Clinic navigation and home visits to improve asthma care in low income adults with poorly controlled asthma: before and during the pandemic

Abstract

Asthma-related deaths, hospitalizations, and emergency visits are more numerous among low-income patients, yet management guidelines do not address this high-risk group’s special needs. We recently demonstrated feasibility, acceptability, and preliminary evidence of effectiveness of two interventions to improve access to care, patient-provider communication, and asthma outcomes: 1) Clinic Intervention (CI): study staff facilitated patient preparations for office visits, attended visits, and afterwards confirmed patient understanding of physician recommendations, and 2) Home Visit (HV) by community health workers for care coordination and informing clinicians of home barriers to managing asthma. The current project, denominated “HAP3,” combines these interventions for greater effectiveness, delivery of guideline-based asthma care, and asthma control for low-income patients recruited from 6 primary care and 3 asthma specialty practices. We assess whether patients of clinicians receiving guideline-relevant, real-time feedback on patient health and home status have better asthma outcomes.

In a pragmatic factorial longitudinal trial, HAP3 enrolls 400 adults with uncontrolled asthma living in low-income urban neighborhoods. 100 participants will be randomized to each of four interventions: (1) CI, (2) CI with HVs, (3) CI and real-time feedback to asthma clinician of guideline-relevant elements of patients’ current care, or (4) both (2) and (3). The outcomes are asthma control, quality of life, ED visits, hospitalizations, prednisone bursts, and intervention costs. The COVID-19 pandemic struck 6.5 months into recruitment. We describe study development, design, methodology, planned analysis, baseline findings and adaptions to achieve the original aims of improving patient-clinician communication and asthma outcomes despite the markedly changed pandemic environment.

1.0. INTRODUCTION

Asthma-related deaths, hospitalizations, and ED visits are more numerous among poor persons, people of color, and older adults, many who suffer from additional chronic diseases. The Guidelines for the Diagnosis and Management of Asthma, last fully revised in 2007 (Expert Panel Report 3)¹ with a 2020 Focused Update² have not addressed health equity mostly because there was little research on social determinants of health in the management of asthma. Patients with current or past tobacco exposure also are typically excluded from asthma research and therefore the guidelines.

Reports reveal that many clinicians do not adhere closely to asthma Guidelines.^3–7 The reasons are unclear- whether due to Guidelines complexity, patients’ inability to reduce out-of-clinic exposures, clinicians’ perception of lack of efficacy, time, or assistance needed to complete these tasks, or other reasons. The element least likely to be followed surrounds action plans.^4–9

In 2003 the Institute of Medicine, now the National Academy of Medicine, focused on inequities in health outcomes and the need to improve patient-clinician communication and access to care.¹⁰ Recently the Center for Disease Control and Prevention and others have stressed the importance of identifying the social determinants of health that influence patient health.¹¹ Not all clinicians fully appreciate these barriers to health in their individual patients because 1) most do not make house calls and 2) medical appointments tend to be short, focused on medications, labs, and diagnosis. In addition, patients might not disclose barriers to access. These circumstances promote and sustain inequities and adverse outcomes. This study investigates the structure of patient–clinician interaction and whether changing it can reduce inequities and improve outcomes.

Previously, we had conducted studies demonstrating feasibility, acceptability, and preliminary evidence of effectiveness of two separate interventions to improve access to care, patient-provider communication, and asthma outcomes. In the first, a Clinic Intervention (CI), a Patient Advocate facilitated a patient’s preparations for office visits, attended the visits, and promoted review of the agreed upon recommendations emerging from the visit.^12,13 In the second, a protocol involving home visits (HVs), a community health worker reviewed care coordination recommendations and together with the patient informed the clinician of neighborhood and home barriers to managing asthma.^14,15 The project described here explores whether these interventions can be combined for greater effectiveness, better delivery of guideline-based asthma care, and improved outcomes in low-income patients. We called the study HAP3 (Helping Asthma Patients 3) (Piloted and earlier versions had been called HAP¹⁶ and HAP2¹³) and the interventionist a Navigator.

We hypothesize improved asthma outcomes as a result of enhanced patient-clinician communication, clinician attention to home and neighborhood environmental exposures, and clinician consideration of the guidelines at a program cost offset by lower patient health care utilization. We describe the study design, methodology, planned analysis, baseline characteristics, and asthma status. Unexpectedly, the COVID-19 pandemic struck 6.5 months after recruitment began. In addition, we present the necessary modifications in this markedly changed pandemic environment in order to carry out the original aims of improving patient-clinician communication and asthma outcomes for patients already at high risk.

2. METHODS

2.1. Design, overview, and aims

In this randomized pragmatic, factorial trial, adults with uncontrolled asthma living in low-income urban neighborhoods are recruited, consented, enrolled, and after 3 months randomized to 15 months in one of four assignments: (1) CI only, (2) CI + HV, (3) CI + clinician feedback, and (4) CI + HV + clinician feedback. For CI, a Navigator facilitates preparation for office visits, attends visits, and confirms patient understanding of recommendations.^12,13 During the HV, the Navigator reviews doctor’s recommendations, and with permission and collaboration of the patient informs the clinician of neighborhood and home barriers to health and asthma management.^14,15 Navigators are community health workers: persons familiar with or members of the community served, who have worked or lived with patients of the community, familiar with asthma, comfortable making home visits and working in medical clinics.

The goal of this study is to enroll 100 participants per treatment arm, observing them for 3 months before randomization. This 3-month observation produces a baseline control period for all patients that will not only allow evaluation of each patient’s ability to adhere to study protocols but will also serve as patient-specific reference for the intervention results. Once randomized, patients are evaluated quarterly for 12 months and then observed for three additional months to assess sustainability of the intervention (Figure 1). Our plan is to estimate not only improvements for each patient over baseline in asthma outcomes: asthma control, quality of life, asthma–related hospitalizations, ED visits, and prednisone bursts (an increase of at least 30 mg daily for at least 3 days), but also to assess sustainability of each treatment option.

Figure 1: HAP3 Study Design.

Asthma control is the primary outcome.^35,43 The first data collection visit includes enrollment and collection of baseline data, the second randomization at 3 months. Patients are observed for 3 months before randomization and 3 months after completion of the protocol. Thus, the visit at 3 months is “Time 0.” The intervention spans a year.

Altogether there are 7 planned data collection visits, each asking about outcomes, e.g. asthma control, asthma-related quality of life, hospitalizations, ED visits, prednisone bursts. The 5-item version of the Asthma Control Questionnaire (ACQ) has been shown in large trials to be without loss of validity or change in interpretation compared ^35,43 to the original ACQ. The minimally important clinic difference is 0.5. Because of the pandemic we could not use the 7-item version which requires spirometry. We could not follow FEV1 as an outcome as originally planned.

The protocol was approved by the University of Pennsylvania Institutional Review Board and registered with ClinicalTrials.gov (NCT04023422). A data safety monitoring board reviews progress and data every 6 months. A Community Advisory Board consisting of 4–5 patients or family members of asthma patients living in neighborhoods from which we recruit meets quarterly to contribute their perspective and advice.

2.2. Conceptual development and design choices

A pragmatic trial, that is, a trial in a real-world setting is essential to address questions of communication and health equity. In real-world studies of behavior such as reported here, randomization controls for the likelihood of bias from known and unknown confounders, and their influences on behavior. The factorial design allows testing of more than one intervention.¹⁷

2.2.1. Development of the Clinic Intervention

CI was inspired by the Patient Navigator proposed by Dr. Harold P. Freeman, a Harlem surgeon who wished to overcome barriers to early diagnosis and treatment of cancer in women living in poverty.^18–20 We adapted this Patient Advocate to a chronic disease, asthma, and to our patient population: low-income adults with uncontrolled asthma and comorbidities using focus groups of patients and physicians.^21,22 We piloted this intervention as the Helping Asthma Patients (HAP)¹⁶ and further modified it in a second study, HAP2 (R18 HL116285).¹³ The final version is depicted in Figure 2.

Figure 2:

Activities of the Navigator for the Clinical Intervention (CI) and Home Visits (HV) The home visits ideally were to be accomplished in the 1st 3 months, the clinic visits in the 1^st 6 months. Data collection is scheduled for every 3 months.

*Asthma Control is measured by the Asthma Control Questionnaire (ACQ-5). It has 5 items, each scored from 0–6 with a lower score indicating better control. In the original questionnaire there were 7 items, one item being FEV1.

**Hospitalizations, ED visits are obtained by patient report documented in almost all cases by the electronic health record. There were rare cases of such an event happening at another institution that did not have an electronic health record.

$†$Prednisone burst is a new or increased dose of prednisone of at least 30 mg/d for at least 3days

Social Cognitive Theory (SCT)^23–25 guides the impact of the Navigator on patients’ self-efficacy and confidence in a behavior like self-managing asthma. SCT stresses that individuals and the environment interact continuously and influence each other (reciprocal determinism). SCT posits that learning is enhanced among individuals with high levels of self-efficacy. The interventions act to change the social environment and social support of patients and clinicians, increasing the likelihood of attaining goal behavior consistent with the Guidelines.^1,2,23 The Navigator also provides social modeling, verbal persuasion, and mastery experiences,²³ increasing the self-efficacy of both patients and clinicians and following Guideline elements (e.g., use of an action plan and proper inhaler technique, limiting exposure to tobacco.)^1,2

2.2.2. Experience with Home Visits

During HVs the Navigator reviews recommendations from the CI, combining both home and office recommendations. The HV protocol (Figure 2) was adapted from protocols developed by Dr. Bryant-Stephens with promising impact for children with uncontrolled asthma^26,27 and piloted in the ARC Study (PCORI AS-1307–05218), comprising home visits from a community health worker and access to the patient portal to improve asthma outcomes.¹⁴ HVs considered the multilevel model of structural disparities and social determinants of health that framed our intent to understand and address the barriers to health faced by low-income and underrepresented patients.

We hypothesize that improved patient-centered asthma outcomes can be achieved in these older patients through enhanced communication of patient and clinician, and clinician attention to home, neighborhood, and other environmental exposures, and a team-based, targeted intervention for the delivery of guideline-based care (e.g. implementation of an action plan with presence of medications and instruction in use of inhalers, and tobacco smoke reduction at home).

2.3. Study Goal

The goal of HAP3 is to estimate whether the effect of one or a combination of home visits and physician feedback improves outcomes important to patients all of whom receive a clinic intervention. The evaluation of the effect of CI represents a simple, uncontrolled longitudinal contrast, susceptible to confounding. The CI group serves as a randomized active control group for the remaining study arms. Asthma control is the primary outcome; the secondary outcomes are annualized hospitalizations, ED visits, prednisone bursts, and cost.

2.4. Target Population and Enrollment Criteria.

The target population is urban low-income adults with uncontrolled moderate or severe asthma and their clinicians. Specifically, participants are at least 18 years and live in a Philadelphia neighborhood in which at least 20% of households have incomes below the federal poverty level. They have a doctor’s diagnosis of asthma, are prescribed an inhaled corticosteroid and are patients of a participating clinic. Within 12 months before enrollment they required prednisone, an ED visit, or hospitalization for asthma. Patients are excluded if they have severe psychiatric or cognitive problems making it impossible to understand or carryout the protocol.

Smoking history:

There is a clinical overlap of asthma with COPD.²⁸ We did not wish to exclude all present or past smokers if they had not had extensive tobacco exposure or did not clearly have COPD. These patients often are excluded from clinical trials. However, we also did not want to include those with severe COPD or such irreversible obstruction they would not potentially benefit from the interventions. After discussions with our data safety monitoring board, we included smokers with not more than 15 pack-years of smoking history if at most 40 years and not more than 10 pack years if older than 40 years to include a representative group of patients overall generalizable to the typical asthma population.

2.5. Participating practices.

We recruit from six primary care (general internal medicine and family medicine) and three asthma specialty (pulmonary and allergy-immunology), practices of the University of Pennsylvania Health System. We visited all participating medical practices describing our project at conferences of clinic staff including physicians, nurses, practice managers, and patient service representatives. The original 6 clinics each have a co-investigator to act as a stakeholder for the practice. Each clinic names a clinic champion to promote communication with the Navigator and team integration. The champion can be a medical assistant, nurse, manager, patient service representative, or physician.

2.6. Navigator

In the HAP2 predecessor study, the intervention was delivered by a recently graduated college student interested in a career in healthcare. Patients had liked working with these young adults. Patients also liked working with community health workers, accepted by patients for sharing cultural, economic and linguistic characteristics, knowing the community, and being able to build trust with underserved patients.

For this project, we incorporated both characteristics to find someone interested in providing help to patients, familiar with Philadelphia neighborhoods and medical practices, experience working in these settings, and familiar with patients with asthma. Each research coordinator trains for 3 weeks to deliver the CI and the home visit using manuals describing recruitment, protocol, and data collection. Training also includes 40 hours of intensive review of asthma pathophysiology and education; asthma medications, spirometry; human subjects research; cultural competence and implicit bias; interpersonal skills; relating to clinicians; recognition of adverse and serious adverse events, data collection, use of the electronic health record and REDCap;²⁹ administrative tasks required of patients; procedures for reviewing medical records, screening, enrolling, obtaining consent. Scripts for recruiting phone calls, data collection, and encounters are employed.

Specific training for the CI and the HV includes use of an asthma action plan, listening skills to allow uniform note-taking across patients during their medical appointments, ways to identify community resources in response to patients’ questions, and proficiency at the teach-back technique at a language level that is also culture-appropriate to the patient. Navigators also are coached on an approach to smoking disorder, resources for patients for housing, nutrition and working,³⁰ Social Cognitive Theory/self-efficacy, and shared decision-making. The research team meets regularly with the Community Advisory Board for their perspective and advice. Any research coordinator who believes a patient’s asthma or health is unstable is coached to notify the principal investigators and the patient’s clinician immediately.

Patients work with a Navigator who delivers the CI, the randomized interventions (home visits and guideline-based feedback to the clinician) and another researcher acting as Data Collector (DC) (Figure 2). Because the DC collects all data after randomization by phone, the patient should not feel pressured to please their Navigator with their responses. At enrollment the Navigator helps the participant schedule an appointment with the patient’s asthma clinician within 3 months if not already scheduled. This baseline attempt to schedule a visit helps ensure that all patients, regardless of randomized treatment allocation, will have a CI visit with the clinician and makes it more likely patients will have a CI appointment within 3–6 months as recommended in the EPR-3 Guidelines.

2.7. Randomization:

After 3 months of observation and satisfaction of all eligibility requirements, patients are randomized to one of the four treatment groups (see Section 2.1). If the patient is randomized to receive Home Visits, these are scheduled. Randomizations are stratified by practice (See Section 2.5.)

2.8. Clinic Intervention (CI), the active control group

for all randomized patients, is the patient advocate protocol (Sections 2.1, 2.2.1.) with input from patient and clinicians,²¹ tested for feasibility and acceptability in HAP,¹⁶ and HAP2 Studies.¹²

By phone/text a few days before the upcoming appointment with the asthma clinician, the Navigator reminds the patient of the upcoming appointment. Before the appointment by phone or in person, the Navigator helps the patient focus on 2 or 3 questions that the patient wants addressed during the visit with the clinician. The Navigator ensures the patient has lists of medications, referrals, needed records available, and assists with a draft of an asthma action plan for clinician approval. On the appointment day the Navigator meets the patient in the waiting room and complete preparations for the visit. With permission of the patient and clinician the Navigator attends the visit. Immediately after the visit, the Navigator and patient review what was agreed upon at this visit (“teach back”).

To overcome the problem in HAP2 that some patients did not have a follow-up appointment and therefore no CI visit, the Navigator offers to help make another appointment within 3 months and reminds the patient of an upcoming scheduled appointment. If the patient does not answer calls, the Navigator tries again at a different time of the day or week or by contacting one of three contacts provided by the patient at enrollment. If a patient has not had a planned visit within 3 months, the Navigator contacts the patient and offers to help with scheduling.^31,32 Once approved by the clinician the Navigator reviews the action plan as part of the “teach back” with the patient.

The principal investigators introduce the protocol to participating clinics briefly at physician/staff meetings. Before a CI, Navigators email clinicians to inform them of the upcoming visit. At each clinic visit the Navigator introduces her/himself, obtains permission to attend, reviews recommendations and checks on patient monthly.

2.9. Home Visit (HV):

The same Navigator who conducts the CI will make the patient’s home visits, if the patient is so randomized, ideally in the first few weeks of participation (Figure 1). The Navigator asks the patient to produce their asthma medications and completes a checklist about the presence of pests and damp damage, to describe any environmental triggers of asthma in the home. The Navigator inquires about smokers in the home and offers smoking cessation counseling to patients through motivational interviewing, quit-line resources (Navigators attend a lecture on smoking cessation). In previous home visits for children living in the same urban Philadelphia neighborhoods, we found that half of families had at least one smoker in the home.³³ For other family members who smoke, the Navigator will give quit-line information. If the patient or family member who smokes agrees, the Navigator will follow a Connect strategy and have the quit-line proactively contact the smoker to initiate treatment (https://www.cdc.gov/tobacco/quit_smoking/how_to_quit/index.htm, CDC, How to Quit Smoking, Accessed August 10, 2021).³⁴ Each home visit will have tailored asthma education utilizing lesson plans based on the Expert Panel Report (EPR-3 and 2020FU) Guidelines.³ At subsequent home visits the Navigator will review activities and recommendations from the previous visit.

At the conclusion of each home visit, the Navigator and patient write a summary of the home environment, and the understanding of the action plan, into a brief message to the patient’s asthma clinician via the electronic health record (EHR).

2.10. Clinician Feedback

Guideline-relevant feedback using the Clinician Feedback form (See Supplement) will be given to clinicians of patients who are so randomized. At each appointment with the clinician, the Navigator and the patient will complete the form on paper and give it to the clinician and a copy placed in the study records. The form collects information on patient symptoms, use of controller medications, recent vaccination, hospitalizations, ED visits, possession of an action plan etc.

3.0. Data Collection and compensation:

There are 7 data collection visits (Figure 1). Originally, the first (enrollment), second (randomization), fifth, and final (7^th) data collection visits were in person and included spirometry; the others were conducted by phone. The data collector is a research team member who is not the patient’s Navigator. The data collector queries about asthma control and other endpoints. All patients receive the same compensation, given for completion of data collection tasks, a maximum of $245 over 18 months.

3.1. Endpoints.

The primary endpoint is asthma control,³⁵ measured by the validated 5-item Asthma Control Questionnaire that assesses asthma symptoms and functioning over a 1-week recall. Each item scores from 0= total control to 6= extremely uncontrolled. The summary score is the average across the items, and >1.5 is considered inadequate control. Asthma hospitalizations, emergency visits, prednisone bursts (addition of at least 30 mg or increase in prednisone dose for at least 3 days) are obtained via self-report and confirmed by the EHR.

Secondary endpoints include the Mini Asthma Quality of Life Questionnaire (AQLQ)^36,37 The score is the mean of all 15 items responses (0=total control, 6=extremely uncontrolled) A within-person change of 0.5 is considered clinically meaningful. The minimally important clinical difference is 0.5. The AQLQ has been shown to be a useful indicator of asthma-related quality of life in low-income adults.³⁸ Spirometry, performed according to American Thoracic Society procedures for FEV1 and FVC, was discontinued in the COVID-19 pandemic.

Cost and cost-offset analysis: We will calculate the costs associated with each of the interventions, and conduct a cost-offset analysis to determine which intervention costs are offset by savings attributable to reductions in ED, hospitalization or other visits for asthma control and other outcomes.

3.2. Guideline targets

include completion of an action plan, approved by the patient’s asthma clinician, Inhaler technique is graded using a 7-point scale for a metered dose inhaler and a 6-point scale for a dry powder inhaler, testing the patient on the inhaler used for their inhaled steroid.³⁹ Smoking assessment is self-reported at enrollment visit.

3.3. Covariates

include socio-demographics: educational attainment (years of formal education completed), type of health insurance, primary language, social determinants of health (food insufficiency, loss of utilities, loss of job, transportation problems etc.), comorbidities (patient reported and verified in the medical record including hypertension, diabetes, obesity, cancer, smoking history, etc.) and exposure to community violence.^40,41 Self-efficacy is measured as 1) perceived confidence in completing medical forms and 2) confidence in adherence to essential inhaled steroid regimens. Covariates are collected to demonstrate balance across the four treatment arms, as potential factors on which to standardize results in the event of loss to follow-up, and to support exploratory analyses of heterogeneity of treatment effect.

4.0. Analysis Plan:

The primary analysis will be “as randomized” (intent-to-treat), with the assumption that any dropout visits are missing at random, after controlling for observed imbalances of patient factors across the treatment arms. Potential confounding can arise when there is incomplete follow-up. All participants, regardless of dropout or adherence, will be included in the analysis. As a sensitivity analysis, we will consider whether results might be biased by post-randomization factors such as irregular visit times and data collection schedules.

Service use (hospitalizations, ED visits prednisone bursts) rates will be calculated according to the observed follow-up time. For example, if a patient is followed for 15 months and has 3 ED visits, then the rate for the patient will be estimated as 3*(12/15) = 12/5 = 2/4. We refer to this as annualized rate.

Analyses will use generalized linear mixed effects models with random intercepts and slopes for patient and time, main effects for the intervention components, and two-way time-by-intervention interactions. Past experience suggests that log-gamma models are required to handle skewed measures of asthma control and quality of life and Poisson models for service use measures. Flexible splines for time from randomization to observation will allow for use of actual data collection times for modeling time-by-intervention interaction without artificial categorical time groupings to estimate expected values of outcomes at planned follow-up times versus date of randomization. Model results will be transformed into units on the asthma control and quality measurement scales, and into rate differences for service use measures.

We do not expect that the intervention will be equally effective across all patient subgroups defined by pre-randomization factors. For that reason, we will attempt to identify subgroups of patients who respond particularly well to the intervention and subgroups for whom the intervention is less effective, if at all. Asthma severity will include hospitalizations and ED visits for asthma, FEV1, asthma control, and asthma-related quality of life, all observed in the 3-month pre-randomization period. Other candidate measures under consideration include sex as a high proportion of these patients are female. Patient educational attainment (years of formal education completed), other socio-demographics, social determinants of health, comorbidities, and health literacy will be considered.

Direct costs of intervention will be measured by the resource costing method.⁴² Time spent and materials utilized for training and supervising Navigators and delivering the services to patients will be tracked and collected. Potential cost-offsets (decreased costs of emergency visits, hospitalizations and office visits) associated with each of the interventions from the perspective of the payer (healthcare sector) will be estimated.

4.1. Power determination:

A factorial design with patients randomized to four groups efficiently permits simple and powerful comparisons between separate components of the intervention. A power analysis based on a two-sided, type 1 error of 0.05, two-sample t-test, and 10% dropout from 400 randomized patients, demonstrates more than 80% power to detect a minimum difference of 0.3 units of standard deviation for comparison within each component and 0.5 units of standard deviation for the interaction of both

5.0. RESULTS

As of February 1, 2022, 254 patients were enrolled and 220 randomized after 3 months’ observation. Figure 1 describes the design. Figure 2 depicts the details of the home visit and clinical intervention. Figure 3 (CONSORT) provides details of recruitment and randomization. The first patient was enrolled August 20, 2019, the University and all activities were shut down for the COVID-19 pandemic on March 12, 2020. Table 1 describes 254 enrolled patients. Mean age was 46 years, 85% were female. 46% had hypertension, 70% were obese (BMI> 30),. and 39% were “ever smokers,” of these 11% were current smokers. 48% received Medicaid,18% received Medicare only. Randomization equally distributed the patient characteristics across randomized groups. Mean asthma control score was 2.1. Mean quality of life was 4.2 (Table 2). These pre-randomization outcomes (Table 2) were equally distributed across the randomized groups.

Fig 3:

Participant flow (CONSORT) diagram.

Table 1.

Characteristics of 254 enrolled participants of whom 213 have so far been randomized. n (%)

Characteristic (n, %)	Total n= 254 +	Randomized n=220	Enrolled, not yet randomized N=34
Age (yrs)
Mean ± SD	46 ± 14	46 ± 15	44 ± 15
Maximum	78	78	78
Minimum	19	19	21
Sex reported at birth
Female	217 (85%)	192 (87%)	25 (74%)
Male	37 (15%)	28 (13%)	9 ( 27% )
Ethnicity (self-reported)
Hispanic/Latinx	13 (5%)	12 (6%)	1 (3%)
Non-Hispanic/non-Latinx White	14 (6%)	13 ( 6%)	1 (3%)
Self-reported race (n, %)
African American/ Black	218 (86%)	186 (85%)	33 (97%)
White	17 (7%)	16 (7%)	1 (3%)
Declined to answer	12 (5%)	12 (6%)	0 (0%)
Insurance type
Medicaid	121 (48%)	111 (51%)	10 (29%)
Medicare only	46 (18%)	38 (17%)	8 (24 %)
Commercial	85 (34%)	71 ( 32% )	14 (41% )
Self-pay	2 (1 % )	0 (0 %)	2 ( 6%)
Educational Attainment
< 12 years	28 (11%)	25 ( 11%)	3 (9%)
Completed 12 years	77 (30%)	64 (29%)	13 (38%)
Some college or trade school	88 (35%)	75 (34%)	15%
College graduate	61 (24%)	56(26%)	5 (21)
Comorbidities
Hypertension	117 (46%)	102 (46%)	15 (44%)
Diabetes	53 (21%)	46 (21% )	7 (21%)
Ever smoker	98 (39%)	84 (38% )	14 (41%)
Current smoker	29 (11%)	28 (13%)	1 (3%)
Exposed to tobacco at home	99 (39%)	85 (39%)	14 (41%)
BMI ( Mean ± SD)	35 ± 9	35 ± 9	38 ± 11
Disease Management
FEV1 (mean ± SD) The first 82 patients	72 ± 22
Lists asthma meds correctly	31 (10%)
Lists number of ICS puffs correctly	31 (10%)
Recite action plan correctly	77 (33%)
Living conditions
Lives alone	47 (20)
At least 5 in household	39 (17)
SDOH social determinants of Health
Without housing	12 (5)
Worry about running out of food	53 (22)
Ran out of food	50 (21)
Lack of transportation	43 {18}
No electricity of gas	13 ( 6)
No telephone or cellphone	24 (10)
Violence in neighborhood	55 (23)
Other Significant life changes
Changed health insurance	14 (7)
Moved	14 ( 7)
Employment changed	23 (12)
Taken off work for asthma	41 (38)
Interfered with work around the house	148 (76)
Paid at least $50 for meds out of pocket in last month	30 (15)

Table 2.

Time zero values of 220 enrolled randomized participants at time of randomization.

Characteristic	Total N= 220	Group A N=54	Group B N=55	Group C N=55	Group D N=56
Asthma Control *	2.1 ± 1.2	1.8 ± 1.3	1.8 ± 0.9	2.4 ± 1.2	2.2 ± 1.3
Hospitalizations in 1st 3 months**	2	0	2	1	0
ED for asthma in 1st 3 months**	4	2	4	4	2
Prednisone bursts in 1st 3 months†	5	5	3	2	2
Asthma-related quality of life 37,38	4.2 ± 1.5	4.5 ± 1.5	4.4 ± 1.4	3.8 ± 1.4	4.1 ± 1.5

^*Measured by the Asthma Control Questionnaire (ACQ-5). It has 5 items, each scored from 0–6 with a lower score indicating better control. In the original questionnaire there were 7 items, one item being FEV1.

^**Hospitalizations, ED visits are obtained by patient report documented in almost all cases be electronic health record. There were rare cases of such an event happening at another institution that did not have an electronic health record.

^†Prednisone burst: A new or increased dose of prednisone of at least 30 mg/d for at least 3days

6. DISCUSSION

Physicians no longer make house calls which were an important mechanism for understanding patients’ lives and their social determinants of health. Our hypothesis is that such understanding is essential for making asthma management recommendations to patients. Asthma shares elements of disability with many diseases (chronic symptoms, need for daily medications that may be expensive and inconvenient to take while holding a job and caring for a family). It is a good model for studying the impact of HVs, CIs and providing guideline-relevant information to clinicians. HVs followed through the COVID pandemic allow understanding of how such natural events influence self-management and ultimately health. Considering asthma management guidelines, it is clear that the current guidelines require adjustments to account tor patients’ ages, chronic diseases and disabilities including obesity, social environment, and poverty. Current guidelines implicitly assume that asthma patients are exclusively young otherwise healthy adults and children. Barriers to access to healthcare caused by social determinants of health must be also considered. The COVID pandemic has taught us the importance of these considerations.

Partly because of the absence of house calls, clinicians and patients tend to speak different “languages”: Clinicians are concerned with a biomedical model of disease while patients focus on food and housing and other SDOH. The biomedical model includes medications which many patients cannot afford. Clinicians sometimes fail to understand that many patients do not have access to telemedicine and electronic communication, even as more and more health communication occurs via these channels.

Navigators, ideally familiar with both the clinical world and the patients’ environment, can potentially help clinicians account for patients’ environment in their recommendations for guideline based care. Navigators living in the same environment as participants are subject to the same socioeconomic barriers to health and, indeed, such illness and stress has affected Navigators as well as participants.

This study has limitations. While there were few dropouts in the run-in period, patients hesitated to commit to enrollment, especially during the pandemic when schools and businesses were shut down. Long-term larger studies are necessary to learn how to adapt interventions for these external circumstances. Table 3 describes actions that we took to adapt to this pandemic. How to accommodate such changes into study design needs further study and may include examination of subgroups to identify what patient characteristics are needed to demonstrate a return on investment.

Table 3.

Modifications to the HAP3 Protocol related to the COVID-19 pandemic for safety of participants and researchers, to be as inclusive as possible of participants, and to collect important pandemic-related information.

Modification	Comments
Safety
Changed from in-person clinic visits to virtual visits when clinics shut down because of COVID-19 pandemic, March 12, 2020.	Initially all visits were virtual. Gradually patients were given a choice. Starting in summer 2021.
Changed from in-person home visits to virtual visits.	Starting in summer 2021, patients were given a choice of in-person or virtual home visits
Spirometry was discontinued because of the risk of exposing patients and staff to exhaled droplets.	Now vaccines are available. Devices that do not cause dispersion of particles have also become available, We are evaluating newer spirometers for safety and also how devices adjust if at all for the race of the patient.1
Team obtained and were fitted for N95 masks and provided with surgical masks and face shields.
The team changed from weekly meeting to daily huddle meetings where we developed strategies and materials, teaching, and monitored the well-being of team members.	Most important was monitoring the wellness and anxiety level of the team, using this time to learn more about each other and understand how each team member was doing in the pandemic. To discuss political events and our CIs and RVs to ensure consistency across team members of encounter with patients and fine tune our protocol encounters with patients.
Maximizing inclusivity of participants
Removed the zip code requirement.	Broadened neighborhoods from which we recruit to include any neighborhood in Philadelphia or surrounding suburbs with at least 20% of the population below the national poverty level.
Changed smoking inclusion criteria. (This was made before the pandemic struck)	Change smoking inclusion criteria to include patients less than 40 years with no more than 15 pack-years, patients ≥ 40 years more than 10 pack years. (Some patients’ lung function was so poor they were unlikely to improve with physical or biologic changes due to protocol.)
Added three participating clinics: Clinical Care Associates Primary Care Clinic at Pennsylvania Hospital, General Medicine Clinic for Primary Care, University City Family Medicine	IRB approval obtained. We are evaluating additional clinics.
Developed a COVID Questionnaire about patients’ experiences and received IRB approval.	We used this COVID-19 questionnaire
Designed buttons identifying research team members	In order to make our team members more visible to clinic staff and patients
We regularly visited all clinic sites and discussed protocol, asked for suggestions, spoke with staff.	We attended health fairs and other health-related events in West Philadelphia
Include patients who have had a hospitalization, or ED visit, or prednisone burst for asthma in the past year even if they are not currently taking inhaled corticosteroids	Our goal was to recruit patients with moderate or severe asthma, Omitting the requirement for inhaled steroid did not change the severity of asthma we were recruiting, but allowed us to recruit more participants.
Received IRB approval for participants to receive $10 for successfully identifying a participant who enrolled. (Max $30)
Collecting new Pandemic-related information
We created a COVID Questionnaire about patients’ experiences	IRB approval received.
We updated posters, brochures, and recruiting materials
We developed a post card to send to hard-to-reach participants. We dropped off the postcards at patients’ homes or emailed the postcards to patients
We held regular meetings of Community Advisory Board
Two research coordinators accepted other jobs (Spring 2021). Research coordinators (2/4) (two out of four) required leave from work because of illness and two new research coordinators were hired.	Coordinators were members of the community we studied and had similar difficulties to those of the patients they followed: difficulty with transportation, childcare requirements, tutoring and supervising children, stress, depression, and exposure to community violence.
We are collecting qualitative data on the experiences of patients, clinicians, and Navigators

^1.Ramsey NB, Apter AJ, Israel E, Louisias M, Noroski LM, Nyenhuis S, Ogbogu PU, Perry TT, Wang J, Davis CM. Deconstructing the Way We Use Pulmonary Function Test Race-Based Adjustments. J Allergy Clin Immunol Pract 2022; in press. PMID: 35184982

Navigators have strong ties to the community, and many are members of the community and experience effects of coronavirus pandemic: financial insecurity, food insecurity, exposure to violence, anxiety and depression all making their lives stressful and challenging. These unusual circumstances led to unanticipated stressors on the Navigators and a need to redesign the study to address their needs.

7. CONCLUSION:

HAP3 is an 18-month pragmatic, randomized factorial study in a patient population that suffers disproportionately from asthma morbidity. We are enrolling adults from under-served populations, while adjusting our protocol to the COVID-19 pandemic.

Abbreviations:

ACQ

Asthma Control Questionnaire

AQLQ

Asthma Quality of Life Questionnaire

CI

Clinic Intervention

DC

Data Collector

ED

Emergency Department

EHR

Electronic Health Record

EPR-3

Expert Panel Report-3

2020FUV

2020 Focused Update

HAP

Helping Asthma Patients

HV

Home Visit

SCT

Social Cognitive Theory