An official website of the United States government
NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
John P. Elder, PhD, MPH, Leticia Ibarra, MPH, Deanna Rossi, José Luis Olmedo, Nadia Campbell, MPH, Esther Bejarano, Chii-Dean Lin, PhD, and Guadalupe X. Ayala, PhD, MPH.
Author Information and AffiliationsAsthma is a prevalent chronic disease in rural and border communities, with active asthma prevalence rates of 21% among 5- to 17-year-olds in Imperial County, California, near the border with Mexico, compared with the statewide prevalence of 12% among all California residents. Risk factors for poor asthma control include ineffective disease management in the home, school, and community; suboptimal health care; and poor air quality. Comprehensive approaches are needed that recognize these multiple risk factors and intervene to reduce them.
The Imperial County Asthma Comparative Effectiveness Research project (ie, Respira Sano [Breathe Easy]) was a multisector, multilevel intervention designed to reach rural border youth aged 6 to 17 years and their caregivers. We sought to determine whether a family intervention, a clinic intervention, or a combined approach (family plus clinic), all nested within a community intervention, would result in differential improvements in youth- and caregiver-reported quality of life (QOL) and asthma symptoms and their consequences compared with a community-only control condition.
Respira Sano was a 2 × 2 factorial study nested within a community intervention. The family intervention provided asthma education and management strategies through home visits and phone calls from community health workers over a 2-month period. The clinic intervention established an Asthma Care Model that included hiring asthma educators, training providers, and making changes to electronic health records to promote quality asthma care. The community intervention promoted awareness of air quality and management techniques through partnerships with schools and media outlets. A cohort of 400 Spanish- or English-speaking youth aged 6 to 17 years with mild to severe persistent asthma and their parent/legal guardian (ie, caregiver) agreed to participate. Youth who were patients of Clinicas de Salud del Pueblo, Inc (CDSDP) were randomly assigned to 1 of the clinic intervention conditions (clinic only or family plus clinic); those who were not patients were randomly assigned to 1 of the no-clinic intervention conditions (family only or control). All youth were randomly assigned to the family intervention or no family intervention. The primary outcomes were QOL (caregiver self-report and in a subsample of age-eligible youth [aged 9-17 years], youth self-report) and asthma symptoms and consequences (caregiver report on youth and youth self-report). Respira Sano was a partnership between academic institutions (San Diego State University and the Institute for Behavioral and Community Health), a federally qualified health center (specifically, CDSDP), a community-based organization (Comité Cívico del Valle), and Impact Assessment, Inc, an environmental public health education and consulting firm. It was conducted in Imperial County, California, between August 16, 2015, and July 15, 2017.
We used mixed-effects models to examine the main effects of the family intervention, main effects of the clinic intervention, and their interaction. There were no significant interactions on the primary outcomes of QOL, asthma symptoms, and asthma consequences (missed school days). There was, however, a main effect of the clinic intervention such that youth who received it vs those who did not had fewer unplanned doctor visits (β = −.17; 95% CI, −0.28 to −0.05; P = .01) as reported by the caregiver.
Multilevel, multisector interventions recognize the complexity of influences on the prevention and control of asthma. The results from Respira Sano suggest that health care-level interventions are critical to improving important expenditure-related outcomes in the affected population.
Several major limitations of the study are important to note, including the quasi-experimental study design, the reliance on self-report for assessment of change in the primary outcomes, and the lack of process evaluation data to minimize risks associated with type III error.
Asthma is a prevalent chronic health condition affecting almost 10% of children <18 years of age in the United States.1 Asthma affects racially/ethnically diverse children and those of lower socioeconomic status (SES) more than their counterparts.1 Health care expenditures attributable to asthma have been established as $5.2 billion annually.2 This is due, in part, to children with asthma incurring more emergency department (ED) visits and hospitalizations than children without asthma.2 Asthma is also associated with lost school days and workdays, including >1.5 million days per year in California, and costs California nearly $345 million per year in school absenteeism alone.3 The burden of a chronic disease is also felt on the family.4,5 This is particularly relevant to residents of Imperial County, California, given that economically it is one of the poorest counties in the United States, with unemployment and poverty rates among the highest in the nation.6 Thus, given the socioeconomic profile of the partner community and those similar to it across the United States, efforts to reduce the burden of asthma are clearly needed.
In addition to causing financial strain, asthma also negatively impacts the quality of life (QOL) of children and their families. Having a chronic physical illness can impair children's psychological, emotional, and social functioning.7 For parents, asthma in children can increase worry and concern, as well as limit the ability to perform work and household activities.8 For families, asthma can disrupt daily activities, causing plans to change and vacations to be canceled.
Targeting improvements in children's and parents' management of asthma,9 provider interventions,10 and the environment11 have all been identified as effective ways to promote a degree of asthma control. This approach is responsive to the calls for multisector, multilevel interventions that target multiple levels of influence simultaneously, and builds on the strengths of evidence-based interventions that target each level of influence (eg, family vs school/community) and evaluate their synergistic effects to achieve the study outcomes.
Interventions delivered by community health workers (CHWs), or promotoras, are effective at promoting disease control and healthy behaviors, among several other achievements.12 CHWs are uniquely able to engage families and ensure excellent fidelity to intervention delivery.13 Our experience suggests that the latter will occur with greater likelihood when the modality of delivery is home visits; thus, home visits by CHWs is a patient-centered approach. Negative asthma sequelae are largely preventable if children and caregivers engage in asthma management behaviors, such as adherence to using controller medications, monitoring symptoms, and reducing exposure to environmental triggers.
Clinic-based interventions, such as the one described here, represent the latest innovations in systems science applied to health care delivery methods for improving the way the health care system works to promote asthma control among pediatric patients and their caregivers.14 For example, effective clinic-based interventions have the potential to reduce asthma symptom-days14 and increase the number of symptom-free days.15 Clinical guidelines encourage health care providers to discuss the following with patients at every follow-up asthma visit: the expectations of the visit, the goals of treatment, medications, and QOL.16 They also emphasize the importance of jointly determining the goals of treatment with the patient and family.
Community-based interventions are those that affect policy, system, and/or environmental changes that ultimately influence change in the health behavior and/or health outcome.17 This can involve working with organizations in the community (eg, schools) to support asthma control. Among the strategies identified as potentially effective at promoting asthma control is the use of school flags to signal air quality and decision-making regarding activity outside.18
The primary aim of Respira Sano was to identify effective strategies for improving the QOL and asthma control of Latino youth with asthma and their caregivers. We sought to determine whether a combined approach (family plus clinic) was more effective than a family-only intervention or a clinic-only intervention, all within the context of a community-wide intervention that involved schools and media outlets. Respira Sano was informed by components of the National Asthma Education and Prevention Program Expert Panel Report 3 (EPR-3) Asthma Guidelines19 while being specific to the needs of a community on the California-Mexico border. The family, clinic, and community interventions were based on the latest evidence for effecting changes in health outcomes, family support, systems of care, and community support. Youth asthma management interventions that include a family component produce greater effects than other treatment modalities, but caregivers need guidance and support from a prepared practice team and within a supportive environment. The chronic care model specifies changes to 4 dimensions of the health care organization (self-management support, delivery system design, decision support, and clinical information systems) that were hypothesized to activate and inform youth and their families and to prepare a proactive provider.20,21 There is evidence that system changes are needed to achieve lasting changes in health care delivery. As such, the clinic intervention involved the development of an Asthma Care Model, consistent with other chronic care models for diabetes and obesity found to be effective at changing both health care and health outcomes.
The presence and willingness of many stakeholders to address childhood asthma is vital for the success of a multisector intervention. Table 1 provides an overview of the various stakeholder groups, their perspectives, and their member characteristics. Figure 1 provides a timeline of when the stakeholders were engaged in intervention activities. The text following the figure provides information on stakeholder roles and responsibilities, as well as frequency of meetings.
Stakeholder Groups, Roles and Responsibilities, and Member Characteristics.
Study Timeline When Stakeholders Engaged in Intervention.
The project was led by the Institute for Behavioral and Community Health (IBACH) at San Diego State University (SDSU), in partnership with Clinicas de Salud del Pueblo, Inc (CDSDP), Comité Cívico del Valle (CCV), and Impact Assessment, Inc (IAI). They prepared the grant application as a team and thus identified their assets and roles in its preparation. For example, CDSDP and IBACH brought >10 years of experience conducting clinic- and community-based interventions to prevent and control diabetes and obesity.22,23 In addition, they, as well as CCV, brought years of experience using the CHW model for promoting individual, family, and community change.17,24-27 Similarly, IAI brought strengths in provider and community interventions.
In terms of their roles, IBACH had the primary responsibility for achieving the technical aspects of the project and providing scientific direction while also complying with the financial and administrative policies and regulations associated with the award. CDSDP and IAI shared leadership on the clinic component. IAI led the development of training for clinic staff, collaborated on the development and training of the CHWs, and helped manage process evaluation data. IAI also identified expert pediatric pulmonologists to assist with training for clinicians. CDSDP led the implementation of the Asthma Care Model. In addition, CDSDP shared leadership with CCV on the CHW component given their expertise in this area, while CCV took the lead on the community component given their record of accomplishments with school-based interventions and environmental health leadership. All parties contributed to the design of media intervention. All organizations shared leadership on collecting and managing process and impact evaluation data. IBACH took primary responsibility for data preparation and site-specific analyses. The investigators communicated monthly, either by phone or in person, to discuss the study design, recruitment, intervention, and evaluation activities. Monthly meetings allowed for an in-depth discussion of study implementation and findings.
Respira Sano engaged 2 advisory groups, specifically the Parent-Youth Advisory Group (PYAG) and the Community Stakeholder Advisory Group (CSAG). Youth with asthma (n = 15; aged 4-22 years) and their caregivers (n = 11; aged 25-64 years) were engaged through the PYAG. Nine members representing health care, schools, environmental health researchers, city government and policymakers working on air quality issues, community advocates, and the Imperial County Public Health Department were engaged through the CSAG.
For the PYAG, the lead agencies (CCV and CDSDP) invited 8 to 12 caregivers of youth aged 6 to 17 years with asthma to be part of a group that guided the development of the research study from its inception through writing of the grant application. PYAG members had to be residents of Imperial County, and the youth had to have physician-diagnosed asthma. At the initial meetings, we described the PCORI funding announcement requirements and their role in developing the asthma intervention. Information was sought on asthma outcome priorities and on what services were needed and/or preferred from health care providers and school personnel. A follow-up meeting involved the review and approval of key sections of the grant application, namely the research strategy sections on burden, how the results would likely improve health care outcomes, and how and where the study would be conducted, as well as patient-centeredness, commitment of patient and stakeholder engagement, and potential facilitators and barriers to dissemination and implementation. From these meetings, we learned that youth were interested in improved lung function (“breathe easier”) and in receiving better care from their doctors. Caregivers were interested in achieving better asthma control and experiencing fewer asthma symptoms themselves and in their children. Youth and caregivers agreed that working with families, clinics, and schools would be necessary to tackle asthma in Imperial County. They felt that working with families individually (families in their homes and families with their health care providers), in addition to working with them in groups, would help them attain the skills and support to achieve asthma control. Throughout the study duration, the PYAG met quarterly to further develop the intervention, monitor its implementation, and provide ongoing guidance on every aspect of the research study. Incentives such as meals and incentives that helped defray the cost of gas and babysitting were associated with attending meetings. Similarly, meeting times in the early evening, after school, were most desired by this group.
For the CSAG, quarterly meetings were held with representatives, including CDSDP providers and staff, to get their feedback when writing the grant application and to facilitate their oversight of the research similar to the PYAG. Clinicians were concerned with their patients achieving good asthma control and not needing to go to the ED. Although the asthma outcome priorities varied slightly across PYAG and CSAG, everyone agreed that QOL was an all-encompassing construct that represented all of their priorities. To promote engagement in these meetings, we emailed the minutes to participants to keep them apprised of study progress.
Discussion themes varied by each year of the study (see Table 2), but discussions regarding family recruitment occupied a significant amount of time. Both groups were keenly aware of our need to maintain our recruitment rates and goal of study quality. After sharing our recruitment materials with stakeholders, we obtained critical feedback on how our recruitment materials were perceived in the community and how we might improve them. They also ensured that materials and measures optimally balanced technical correctness and cultural sensitivity. For example, by sharing intervention materials with our patient stakeholders, we obtained confirmation that the materials would be well received once critical rewording changes had been made.
Themes Discussed at PYAG and CSAG Meetings.
Finally, when we combined these 2 advisory groups, we often provided simultaneous English/Spanish translations, which can be challenging because it adds time to meetings. We held meetings together to learn about study results and to allow stakeholders to provide input on the best ways to disseminate study findings. At these meetings, we introduced the concept of dissemination and began a discussion of how to present findings in a meaningful and useful way. All partners were satisfied that the partner agencies could disseminate the results to others via their professional, medical, and community networks; however, they suggested that further dissemination to community members could be achieved via media outlets such as television and radio news and community programs, as well as door-to-door flyers. The investigative team continues to meet to review analyses, discuss plans for data analyses, and disseminate study results.
The Respira Sano partners were dedicated to seeing the intervention sustained. CDSDP provided ongoing training and support to maintain the systems in place to provide quality asthma care. They institutionalized an asthma care plan that involves spirometry testing, asthma education delivered by a patient care coordinator (PCC) and CHWs, and a quality assurance monitoring system. Among other things, the PCC and CHWs educate families on what kind of symptoms can be managed at home or in a primary care setting vs the ED. These system changes are supported by alerts and other prompting mechanisms in their electronic health record (EHR) system that occur when a child with asthma presents to the clinic. CDSDP also is participating with local hospitals in a task force to address the high rate of ED visits. CCV will continue to host the annual Environmental Health Leadership Summit (see the Methods section) with support from many funders and collaborators. In addition, they will continue to support the use of the home visitor model statewide. The CHWs participated in a statewide assessment of asthma home visiting programs conducted by impact assessment staff in partnership with the California Department of Public Health. This assessment highlighted strategies used by successful asthma home visiting programs. Additionally, the CHWs participated in a train-the-trainer program hosted by the California Department of Public Health.
Respira Sano was designed to improve QOL, asthma symptoms, and consequences of asthma among Latino youth with asthma and their caregivers. We did this by testing 2 intervention approaches (a family intervention and a clinic intervention), plus their combined effects, all within the context of a community intervention that involved working with schools and media outlets. A cohort of 400 youth aged 6 to 17 years with mild to severe persistent asthma and their caregivers were recruited and consented to participate; they completed the baseline assessment, were assigned to 1 of 4 conditions, and were followed for 12 months. The 4 conditions were as follows:
Respira Sano was a 2 (family intervention: yes/no) × 2 (clinic intervention: yes/no) factorial study nested within a community intervention (see Table 3). We were interested in whether it was necessary to engage in a multisector approach to reach families (testing the combined effects of both the family and clinic interventions) or whether family and clinic intervention approaches could work well alone (testing the main effects of each) in improving QOL, asthma symptoms, and consequences among youth and caregivers, all within the context of a community intervention. Randomization strategies are described in detail in the next section. Evaluation activities occurred at 4 levels at 2 time points (baseline and 12 months): youth, caregiver, clinic, and community. In this report, we describe the results from the primary outcomes and 1 secondary outcome (lung function) as reported by the youth and caregivers.
Study Design.
Our target population lives in Imperial County, which is on the US-Mexico border in Southern California. Imperial County is a designated rural area with an average density of 109 people per 1000 square miles6 and has 17 cities and towns (incorporated and unincorporated). To minimize travel costs, we focused on the youth living in the cities/towns in close proximity to each other. The residents (N = 174 528) are primarily of Mexican origin.
For Respira Sano, youth (and their caregivers) were identified through EHR queries from CDSDP and school nurses in elementary and middle schools, as well as during community events and by word of mouth. Media outlets shared information with the public about the opportunity to enroll. Potentially eligible families received a 15-minute screening either face to face or by phone, with the vast majority being screened by phone. The eligibility screening form assessed eligibility and reasons for ineligibility. All attempted screenings were coded as never reached, eligible, or ineligible. We attempted to schedule appointments to conduct the baseline assessment protocol with everyone who was coded as eligible.
Youth and their caregivers were eligible to participate if they met the following criteria: (1) youth identify as Latino/Hispanic/Mexican origin; (2) youth are between 6 and 17 years of age and caregivers are aged ≥18 years; (3) youth and caregiver predominantly speak English or Spanish; (4) youth and caregiver live together at least 4 days a week, and the caregiver is the youth's legal guardian; (5) youth and caregiver live in Imperial County and plan to stay for the duration of the study; and (6) youth has mild to severe persistent asthma as determined based on caregiver report of medication use and frequency of asthma symptoms in the past 12 months. Youth were classified as having moderate to severe persistent asthma if the caregiver reported that (1) the youth currently takes 2 controller medications or at least 1 combination medication; or (2) the youth had any of the following symptoms ≥5 times per month in the past 12 months: wheezing with a cold, a cough that would not go away, wheezing with exercise or running or playing hard, or sleep disturbance caused by wheezing, coughing, chest tightness, or shortness of breath. All other youth were classified as having other than moderate to severe persistent asthma. Youth and their caregivers were ineligible to participate in the study if they had participated in another asthma education program or research study in the past year.
Bilingual/bicultural research assistants were trained and certified in both English and Spanish for all recruitment protocols. They followed strict protocols in implementing the recruitment protocol, which consisted of screening the caregiver and youth for inclusion and exclusion criteria and obtaining informed consent from eligible caregivers and assent from youth. The research assistants then completed the baseline evaluation protocol as described in the “Data Collection and Sources” section.
Once the baseline evaluation protocol was completed, families were allocated to condition by the project coordinator using a 2-step process: (1) families whose children were CDSDP patients were placed on 1 list, and families whose children were not CDSDP patients were placed on a second list; and (2) families on each list were randomly assigned to either the family or no family intervention using a computerized random number generator. This resulted in a balanced sample of youth in each of the cells represented in Table 3.
If assigned to the family intervention, the caregiver received a call from the assigned CHW to begin the intervention. The clinic intervention was ongoing; thus, if the youth was a patient of CDSDP, the family was introduced to the clinic intervention by the clinic asthma educator. The community intervention was ongoing and designed to reach individuals connected to the schools and in the community.
Finally, all caregivers were contacted by Respira Sano evaluation staff to schedule the 12-month evaluation visit at a time most convenient to the youth and caregiver. A window of 3 additional months was provided to families to complete the evaluation protocol.
CCV employed 4 full-time CHWs and a full-time coordinator to oversee the program, which involved four 1-hour home visits, four 10-minute phone calls, and 4 mailed newsletters to each family over a 2-month period. The CHWs were current or former employees of CCV, with experience conducting similar outreach efforts. It was important that the youth and their caregivers participated in the home visits together because we wanted the entire family to benefit from the intervention and work together toward better asthma control. However, phone calls primarily occurred with the caregivers rather than the youth. Through these contacts, CHWs worked with families to improve communication about asthma in the home, with school personnel, and with the youth's health care providers. The CHWs provided strategies to (1) manage asthma at home, at school, and in the community; (2) reduce environmental triggers; and (3) engage in community efforts to reduce pollutants that exacerbate asthma. CHWs shared tools and resources with the families, including a backpack of educational materials, a peak flow meter, and a bucket of asthma-safe cleaning tools. Home visits were conducted in English, Spanish, or bilingually depending on the preferences of the family.
CHWs received 60 hours of training on asthma, appropriate use of medications, indoor and outdoor environmental triggers, and effective communication strategies between family members and with health care providers. They also received training on documentation of intervention contacts and no shows/no answers, popular education methodology, public speaking, and advocacy. The lead asthma educator from CDSDP and the Healthy Housing educator from IAI were among the trainers. The following topics were covered: appropriate use of medications; self-management and goal setting, including following an asthma action plan; home assessment for triggers; trigger reduction strategies, including asthma-safe cleaning; information about local air-quality issues; and advocacy opportunities.
The goals of the clinic intervention were to increase adherence to clinical asthma care, including attendance at 4 visits annually with a physician and/or asthma educator. The clinic intervention took place within 3 of CDSDP's 15 clinic sites in Imperial and Riverside counties. Federally qualified health centers (FQHCs) provide primary care for adults and youth, prenatal care, and dental services for families throughout this area. CDSDP spent the first year initiating quality improvement practices and policies to provide the most comprehensive quality asthma care possible. The quality improvement strategies included 5 components:
Once trainings were completed, new policies and systems were implemented and closely monitored. Medical teams conducted self-assessments, and records were reviewed to ensure completion of spirometry, asthma action plans, flu shot administration, and appropriate self-management education. The medical director and the director of programs coached staff who needed additional support in implementing the new asthma care practices and policies.
The Respira Sano community intervention consisted of 3 distinct components:
We originally identified 3 outcomes: (1) youth- and caregiver-reported asthma control, (2) youth QOL, and (3) youth lung functioning. Following intervention development and stakeholder group considerations, our primary outcomes were identified as follows: (1) caregiver report on youth's asthma symptoms and consequences, (2) youth report on QOL and asthma symptoms and consequences (among a subsample of youth old enough to respond to interview questions based on established validity data [9- to 17-year-olds]), and (3) caregiver QOL. One secondary outcome was lung function using spirometry (see the “Secondary Outcome” section). Reported Cronbach α coefficients are derived from study participant data suggesting strong internal consistency of most study measures.28 Finally, we examined differences in the outcomes stratified by severity status (moderate to severe persistent asthma vs mild asthma) and found no differences in the results. Thus, we report on the overall findings:
Lung function was assessed using spirometry, in part to confirm asthma severity classification, but also to determine whether youth improved their lung function according to study condition. Spirometry tests were conducted using the EasyOne Model 2001 spirometer. Youths' height and weight were measured at the time of the spirometry test. Best trials of the tests were used for analyses. Percentage predicted values were based on data collected by Hankinson et al.33 We only retained test results that received a quality grade of A to C, which indicate reliable results. Given this quality grade restriction, 51% (n = 204) and 33.5% (n = 134) of the spirometry data were dropped at baseline and 12 months, respectively. There was no differential dropped rate by study condition at either time point.
Our target community was Imperial County, located on the California-Mexico border between San Diego, California, and Yuma, Arizona. Asthma is a particular challenge in rural and US-Mexico border communities where rates of asthma are much higher than in other parts of the United States. For example, approximately 21% of 5- to 17-year-olds had active asthma in Imperial County compared with 12% in California.34 In 2014, Imperial County youth visited the ED for asthma at more than twice the rate of youth in all of California; similar differences were observed for asthma-related hospitalizations.34 In Imperial County, 77% of people with asthma did not receive a written asthma management plan from a health care provider. One factor contributing to high asthma rates is poor air quality; there were >100 days in 2015 where particulate matter (PM10) exceeded the state standard at 2 monitoring sites in the county.34 This is why the community intervention focused on educating residents about air quality and strategies to prevent exposure to environmental toxins.
The family intervention (4 home visits) was designed to be delivered over a 2-month period. This was determined to be a feasible and effective method for modifying behaviors and environmental changes in the home environment. The clinic intervention was designed to be ongoing, given that it involved changes in the delivery of asthma care at the 3 clinic sites. The community intervention was delivered over a 24-month period and included the school flag program and the media campaign. Evaluation changes were examined over a 12-month period, a period that did not introduce confounding by seasonality in the analyses and potentially long enough to examine possible maintenance of improvements in QOL and asthma symptoms and their consequences.
Bilingual/bicultural research assistants were trained and certified in both English and Spanish on all evaluation protocols. They were trained on all study-specific evaluation protocols, as well as culturally competent interviewing, social desirability and response bias, confidentiality, and ethical behavior. In addition, they completed the SDSU IRB online training and received a certificate before data collection. Training continued during data collection in the form of booster sessions. Assessment of research assistants' adherence to protocols and high interrater agreement (eg, r > 0.90 on measures of height and weight for spirometry purposes) on several participating families occurred before field data collection at each evaluation time point. Quality control was an ongoing activity designed to ensure the highest quality data possible. Before entering or scanning forms into the database, we examined study forms for completeness and accuracy. We also used a tracking system and checklists to ensure that all data were compiled for all assessments. We used range and logic checks to minimize errors. Research assistants, blinded to condition assignment, followed strict protocols in implementing the baseline and 12-month follow-up evaluation protocols with participants. This consisted of collecting measured data (ie, lung function using spirometry, height, weight) from youth, and self-reported data using an interviewer-based structured questionnaire with both youth and caregivers.
All data collection occurred in person and in the language preferred by the participant (English or Spanish). Data collection occurred primarily in 3 locations: study offices located centrally in one of the cities, in one of several rooms at the clinic sites, and in participants' homes. Evaluation protocols were designed to occur in many different locations, and adherence to standard administration was tested during the certification process to maximize its implementation during the study. Youth and caregivers were remunerated $20 for completing the baseline protocol and $30 for completing the 12-month postbaseline protocol.
Efforts to reduce attrition occurred during 3 phases: during the baseline evaluation time point, in the interim between the evaluation time points, and just before the 12-month evaluation time point. During baseline, staff were trained to engage the youth to foster retention (eg, informal vs formal interaction style), including by providing refreshments and entertainment (eg, videos, books, games) at all evaluation visits. Between the baseline and 12-month follow-up, the study team sent cohort maintenance postcards to families. Just before the 12-month time point, the study team conducted multiple reminder calls and sent emails. In addition, they provided weekend appointment slots and drove to homes to leave notes on their doors. Alternative contact information (eg, neighbor, other family member) was collected at baseline and used with hard-to-reach families. Before contacting families for their 12-month evaluation time point, we sent a mailer that included a thank-you letter and a magnetic notepad with the study's phone number. The letter reiterated the importance of the family's role in the study, informed families that their 12-month evaluation time point was upcoming, and asked families notify the study team of any new contact information. We received some returned mailers that indicated a new street address or a move but no new address. In the latter case, evaluation staff contacted intervention staff to verify current contact information.
When preparing the PCORI application for research funding, we determined the number of youth needed in the cohort to test 1 of the primary aims; we used the power function for the 2-sample t test in R to detect an effect size of 0.3, 0.6, and 0.9 on youth QOL using a 7-point scale31 with 80% and 90% power when comparing the family plus clinic intervention and the control (community only) condition. The effect size reflects how many SDs there are for the mean differences between the 2 groups. For an effect size of 0.3, it means the mean difference between the 2 groups is 0.3 SD. Then we inflated these sample size estimates to consider any possible clustering by clinic. We also inflated the sample size to consider an attrition rate of 25% over the 12 months. This formula led to the approximate sample size of 400 needed to detect a small effect of 0.3 in our study with 90% power after considering possible intracluster correlation (0.5) and attrition (25%).
To describe the sample characteristics for all 4 study conditions, we conducted univariate analyses on our primary outcomes and the baseline covariates. We conducted cross-condition comparisons on all measured baseline covariates to identify potential systematic differences among the youth and caregivers. Continuous variables were compared using 1-way analysis of variance, and categorical variables were compared using χ2 tests. We adjusted for measured covariates that were not balanced in the subsequent multivariate models.
To assess intervention effects on our primary outcomes at 12 months postbaseline, within the context of our 2 × 2 factorial design, we examined the following using mixed-effects models: (1) the main effects of the family intervention, comparing outcomes between those who were randomly assigned to the family intervention (family and family plus clinic conditions) vs those who were not; (2) the main effects of the clinic intervention, comparing outcomes between those who were clinic pediatric patients and thus had the potential to receive the clinic intervention (clinic and family plus clinic conditions) vs those who were not clinic patients; and (3) the interaction of the 2 main effects: family intervention × clinic intervention.
In addition, demographic covariates whose prevalence was significantly different across conditions were included in the model to adjust for potential confounding effects (ie, child age, caregiver education and employment, household poverty status); baseline outcome values also were adjusted in the models. Although the sample size was relatively large (N = 400), diagnostic analyses on the residuals were conducted to confirm that assumptions for fitting the mixed-effects models were met. All of our primary outcomes were continuous, but some outcomes were right skewed; these variables were log-transformed to satisfy the normal distribution assumption of the models. Regarding the missing values, the other covariate with missing values was poverty status (4.5%). As expected from the small rates of missing data, the results including missing cases and excluding missing cases were very similar. The results reported in this report include participants with partial missing outcome values. Finally, 3 models were built for each primary outcome:
The P value for family × clinic interaction is reported from model 2, and the β coefficients, 95% CIs, P values, and adjusted means for the main effects of the family and clinic interventions are reported from model 3. The estimated coefficients, 95% CIs, and the adjusted means from the log-transformed variables were back-transformed into the original scale for ease of interpretation.
Given the number of primary outcomes, potentially increasing the likelihood of type I error, we considered a per family-wise error rate (in this study, we define each response variable as a family), using the Bonferroni adjustment so that the overall family-wise error rate is maintained at a P ≤ .05 level. Table 4 summarizes the outcomes under each family and how the α was adjusted for each outcome.
α Adjustments to Control for Type I Error in Analyses of Primary Outcomes.
The statistical software SAS v.9.4 (SAS Institute) was used for all analyses. The mixed-effects models were conducted using the SAS PROC MIXED procedure.
No significant actions were taken by our IRB, data and safety monitoring board, or other regulatory oversight body during the study. IRB renewal was obtained through October 2018. Given how the interventions were ultimately designed, we modified our aims to examine youth-reported QOL and asthma symptoms and consequences, caregiver report of youth's asthma symptoms and consequences, and caregiver QOL. Second, we requested and received IRB approval to expand our inclusion criteria to allow participation of youth with mild persistent asthma. We made this change for several reasons: (1) recognition that those with mild persistent asthma would benefit equally from the family and clinic interventions; (2) simplifying the process for health care providers in terms of their actions with youth who had more than mild intermittent symptoms; and (3) anticipated challenges with recruitment. We did not want to modify our approach too late in the recruitment process and thus bias our approach. Therefore, we began including youth with mild persistent asthma within 1 month of the start of enrollment.
Of the 1890 families processed for recruitment, 54.4% (n = 1028) were eligible for recruitment. Of these, 82.0% (n = 843) completed the recruitment screener. Among those not eligible were 85 youth who did not meet our screening criteria for active asthma or physician-diagnosed asthma. Five youth had an illness that met ineligibility criteria. A total of 63% of the families (n = 533) were eligible for study enrollment. From among these, we met our required sample size of 400 youth/caregiver dyads. A detailed CONSORT diagram is presented in Figure 2. The Respira Sano study completed 388 twelve-month assessments, representing a 97% follow-up rate. Reasons for not completing the 12-month assessment were that 4 participants refused, 2 moved outside the study area, 4 were never reached, and 2 were reached but failed to complete their appointments.
Respira Sano CONSORT Diagram.
Table 5 shows the caregiver-reported youth characteristics and youth-reported characteristics at baseline; Table 6 shows the caregiver characteristics. The evaluation protocol was administered to caregivers of all dyads enrolled in the study. Our youth interview was only administered to youth aged ≥9 years (n = 244); this was due to lack of scale validity data on younger children.
Caregiver Report on Youth and Youth Self-reported Characteristics at Baseline Overall and by Condition.
Caregiver Characteristics and PAQLQ Score at Baseline Overall and by Condition.
The average (SD) age of the youth was 11 (3.2) years. Youth in the clinic and family plus clinic conditions were about a year older than those in the other 2 conditions (P ≤ .01). All youth were Latino, and more than a third reported being bilingual (English/Spanish). All had health insurance in the past year except 1 youth. Approximately two-thirds of the youth were diagnosed with moderate to severe persistent asthma. Asthma symptoms were reported as occurring more than twice a month but less than twice a week. The number of unplanned doctor visits during the past year was high for all youth (about 5 visits per year) and higher by an additional year in 2 conditions (P ≤ .01). ED visits were also high, averaging between 1 and 2 visits in the past year. In the subsample with youth self-report, there were no differences by condition. The average (SD) age of youth was 13 (2.5) years; 62% were boys. All youth identified as Latino, although only 14% were born outside the United States (13.9%). None of the youth reported smoking. Their QOL was very high; they reported only being bothered a bit or just once in a while with symptoms. However, they also reported experiencing symptoms about 5 times during the past 2 weeks and missing almost 2 school days during the past 3 months.
As shown in Table 6, the average (SD) age of the caregiver was 39 (8.3) years. Most of the caregivers were female (94.5%) and the mothers of the youth (90.5%). Most of the caregivers identified as Latina (98.8%), were born outside the United States (64.3%), and spoke Spanish as their primary language (72.3%). Nearly three-quarters were married and lived in a household with about 5 people. Almost half of the caregivers were employed; those in the clinic condition were less likely to be employed compared with the other conditions (P ≤ .05). Three-quarters of the caregivers had a high school degree or more; however, a significantly lower proportion of caregivers in the clinic condition had obtained a high school degree or more (P ≤ .05). Nearly two-thirds of the households were living below the poverty threshold, with those enrolled in the clinic and family plus clinic conditions more likely to be living below the poverty threshold compared with the other 2 conditions (P ≤ .001). Most caregivers had health insurance, and few reported smoking. Similar to their children, their QOL was high, with an average of 5.5 on a 7-point scale.
Outcome analyses examined changes over the 12-month period and were organized as follows: (1) caregiver-reported youth asthma symptoms and consequences, (2) youth self-reported QOL and asthma symptoms and consequences (subsample), and (3) caregiver QOL. Table 7 shows the β coefficients, 95% CIs, P values, and predicted values (adjusted mean) of each outcome. In addition, adjusted α levels are indicated with footnote marker superscript “c” in this table.
Evaluation of Intervention Efficacy on Primary Outcomes Including Interactive and Main Effects of the Family and Clinic Interventions, All Nested Within a Community Intervention: Adjusted Means.
There was neither a significant interaction between the family and clinic interventions nor a significant main effect for the family or clinic interventions on the forced expiratory volume in the first second of expiration (FEV1)-to-forced vital capacity ratio. However, there was a significant interaction effect between the family and clinic interventions on FEV1 % predicted (P ≤ .05; see Figure 3). Being in the clinic intervention condition had a differential impact on youth who were (or were not) in the family condition. Among youth who were in the family condition, those who were in the clinic condition had higher FEV1 % predicted than those not in the clinic condition. Among youth who were not in the family condition, those who were in the clinic condition had lower FEV1 % predicted than those who were not in the clinic condition.
Significant Interaction Between the Family and Clinic Interventions on FEV1 % Predicted: Mean at 12 Months Adjusted for Baseline Values (n = 139).
The Respira Sano study was designed to test whether a family intervention, a clinic intervention, or the combination of the 2, all nested within a community intervention, improved youth- and caregiver-reported QOL and asthma symptoms and their consequences over the period of 12 months. Findings indicated that the clinic intervention resulted in fewer unplanned doctor visits due to youth's asthma as reported by caregivers. Measured lung function produced results with low reliability, forcing us to drop between a third and a half of the tests at the 2 measurement points. Nevertheless, using the valid data, there was a significant interaction between the family and clinic interventions, suggesting that the combination of the 2 interventions was better than either intervention alone to promote lung functioning. However, the combined condition (family plus clinic intervention) was no more effective at improving lung function than the community-only intervention.
A comprehensive approach to asthma control is appropriate for addressing childhood asthma. This may be particularly true in economically depressed and rural areas. In the current study, results generally supported a strong clinical intervention reinforced by a community intervention. There was no effect of the family intervention. This is important because states like California consider Medicaid reimbursement for asthma home visiting services. This study contributes to the body of knowledge supporting the benefits of quality clinical asthma care.
Several factors must be considered when drawing conclusions from the study findings: the study design, the target population, and the setting, among others.35 Regarding the study design, in this 2 × 2 factorial study, families were in a condition based on 2 factors: (1) randomization to the family intervention or no family intervention, and (2) where the youth received health care services; if the youth was a patient of the health care system delivering the clinic intervention, then he or she was in 1 of 2 clinic intervention conditions (clinic alone or clinic plus family). Our approach ensured balance in demographic factors across the 2 family intervention conditions, thus meeting an important assumption of randomization. However, given that involvement in the clinic intervention was predicated on the youth being a patient at the clinic, this introduced potential biases to our design. As a result, our approach did not remove differences observed in SES indicators by clinic intervention condition; the caregivers/households of youth patients of CDSDP were less likely to have a high school degree, and more likely to be unemployed and to live in poverty than those who were not patients of CDSDP. Similarly, our study design and resultant condition differences limit inferences to a population with lower income who use FQHCs for their health care needs. Regarding the target population and setting, we worked in a rural border environment, one of the least densely populated and poorest counties in California.36 Imperial County abuts the large Mexican city of Mexicali; thus, it is not unexpected that most of our caregivers were born outside the United States (primarily in Mexico). Given these 2 factors, the cultural milieu may have more in common with other rural border areas from Texas to California than with other Latino or immigrant populations elsewhere in the United States. For example, CHW interventions are differentially effective based on a combination of their approach and the geographic location of intervention delivery in the United States.12
This is the fourth large intervention involving multilevel, multisector changes targeting health that we and our partners have conducted over the past 10 years.37-39 From these efforts, we have discovered the need to institutionalize intervention strategies to ensure that they are sustained. This has involved adopting organizational policies to reinforce adherence to new practices, as well as modifying aspects of the physical environment.40 Intervention strategies need to consider local cultural and social factors in their adoption and adaptation processes.41
In this study, no subanalyses were conducted to examine whether the intervention had a differential effect on a subpopulation of interest.
This study had several limitations. First, from a measurement perspective, reliability of the lung function test limited the conclusions that we can draw from these findings. As such, we now rely on youth and caregiver report for our primary outcomes. Second, this was a quasi-experimental study with randomization only occurring for the family intervention component. Thus, analyses comparing the clinic vs non-clinic sample were subject to potential selection bias and other unmeasured confounders. Third, although 1 year (12 months) is an appropriate follow-up period to minimize the effects of seasonality on reported asthma symptoms, it may be too long to capture the immediate effects of a 2-month family intervention and too short to meaningfully predict sustained improvements or to determine whether additional improvements will be observed. Finally, relaxing the criteria to include youth with any level of persistent asthma (mild, moderate, or severe) likely increased the overall amount of variance observed in the outcomes, making it more difficult to find an effect. However, this limitation was outweighed by our interest in testing relevant strategies with youth and families who need them.
From an intervention perspective, trials of this type are inherently delivered with the real-world constraints of everyday practice. For example, we know that scheduling home visits is challenging and can lead to compromises in intervention fidelity.42 Resource limitations may affect the amount of home visiting that can be achieved. Future analyses will examine these and other issues consistent with RE-AIM (Reach, Effectiveness, Adoption, Implementation, and Maintenance) and other intervention evaluation frameworks.43,44 Second, delivery of the clinic intervention may have been affected by the heavy patient caseload carried by many of the health care providers in our partner clinics. Third, it is possible that changes were due to other asthma control initiatives occurring in the county (eg, Imperial Valley Asthma Coalition). However, these county-level initiatives would have affected children in all conditions equally, thus minimizing this threat to internal validity. Finally, it is possible that among participating families, there was some “cross talk” between those in different conditions.
Study investigators have worked with research partners on subsequent grant submittals, given the positive working relationship experienced during this PCORI-funded study. One of our research partners obtained additional funding through the Imperial County Public Health Department to continue conducting asthma-related home visits. Some study participants asked us if we could continue with asthma support groups, and others gave testimonials about their positive experiences.
Regarding the broader institution (ie, SDSU), we have begun collaborations with 2 other investigators by including them in our discussions of study findings. These investigators have expressed an interest in continuing to work on asthma control in Imperial County, and we are excited to have them involved. We have introduced other faculty at SDSU (and the SDSU satellite campus in Calexico) to organizations based in Imperial County with whom we have solid relationships. These faculty were interested in research in air quality, LGBT issues, and promoting physical activity.
Health care organizations, clinic managers, and clinicians will find the outcomes of the Respira Sano study useful as they develop and integrate quality improvement measures in their own settings. The integration of asthma-focused components in the EHR, training for health care providers and staff clinic-wide, and the implementation of several new practices led to significant improvements. Although there was a learning curve to standardize these practices, they are replicable in other settings. Other FQHCs—many of which face some of the same challenges faced by our clinic partners, such as provider and staff turnover and patients who frequently miss scheduled visits—should take note of the positive impact on asthma outcomes demonstrated by the present health care-level intervention. Health care insurers and other managed care organizations will also see the benefits of quality clinical care in conjunction with home visiting services. Although we did not measure health care costs, a reduction in costly ED visits is significant when considering reimbursement for clinic and home visiting services. We were not able to show an equivalent positive result through the family-based intervention, however. This aspect of a comprehensive intervention strategy needs additional development and investigation, which could ultimately increase the effectiveness of a multilevel approach.
Research reported in this report was funded through a Patient-Centered Outcomes Research Institute® (PCORI®) Award (AS-1308-05876). Further information available at: https://www.pcori.org/research-results/2013/comparing-programs-improve-asthma-control-and-quality-life-latino-youth-living
Elder JP, Ibarra L, Rossi D, et al. (2021). Comparing Programs to Improve Asthma Control and Quality of Life for Latino Youth Living in Rural Areas and Their Caregivers—The Respira Sano Study. Patient-Centered Outcomes Research Institute (PCORI). https://doi.org/10.25302/04.2021.AS.130805876
The [views, statements, opinions] presented in this report are solely the responsibility of the author(s) and do not necessarily represent the views of the Patient-Centered Outcomes Research Institute® (PCORI®), its Board of Governors or Methodology Committee.
This book is distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License which permits noncommercial use and distribution provided the original author(s) and source are credited. (See https://creativecommons.org/licenses/by-nc-nd/4.0/
Your browsing activity is empty.
Activity recording is turned off.
See more...
