Abstract
Rationale
Pulmonary arterial hypertension (PAH) is a heterogeneous disease within a complex diagnostic and treatment environment. Other complex heart and lung diseases have substantial regional variation in characteristics and outcomes; however, this has not been previously described in PAH.
Objectives
To identify baseline differences between U.S. census regions in the characteristics and outcomes for participants in the Pulmonary Hypertension Association Registry (PHAR).
Methods
Adults with PAH were divided into regional groups (Northeast, South, Midwest, and West), and baseline differences between census regions were presented. Kaplan-Meier survival analyses and Cox proportional hazards were used to estimate the association between region and mortality in unadjusted and adjusted models.
Results
Substantial differences by census regions were seen in age, race, ethnicity, marital status, employment, insurance payor breakdown, active smoking, and current alcohol use. Differences were also seen in PAH etiology and baseline 6-minute walk distance test results. Treatment characteristics varied by census region, and mortality appeared to be lower in PHAR participants in the West (hazard ratio, 0.60; 95% confidence interval, 0.43–0.83, P = 0.005). This difference was not readily explained by differences in demographic characteristics, PAH etiology, baseline severity, baseline medication regimen, or disease prevalence.
Conclusions
The present study suggests significant regional variation among participants at accredited pulmonary vascular disease centers in multiple baseline characteristics and mortality. This variation may have implications for clinical research planning and represent an important focus for further study to better understand whether there are remediable care aspects that can be addressed in the pursuit of providing equitable care in the United States.
Keywords: epidemiology, regional variation, pulmonary vascular disease, social determinants of health
Pulmonary arterial hypertension (PAH) is a heterogeneous entity that comprises many different underlying etiologies and associated conditions united by shared aspects of pathophysiology and treatment. PAH can be easily mistaken for other forms of pulmonary hypertension, and it is not uncommon for the diagnosis to shift as patients move between providers (1). Compounding the complexities in diagnosis and pathophysiology are the number of therapies available for PAH that have rapidly expanded over the past 2 decades. Even the approach to how these medications are deployed is in a state of flux (2, 3).
Given the complexity of diagnosis, variability in local and demographic risk factors leading to PAH, and an increasingly nuanced treatment environment, it stands to reason that regional variation in care and outcomes for patients with PAH may exist. Other similarly complex lung diseases with an emerging and changing treatment landscape, such as cystic fibrosis, have been shown to have substantial regional differences in outcomes and therapeutic approach (4, 5).
Although there has been some recent focus on regional variation in PAH from the vantage of the methamphetamine epidemic and methamphetamine-associated PAH, relatively little is known about regional differences in PAH in the United States more broadly (6, 7). A better understanding of regional differences may offer insight as natural experiments to understand important foci of care, may identify disparities in care, and may have implications for resource allocation and clinical research. In addition, there is regional variability in access to expert centers, which may impact the timing of referral to a center and the severity of illness when patients are initially seen at a pulmonary hypertension center (8).
This study was designed to explore regional differences in participants with PAH who were seen at accredited pulmonary hypertension centers and enrolled in the national Pulmonary Hypertension Association Registry (PHAR) to begin understanding regional differences in PAH within the United States.
Methods
Study Population
The PHAR is an ongoing institutional review board–approved, multicenter, longitudinal registry of patients with PAH in the United States that has been described previously (University of Pennsylvania, single institutional review board, Protocol number 822830) (9). Briefly, the PHAR was launched in 2015 and collects demographic, medical history, diagnostic, medication, and quality-of-life information on participants with PAH who are newly evaluated at an accredited pulmonary hypertension center (within 6 months of their first visit at an accredited center). The current analysis includes adult patients (⩾18 yr old) who were determined by their enrolling center to have PAH and who consented to participate in the PHAR. Participants with pulmonary veno-occlusive disease, persistent pulmonary hypertension of the newborn, congenital diaphragmatic hernia, chronic thromboembolic pulmonary hypertension, and/or non–Group 1 pulmonary hypertension were excluded from these analyses. All characteristics presented, apart from mortality and utilization of hospital-based care, were collected at the baseline PHAR visit for these analyses. Information on mortality and utilization of hospital-based care was collected during PHAR follow-up visits for those participants with at least one follow-up visit. The dataset analyzed in this study was locked on October 4, 2022.
Variables and Outcomes
Exposure of Interest
The exposure of interest was U.S.-defined census region. Each accredited center was assigned to a defined census region, including the Northeast (Region 1), Midwest (Region 2), South (Region 3), and West (Region 4) (10). (See Figure 1.)
Figure 1.
Map of U.S.-defined census regions and study sample.
Demographics
Age, sex at birth, body mass index (BMI), race, ethnicity, marital status, employment status, insurance payor, current and past cigarette use, and current alcohol use data were collected. Apart from BMI, these variables were all self-reported. Insurance payor was categorized as private, Medicare, Medicaid, other, no coverage, or unknown. Many participants endorsed having more than one form of insurance. PAH etiology was collected and adjudicated by the enrolling site.
Hemodynamics and PAH Treatment
The most recent hemodynamics were reported at the time of enrollment, including right atrial pressure, pulmonary artery pressure, wedge pressure, heart rate, and cardiac output. Cardiac index, pulmonary vascular resistance, and pulmonary vascular compliance were calculated by the coordinating center.
Active treatment information was also collected at the time of PHAR enrollment. Binary (yes-no) data were collected for the following variables: PAH-directed treatment for more than 6 months; treatment with phosphodiesterase 5 inhibitors, endothelin receptor antagonists, and/or prostacyclin-targeted therapies within 6 months. Notably, enrollment in PHAR occurred within 6 months of arrival to an expert center. As such, data regarding PAH-directed treatment should be conceptualized as a combination of treatments initiated in the community with the possibility for some contribution of early treatment choices initiated at the accredited center.
Baseline Measures of Disease Severity
Baseline severity at the time of PHAR enrollment was estimated using several approaches. Biomarkers included either B-type natriuretic peptide (BNP) or N-terminal prohormone BNP (NT-proBNP) and 6-minute walk distance. Symptom burden and health-related quality of life were considered using the New York Heart Association functional class and the EmPHasis-10 score (The emPHasis-10 is a short questionnaire for assessing HRQoL in pulmonary arterial hypertension) (11). Finally, composite assessments of risk were estimated using the REVEAL (Registry to Evaluate Early and Long-Term PAH Disease Management) Lite risk score (maintained as a continuous variable) and the European Respiratory Society risk score (which classified participants as low, intermediate-low, intermediate-high, or high risk) (3, 12).
Emergency Room Utilization and Hospitalizations
Utilization of hospital-based care (admissions or emergency services) was estimated for all participants during follow-up and presented as the rate per 100 person-days of follow-up.
Outcome
Mortality was measured as the outcome of interest. Risk time began to accrue after enrollment in PHAR and continued until the participant died, was lost to follow-up, or administratively censored at the final time that the participant was known to be alive.
Statistical Analysis
Descriptive statistics were presented as appropriate, using means and standard deviations or percentages. These measures were meant to provide context for observed differences, but because they were not independently planned hypotheses and no correction was made for multiple comparisons, formal statistical testing was not presented. Nonparametric and parametric estimates of mortality were considered. Unadjusted nonparametric estimates of mortality were presented using Kaplan-Meier survival analysis. The hazard of mortality was estimated using Cox proportional hazards. Unadjusted and adjusted models were considered. All adjusted models accounted for differences in age, sex at birth, and BMI. To consider the impact of regional differences in etiology, social determinants of health, severity at enrollment into the PHAR, medication use at enrollment into the PHAR, and differences in incident versus prevalent disease at enrollment, we considered additional adjusted models. These included the standard adjustments alongside adjustments for PAH etiology, social determinants (including race, ethnicity, insurance status, employment, income, and whether a participant was born outside the United States), REVEAL Lite risk score, medications used within the 6 months, or treatment for PAH for more than 6 months, respectively. A final model considered the possibility that all covariates in aggregate might explain regional mortality differences. The possibility of effect modification was considered for the relationship with mortality by age, PAH etiology, or prevalent disease. Age-standardized case-fatality rates and case fatality by strata of age were presented to further evaluate the possibility that differences in population-based age structure by region impacted the results (13). Analyses were performed using STATA (v15.1), and a P < 0.05 was considered statistically significant.
Results
Baseline Characteristics
A total of 1,578 participants across 57 participating pulmonary hypertension centers were included (Figure 1). Several demographic differences were identified (Table 1). Participants in the West tended to be younger. In addition, race and ethnicity varied across U.S. census regions: the Northeast and Midwest had a largely White population, the South had the greatest proportion of Black participants, and the West had the largest proportion of Asian participants and of participants who reported Hispanic ethnicity. The West also had the greatest proportion of participants born outside the United States. Marital status and employment rate also showed variation, as did self-reported insurance payor mix. The Northeast and Midwest had the highest proportion of patients with private coverage. Sex at birth and BMI appeared to be relatively similar by region.
Table 1.
Participant characteristics by U.S. census region
| Characteristic | U.S. Census Region |
|||
|---|---|---|---|---|
| Northeast | Midwest | South | West | |
| Demographics | ||||
| Age, yr | 56.3 ± 17.3 | 55.3 ± 16.5 | 56.3 ± 16.3 | 50.5 ± 17.7 |
| Female sex at birth, % | 75.4 | 76.7 | 77.5 | 74.2 |
| Body mass index, kg/m2, mean ± SD | 29.6 ± 6.9 | 30.3 ± 7.9 | 29.3 ± 7.3 | 29.3 ± 6.6 |
| Race, % | ||||
| White | 82.7 | 80.9 | 68.9 | 68.7 |
| Black or African American | 8.1 | 10.5 | 24.4 | 4.3 |
| Asian | 0.7 | 2.4 | 1.8 | 7.9 |
| Other | 8.5 | 6.3 | 4.9 | 19.1 |
| Hispanic ethnicity, % | 6.2 | 2.4 | 5.9 | 22.8 |
| Born outside the United States, % | 7.9 | 6.6 | 8.2 | 15.8 |
| Married, % | 54.0 | 55.7 | 47.2 | 40.3 |
| Employed, % | 27.7 | 28.5 | 25.6 | 21.4 |
| Self-reported insurance/payors, % | ||||
| Private | 59.5 | 56.1 | 53.1 | 38.7 |
| Medicare | 48.0 | 44.2 | 46.7 | 36.3 |
| Medicaid | 11.6 | 14.3 | 15.9 | 16.7 |
| Other (Indian Health, military, etc.) | 11.4 | 14.7 | 14.7 | 34.3 |
| No coverage | 0.7 | 2.7 | 2.5 | 2.1 |
| Don’t know | 2.6 | 0.9 | 1.4 | 2.7 |
| PAH etiology | ||||
| Idiopathic/familial, % | 45.6 | 48.4 | 51.9 | 30.6 |
| Toxin, % | 3.6 | 5.3 | 3.5 | 29.4 |
| Connective tissue disease, % | 33.6 | 33.2 | 32.3 | 26.5 |
| Portopulmonary hypertension, % | 9.8 | 7.4 | 4.1 | 5.8 |
| Congenital, % | 6.2 | 4.8 | 5.1 | 6.5 |
| Other, % | 1.3 | 0.9 | 3.1 | 1.1 |
| Health behaviors | ||||
| 100+ cigarettes in lifetime, % | 48.7 | 47.3 | 41.3 | 45.5 |
| Smoked in the past 30 days, % | 3.9 | 6.2 | 8.1 | 8.7 |
| Current alcohol use, % | 38.8 | 36.1 | 26.3 | 33.5 |
| Vaccinations at baseline | ||||
| Never received flu vaccine, % | 56.0 | 32.1 | 38.2 | 41.1 |
| Never received pneumococcal vaccine, % | 64.8 | 43.3 | 43.8 | 48.8 |
Definition of abbreviation: PAH = pulmonary arterial hypertension.
Data are presented as mean ± SD or as percentages.
Bold indicates conceptual domains.
The etiology of PAH was also different by region. Most notably, the West had a considerably larger proportion of patients who were diagnosed with toxin-mediated PAH and a smaller proportion of participants who were diagnosed with idiopathic or familial PAH.
Self-reported health behaviors varied across census regions. The West had the highest percentage of participants who had smoked in the past 30 days, and the Northeast had the lowest. By contrast, the Northeast had the highest proportion of participants reporting current alcohol use, whereas the South had the lowest. The Northeast also had the highest percentage of participants who reported never having received a flu or pneumococcal vaccine before enrolling in the PHAR.
Hemodynamics and PAH Treatment
Reported hemodynamics did not suggest clear regional differences at the time of enrollment into the PHAR, but several regional variations in treatment patterns were identified (Table 2). Notably, the West had the highest proportion of participants who had been treated with PAH-specific therapies for more than 6 months before enrollment, whereas the Midwest had the lowest proportion. Within 6 months of enrollment in the PHAR, there were also differences in phosphodiesterase 5 inhibitor use, endothelin receptor antagonist use, and use of prostacyclin-targeted therapies. Notably, participants from the Midwest reported the highest use of parenteral prostacyclin therapies, whereas those from the South reported the highest use of oral prostacyclin-targeted therapy.
Table 2.
Participant hemodynamics, treatment, and severity by census region at the time of enrollment into the Pulmonary Hypertension Association Registry
| Variable | U.S. Census Region |
|||
|---|---|---|---|---|
| Northeast | Midwest | South | West | |
| Hemodynamics at baseline | ||||
| Right atrial pressure, mm Hg | 10 ± 6 | 11 ± 7 | 10 ± 7 | 10 ± 6 |
| Mean pulmonary artery pressure, mm Hg | 48 ± 13 | 50 ± 13 | 48 ± 14 | 50 ± 15 |
| Wedge pressure, mm Hg | 11 ± 5 | 12 ± 6 | 11 ± 6 | 11 ± 5 |
| Cardiac index, ml/min/m2 | 2.4 ± 0.8 | 2.3 ± 0.7 | 2.4 ± 0.8 | 2.3 ± 0.8 |
| Pulmonary vascular resistance, Wood units | 9.4 ± 5.0 | 10.0 ± 5.2 | 9.7 ± 5.6 | 10.7 ± 5.8 |
| Pulmonary vascular compliance, ml/mm Hg | 1.4 ± 0.8 | 1.3 ± 0.8 | 1.3 ± 0.7 | 1.4 ± 1.0 |
| PAH treatment | ||||
| Treatment greater than 6 mo, % | 21.2 | 16.8 | 24.5 | 28.9 |
| PDE5 inhibitor use, % | 78.5 | 84.1 | 72.4 | 81.9 |
| ERA use, % | 56.7 | 51.2 | 59.9 | 54.9 |
| Oral prostacyclin use, % | 5.5 | 6.3 | 14.3 | 10.2 |
| Parenteral prostacyclin use, % | 25.7 | 29.4 | 19.0 | 17.8 |
| Digoxin use, % | 7.2 | 3.6 | 6.3 | 12.7 |
| Anticoagulant use, % | 23.1 | 21.3 | 19.6 | 23.8 |
| Supplemental oxygen use, % | 35.2 | 40.7 | 46.2 | 38.6 |
| Markers of severity at baseline | ||||
| Six-minute walk distance, m | 335 ± 141 | 331 ± 127 | 313 ± 125 | 353 ± 116 |
| Natriuretic peptide (as available) | ||||
| BNP (n = 854) | 371 ± 471 | 243 ± 329 | 313 ± 549 | 293 ± 623 |
| NT-proBNP (n = 737) | 1,918 ± 5,091 | 1,347 ± 1,924 | 1,742 ± 3,305 | 1,642 ± 3,273 |
| NYHA functional class, % | ||||
| I/II | 45.6 | 42.8 | 44.7 | 42.5 |
| III | 49.8 | 51.7 | 48.8 | 49.7 |
| IV | 4.6 | 5.5 | 6.5 | 7.8 |
| ERS risk score, % | ||||
| Low risk | 20.6 | 17.4 | 17.2 | 22.8 |
| Low-intermediate risk | 40.6 | 43.1 | 39.3 | 40.3 |
| High-intermediate risk | 32.6 | 34.5 | 37.8 | 33.6 |
| High risk | 6.2 | 5.0 | 5.7 | 3.3 |
| REVEAL Lite risk score, mean ± SD | 6.4 ± 2.6 | 6.4 ± 2.4 | 6.6 ± 2.5 | 6.2 ± 2.4 |
| EmPHasis-10 score, mean ± SD | 23 ± 12 | 25 ± 12 | 25 ± 13 | 26 ± 12 |
| ER visits and hospitalizations during follow-up | ||||
| ER visits per 100 person-days | 0.7 ± 3.0 | 0.5 ± 1.1 | 0.6 ± 1.6 | 0.5 ± 1.0 |
| Hospitalizations per 100 person-days | 0.5 ± 2.1 | 0.4 ± 0.6 | 0.5 ± 1.5 | 0.3 ± 0.8 |
| Hospital days per 100 person-days | 2.7 ± 5.5 | 2.6 ± 6.0 | 3.3 ± 7.9 | 2.4 ± 8.1 |
Definition of abbreviations: BNP = B-type natriuretic peptide; EmPHasis-10 = is a short questionnaire for assessing HRQoL in pulmonary arterial hypertension; ER = emergency room; ERA = endothelin receptor antagonist; ERS risk score = European Respiratory Society four-component risk score; NT-proBNP = N-terminal prohormone BNP; NYHA = New York Heart Association; PAH = pulmonary arterial hypertension; PDE5 = phosphodiesterase 5; REVEAL = Registry to Evaluate Early and Long-Term PAH Disease Management.
Data are presented as mean ± SD or percentage. Data were gathered at time of enrollment into the Pulmonary Hypertension Association Registry (within 6 months of establishing care at an accredited PAH center). BNP + NT-proBNP is more than the sample size, as some participants had both measures.
Bold indicates conceptual domains.
Measures of Disease Severity
Most measured metrics of severity did not suggest regional differences at the time of enrollment with similar baseline BNP or N-terminal proBNP, New York Heart Association functional class, REVEAL Lite score, European Respiratory Society risk score, and EmPHasis-10 score. Only the 6-minute walk distance differed by region, with participants in the West having the farthest walk distance at baseline (Table 2).
Emergency Room Utilization and Hospitalizations
There were no substantive differences between the rate of emergency room visits, hospitalizations, or number of hospital days between regions (Table 2).
Mortality
Analyses included 3,495 person-years, with a median follow-up of 1.9 years (interquartile range (IQR) = 1.0–3.2 years). This included 310 deaths, for an overall mortality rate of 8.8 deaths per 100 person-years. Follow-up time by region was similar, such that the Northeast had 714 person-years of follow-up, with a median follow-up of 2.0 years (IQR = 1.0–3.3 yr) and a mortality rate of 10.9 deaths per 100 person-years. The Midwest included 674 person-years, with a median follow-up of 1.7 years (IQR = 1.0–2.9 yr) and a mortality rate of 7.9 deaths per 100 person-years; The South included 1,008 person-years, with a median follow-up of 1.7 years (IQR = 1.0–3.0 yr) and a mortality rate of 10.9 deaths per 100 person-years. Finally, the West included 1,099 person-years, with a median follow-up of 2.2 years (IQR = 1.3–3.4 yr) and a mortality rate of 6.3 deaths per 100 person-years.
Regional differences in survival and the hazard of mortality after enrollment in PHAR were identified (Figure 2; Table 3). In unadjusted models, mortality appeared to be lowest in the West. This finding was not significantly different after accounting for regional differences in age, sex, and BMI in an adjusted model. Further accounting for regional differences in PAH etiology, social determinants of health, baseline severity (as assessed by the REVEAL Lite risk score), medication use on enrollment, health behaviors, the proportion of prevalent disease at enrollment, and all covariates in aggregate did not change the impression of better survival in the West (Table 3). Age, etiology, severity, and prevalent disease did not appear to be effect modifiers in this relationship (for the interaction, Ps = 0.29, 0.49, 0.91, and 0.50, respectively).
Figure 2.
Survival after enrollment into the Pulmonary Hypertension Association Registry among participants with pulmonary arterial hypertension in the United States.
Table 3.
Hazard of mortality by census region in staged models to evaluate the impact of differences in demographics, PAH etiology, PAH severity, and PAH treatment regimen, and prevalent PAH (treatment for more than 6 months at enrollment)
| Variable | Hazard Ratio for Mortality by U.S. Census Region |
P Value | |||
|---|---|---|---|---|---|
| Northeast | Midwest | South | West | ||
| Unadjusted model (n = 1,575) | Referent | 0.73 (0.53–1.07) | 1.01 (0.76–1.36) | 0.57 (0.41–0.77) | 0.0007 |
| Adjusted model* (n = 1,523) | Referent | 0.74 (0.52–1.06) | 0.97 (0.72–1.31) | 0.60 (0.43–0.83) | 0.005 |
| Adjusted model* + PAH etiology (n = 1,523) | Referent | 0.75 (0.52–1.07) | 1.01 (0.75–1.37) | 0.64 (0.46–0.90) | 0.02 |
| Adjusted model* + social determinants† (n = 1,511) | Referent | 0.75 (0.52–1.07) | 0.95 (0.70–1.30) | 0.56 (0.39–0.81) | 0.007 |
| Adjusted model* + REVEAL Lite Score (n = 1,459) | Referent | 0.72 (0.50–1.04) | 0.93 (0.68–1.26) | 0.59 (0.41–0.83) | 0.009 |
| Adjusted model* + medications‡ (n = 1,523) | Referent | 0.69 (0.48–0.98) | 0.93 (0.69–1.25) | 0.60 (0.43–0.84) | 0.007 |
| Adjusted model* + health behaviors§ (n = 1,500) | Referent | 0.69 (0.48–1.00) | 0.85 (0.63–1.16) | 0.55 (0.39–0.77) | 0.003 |
| Adjusted model* + treatment >6 mo (n = 1,523) | Referent | 0.75 (0.52–1.06) | 0.95 (0.71–1.28) | 0.58 (0.42–0.82) | 0.004 |
| Adjusted model* + all covariates | Referent | 0.68 (0.46–1.01) | 0.85 (0.60–1.20) | 0.59 (0.39–0.89) | 0.05 |
Definition of abbreviations: PAH = pulmonary arterial hypertension; REVEAL = Registry to Evaluate Early and Long-Term PAH Disease Management. Data are presented as hazard ratio (95% confidence interval) with covariates assessed at the baseline visit for enrollment in the Pulmonary Hypertension Association Registry. P values represent an analysis of variance comparison of the marginal linear prediction for the region relative to the referent group, with 95% confidence intervals in each individual region to provide region-specific comparisons.
Adjustment accounted for differences in age, sex at birth, and body mass index.
Social determinants of health included race, ethnicity, insurance status, employment, income, and whether a participant was born outside the United States.
Medications included phosphodiesterase 5 inhibitors, endothelin receptor antagonists, and prostacyclin-targeted therapy.
Health behaviors included smoking 100+ cigarettes, cigarette smoking within the past 30 days, current alcohol use, and lifetime vaccination status.
Age-Standardized Case Fatality Rate
Age-standardized case fatality rates were calculated for each census region as another approach to account for differences in age. This similarly demonstrated the highest case fatality rate in the Northeast and South (Table 4).
Table 4.
Age-standardized case fatality rates
| U.S. Census Regions and Age Groups | No. of Deaths | No. of Person-Years | Case Fatality Rate* | Age-standardized Case Fatality Rate† |
|---|---|---|---|---|
| Northeast | 10.9 | |||
| >18–40 | 6 | 180 | 3.3 | — |
| >40–60 | 17 | 221 | 7.7 | — |
| >60–80 | 51 | 293 | 17.4 | — |
| >80 | 4 | 20 | 20.0 | — |
| Midwest | 7.9 | |||
| >18–40 | 3 | 124 | 2.4 | — |
| >40–60 | 14 | 258 | 5.4 | — |
| >60–80 | 34 | 280 | 12.2 | — |
| >80 | 2 | 12 | 17.3 | — |
| South | 10.9 | |||
| >18–40 | 12 | 185 | 6.5 | — |
| >40–60 | 36 | 363 | 9.9 | — |
| >60–80 | 56 | 411 | 13.6 | — |
| >80 | 6 | 49 | 12.3 | — |
| West | 6.3 | |||
| >18–40 | 10 | 271 | 3.7 | — |
| >40–60 | 27 | 512 | 5.3 | — |
| >60–80 | 25 | 295 | 8.5 | — |
| >80 | 7 | 21 | 32.8 | — |
Standard pooled population by age group: >18–40 (20.0%), >40–60 (37.0%), >60–80 (39.5%), and >80 (3.6%).
Bold indicates conceptual domains.
Calculated per 100 person-years.
No. of deaths per 100 age-standardized person-years.
Discussion
To our knowledge, this is the first study to describe regional differences in participants with PAH in the U.S. Recognition of such differences may help contextualize existing research, allow more efficient and targeted research planning in the future, and serve as a first step in understanding relevant disparities and variations in care. Although many of the identified differences may simply reflect broader regional differences independent of pulmonary vascular disease, some may uniquely contribute to observed differences in mortality for participants with PAH.
U.S. census region may be a proxy for a myriad of exposures that contribute to health outcomes as social determinants of health. Similar to other cardiopulmonary diseases, previous research with PHAR participants supports the association between social determinants and health outcomes in PAH (14, 15). Several of the notable regional differences are well recognized as markers or drivers for these social determinants of health in cardiovascular disease (16). This includes regional differences in self-reported race, ethnicity, country of birth, marital status, employment status, insurance payor, current alcohol use, and previous vaccination. Notably, U.S. and European guidelines strongly recommend vaccination for influenza and pneumococcus given the poor outcomes among PAH patients who develop pulmonary infections (3, 17). Additional study is warranted to fully assess the impact of these variables on PAH outcomes.
To our knowledge, our study is the first to demonstrate variation in PAH mortality across U.S. census regions. These differences were primarily driven by decreased mortality in the West, with a more modest trend toward decreased mortality in the Midwest (a significant association in only one model). The explanation for the observed difference in mortality is not readily apparent. As noted, there were regional differences in social determinants of health and health-related behaviors. This would be an intuitive explanation for differences in mortality. In some cases, these differences would seem well aligned with the results (e.g., the Northeast, where both a high rate of participants without a previous vaccination and higher mortality are reported), whereas other instances seem misaligned (e.g., the West, where the lowest mortality, but also the highest proportion of current smokers, is reported). These misalignments and the lack of change in the relationship with adjustment reinforces the observation that relationships between social determinants and health are quite complex and may not explain the differences in observed mortality (18).
A further consideration would be residual and unmeasured confounding. We adjusted for differences in demographics (age, sex at birth, and BMI), PAH etiology, social determinants of health, REVEAL Lite risk score, PAH medication regimen on enrollment in the PHAR, and a proxy for prevalent disease at enrollment. We also looked for, and did not find, evidence for effect modification in many of these key metrics; nevertheless, it is notable that the West had a substantially younger age, a different composition of PAH etiologies, and the greatest proportion of participants who were treated for more than 6 months at baseline. It remains possible that these observed differences reflect a range of relevant unmeasured factors that were not accounted for or collected in the PHAR.
Similarly, although estimates of severity and hemodynamics were similar across regions at baseline, these measures may not fully capture the burden of illness. In conjunction, the structure of care is likely different by region and includes factors such as distance between centers, open/closed health systems, and the potential for assessments and events to variably bleed into multiple different electronic medical records. These differences may underpin and cause some of the observed differences in mortality; however, these factors may also contribute to an ascertainment bias that complicates the assessment of severity and mask the “true” relationship with follow-up measures such as emergency room visits and hospitalizations.
A separate consideration that may impact observed differences in mortality is regional variation in what is labeled as PAH. Several studies have detailed the frequency with which a patient’s diagnosis changes when moving between physicians (1, 19, 20). Although these studies have focused on differences in diagnosis as patients move to PAH centers, it is conceivable that differences in diagnostic formulation may exist between well-trained pulmonary vascular physicians as well. A physician’s local practice environment, including their local practice group and regional conferences, could lead to subtle but important differences in diagnostic and management approach such as those that have been reported in several other fields (4, 21–23). For example, using current guidelines, a well-informed pulmonary vascular physician could label the same patient as having PAH with cardiovascular comorbidities (Group 1 PAH diagnosis) or heart failure with preserved ejection fraction and combined pre- and postcapillary pulmonary hypertension (Group 2 pulmonary hypertension diagnosis) (3). The PHAR allows each center to determine the diagnosis for its participants. If the diagnostic approach diverges subtly by region, then it could contribute to the observed variation in mortality. Given the older ages of the participants in the Northeast and the higher proportion of toxin-associated PAH (a subtype closely akin to idiopathic PAH) in the West, regional variation in the amount of “pure” PAH evaluated may translate to differential practice patterns or willingness to enroll older patients with more comorbid illnesses.
Finally, it is important to acknowledge that PHAR comprises patients with PAH who volunteered to participate in research. We suspect these research participants are largely representative of their parent population, although this cannot be automatically assumed to be true (24, 25). If PHAR participants are not representative of the broader population with PAH, then the value of this work in beginning to understand the presence, absence, or drivers of regional disparities is more limited. That said, the identified relationships may continue to have value insofar as this cohort describes participants in PAH research. Differences in etiology, social determinants of health, age, and mortality are vitally important aspects to designing high-quality clinical research and are foundational to tasks such as sample size selection for an adequately powered study (26).
It is important to note that we describe several limitations that impact the inference on our findings. A minority of participants were lost to follow-up (7.4%). While this was similar by region and lost-participants did not contribute risk-time after their vital status was unknown, this may have subtly contributed to differences in measured outcomes. Critically, we neither assert nor intend to imply that regional differences in mortality relate to better or worse care after a participant has arrived at an accredited center. This was not analyzed nor a focus of this work, given concerns about ascertainment bias during PHAR follow-up. Although our work does not directly inform the hypothesis, we think it is unlikely that there is a simple cause-and-effect relationship such that moving between regions would impact a patient’s survival. Instead, the numerous regional differences seen at baseline seem to suggest a suite of exposures and personal characteristics that may already be present on arrival to a center. Future work can try to better understand which, if any, of these factors contribute causally to the observed differences in mortality.
Conclusions
In conclusion, the present study suggests significant regional variation among participants at accredited pulmonary vascular disease centers in multiple baseline characteristics and mortality. This regional variation should be investigated further to better understand the drivers for these differences and consider whether they represent remediable aspects in the pursuit of providing equitable care in the United States.
Acknowledgments
Acknowledgment
The Pulmonary Hypertension Association Registry (PHAR) is supported by Pulmonary Hypertension Care Centers, Inc., a supporting organization of the Pulmonary Hypertension Association. The authors thank the other investigators; the staff; and, particularly, the participants of the PHAR for their valuable contributions. A full list of participating PHAR sites and institutions can be found at www.PHAssociation.org/PHAR.
Footnotes
Supported by the National Heart, Lung, and Blood Institute grant R01HL152724 and the Pulmonary Hypertension Association.
Author Contributions: C.F. and P.J.L. assumed responsibility for the content of this manuscript, including the data and analysis and were involved in each step of development. T.D.M., M.R.L., J.J.R., C.E.V., R.J.W., and R.T.Z. contributed to study design and to data collection and interpretation. N.A.K. and L.J.O. contributed to study design and data interpretation. All authors drafted and revised the manuscript.
This article has a supplement including the full list of the Pulmonary Hypertension Association Registry (PHAR) investigators who collaborated on this project and were involved in the design of the PHAR and data collection. They are included in Table E1 in the data supplement. This list is accessible from this issue’s table of contents at www.atsjournal.org.
Author disclosures are available with the text of this article at www.atsjournals.org.
References
- 1. Saunders H, Helgeson SA, Abdelrahim A, Rottman-Pietrzak K, Reams V, Zeiger TK, et al. Comparing diagnosis and treatment of pulmonary hypertension patients at a pulmonary hypertension center versus community centers. Diseases . 2022;10:5. doi: 10.3390/diseases10010005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Galiè N, Humbert M, Vachiery JL, Gibbs S, Lang I, Torbicki A, et al. ESC Scientific Document Group 2015 ESC/ERS guidelines for the diagnosis and treatment of pulmonary hypertension: The Joint Task Force for the Diagnosis and Treatment of Pulmonary Hypertension of the European Society of Cardiology (ESC) and the European Respiratory Society (ERS): Endorsed by: Association for European Paediatric and Congenital Cardiology (AEPC), International Society for Heart and Lung Transplantation (ISHLT) Eur Heart J . 2016;37:67–119. doi: 10.1093/eurheartj/ehv317. [DOI] [PubMed] [Google Scholar]
- 3. Humbert M, Kovacs G, Hoeper MM, Badagliacca R, Berger RMF, Brida M, et al. ESC/ERS Scientific Document Group 2022 ESC/ERS guidelines for the diagnosis and treatment of pulmonary hypertension. Eur Heart J . 2022;43:3618–3731. doi: 10.1093/eurheartj/ehac237. [DOI] [PubMed] [Google Scholar]
- 4. Kopp BT, Nicholson L, Paul G, Tobias J, Ramanathan C, Hayes D., Jr The geographic impact on hospitalization in patients with cystic fibrosis. J Pediatr . 2016;170:246–252.e4. doi: 10.1016/j.jpeds.2015.11.012. [DOI] [PubMed] [Google Scholar]
- 5. Ramos KJ, Sykes J, Stanojevic S, Ma X, Ostrenga JS, Fink A, et al. Survival and lung transplant outcomes for individuals with advanced cystic fibrosis lung disease living in the United States and Canada: an analysis of national registries. Chest . 2021;160:843–853. doi: 10.1016/j.chest.2021.04.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Kolaitis NA, Zamanian RT, de Jesus Perez VA, Badesch DB, Benza RL, Burger CD, et al. Pulmonary Hypertension Association Registry Investigators Clinical differences and outcomes between methamphetamine-associated and idiopathic pulmonary arterial hypertension in the Pulmonary Hypertension Association Registry. Ann Am Thorac Soc . 2021;18:613–622. doi: 10.1513/AnnalsATS.202007-774OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Prohaska CC, Machado RF. Regional variation in methamphetamine-associated pulmonary arterial hypertension: who’d better call Saul? Ann Am Thorac Soc . 2021;18:584–585. doi: 10.1513/AnnalsATS.202011-1415ED. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Sahay S, Melendres-Groves L, Pawar L, Cajigas HR, Pulmonary Vascular Diseases Steering Committee of the American College of Chest Physicians Pulmonary hypertension care center network: improving care and outcomes in pulmonary hypertension. Chest . 2017;151:749–754. doi: 10.1016/j.chest.2016.10.043. [DOI] [PubMed] [Google Scholar]
- 9. Gray M, Kawut S. The Pulmonary Hypertension Association Registry: rationale, design, and role in quality improvement. Adv Pulm Hypertens . 2018;16:185–188. [Google Scholar]
- 10.2021. https://www.census.gov/programs-surveys/economic-census/guidance-geographies/levels.html
- 11. Borgese M, Badesch D, Bull T, Chakinala M, DeMarco T, Feldman J, et al. PHAR Study Group EmPHasis-10 as a measure of health-related quality of life in pulmonary arterial hypertension: data from PHAR. Eur Respir J . 2021;57:2000414. doi: 10.1183/13993003.00414-2020. [DOI] [PubMed] [Google Scholar]
- 12. Benza RL, Kanwar MK, Raina A, Scott JV, Zhao CL, Selej M, et al. Development and validation of an abridged version of the REVEAL 2.0 risk score calculator, REVEAL Lite 2, for use in patients with pulmonary arterial hypertension. Chest . 2021;159:337–346. doi: 10.1016/j.chest.2020.08.2069. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Weiss NS, Koepsell TD. Epidemiologic methods: studying the occurrence of illness. 2nd ed. Oxford: Oxford University Press; 2014. [Google Scholar]
- 14. Bernardo RJ, Lu D, Ramirez RL, III, Hedlin H, Kawut SM, Bull T, et al. Hispanic ethnicity and social determinants of health in pulmonary arterial hypertension: the Pulmonary Hypertension Association Registry. Ann Am Thorac Soc . 2022;19:1459–1468. doi: 10.1513/AnnalsATS.202109-1051OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Grinnan D, Kang L, DeWilde C, Badesch D, Benza R, Bull T, et al. PHAR Investigators Prediction of patient outcomes through social determinants of health: the Pulmonary Hypertension Association Registry (PHAR) evaluation. Pulm Circ . 2022;12:e12120. doi: 10.1002/pul2.12120. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Son H, Zhang D, Shen Y, Jaysing A, Zhang J, Chen Z, et al. Social determinants of cardiovascular health: a longitudinal analysis of cardiovascular disease mortality in US counties from 2009 to 2018. J Am Heart Assoc . 2023;12:e026940. doi: 10.1161/JAHA.122.026940. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Klinger JR, Elliott CG, Levine DJ, Bossone E, Duvall L, Fagan K, et al. Therapy for pulmonary arterial hypertension in adults: update of the CHEST guideline and expert panel report. Chest . 2019;155:565–586. doi: 10.1016/j.chest.2018.11.030. [DOI] [PubMed] [Google Scholar]
- 18. Eckersley R. Beyond inequality: acknowledging the complexity of social determinants of health. Soc Sci Med . 2015;147:121–125. doi: 10.1016/j.socscimed.2015.10.052. [DOI] [PubMed] [Google Scholar]
- 19. Deaño RC, Glassner-Kolmin C, Rubenfire M, Frost A, Visovatti S, McLaughlin VV, et al. Referral of patients with pulmonary hypertension diagnoses to tertiary pulmonary hypertension centers: the multicenter RePHerral study. JAMA Intern Med . 2013;173:887–893. doi: 10.1001/jamainternmed.2013.319. [DOI] [PubMed] [Google Scholar]
- 20. Emmons-Bell S, Johnson C, Boon-Dooley A, Corris PA, Leary PJ, Rich S, et al. Prevalence, incidence, and survival of pulmonary arterial hypertension: a systematic review for the Global Burden of Disease 2020 study. Pulm Circ . 2022;12:e12020. doi: 10.1002/pul2.12020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Baird M, Daugherty L, Kumar KB, Arifkhanova A. Regional and gender differences and trends in the anesthesiologist workforce. Anesthesiology . 2015;123:997–1012. doi: 10.1097/ALN.0000000000000834. [DOI] [PubMed] [Google Scholar]
- 22. Stafford RS, Misra B. Variation in routine electrocardiogram use in academic primary care practice. Arch Intern Med . 2001;161:2351–2355. doi: 10.1001/archinte.161.19.2351. [DOI] [PubMed] [Google Scholar]
- 23. Song Y, Skinner J, Bynum J, Sutherland J, Wennberg JE, Fisher ES. Regional variations in diagnostic practices. N Engl J Med . 2010;363:45–53. doi: 10.1056/NEJMsa0910881. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Rosenberg AR, Skapek SX, Hawkins DS. The inconvenience of convenience cohorts: rhabdomyosarcoma and the PAX-FOXO1 biomarker. Cancer Epidemiol Biomarkers Prev . 2012;21:1012–1018. doi: 10.1158/1055-9965.EPI-12-0207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Ingram D, Alamri AK, Penn BA, Mayeux JD, Ma CL, Clapham KR, et al. Characteristics of patients with pulmonary arterial hypertension in a Pulmonary Hypertension Association-accredited comprehensive care center: a contrast in features when compared with US national registry data. Cureus . 2022;14:e31764. doi: 10.7759/cureus.31764. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Mehta C, Gao P, Bhatt DL, Harrington RA, Skerjanec S, Ware JH. Optimizing trial design: sequential, adaptive, and enrichment strategies. Circulation . 2009;119:597–605. doi: 10.1161/CIRCULATIONAHA.108.809707. [DOI] [PubMed] [Google Scholar]