination via text receive the same message as those in text message only condition, plus a note that a patient navigator will be contacting them to assist in connecting them to assistance for testing/vaccination, address any concerns for testing/vaccination, and address social determinants of health. Patient navigators were trained to provide information regarding testing/vaccination, and to address motivation and practical barriers to testing and vaccination. Navigation calls could also address social determinants of health that the patient may be experiencing or anticipates they will experience as a result of a positive test. Assistance may include transportation, housing and rent relief, support for paying for utilities, food banks and food delivery services, and childcare. Patients in TM + PN who declined testing/vaccination receive the same message as those who decline in the TM only group. All messages contain a one-touch option to opt out of future messages. All interventions are available in both English and Spanish, and text messages were developed using plain language guidelines [26]. Messages translated to Spanish were adapted to account for cultural and grammatical differences and to ensure Spanish language messages remained consistent with plain language guidelines. Interventions were delivered in Spanish to patients that had Spanish indicated as their preferred language in the EHR.

Research–practice partnership

The study was conducted in partnership with the Utah Health Resources and Services Administration Primary Care Association (the Association for Utah Community Health), CHCs across Utah, the Utah Department of Health and Human Services, and a university academic medical center. The partnership was established in 2017 and has been leveraged for projects to address tobacco use [27], colorectal cancer screening, HPV vaccination, and nonopioid pain management for patients seen at CHCs. Additional description of the partnership has been described elsewhere [27].

Setting and participants

Seven of the 13 CHC systems in Utah volunteered for participation. The seven participating CHC systems have a patient population of 165,077 patients of which approximately 30% are uninsured and 44% are Hispanic/Latino. Participating CHCs systems range in size from 1 to 7 clinics.

Procedures

Community engagement process

The selection of pilot sites, priority patient cohorts, study procedures, and intervention adaptations were informed by a multi-method, pragmatic engagement approach. Engagement activities included consultation meetings with the Primary Care Association in Utah (Association for Utah Community Health) and the Utah Department of Health; meetings with CHC system leadership; and separate Patient Advisory Committee and Scientific Advisory Committee meetings.

Consultation with the Primary Care Association and the Utah Department of Health

The Association for Utah Community Health assists CHCs in monitoring real-time capacity for COVID-19 testing and vaccination, interest in project interventions, and staff capacity for project participation. The Association for Utah Community Health also administered a needs assessment survey to CHC systems regarding COVID-19 priorities, needs, and health information technology capacity. The Utah Department of Health and Human Services provides data on state priorities and policies, including geographically targeted infection and vaccination hot spots, availability of testing centers, availability of vaccines, and routinely collected state-wide survey data on vaccine intentions and hesitancy, all of which are reviewed weekly and helped guide intervention content and tailoring.

Meetings with CHC system leadership

Leadership of CHC systems were approached to understand current CHC system capacity and needs by the research team. These meetings were 30–45 min in length and were attended by 2–3 researchers, 2–5 CHC staff, and 1–3 Association for Utah Community Health representatives. Each meeting included: (a) brief description of available interventions of the project (i.e., AAC, TM, and TM + PN for testing and/or vaccine), (b) an example of practical ways the interventions could be adapted for the CHC system (e.g., targeting priority patient cohorts; messaging specific to where the CHC system wanted to send patients for testing), and (c) time allotted to answer CHC-specific questions about project implementation.

Patient Advisory Committee and Study Advisory Committee meetings

The Patient Advisory Committee consisted of patient representatives from participating CHC systems. The meetings were used to understand patient attitudes and needs regarding COVID-19, and obtain patient feedback on proposed interventions and adaptations. The Study Advisory Committee consisted of patients from multiple CHC systems, clinic staff from each participating CHC system, representatives from the Utah Department of Health and Human Services, representatives from the Primary Care Association, and the research team. These meetings were used to engage stakeholders in understanding CHC system priorities, needs, and capacity; obtain input and feedback on proposed study procedures and intervention content; and promote study engagement among individual CHC systems.

Rapid-cycle design for adaptation

SCALE-UP Utah utilized a rapid-cycle design with continuous improvement cycles (Fig. 1). Interventions were designed, adapted, implemented, evaluated, and continually refined on a small scale in multiple iterative pilots, using short time frames (e.g., ≤1 month), after which they are disseminated to other clinics/systems. The purpose of this approach was to provide CHC clinics and patients effective, relevant interventions as quickly as possible, and to allow the interventions to be adapted to rapidly changing contexts in testing and vaccination throughout the duration of the project. Pilots were conducted at a single site based on system priorities, capacity, and need (e.g., infection hot spots).

Fig 1.

Plan-Do-Study-Act cycle.

The rapid-cycle design process consisted of three general steps: (a) Assess community context and determine relevant theoretical models and framework, (b) Determine core and modifiable components of interventions, and (c) Conduct rapid-cycle design iterations for intervention adaptation using continuous improvement cycles.

Assess community context and determine relevant theoretical models and frameworks

Stakeholder and community partner priorities, assets, and needs were evaluated throughout the entire project period to capture changing context around COVID-19 (i.e., consultation meetings with the Primary Care Association and the Utah Department of Health and Human Services; meetings with CHC system leadership; and Patient and Study Advisory Committee meetings). Relevant theories and frameworks of behavioral science, implementation science, and social determinants of health were selected by the research team. The theories and frameworks were selected to inform intervention development for populations experiencing adverse social determinants of health, identify barriers and facilitators to implementation of interventions to address COVID-19 (the Consolidated Framework for Implementation Research [CFIR]) [28], mechanisms of behavior change to improve COVID-19 testing and vaccination behavior among patients (Social Cognitive Theory) [29, 30], and relevant process and outcome data to assess throughout project duration (Reach, Effectiveness, Adoption, Implementation, Maintenance [RE-AIM]) [31]. The conceptual model is presented in Fig. 2.

Fig 2.

Conceptual model.

Determine core and modifiable components of interventions.

Using selected theoretical models, core components of the intervention were identified that, if adapted, would change the function [32] of the intervention. The remaining modifiable components could be adapted or tailored without compromising the function of the intervention.

Rapid-cycle design iterations for adaptation following continuous improvement cycles

Adaptations to the study were created using a series of rapid-cycle design iterations informed by Plan-Do-Study-Act (PDSA) cycles (Fig. 1) [33]. Given the dynamic nature of the COVID-19 pandemic, a critical component of the improvement cycles was ongoing, near real-time evaluation of context (e.g., infection hot spots, CHC system capacity, stakeholder priorities, local and national policies, vaccine availability). This information was incorporated throughout the PDSA cycles, and adaptions to study procedures and interventions were created using full and partial PDSA cycles. PDSA cycles include four components. (a) Plan. The research team conducted meetings and surveys with stakeholders and potential adopters/implementers (e.g., CHC staff/patients) to design and review the initial interventions. (b) Do. Interventions were implemented in a single CHC, with a targeted patient cohort that was selected based on CHC and Utah Department of Health and Human Services priorities. (c) Study. Descriptive analyses of data from interventions were examined by the research team and stakeholders. (d) Act. Refinements were made to interventions based on process and outcome data, feedback from potential adopters/implementers, and changes in context (e.g., infection hot spots). Data collected through community engagement activities were synthesized and presented to study stakeholders at a weekly stakeholder meeting consisting of members from the research team, the Association for Utah Community Health, and the Utah Department of Health and Human Services. Intervention refinements were proposed and implemented after discussion among the meeting participants and CHC leadership. If no refinements were necessary, interventions were disseminated to other cohorts/CHCs. Information exchange included data reports from the Association for Utah Community Health and Utah Department of Health and Human Services; meeting minutes and email exchanges with CHC leadership and staff; and meeting minutes from the Patient and Study Advisory Committees.

RESULTS

Throughout the first 12 months of SCALE-UP Utah, multiple adaptations were made to the study design, intervention content, and intervention cohorts. Adaptations were made to address CHC capacity, COVID-19 scientific and policy guidance, and the disproportionate burden of COVID-19 among underserved populations. Figure 3 provides a visual depiction of the timeline of some exemplar adaptions relative to the context of COVID-19 in Utah. The number of COVID-19 cases per day across the state was approximately 400 when the SCALE-UP Utah grant application was funded, and by the time SCALE-UP Utah was ready to implement interventions (<5 months later) the number of cases had reached a maximum at ~4,500/day [34]. Consequently, many CHC systems were unable to increase testing for patients due to constraints (e.g., inadequate amounts of personal protective equipment, limited personnel due to COVID-19 infections among staff members). In addition, the CHCs in this project were geographically dispersed across the state (including urban, rural, and frontier clinics), and had wide variability in local rates of COVID-19, demand for testing/vaccines, availability of testing/vaccine resources at the CHC (e.g., personnel/equipment), and availability of testing locations outside of the CHC system (e.g., state-run locations, mobile testing sites, at-home testing, etc.). To date, 114,721 unique patients seen at participating CHC have been sent at least one text message. Patients were 55% female (n = 62,769), 44% Hispanic/Latino (n = 50,752), 28% uninsured (n = 32,691), with a mean age of 42 years (SD = 16.1). Thirty-five percent of patients (n = 38,740) indicated Spanish as their primary language in the EHR and received interventions in Spanish.

Fig 3.

Timeline of study adaptations and COVID-19 testing, confirmed cases, and vaccinations. A: Grant proposal for SCALE-UP Utah submitted. B: Funding awarded for SCALE-UP Utah. C: First intervention delivered; interventions focus solely on texting. D: Vaccine receives temporary authorization. E: Interventions adapted to target vaccines as requested by the CHCs. F: Interventions adapted to include messages for testing in hot spot infection areas among unvaccinated patients only. G: Interventions adapted to include at-home testing offered for unvaccinated patients with symptoms. H: Interventions adapted to offer vaccine and testing concurrently; testing targeted symptomatic and asymptomatic. I: Interventions adapted offer at-home testing for all vaccinated and unvaccinated individuals.

Intervention functions

The PHM intervention functions are described in Table 1. Intervention functions were used to create adaptations that were fidelity consistent.

Table 1.

PHM intervention functions

Intervention	Function
Text message intervention	• Proactively reach out to patients.
	• Provide multiple opportunities for screening/testing.
	• Provide recommendations for testing.
	• Provide information for where to get tested.
Patient navigation intervention	• Proactively reach out to patients.
	• Provide multiple opportunities for screening/testing.
	• Provide recommendations for testing.
	• Provide information for where to get tested.
	• Address motivation and practical barriers to testing.

PHM population health management.

Study design adaptation

Vaccination

The original study was proposed and funded prior to FDA emergency use authorizations (EUAs) for COVID-19 vaccines. After the initial EUAs were approved, study partners identified vaccine administration as a top priority, and the study design was adapted to address vaccination in addition to testing. This included using vaccination status to target testing interventions (e.g., providing interventions only to those who indicated they were not vaccinated/did not want to receive the vaccine), sending messages specific to vaccination events if requested by the CHC system, and adding vaccine completion as a secondary outcome.

Intervention adaptation

Intervention content

The content of the TM and PN interventions was adapted based on CHC system and CHC patient preference, advances in scientific and policy guidance, and local infection rates. Examples of TM content are provided in Fig. 4. At the beginning of the project, PHM interventions screened for COVID-19 symptoms and/or close contact to a confirmed case, and offered a range of community-based testing resources for patients who screened positive. During this time, many of the CHCs were not set up as formal testing sites and/or did not have the capacity to increase testing at their site. Therefore, the interventions directed patients who screened positive to be tested at the nearest public health testing sites, which were available throughout the state. As cases began to plateau and even decrease, CHCs had more capacity to test onsite and intervention content was adapted to direct patients to their local CHC clinics for testing. Upon vaccine EUA, a rapid shift was needed in intervention content to support vaccination efforts, and most CHCs requested to put testing interventions on hold. A few months later, when demand for vaccination reduced substantially, the interventions were adapted to concurrently focus on testing and vaccination. Patients were first offered opportunities for vaccination; those who did not want to receive the vaccine were offered testing opportunities. As at-home test kits became more readily available, interventions were adapted to provide at-home tests for interested, unvaccinated patients to use as needed. At-home test kits were selected to reach rural/frontier populations that may have difficulty accessing testing sites, and to address the time and transportation constraints experienced by CHC patients.

Fig 4.

Text message examples and adaptations. Panels A–D depict the adaptations to text messages throughout the study duration for a single CHC system.

Intervention content was also adapted based on patient feedback, geographic location, and other demographic characteristics. Based on responses from the Patient Advisory Committee, information was added to interventions to emphasize that testing and vaccination were free and confidential; this adaptation was made to address concerns that CHC patients may have about being able to afford either tests or vaccines, and confidentiality regarding positive tests or residence status. Additionally, intervention content was adapted based on real-time infection rates for an individual’s geographic area. When cases were surging in specific areas of the state, text messages were adapted to describe the real-time case rates in the patient's zip code (Fig. 4: Panel D).

Intervention timing and selection

CHC systems varied in the content and timing of interventions based on CHC capacity and preference. For example, one CHC system started with testing messages that referred patients to testing outside of the clinic in February 2021. At the end of April, the intervention content was adapted to include vaccine messages for patients as eligibility changed and vaccines became available. By the end of June, the intervention content changed such that patients were first offered opportunities for vaccination, then patients who did not want to receive the vaccine were offered testing opportunities (both at home and at the CHC system). By mid-August the CHC system stopped administering vaccinations; thus, patients no longer received vaccination intervention content. To date, CHC systems have only selected to implement PHM interventions, and no systems have implemented AAC.

Intervention cohort adaptation

Patient cohort selection

At the beginning of the study, patient cohorts were identified based on risk factors such as age, race/ethnicity, and primary language. In the first few months after the COVID-19 vaccine received EUA, messages promoting vaccination were sent to a patient cohort defined as patients whose primary language was Spanish (as indicated in the EHR) to address the disparity in vaccination rates among racial/ethnic minorities. As the project progressed, patient cohorts were adapted based on recommendations from the Utah Department of Health and Human Services, CHC priorities (e.g., zip codes of infection hot spots), and changes in COVID-19 testing and vaccination procedures. For example, in the months following vaccine authorization, interventions for testing were sent only to unvaccinated patients. After the Delta variant surge, the Utah Department of Health and Human Services recommended interventions be offered to vaccinated and unvaccinated patients, and the team planned to adapt the cohorts to include vaccinated and unvaccinated patients. However, prior to deployment of the adapted cohorts, there was a nationwide shortage of test kits; thus, unvaccinated patients were identified as the cohort with the highest priority for the limited number of available tests.

Intervention targeting for patient cohorts

In some scenarios, intervention content was targeted within patient cohorts. Within some cohorts, specific information about confidentiality for people who may be undocumented was included in the interventions for patients with Spanish as their primary language. Additionally, interventions in some patient cohorts were targeted using the Extended Parallel Processing Model (EPPM) [35, 36]. The EPPM provides guidance for effective communication of health and risk information based on two key variables: threat (comprised of perceived susceptibility to and severity of a health risk) and efficacy (comprised of response efficacy, or beliefs about a proposed solution, and self-efficacy, or beliefs about one’s ability to perform the recommended solution). Based on EPPM, cohorts of unvaccinated patients were divided into four categories for targeting interventions: low threat/low efficacy beliefs (i.e., receive messages about risk of disease and efficacy of vaccines); low threat/high efficacy beliefs (i.e., receive messages about risk); high threat/low efficacy beliefs (i.e., receive messages about efficacy of vaccines/testing); and high threat/high efficacy beliefs (i.e., receive call to action for vaccines and testing). Patient information obtained from the EHR and area-level survey data from the Utah Department of Health and Human Services were used to determine the category for each patient. Patient address was matched with area-level data on population-level vaccine uptake obtained from the Utah Department of Health and Human Services as a proxy for efficacy (e.g., lower area-level vaccine uptake was an indicator of lower efficacy). Data obtained from state surveys indicated that younger individuals felt less threatened by COVID-19 than older individuals (>30), thus patient age was used as a proxy for threat. Regardless of category, each patient received the same offer for vaccination (if in the TM condition) and additional support of a patient navigator (if in the TM + PN condition).

DISCUSSION

SCALE-UP Utah utilized a rapid-cycle design and adaptation process with iterative PDSA cycles that incorporated near real-time evaluation of COVID-19 pandemic context (e.g., infection hot spots, CHC capacity, stakeholder priorities, local/national policies, vaccine availability). Given the dynamic nature of COVID-19, the rapid-cycle design ensured that CHC systems and patients were receiving relevant and timely interventions, while still enabling rigorous evaluation of interventions. With the remaining uncertainty of the COVID-19 pandemic, other healthcare systems may be able to adopt this process to address changing policy, scientific evidence, and public health guidance regarding COVID-19 (e.g., vaccine booster shots; testing recommendations; future infection surges due to variants). Furthermore, the rapid-cycle design process has the potential to be used in interventions to address other chronic and infectious diseases that contribute to health inequities. Although COVID-19 presented an extreme case of dynamic changes in scientific and policy recommendations, the process could be used to address changes in clinical guidelines for chronic disease prevention and control that occur over longer time frames.

The COVID-19 pandemic has highlighted long standing inequities in the underlying physical, social, and economic conditions that have contributed to health disparities across the USA [37–39]. These inequities, coupled with research to suggest that non-targeted interventions may inadvertently exacerbate health inequities, has prompted numerous calls specifically to address health inequities in implementation science research [40–44]. SCALE-UP Utah considered health equity from the conception of the project proposal through implementation. For example, the project was conducted in partnership with CHCs to reach populations that experience a wide range of health inequities, including a disproportionate burden of COVID-19; moreover, the interventions for SCALE-UP Utah were selected because they leverage ubiquitous health information technology (i.e., text messages and phone calls) to reach populations that typically have lower access to resources. Intervention content for PHM was designed for populations who may be experiencing health inequities. For example, content was created based on plain language guidelines and best practices [26], and addressing adverse social determinants of health such as transportation barriers, housing, and food insecurity was integrated into PN procedures. Adaptations throughout the trial also considered health equity; for example, at-home test kits were incorporated into the project to address time and transportation barriers that underserved populations may experience, and the at-home test kits that were selected for the project were specifically chosen for their minimal technology requirements. Importantly, using a PHM approach conducted in partnership with CHCs leveraged the CHCs’ ability to reach underserved populations and concurrently addressed barriers to implementing interventions to address COVID-19 for CHCs. Despite ubiquitous adoption of health information technology, there is a growing digital divide between high-resource healthcare systems and socioeconomically disadvantaged and rural systems such as CHCs [45, 46]. Thus, our interventions were designed to fill a need in CHC system capacity.

Allison Metz et al. recently outlined six factors essential to successful, equitable implementation: (a) build trusting relationships; (b) dismantle power structures; (c) invest and make decisions to advance equity; (d) develop community-defined evidence; (e) make adaptations; and (f) consider critical perspectives to how implementation science may exacerbate health inequities [44]. SCALE-UP Utah incorporated a number of these elements. The study was conducted leveraging the infrastructure of a research–practice partnership established over ~4 years, relied on shared resources between partners, and incorporated expertise from diverse stakeholder groups (e.g., State Department of Health; Primary Care Association; CHC systems; CHC patients). CHC leadership, clinic practice team members, and patient perspectives led to numerous adaptations to intervention content, and the rapid-cycle design enabled iterative feedback from stakeholders to be incorporated throughout the conduct of the study. The rapid-cycle process enabled CHCs to choose the interventions that were most appropriate for their system, and allowed the CHCs to vary their choices over time (as depicted in Fig. 4). This also improved engagement in the project by delivering interventions useful and timely for the CHC systems. For example, one CHC system requested the project team alert eligible patients about their vaccine eligibility through the PHM interventions. Within hours of sending text messages, the vaccination appointments were filled, which ultimately alleviated days of work from CHC staff who, prior to the PHM interventions, were individually calling eligible patients to offer vaccination. When asked to discuss their involvement with the project, the CHC representative told the group “It is so much more help than I ever imagined it would be. I would put it as one of my top three favorite things.” Additionally, the intervention content was developed and refined with the perspective of CHC patients, including emphasis on the confidentiality and price of COVID-19 testing and vaccination.

Similar to other research conducted during the COVID-19 pandemic [47], SCALE-UP Utah has faced the challenge of balancing scientific rigor and adapting study procedures in the context of a global pandemic. SCALE-UP Utah approached this challenge by using a rapid-cycle design with a systematic process to make adaptations. A critical component of adaptation is ensuring fidelity to the functions of intervention [32], while adapting interventions to fit local context. By identifying the functions of interventions a priori, the SCALE-UP Utah team was able to make proactive decisions about adaptations while still ensuring that the original aims and hypotheses could be tested.

Other key facilitators to the application of a rapid-cycle design approach included the development and implementation of an agile PHM system that allowed for cohorts to be selected, and interventions to be adapted and delivered in a matter of hours. This system, in conjunction with a robust integrated data system for sharing clinical encounter data between participating CHCs and the research team, allowed the interventions to be adapted and implemented in near real time. The research team planned for adaptation from the onset of the trial, and developed systematic documentation to characterize adaptations and contextual factors that influenced adaptation. Additionally, our research team had an established relationship with the Institutional Review Board (IRB) and developed study procedures with the IRB’s frequent input. Finally, the community engagement and research–practice partnership approach was critical to the feasibility of this study design. COVID-19 has presented a unique scenario in which multiple sectors are focused on addressing the pandemic, and has required the coordination of multiple stakeholder perspectives, resources, and effort. The rapid-cycle design for the existing project may not have been possible without this coordination of resources, including leveraging an existing research–practice partnership, utilizing data from multiple entities to drive the interventions (i.e., identify patient cohorts, adapt intervention cohorts), and building on the trust and mutual respect of expertise between the partners. Future research is needed to understand the extent to which principles of community engagement, trust, and shared distribution of power among study stakeholders influences effectiveness of rapid-cycle designs.

Rapid-cycle adaptations require coordination among multiple entities, the ability to frequently collect data to inform adaptations, and the infrastructure to implement adaptations. Therefore, these designs can be challenging to implement. In SCALE-UP Utah, a key pragmatic strength of the study design was that CHCs were able to self-select intervention content and timing that aligned with their priorities. This approach was also challenging, as it created a large number of variations in intervention content, and could unintentionally create inequities in the populations who were eligible to be reached by interventions. For example, if a CHC decided not to use the SCALE-UP interventions to target COVID-19 vaccines, their patient population did not receive interventions that focused on COVID-19 vaccination. Despite these challenges, rapid-cycle adaptations have the potential to improve the relevance and timeliness of interventions, especially for research with dynamic contexts such as COVID-19.

This research is not without limitations. The purpose of this paper was to describe the rapid-cycle design process and exemplars of adaptation; thus, we do not have a comparison group to evaluate the effectiveness of adaptation compared with no adaptation. Additionally, our study was conducted in a single state and may not generalize to CHC systems in other states, or healthcare systems that serve different patient populations. However, research suggests that resource constraints such as those experienced by CHC systems can lead to innovative adaptations and solutions to problems that have the potential to be generalized to other low- and high-resource settings [48, 49]. Finally, SCALE-UP Utah was conducted with funding by an extramural funding agency, which may limit the generalizability of similar programs to be implemented in safety-net healthcare systems without external support. More research is needed to understand sustainability of these approaches beyond extramural funding support, as well as what support (financial or otherwise) is needed for safety-net healthcare systems to develop, implement, and sustain these approaches in routine practice.

CONCLUSION

Rapid-cycle designs can provide relevant and timely interventions to address evolving research questions, such as those related to COVID-19, while still enabling rigorous evaluation. Iteratively improving interventions based on real-time data collection may increase intervention relevance and timeliness, especially for resource constrained settings like safety-net healthcare systems. Future research should consider incorporating rapid-cycle designs as a way to adapt interventions to changes that occur over time.

Funding

This work was supported through the National Center for Advancing Translational Sciences of the National Institutes of Health (UL1TR002538; 3UL1TR002538-03S4), the National Cancer Institute (P30CA042014), and the Huntsman Cancer Foundation.

Compliance with Ethical Standards

Conflict of Interest: All authors declare that they have no conflicts of interest.

Ethical Approval: All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The protocol for this ongoing study was approved by the University of Utah Institutional Review Board (#00136001).

Informed Consent: This study received a waiver of consent from the University of Utah Institutional Review Board, informed consent was therefore not required.

Welfare of Animals: This article does not contain any studies with animals performed by any of the authors.

Study Registration: The study was preregistered at clinicaltrials.gov.

NCT04939532 (https://clinicaltrials.gov/ct2/show/NCT04939532?term=David+wetter&draw=2&rank=2)

NCT04939519 (https://clinicaltrials.gov/ct2/show/NCT04939519?term=David+wetter&draw=2&rank=3)

Analysis Plan: The analysis plan was not formally preregistered.

Analytic Code Availability: There is no analytic code associated with this study.

Materials Availability: Materials used to conduct the study are not currently publicly available. Materials may be requested by emailing the corresponding author.

Data Availability

Deidentified data from this study are not available in a public archive. Deidentified data from this study will be made available (as allowable according to institutional IRB standards and data use agreements) by emailing the corresponding author.