A Community Participatory Approach to Creating Contextually Tailored mHealth Notifications: myBPmyLife Project

Just-in-time adaptive interventions (JITAIs) are a novel approach to mobile health (mHealth) interventions, sending contextually tailored behavior change notifications to participants when they are more likely to engage, determined by data from wearable devices. We describe a community participatory approach to JITAI notification development for the myBPmyLife Project, a JITAI focused on decreasing sodium consumption and increasing physical activity to reduce blood pressure. Eighty-six participants were interviewed, 50 at a federally qualified health center (FQHC) and 36 at a university clinic. Participants were asked to provide encouraging physical activity and low-sodium diet notifications and provided feedback on researcher-generated notifications to inform revisions. Participant notifications were thematically analyzed using an inductive approach. Participants noted challenging vocabulary, phrasing, and culturally incongruent suggestions in some of the researcher-generated notifications. Community-generated notifications were more direct, used colloquial language, and contained themes of grace. The FQHC participants' notifications expressed more compassion, religiosity, and addressed health-related social needs. University clinic participants' notifications frequently focused on office environments. In summary, our participatory approach to notification development embedded a distinctive community voice within our notifications. Our approach may be generalizable to other communities and serve as a model to create tailored mHealth notifications to their focus population.

Correction: Community-Led, Cross-Sector Partnership of Housing and Health Care to Promote Aging in Place (Unite Health Project): Protocol for a Prospective Observational Study

[This corrects the article DOI: 10.2196/47855.].

Determinants of Bluetooth-enabled Self Measured Blood Pressure monitoring in Federally Qualified Health Centers

Background

In 2021, the Health Resources and Services Administration (HRSA) launched the National Hypertension Control Initiative (HTN Initiative) with the goal to enhance HTN control through Bluetooth-enabled self-measured blood pressure (BT SMBP) monitoring and use this data to inform clinical decisions in Federally Qualified Health Centers (FQHCs) with large proportion of their population with uncontrolled BP. We sought to understand the experience of Michigan-based FQHCs in implementing the HTN initiative.

Methods

Staff from three Michigan-based FQHCs were invited to participate in semi-structured interviews from September to November 2022. Interviews were conducted in-person and were based on the Tailored Implementation of Chronic Diseases framework. Content analysis was performed by three coders.

Results

Ten staff participated in interviews (FQHC 1: n=6, FQHC 2: n=1, FQHC 3: n=3). The FQHCs differed in their stage of implementation and their approach. FQHC 1 created a large-scale, community health worker driven program, FQHC 2 created a small-scale, short term, BP device loan program, and FQHC 3 created a primarily outsourced, large-scale program through a contracted partner. Positive staff attitudes and outcome expectations, previous experience with SMBP grants, and supportive clinic leadership were identified as facilitators to implementation; Patients high social needs, SMBP-related Technology, and insufficient workforce and staff capacity were identified as barriers.

Conclusion

BT SMBP among FQHC patients is promising but challenges in integrating SMBP data into clinic workflow, workforce capacity to support the high social needs of participants and to assist in reacting to the more frequent BP data remain to be overcome.

Community-Led, Cross-Sector Partnership of Housing and Health Care to Promote Aging in Place (Unite Health Project): Protocol for a Prospective Observational Study

BACKGROUND

For many older Americans, aging in place is their preferred living arrangement. Minoritized and socioeconomically disadvantaged older adults are up to 3 times more likely to experience disability than other groups, which increases their likelihood of being unable to age in place. Bold ideas to facilitate aging in place, particularly among vulnerable populations, are needed. One such idea is the Unite care model, a community-initiated, academic-supported, cross-sector initiative that combines 2 sectors: housing and health care. The Unite care model colocates a federally qualified health center clinic on an older adult affordable housing campus in Flint, Michigan.

OBJECTIVE

There are two aims to this study. Aim 1 is to evaluate the implementation of the Unite care model in terms of acceptability, adoption, and penetration. Aim 2 is to determine which older adults use the care model and whether the care model promotes aging in place through risk factor reduction and improvement in the physical and social environment.

METHODS

We will assess the care model using a concurrent, exploratory mixed methods design. For aim 1, acceptability will be assessed through semistructured interviews with key stakeholder groups; adoption and penetration will be assessed using housing and health care records. For aim 2, residents residing in the Unite clinic building will participate in structured outcome assessments at 6 and 12 months. Risk factor reduction will be measured by change in systolic blood pressure from baseline to 12 months and change in the physical and social environment (item counts) will also be assessed from baseline to 12 months.

RESULTS

Data collection for aim 1 began in July 2021 and is anticipated to end in April 2023. Data collection for aim 2 began in June 2021 and concluded in November 2022. Data analysis for aim 1 is anticipated to begin in the summer of 2023 and analysis for aim 2 will begin in the spring of 2023.

CONCLUSIONS

If successful, the Unite care model could serve as a new care model to promote aging in place among older adults living in poverty and older Black Americans. The results of this proposal will inform whether larger scale testing of this new model of care is warranted.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)

DERR1-10.2196/47855.

Abstract 86: Understanding Patterns Of Missingness In Acute Ischemic Stroke Trials: A Secondary Analysis Of Pooled Patient-level Follow-up Data

Background: Understanding the magnitude and causes of missing data is crucial to ensuring scientific integrity of clinical trials. Missing data threaten statistical and ecological validity, as results will only generalize to those who remain represented in the sample.

Goal: Identify factors associated with loss at follow-up in acute ischemic stroke trials.

Methods: We harmonized patient-level data from multiple NIH-funded acute ischemic stroke trials including NINDS IV-tPA, ALIAS part 2, SHINE, FAST-MAG, IMS-III, POINT, and DEFUSE 3, all of which had a 90-day study outcome. The primary outcome was the proportion of missing modified Rankin Scale (mRS) scores at 90 days. We compared patients with and without a 90-day mRS score among a variety of baseline patient characteristics.

Results: Among 9580 subjects, 459 (4.8%) were missing their 90-day mRS. Age and race were associated with missingness. Compared to those with complete data, participants with missing data were younger (62 vs 66 years, p<0.001). Non-Hispanic Black participants represented 18.8% of those with complete data, but 25.3% of those with missing data (p=0.001). History of cardiac disease —atrial fibrillation, coronary artery disease, myocardial infarction—and better baseline NIHSS score were associated less missingness (p’s≤.002 and p<.001, respectively). History of stroke/TIA, however, was positively associated with missingness (50.3% vs. 35.4%, p<.001).

Conclusion: Bias due to missingness is an important consideration for all clinical trials, including genetic studies. Disparities in the availability of genetic data compounded by missing phenotypic data creates a twofold disadvantage for underrepresented groups. Long term, we need to understand the biomedical and sociological causes of these associations and develop recruitment and retention strategies to ensure the generalizability of trial results.

Abstract WP193: Neighborhood-level Ethnic And Racial Segregation And Incident Stroke Risk

Background: Neighborhood-level ethnic and racial segregation has been linked to health status. We explored whether community level segregation is associated with incident stroke.

Methods: We used data from Multi-Ethnic Study of Atherosclerosis (MESA). The exposure was neighborhood racial/ethnic segregation measured separately for self-identified White, Black and Hispanic participants and calculated as the Getis-Ord Gi* statistic based on the geocoded address of the participant. The Gi* statistic was categorized using the validated cut points of low, medium, and high neighborhood racial/ethnic segregation (high segregation means that the participant’s neighborhood primarily contained individuals of their race/ethnicity). The primary outcome was incident stroke (ischemic and hemorrhagic) during follow-up. We fit time-to-event Cox models with an interaction between race*segregation and adjustment for age, sex, hypertension, diabetes, hyperlipidemia, and current smoking. The results were further evaluated using marginal effects after logistic regression.

Results: We included 5,411 MESA participants (mean age 62.0±10.1, 47.4% male, 44.4% White, 31.0% Black, 24.6% Hispanic), of which 111 (2.1%) had incident stroke during 8.0±1.3 years of follow-up. In the low, medium, and high neighborhood segregation categories there were 1,603 (29.6%), 1,667 (30.8%), and 2,141 (39.6%) individuals. High neighborhood segregation was more common in Black (45.7%) and Hispanic (39.7%) individuals than White (14.6%) (p<0.001). In Cox regression, the interaction between race and segregation was significant (p<0.1). Compared to low segregation, high segregation among White individuals was associated with 3.1% LOWER probability of incident stroke (p<0.001, from 3.4% to 0.3%), among Black individuals with 1.2% HIGHER probability of incident stroke (p=0.046, from 0.7% to 1.9%), and in Hispanic individuals there was an insignificant difference (p=0.842).

Conclusions: For White individuals, a high level of segregation is associated with a lower probability of stroke, but for Black individuals with a higher probability of stroke. These findings call for further research into the mechanisms by which neighborhood segregation may mediate stroke risk.

Engagement in Self-Measured Blood Pressure Monitoring among Medically Under-resourced participants: A Digital Framework Qualitative Study from the Reach Out Study (Preprint)

BACKGROUND

Mobile health (mHealth) interventions serve as a scalable opportunity to engage people with hypertension in self-measured blood pressure (SMBP) monitoring, an evidence-based approach to lowering blood pressure (BP) and improving BP control. Reach Out is an SMS text messaging-based SMBP mHealth trial that aims to reduce BP among hypertensive patients recruited from the emergency department of a safety net hospital in a low-income, predominately Black city.

OBJECTIVE

As the benefits of Reach Out are predicated on participants' engagement with the intervention, we sought to understand participants' determinants of engagement via prompted SMBP with personalized feedback (SMBP+feedback).

METHODS

We conducted semistructured telephone interviews based on the digital behavior change interventions framework. Participants were purposively sampled from 3 engagement categories: high engagers (≥80% response to SMBP prompts), low engagers (≤20% response to BP prompts), and early enders (participants who withdrew from the trial).

RESULTS

We conducted interviews with 13 participants, of whom 7 (54%) were Black, with a mean age of 53.6 (SD 13.25) years. Early enders were less likely to be diagnosed with hypertension prior to Reach Out, less likely to have a primary care provider, and less likely to be taking antihypertensive medications than their counterparts. Overall, participants liked the SMS text messaging design of the intervention, including the SMBP+feedback. Several participants across all levels of engagement expressed interest in and identified the benefit of enrolling in the intervention with a partner of their choice. High engagers expressed the greatest understanding of the intervention, the least number of health-related social needs, and the greatest social support to engage in SMBP. Low engagers and early enders shared a mixed understanding of the intervention and less social support compared to high engagers. Participation decreased as social needs increased, with early enders sharing the greatest amount of resource insecurity apart from a notable exception of a high engager with high health-related social needs.

CONCLUSIONS

Prompted SMBP+feedback was perceived favorably by all participants. To enhance SMBP engagement, future studies could consider greater support in the initiation of SMBP, evaluating and addressing participants' unmet health-related social needs, as well as strategies to cultivate social norms.

Biological and Social Determinants of Hypertension Severity Before vs After Intracerebral Hemorrhage

BACKGROUND AND OBJECTIVES

Although blood pressure (BP) control is considered the most effective measure to prevent functional decline after intracerebral hemorrhage (ICH), less than half of survivors achieve treatment goals. We hypothesized that long-term (i.e., pre-hemorrhage) hypertension severity may be a crucial factor in explaining poor BP control after ICH. We investigated changes in hypertension severity after vs before ICH using Latent Class Analysis (LCA), and identified patient characteristics predictive of individuals' BP trajectories.

METHODS

We analyzed data for ICH survivors enrolled in a study conducted at Massachusetts General Hospital (MGH) from 2002 to 2019 in Boston MA, a high-resource setting with near-universal medical insurance coverage. We captured BP measurements in the 12 months preceding and following the acute ICH hospitalization. Using LCA we identified patient groups (classes) based on changes in hypertension severity over time in an unbiased manner. We then created multinomial logistic regression models to identify patient factors associated with these classes.

RESULTS

Among 336 participants the average age was 74.4 years, 166 (49%) were male, and 288 (86%) self-reported White race/ethnicity. LCA identified 3 patient classes, corresponding to minimal (n = 114, 34%), intermediate (n = 128, 38%) and substantial (n = 94, 28%) improvement in hypertension severity after vs before ICH. Survivors with undertreated (Relative Risk Ratio [RRR] 0.05, 95% Confidence Interval [CI] 0.01-0.23) or resistant (RRR 0.03, 95% CI 0.01-0.06) hypertension before ICH were less likely to experience substantial improvement afterwards. Residents of high-income neighborhoods were more likely to experience substantial improvement (RRR 1.14 per $10,000, 95% CI 1.02-1.28). Black, Hispanic and Asian participants with uncontrolled hypertension before ICH were more likely to experience minimal improvement after hemorrhagic stroke (interaction p < 0.001).

DISCUSSION

Most ICH survivors do not display consistent improvement in hypertension severity after hemorrhagic stroke compared to prior. BP control after ICH is profoundly influenced by patient characteristics predating the hemorrhage, chiefly pre-stroke hypertension severity and socio-economic status. Of note, neighborhood income was associated with hypertension severity after ICH in a high-resource setting with near-universal healthcare coverage. Furthermore, these findings likely contribute to previously documented racial/ethnic disparities in BP control and clinical outcomes following ICH.

Reach out behavioral intervention for hypertension initiated in the emergency department connecting multiple health systems: study protocol for a randomized control trial

Background

Hypertension is the most important modifiable risk factor for cardiovascular disease, the leading cause of mortality in the United States. The Emergency Department represents an underutilized opportunity to impact difficult-to-reach populations. There are 136 million visits to the Emergency Department each year and nearly all have at least one blood pressure measured and recorded. Additionally, an increasing number of African Americans and socioeconomically disadvantaged patients are overrepresented in the Emergency Department patient population. In the age of electronic health records and mobile health, the Emergency Department has the potential to become an integral partner in chronic disease management. The electronic health records in conjunction with mobile health behavior interventions can be leveraged to identify hypertensive patients to impact otherwise unreached populations.

Methods

Reach Out is a factorial trial studying multicomponent, behavioral interventions to reduce blood pressure in the Emergency Department patient population. Potential participants are identified by automated alerts from the electronic health record and, following consent, receive a blood pressure cuff to take home. During the initial screening phase, they are prompted to submit weekly blood pressure readings. Responders with persistent hypertension are then randomized into one of three component arms, consisting of varying intensity levels: (1) healthy behavior text messaging (daily vs. none), (2) blood pressure self-monitoring (daily vs. weekly), and (3) facilitated primary care provider appointment scheduling and transportation (yes vs. no). If participants are randomized to receive facilitated primary care provider appointment scheduling and are not established with a primary care provider, care will be established at a local Federally Qualified Health Center. Participants are followed for 12 months.

Discussion

The Reach Out study is designed to determine which behavioral intervention components or ‘dose’ of components contributes to a reduction in systolic blood pressure after 1 year (Aim 1). The study will also assess the effect of primary care provider appointment assistance on total primary care follow-up visits of hypertensive patients treated in an urban, safety net Emergency Department (Aim 2). Ideally, the Reach Out system will contribute to hypertension management, serving as a model for safety net hospitals and Federally Qualified Health Centers to improve chronic disease management in underserved communities.

Trial registration

This study was registered at clinicaltrials.gov, identifier NCT03422718. The record was first available to the public on January 30, 2018 prior to the enrollment of patients on March 25, 2019.

Feasibility of Emergency Department-initiated, Mobile Health Blood Pressure Intervention: An Exploratory, Randomized Clinical Trial

OBJECTIVES

We aimed to assess the feasibility of a text messaging intervention by determining the proportion of emergency department (ED) patients who responded to prompted home blood pressure (BP) self-monitoring and had persistent hypertension. We also explored the effect of the intervention on systolic blood pressure (sBP) over time.

METHODS

We conducted a randomized, controlled trial of ED patients with expected discharge to home with elevated BP. Participants were identified by automated alerts from the electronic health record. Those who consented received a BP cuff to take home and enrolled in the 3-week screening phase. Text responders with persistent hypertension were randomized to control or weekly prompted BP self-monitoring and healthy behavior text messages.

RESULTS

Among the 104 patients enrolled in the ED, 73 reported at least one home BP over the 3-week run-in (screening) period. A total of 55 of 73 reported a home BP of ≥140/90 and were randomized to SMS intervention (n = 28) or control (n = 27). The intervention group had significant sBP reduction over time with a mean drop of 9.1 mm Hg (95% confidence interval = 1.1 to 17.6).

CONCLUSIONS

The identification of ED patients with persistent hypertension using home BP self-monitoring and text messaging was feasible. The intervention was associated with a decrease in sBP likely to be clinically meaningful. Future studies are needed to further refine this approach and determine its efficacy.

Abstract WP409: Heterogeneous Patterns of Depressive Symptoms Among Stroke Survivors: Implications for Screening

Introduction: Although routine screening for post-stroke depression is recommended by the AHA/ASA, the optimal screening method remains unclear. A high rate of false-positives is a potential issue of existing screening tools, as indicated in a recent meta-analysis. This study aims to examine patterns of depressive symptoms that may provide information to improve screening accuracy.

Methods: The study sample consists of 613 first-ever stroke patients from the Brain Attack Surveillance in Corpus Christi project (2011-2015), a population-based stroke surveillance study. Depressive symptoms at 90 days after stroke were assessed by the 8-item Patient Health Questionnaire (PHQ-8). Symptom patterns were identified using latent class analysis. Depression status (PHQ-8≥10) by class was examined.

Results: The sample was mainly composed of non-Hispanic Whites (38.3%) and Mexican Americans (57.3%), and equally distributed by sex. Mean age was 65.7 (SD=11.0). Prevalence of depression was 26.6% at 90 days after stroke. The best-fitting model yielded 4 classes (Figure). Notably, one class was characterized by symptoms that often overlap between depression and stroke, as opposed to psychological symptoms (termed overlapping symptom class). Percentages of participants classified as having depression were 84.1% in the psychological symptom class, 100.0% in the high-risk class, and 51.7% in the overlapping symptom class. The overlapping symptom class accounted for 28.8% of participants classified as having depression.

Conclusion: Existing depression screening tools developed in non-stroke populations may have poor test characteristics in individuals with stroke. Additional research is needed to develop highly sensitive and specific screening tools to identify individuals with post-stroke depression in need of further evaluation and treatment. Understanding symptom patterns may increase our ability to personalize treatment for individual stroke survivors.

Abstract WMP89: Examining the Constructs of the Patient Health Questionnaire (PHQ-8) in the Stroke Population

Introduction: Most existing screening tools used for post-stroke depression were not originally developed for stroke patients, but the same criteria for identifying depression are usually applied. For instance, the common use of the total score of the Patient Health Questionnaire (PHQ) assumes a unidimentional construct of depression, which may result in distorted performance in the presence of the complexity of stroke. This study aims to test a bifactor measurement model (a depression factor and a stroke-specific factor) using the PHQ-8 in a population-based sample of stroke patients.

Methods: The study sample consisted of 613 first-ever stroke patients from the Brain Attack Surveillance in Corpus Christi project (2011-2015). Depressive symptoms at 90 days after stroke were assessed by the PHQ-8. A bifactor measurement model was proposed based on existing knowledge of stroke symptoms, and investigation of depressive symptoms in the study sample. The model included a general depression factor that influences all eight symptoms, and a stroke-specific factor that influences symptoms that frequently overlap between depression and stroke (Figure). The model fit was evaluated using confirmatory factor analysis, and compared with the unidimensional model.

Results: The sample was equally distributed by sex, with a mean age of 65.7 (SD=11.0). Fifty-seven percent were Mexican Americans and 38.3% were non-Hispanic Whites. The bifactor model showed statistically significant better fit than the unidimensional model (P<0.001).

Conclusion: Future research should explore whether a bifactor measurement model for the PHQ-8, which may result in a different scoring and classification scheme, improves the accuracy of depression screening among stroke survivors.

Abstract WP498: A Nationwide Analysis of Utilization of Life-Sustaining Treatments Following Aneurysmal Subarachnoid Hemorrhage

Introduction: Aneurysmal subarachnoid hemorrhage (aSAH) carries high mortality and disability rates, with many patients receiving life-sustaining treatments (LST). We examined utilization of LST by demographic and regional characteristics for aSAH patients using a nationwide dataset.

Methods: We used Cerner Health Facts, a database of de-identified, longitudinal electronic health record information. Patients with aSAH were selected by ICD-9/10 codes for inpatient admission dates between 1/2004-6/2018. Available demographics included census region, age, race, sex, and marital status. Clinical information included admission Glasgow Coma Scale (GCS) score and aneurysm treatment modality. Primary outcome was utilization of LST, defined as cardiopulmonary resuscitation, ventilation, hemodialysis, ventriculostomy, tracheostomy, or gastrostomy. Descriptive statistics were calculated and associations with LST examined using chi-squares for categorical variables and Kruskal-Wallis tests for continuous variables. A stepwise logistic regression model selection procedure included all available data, with the final model selected by Akaike information criteria (AIC). Significance was set at p ≤0.05.

Results: There were 34,021 patients with aSAH, with a mean age of 57.9±19.8 years. The sample was 72% Caucasian and 55% female, with a median GCS of 15 (IQR 10-15). Approximately 18% (n=6167) received one or more LST. Utilization varied by census region, age, race, marital status, and GCS ( p for each <1E-8), with those most likely to receive support being from the West, in their 60s, non-white, single, and having lower GCS scores. Sex and aneurysm treatment method were not associated with LST ( p =0.12 and 0.77, respectively). With stepwise logistic regression, age and GCS were found to be significant, with a likelihood ratio test demonstrating independent prognostic value for each factor ( p <0.02).

Conclusions: Demographic factors are associated with LST. Potential confounds include clinical indicators, patient/family preference, and individual hospital practice patterns, and future work should include additional information on these characteristics.

Abstract TP391: Reach Out: An Exploratory Randomized Clinical Trial of Emergency Department-Based Hypertension Management

Introduction: Hypertension (HTN) is the most important modifiable stroke risk factor. The Emergency Department (ED) represents an opportunity to identify and treat HTN in difficult-to-reach populations. Electronic health records (EHR) can alert clinicians and researchers when parameters, such as elevated blood pressure (BP) are met. Our primary objective was to determine what proportion of ED patients would utilize prompted BP self-monitoring and respond to texts with home readings consistent with persistent HTN.

Methods: This was a randomized, controlled trial of patients recruited from a university ED with a volume of about 70,000 adult patient visits per year. Real-time EHR alerts identified patients with systolic BP ≥160 or a diastolic BP ≥100 (stage II HTN) who were likely to be discharged from the ED. At enrollment, all subjects were provided a home BP cuff and were sent weekly text messages to check their BP and text back the values to the study team. Subjects responding to texts and indicating persistent HTN during the 3-week run-in period were randomized to control or a 12-week intervention consisting of healthy behavior text messages and reminders to check BP and text the result back to investigators. We conducted secondary analyses evaluating change in BP over 12 weeks based on the self-reported assessments.

Results: Over 9,300 patients with elevated BP were identified through the EHR alerts over 7 months; data were abstracted on 1,908 patients. Of these, 169 were approached and 104 patients enrolled (64%). Of those enrolled, 73 (70%) texted back at least 1 BP reading during the run-in period of which 55 (53% of the 104) subjects had persistent HTN; this was higher than pre-specified goal of 25.5%. In those randomized (n=55), SBP decreased 8.9 mm Hg in the intervention group compared to an increase of 8.3 mm Hg in controls (p=0.039, see table).

Conclusion: Real-time, ED EHR based patient selection combined with mobile health monitoring of BP is feasible and promising.

Progressive rural-urban disparity in acute stroke care

OBJECTIVE

To explore rural-urban differences and trends in tissue plasminogen activator (tPA) utilization among acute ischemic stroke (AIS) patients and examine the association between primary stroke center (PSC) growth and geographic disparity in tPA use.

METHODS

We used hospital discharge data from the National Inpatient Sample (NIS) from 2000 to 2010 and indicators of tPA utilization and describe temporal trends in geographic disparities in AIS care during PSC growth. The Gini coefficient was used to quantify rural-urban inequity in tPA use at the state level (from 0% to 100% of maximum potential rural-urban inequity) in tPA use.

RESULTS

Of 914,500 cases of AIS between 2001 and 2010, 2.3% (n = 21, 190) received tPA. The rural-urban disparity in tPA worsened: tPA use in urban hospitals quadrupled (1.17%-4.87%) compared to rural hospitals (0.87%-1.59%). Of 33 states with NIS data, 15 reached at least 75% of the maximum rural-urban inequality from 2004 to 2010.

CONCLUSIONS

Geographic disparities in tPA use for AIS are increasing. Greater understanding of the effectors of tPA utilization is necessary to ensure that access to tPA treatment is equitable for all communities in the United States.

Abstract WP299: An Expanded Definition of Caregiving Among Elderly US Stroke Survivors: Have We Understated the Problem?

Introduction: Previous studies exploring stroke caregiving have estimated that the average stroke survivor receives as many as 16 hours per week hours of informal caregiving at a societal cost of as much as $26.8 billion per year. However, these studies have focused on basic and instrumental activities of daily living, and have not included healthcare and transportation activities. We sought to determine whether, and at what cost, caregiving differs between stroke survivors and matched controls using an expanded definition of caregiving.

Methods: A total of 892 stroke survivors were drawn from the 2011 wave of the National Health and Aging Trends Study (NHATS), a nationally representative survey of Medicare beneficiaries. One-to-one propensity matching was performed on demographics and comorbidities to create a sample of matched non-stroke controls. Caregivers were identified by detecting all helpers for NHATS respondents who had help with self-care or mobility activities or with household activities due to health or functioning. Stroke survivors and controls were compared using survey-weighted chi-square tests and Wald tests. The annual cost of caregiving was estimated using reported paid caregiving data and estimates of informal caregiving costs.

Results: On average, stroke survivors received about 10 hours of additional caregiving per week compared to controls (22 hours vs. 12 hours, p<0.01). We estimate that the average annual cost for caregiving for an elderly stroke survivor is approximately $11,300. This extrapolates to a total cost of about $40 billion annually among all Medicare beneficiaries.

Conclusion: Our expanded definition of caregiving suggests that prior studies may underestimate both the time and costs invested in caregiving to older stroke survivors.

Abstract TP142: Factors Influencing Post-stroke Discharge to Institutional Rehabilitation Settings versus Home

Introduction: We assessed relative impacts of sociodemographic, clinical, and geographic factors on discharge to rehabilitation or home in a population-based stroke study.

Methods: Stroke survivors were identified from 2011-2013 in the Brain Attack Surveillance in Corpus Christi (BASIC) Project (death<30 days or nursing home residence excluded). Sociodemographic factors (age, sex, ethnicity, insurance type), clinical factors (risk factors, comorbidities, NIH Stroke Scale), and discharge location (inpatient rehabilitation facility (IRF), skilled nursing facility (SNF), home) were collected from medical records and interviews. Geographic factors were distance to nearest IRF/SNF and number of IRF/SNFs within 5 miles of the survivor’s home. Multinomial logistic regression models were used to calculate probabilities of discharge to IRF, SNF, and home for each factor set, fixing the other two factors at their means. High vs low probabilities (25 vs 75 percentile) of discharge to IRF, SNF, and home were compared to show the impact of each factor set.

Results: Discharge location was available for 796/942 (85%) survivors; 15% were discharged to IRF, 26% SNF, and 59% home. Median age was 68 (IQR 58-79); 64% were Mexican American and 36% non-Hispanic white. High/low probabilities of discharge to IRF differed by 2% for sociodemographic factor variation, 7% for clinical factors, and 9% for geographic factors; probabilities of discharge to SNF differed by 22% for sociodemographic factor variation, 14% for clinical factors, and 6% for geographic factors; and probabilities of discharge home differed by 25% for sociodemographic factor variation, 24% for clinical factors, and 6% for geographic factors.

Conclusion: The probability of discharge to IRF is most affected by variation in clinical factors; discharge to SNF by sociodemographic factors; and discharge home by sociodemographic and clinical factors. Geographic factors do not substantially affect discharge location.

Abstract WP173: Mexican Americans Less Likely to Return to Work Following Stroke

Introduction: Lost earnings are the largest driver of the projected $2.2 trillion in stroke costs over the next 4 decades. Mexican Americans (MA) are more likely to have a working age stroke and suffer greater post-stroke disability than non-Hispanic whites (NHW). Thus, we explored ethnic differences in post-stroke return to work and whether sociodemographics and stroke severity contribute to ethnic differences.

Methods: Ischemic stroke patients were identified from the population-based Brain Attack Surveillance in Corpus Christi (BASIC) study from August 2011-December 2013. Employment status was obtained at baseline and 90-day interviews conducted with patients or proxies. Sequential logistic regression models were built to assess ethnic difference in return to work after accounting for: 1) age (<65 vs ≥65), sex; 2) 90-day NIH stroke score, and 3) education (<high school vs. ≥high school).

Results: Of the 729 MA and NHW stroke survivors who completed the baseline interview, 197 (27%) were working at the time of their stroke of which 125 completed the 90-day outcome interview. There were no ethnic differences in sex or the proportion over the age of 65. MAs had less education (5% vs. 24% college graduate, p<0.01) and greater median 90-day stroke severity (2 vs. 1, p=0.02) than NHWs. Forty-nine (40%) stroke survivors returned to work. MAs were less likely to return to work (OR= 0.45, 95% CI 0.22-0.94) than NHWs. This difference remained after accounting for age and sex (OR=0.45, 95% CI 0.21-0.94). The ethnic difference was attenuated and became non-significant after adjusting for stroke severity (0.59, 95%CI 0.24-1.24) and further attenuated after accounting for education (0.85, 95% CI 0.32, 2.22). In the fully adjusted model, lower stroke severity and higher education were associated with return to work.

Conclusion: MAs are less likely to return to work after stroke than NHWs. This finding is important given that MAs are younger and poorer at the time of their stroke suggesting a crucial public health problem. Future work should consider including return to work as part of patient centered outcomes and efforts to optimize stroke recovery.

Headache neuroimaging: Routine testing when guidelines recommend against them

AIMS

The aim of this article is to determine the patient-level factors associated with headache neuroimaging in outpatient practice.

METHODS

Using data from the 2007-2010 National Ambulatory Medical Care Surveys (NAMCS), we estimated headache neuroimaging utilization (cross-sectional). Multivariable logistic regression was used to explore associations between patient-level factors and neuroimaging utilization. A Markov model with Monte Carlo simulation was used to estimate neuroimaging utilization over time at the individual patient level.

RESULTS

Migraine diagnoses (OR = 0.6, 95% CI 0.4-0.9) and chronic headaches (routine, chronic OR = 0.3, 95% CI 0.2-0.6; flare-up, chronic OR = 0.5, 95% CI 0.3-0.96) were associated with lower utilization, but even in these populations neuroimaging was ordered frequently. Red flags for intracranial pathology did not increase use of neuroimaging studies (OR = 1.4, 95% CI 0.95-2.2). Neurologist visits (OR = 1.7, 95% CI 0.99-2.9) and first visits to a practice (OR = 3.2, 95% CI 1.4-7.4) were associated with increased imaging. A patient with new migraine headaches has a 39% (95% CI 24-54%) chance of receiving a neuroimaging study after five years and a patient with a flare-up of chronic headaches has a 51% (32-68%) chance.

CONCLUSIONS

Neuroimaging is routinely ordered in outpatient headache patients including populations where guidelines specifically recommend against their use (migraines, chronic headaches, no red flags).

Abstract T P387: Stroke Health and Risk Education Project: main results

Objective: The Stroke Health and Risk Education (SHARE) Project is a cluster-randomized, faith-based, culturally-sensitive, theory-based behavioral intervention trial to reduce key stroke risk factor behaviors.

Methods: Ten Catholic churches in Corpus Christi, Texas were randomized to intervention or control group. Subjects were enrolled as friend or family partners who provided reciprocal support to each other. The intervention group received a 1-year multicomponent intervention that included self-help materials (physical activity guide with pedometer, healthy eating guide, blood pressure management photonovella, and motivational short film), 2 tailored newsletters, 5 motivational interviewing counseling calls, and a peer partnership support workshop. Multilevel modeling, accounting for clustering within pairs and parishes, was used to test treatment differences in the average of two measures of change since baseline (ascertained at 6 and 12 months), using standardized questionnaires, in dietary sodium, dietary fruit and vegetable intake, and physical activity of moderate or greater intensity. The primary endpoint was met if any one of the three outcomes was significant based on a pre-specified p-value threshold of 0.05/3.

Results: Of 801 subjects who consented, 760 completed baseline data assessments. The median age was 52; 84% were Hispanic, the remainder were non-Hispanic whites; 64% were women. The intervention group had a greater increase in fruit and vegetable intake than the control group (0.25 (95% CI: 0.08, 0.42) cups per day, p = 0.002), a greater decrease in sodium intake (-123.17 (-194.76, -51.59) mg/day, p=0.04), but no difference in change in moderate or greater intensity physical activity (-27.47 (-526.32, 471.38) MET-minutes per week, p=0.56).

Conclusion: This cluster-randomized trial was effective in increasing fruit and vegetable intake among Hispanic and non-Hispanic white Catholic parishioners, reaching its primary endpoint. The intervention also seemed to lower sodium intake. Church-based health promotions can be successful in primary stroke prevention efforts.

Abstract T MP45: Ethnic Differences in Access to Post-acute Care Following Stroke: The BASIC Project

Background: Mexican Americans (MAs) have worse post-stroke functional and neurologic outcomes than non-Hispanic whites (NHWs). Ethnic differences in utilization and intensity of post-acute care (PAC) are possible explanations. We tested the hypothesis that MAs have less utilization and less intense PAC than NHWs in a bi-ethnic stroke population.

Methods: Stroke patients (n=393) were identified from the BASIC Project (2009-2011). Patients admitted from a nursing home and those discharged to a setting other than home or institutional PAC were excluded. Ethnicity was based on self-report. Utilization of any institutional PAC (inpatient rehabilitation facility (IRF) or skilled nursing facility (SNF)) following hospital discharge was determined from UB-92 claims forms. Intensity of rehabilitation was explored by comparing utilization of IRF versus SNF among those receiving PAC. Hierarchical logistic regression models to account for clustering of patients within hospitals were used to determine the associations between ethnicity and utilization and intensity of PAC. Models were run unadjusted and adjusted for demographics (age, sex), socioeconomic status (insurance, education), clinical stroke factors (NIHSS, stroke type), and pre-stroke factors (comorbidity index, pre-stroke function, pre-stroke cognition).

Results: Sixty-three percent were MA (n=249); 49% (n=194) were women. Median age was 66 (IQR: 57-78); median NIHSS was 4 (IQR: 2-8). Among MAs, 64% were discharged to home, 21% to a SNF and 15% to an IRF. Among NHWs, 53% were discharged to home, 26% to a SNF, and 22% to an IRF. In the unadjusted model MAs were 39% (OR=0.61; 95% CI: 0.39, 0.94) less likely to utilize institutional PAC than NHWs. After adjustment, MAs were 41% (OR=0.59; 95% CI: 0.32, 1.09) less likely to utilize institutional PAC than NHWs. Among those who utilized PAC there were no ethnic differences in intensity (IRF vs SNF) in unadjusted (p=0.36) or adjusted (p=0.63) analyses.

Conclusion: We found a trend towards MAs having less utilization of PAC than NHWs following stroke. Additional research with more refined measures of PAC is needed to understand the role that access to, and effectiveness of, rehabilitative services play in ethnic differences in post-stroke outcomes.

Abstract T P117: Impact of Medicaid Coverage on Utilization of Inpatient Rehabilitation Facilities

Objective: Optimal post-acute care is associated with improved stroke outcomes. Among working age stroke patients discharged to institutional post-acute care, those with Medicaid are less likely to be discharged to an inpatient rehabilitation facility (IRF) than those with private insurance, a finding which may be influenced by state Medicaid coverage. We hypothesized that stroke patients residing in states where Medicaid does not cover IRFs would be less likely to be discharged to an IRF than patients residing in states where Medicaid covers IRFs.

Methods: Working age ischemic stroke patients with Medicaid were identified from the 2010 Nationwide Inpatient Sample (NIS) using ICD-9 CM codes 433.x1, 434.x1 and 436. Medicaid coverage of IRFs (yes versus no) was ascertained for 45 states with NIS data by review of state Medicaid websites. The primary outcome was discharge to IRF (versus other discharge destinations). We fit a hierarchical logistic regression model that included patient-level factors (demographics and stroke severity measures (length of stay, t-PA use and Charlson comorbidity score)), and a state policy variable representing whether a State’s Medicaid pays for IRF, with a random intercept for hospital. Based on this model, we estimated the probability of utilization of IRFs in states with Medicaid coverage of IRFs compared to those without.

Results: Medicaid did not cover IRFs in 4 (TN, TX, SC, WV) out of 45 states. Compared to stroke patients residing in states with Medicaid coverage of IRF, stroke patients hospitalized in states without Medicaid coverage of IRF were less likely to be discharged to an IRF (12.8% (7.5-18.0%) vs. 19.4% (17.0-21.8%), p=0.02) after adjusting for patient and hospital factors.

Conclusion: Working age stroke patients with Medicaid who reside in states where Medicaid does not cover IRFs have less utilization of IRFs than patients residing in states where Medicaid covers IRFs. As the Medicaid population expands under the Patient Protection and Affordable Care Act and the number of working age stroke patients increase, careful attention to state Medicaid policy for post-acute care and analysis of its effects are warranted.