Guideline-discordant inhaler regimens after COPD hospitalization: associations with rurality, drive time to care, and fragmented care – a United States cohort study

Summary Background Many patients receive guideline-discordant inhaler regimens after chronic obstructive pulmonary disease (COPD) hospitalization. Geography and fragmented care across multiple providers likely influence prescription of guideline-discordant inhaler regimens, but these have not been comprehensively studied. We assessed patient-level differences in guideline-discordant inhaler regimens by rurality, drive time to pulmonary specialty care, and fragmented care. Methods Retrospective cohort analysis using national Veterans Health Administration (VA) data among patients who received primary care and prescriptions from the VA. Patients hospitalized for COPD exacerbation between 2017 and 2020 were assessed for guideline-discordant inhaler regimens in the subsequent 3 months. Guideline-discordant inhaler regimens were defined as short-acting inhaler/s only, inhaled corticosteroid (ICS) monotherapy, long-acting beta-agonist (LABA) monotherapy, ICS + LABA, long-acting muscarinic antagonist (LAMA) monotherapy, or LAMA + ICS. Rural residence and drive time to the closest pulmonary specialty care were obtained from geocoded addresses. Fragmented care was defined as hospitalization outside the VA. We used multivariable logistic regression models to assess associations between rurality, drive time, fragmentated care, and guideline-discordant inhaler regimens. Models were adjusted for age, sex, race/ethnicity, Charlson Comorbidity Index, Area Deprivation Index, and region. Findings Of 33,785 patients, 16,398 (48.6%) received guideline-discordant inhaler regimens 3 months after hospitalization. Rural residents had higher odds of guideline-discordant inhalers regimens compared to their urban counterparts (adjusted odds ratio [aOR] 1.18 [95% CI: 1.12–1.23]). The odds of receiving guideline-discordant inhaler regimens increased with longer drive time to pulmonary specialty care (aOR 1.38 [95% CI: 1.30–1.46] for drive time >90 min compared to <30 min). Fragmented care was also associated with higher odds of guideline-discordant inhaler regimens (aOR 1.56 [95% CI: 1.48–1.63]). Interpretation Rurality, long drive time to care, and fragmented care were associated with greater prescription of guideline-discordant inhaler regimens after COPD hospitalization. These findings highlight the need to understand challenges in delivering evidence-based care. Funding 10.13039/100000002NIH10.13039/100006108NCATS grants KL2TR002492 and UL1TR002494.


Introduction
5][6][7][8] Guideline-concordant inhaler regimens, in addition to tobacco cessation, immunizations, and pulmonary rehabilitation, are particularly important for patients following hospitalization for COPD exacerbations.The COPD patients at highest risk for poor subsequent outcomes, including increased mortality and accelerated lung function decline, are those recently hospitalized for a COPD exacerbation. 9,106][17][18] These factors likely impact health care access, but few studies have evaluated whether these factors are associated with prescription of guidelineconcordant COPD inhaler regimens. 12,19his study aimed to evaluate differences in prescription of guideline-discordant inhaler regimens after hospitalization for COPD exacerbation by rurality, drive time to the closest pulmonary specialty care, and whether care was fragmented using the United States Veterans Health Administration (VA) national data.We hypothesized that guideline-discordant inhaler regimens would be more common in patients: 1) living in a rural area, 2) with longer drive times, and 2) with fragmented care.

Data sources
We obtained patient data from the VA Corporate Data Warehouse (CDW).The VA CDW is a network of national databases that incorporates data from multiple data sets throughout the Veterans Health Administration into one standard database structure to facilitate reporting and data analysis at the enterprise level. 20We used VA Inpatient files to extract VA hospitalization data.To capture hospitalizations paid for by the VA but received outside the VA (VA-purchased care), we used the Non-VA Care Program Integrity Tools and Fee Basis files.Data for inhalers prescribed by a VA provider were acquired from VA Pharmacy files.This cohort study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.

Study population
We identified patients hospitalized for acute COPD exacerbation between January 2016 and December 2019

Research in context
Evidence before this study We searched PubMed and Google Scholar for studies reporting on guideline-based inhaler regimens in chronic obstructive pulmonary disease (COPD).We used the following search terms ("inhalers") AND ("guideline" OR "guideline adherence" OR "practice guidelines") AND ("chronic obstructive pulmonary disease OR pulmonary disease, chronic obstructive).Our search was limited to studies among adults which were published between January 1, 2000 and July 30, 2023.Our search identified 39 studies.Many studies were clinical practice guidelines and studies quantifying and evaluating trends in prescription of inhaler therapies in COPD.At least three studies assessed factors associated with prescription of guideline-recommended inhaler regimens in COPD.In observational studies, prescription of guidelinediscordant inhaler regimens is common among patients discharged after hospitalization for COPD exacerbationspatients at highest risk for poor subsequent health outcomes.Living in a rural area and fragmented care are increasingly recognized as potential factors that contribute to poor health outcomes in COPD, however, few studies have evaluated how these factors are associated with receipt of guidelineconcordant COPD management.Additionally, to our knowledge, no prior studies have investigated how travel requirements to receive health services, such as pulmonary specialty care, are associated with receipt of guidelinediscordant COPD inhaler regimens.

Added value of this study
To address gaps in the literature, we used the national Veterans Health Administration (VA) electronic health record data to examine receipt of guideline-discordant inhaler regimens among patients discharged from the hospital for COPD exacerbation by rurality, drive time to the closest pulmonary specialty care, and fragmented care.This work could help tailor health system efforts by contributing to our understanding of how factors such as geographical access to and fragmentation of health care are associated with delivery of guideline-discordant care for patients with COPD.

Implications of all the available evidence
Prescription of guideline-discordant inhaler regimens was common after hospitalization for COPD hospitalization and was associated with living in a rural area, longer drive time to pulmonary specialty care, and fragmented care.Our findings suggest that access to health care and challenges in care coordination are potentially contributing to suboptimal delivery of evidence-based COPD care, highlighting the need to develop effective methods to health care delivery to target COPD patients with these risk factors.

Outcomes
The primary outcome was prescription of guidelinediscordant inhaler regimens at 3 months after hospitalization.We selected 3 months as the time period to allow sufficient time for outpatient clinic follow-up and inhaler optimization after hospital discharge.To assess if inhaler therapy patterns change over time, we also assessed guideline-discordant inhaler regimens at 6 months after hospital discharge as a secondary outcome.

Exposure variables Rurality
Geocoded patient residential addresses, urban and rural designations, and drive times from the patients residential address to VA facilities were collected from the VA Planning Systems Support Group Geocoded Enrollee Files. 29Each patient's residential address is designated by the VA as urban or rural using United States Rural-Urban Commuting Area (RUCA) codes 30 with 'urban' as RUCA codes 1.0 or 1.1 and 'rural' as all others.This definition of 'urban' is narrow, incorporating only census tracts with a metropolitan area core, but is consistent with definitions used in VA policy and other research. 29,31ive time to the closest VA pulmonary specialty care Drive time estimates from each patient's address to the closest VA facility where pulmonary specialty care was available were calculated using geospatial technologies as described by the VA Health Economics Resource Center and were based on expected driving routes, traffic, and average driving conditions.29,32 Fragmented care We defined fragmented care as hospitalization outside the VA, but paid for by the VA (VA-purchased care) among patients who receive primary care and prescriptions from the VA. 18

Covariates
We used the Area Deprivation Index (ADI), which includes census block-level income, education, employment, and housing quality metrics, as a measure of socioeconomic status. 33We categorized ADI into ≤20th percentile, 21st-40th, 41st-60th, 61st-80th, 81st-100th for analysis.We used the Charlson Comorbidity Index (CCI) to assess health status. 34We included geographic region-Midwest, Northeast, South, and West-as a covariate to account for regional variation in practice. 35,36n the U.S. Veteran Health Administration, the majority of patients who are hospitalized for any indication typically receive 1-month supply of medications on discharge.After that first month post-discharge, outpatient VA providers, most commonly primary care VA providers, are responsible for continuing, renewing, or changing prescriptions.This means that, in the case of COPD, the inhalers prescribed at discharge can differ from the ones prescribed starting one month after discharge.In addition, based on our prior work, we found that many patients living in rural areas have their COPD hospitalizations in non-VA hospitals, which we are referring to in this paper as fragmented care. 36We use this term because being hospitalized outside the VA system can frequently result in lack of information regarding hospital course, including newly-prescribed medications reaching VA primary care providers.

Statistical analysis
We evaluated the characteristics of patients who were prescribed guideline-discordant inhaler regimens after hospitalization for COPD exacerbation.We calculated the percentage who received guideline-discordant inhaler regimens (a binary outcome at the patient level) and used the exact (Clopper-Pearson) confidence limits to estimate the 95% confidence interval.We used multivariable logistic regression models to estimate the individual associations between prescription of guideline-discordant inhaler regimens and: 1) rurality; 2) drive time (categorized into ≤30 [referent group], 31-60, 61-90 and > 90 min); and 3) fragmented care.We ran the logistic analyses for each exposure variable as separate models to aid in identification of geographic risk factors and patient populations who could be targeted for future interventions.We also wanted to assess if these exposure variables could be predictive on their own.We additionally performed sensitivity analyses including all three exposure variables in one model (Supplementary Table S6).Multivariable logistic regression models were adjusted for age, sex, race/ ethnicity, CCI, ADI, and health care facility region to minimize the potential confounding effects of sociodemographic factors, underlying comorbidities, and regional variability in clinical practices (Supplementary Tables S7-S11).We used an omnibus likelihood-ratio chi-square test to assess whether a predictor was associated with prescription of guideline inhaler regimens.We additionally assessed multicollinearity among the three exposure variables by examining the variable correlation matrix and variance inflation factors in a multivariable regression model (Supplementary Tables S11-S13).We did not observe high correlation (e.g.coefficients of ≥0.8) among the three exposure variables and the variance inflation factors were not large (values between 1.02 and 1.30), suggesting few issues stemming from multicollinearity.8][39] Finally, we assessed interactions between exposure and explanatory variables which showed significant interactions between rurality and race, rurality and region, and fragmented care and race; there was no significant interaction between drive time and exposure variables (Supplementary Table S15).We then performed stratified analyses for the groupings with significant interactions (Supplementary Tables S16  and S17).All statistical analyses were performed using SAS software, version 9.4 (SAS Institute).

Study oversight
This study was approved and the requirement for informed consent was waived by institutional review boards at the Minneapolis VA Health Care System (VAM-20-00583) and the University of Minnesota (STUDY00011069).

Role of the funding source
The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

Results
The baseline characteristics of the patients are in Table 1.Among 33,785 patients who received their primary care and prescriptions from the VA and had a hospitalization for COPD exacerbation, the mean age was 70.5 ± 8.3 years, 26,893 (79.6%)White, 4660 (13.5%)Black/African American, and 13,606 (40.3%) had CCI of 3 or higher.More than 14,360 (42.5%) were assigned to primary care facilities in the Southern region, 8663 (25.6%)Midwest, 6474 (19.2%)West, and 4219 (12.5%)Northeast regions.More than 12,705 (36.7%) lived in a rural area and 21,692 (64.2%) had to drive more than 30 min to the closest VA pulmonary specialty care.Twenty-nine percent had fragmented care (Table 1).
The most commonly prescribed inhaler regimen was the guideline-concordant combination LAMA + LABA + ICS (15,574; 46.1%), followed by two guideline-discordant regimens ICS + LABA (7884; 23.3%) and short-acting inhalers only (4451; 13.2%) (Fig. 2, Supplementary Table S2, and Supplementary Table S3).In sensitivity analyses evaluating individual inhaler regimens, living in a rural area, longer drive time to the closest pulmonary specialty care, and fragmented care were similarly associated with higher odds of prescription of short-acting inhalers only and ICS + LABA; however, we did not see significant associations with LAMA monotherapy (Supplementary Table S3).
At 6 months after COPD hospitalization, the percentage of patients with guideline-discordant inhaler regimens decreased to 12,188 or 38.3% (95% CI: 37.7%-38.8%,Supplementary Figure S1).Living in a rural area, longer drive time to pulmonary specialty care, and fragmented care remained associated with higher odds of guideline-discordant inhaler regimens at 6 months (Supplementary Table S4).
Female patients had higher odds of guidelinediscordant inhaler regimens after COPD hospitalization compared to male patients (aOR 1.14 [95 CI: 1.02-1.28])(Supplementary Table S14).Compared to patients hospitalized for a COPD exacerbation who were White, Black/African American patients had lower odds of guideline-discordant inhaler regimens (aOR 0.81 [95% CI: 0.76-0.86])(Supplementary Table S14).We did not find significant differences in prescription of guideline-discordant inhaler regimens among other races/ethnicities compared to White patients, which could partially be explained by the very small percentages of patients from some other race/ethnic groups in this predominantly White VA cohort (<1% American Indian/Alaska Native, Native Hawaiian/Pacific Islander, and Asian).Compared to patients living in urban areas, Black/African American and White patients living in rural areas had higher odds of guideline-discordant inhaler regimens (aOR 1.46 [95% CI: 1.22-1.73]and aOR 1.16 [95% CI: 1.10-1.22],respectively) (Supplementary Table S16).Additionally, Black/African American and White patients who had fragmented care had higher odds of guideline-discordant inhaler regimens compared to those who did not have fragmented care (aOR 1.88 [95% CI: 1.62-2.18)and aOR 1.52 [95% CI: 1.44-1.60])(Supplementary Table S17).Compared to patients in urban regions, patients in rural Midwest, Northwest, and West regions had higher odds of guideline-discordant inhaler regimens (Supplementary Table S16).

Discussion
In this national cohort of COPD patients, prescription of guideline-discordant inhaler regimens was common after hospitalization for COPD exacerbation and was associated with living in rural areas, longer drive time to pulmonary specialty care, and fragmented care.Our findings suggest that access to health care and challenges in care coordination are potentially contributing to suboptimal delivery of evidence-based COPD care in this high-risk patient population.Sixty-four patients (0.2%) had missing VISN data.d Fragmented care was defined as hospitalization in a non-VA health care facility, but paid for by the VA (VA-purchased care), among patients who receive primary care and prescriptions at the VA. 18ble 1: Baseline characteristics of patients hospitalized for COPD exacerbation.
Other investigators have reported a wide range of rates of guideline-discordant inhaler regimen prescriptions.Yip et al. reported that 54% of patients discharged after COPD hospitalization in the New-York Presbyterian Healthcare System did not receive long-acting inhalers (defined as no ICS, ICS + LABA, or LAMA). 41Gershon et al. and Poon et al. found that only 12% received guideline-discordant inhaler regimens (defined as not receiving LAMA or ICS + LABA) among patients hospitalized for COPD exacerbation or patients with ≥2 COPD exacerbations in Ontario, Canada. 12,42Among COPD patients discharged from hospitals in Italy, Di Martino et al. reported that 35% received guideline-discordant inhaler prescriptions, defined as not receiving LAMA or LABA. 13These mixed results are likely due to differences in inhaler regimens considered as guideline discordant.Because these studies included patients from different years, the variation in regimens considered as guideline discordant in these studies likely reflect temporal evolution of guideline recommendations.In our study, we defined guideline-discordant inhaler regimens as not receiving LAMA + LABA or LAMA + LABA + ICS based on application of clinical practice guidelines and recommendations from multiple organizations (Supplementary Table S1).Factors that have been associated with guideline-discordant inhaler regimens include older age and higher comorbidity burden; however, geographic factors and fragmented care have not been extensively evaluated. 12,13e found that living in a rural area and/or having a longer drive time to pulmonary specialty care were associated with prescription of guideline-discordant inhaler prescriptions.These findings are consistent with our previous work showing that underutilization of spirometry was associated with longer drive time to a site offering this service, which was more common among patients living in rural areas. 19These studies suggest that geographic barriers to accessing health care services are possible mechanisms by which rural-urban disparities might arise.Importantly, we also show the separate associations between rurality and drive time to accessing care; while they are related, urban residents can also have long drive times, which may impact the quality of their care.Our findings show the need to focus on these two distinct, albeit related, geographic factors.
We also found that fragmented care (hospitalization in a non-VA health care facility, but paid for by the VA, among patients who receive primary care and   prescriptions at the VA) was associated with guidelinediscordant inhaler therapy regimens.We did not find prior studies that examined the association of fragmented care with concordance of COPD prescriptions with guideline recommendations.However, Rinne et al. found in a national VA cohort of hospitalized COPD patients that the odds of readmission after hospitalization for COPD was higher among Veterans who were hospitalized outside the VA (aOR 1.20 [95% CI: 1.03-1.40]). 18Our findings that fragmented care was associated with higher risk of guideline-discordant prescriptions offers a potential causal mechanism for Rinne et al.'s observation.Our findings are not isolated to COPD.7][48][49] These suggest geographic barriers to care are broadly predictive of receiving lower quality of care and having worse outcomes.
The VA is the largest integrated health care system in the U.S., and many VA patients receive all their care through the VA.This includes not just primary care, specialty care, and hospital services, but also low or no cost prescriptions.To mitigate health care access issues related to rurality and drive time to care, the VA already uses multiple strategies such as virtual or telehealth options, addition of satellite clinics in rural areas transportation services to VA health care facilities, travel reimbursement, and lodging options on VA property proximal to health care facility. 19Other potential solutions could be considered.For example, health systems could identify patients living in a rural area, those with longer drive time to care, and recent hospitalization in non-VA facilities and offer additional support when setting up appointments, including specialty care visit.Leveraging the skills of other allied health care providers, such as pharmacists, respiratory therapists, and nursing in providing guideline-adherent care can also be explored.Another consideration is investment in broadband infrastructure to expand telehealth services.We need to test novel, targeted solutions in health care delivery to improve guideline-concordant care for these high-risk populations.Future research is needed to evaluate potential interventions prior to widespread implementation.
However, the VA also pays for veterans in emergency situations (such as an acute COPD exacerbation) to go to a non-VA hospital.This provides more access, options, and flexibility, but can also lead to challenges in sharing information, coordinating care, and maintaining continuity of care. 36,50To alleviate issues related to fragmented care, potential solutions that could be considered include improving efforts to provide services for patients to access care within the VA system creating incentives for patients, providers, and health care systems to reduce fragmented care, and increasing support to coordinate care between VA and non-VA systems.
Historical data have shown that access to and receipt of health care services (e.g.2][53][54] This pattern has been observed specifically in COPD care provided outside the VA.A national U.S. survey showed that Black/African American, American Indian, and Hispanic respondents with COPD were less likely to receive preventive care such as immunizations than White respondents. 55Spirometry testing to confirm a diagnosis of COPD was also significantly more common for White than Hispanic than White (83.6% vs. 59.4%, p = 0.006).In contrast, in our current study assessing care for patients covered by the VA system, we found that minorities did not receive lower quality of care than White patients.In fact, prescription of guideline-discordant inhaler regimens after hospitalization for COPD exacerbation was lower for Black/African Americans vs. White patients, indicating that Black/African patients had received better care.This is consistent with data from Mularski et al., who found that Black/African Americans received higher quality COPD care after adjusting for other covariates including insurance. 56Two other studies among veterans with COPD did not show significant differences in receipt of COPD-related services by race/ethnicity, but no studies have shown that patients from minority groups receive worse quality of care than White patients. 19,57A potential explanation for the difference in findings between studies done within the VA and those done outside the VA population is insurance coverage.All the patients in our study receive health care coverage from the VA, whereas health care coverage in non-VA settings may vary significantly across race/ethnicity, with higher proportion of uninsured or underinsured rates among minority groups.
The literature about sex differences in inhaler prescriptions in COPD includes conflicting results.Some studies report that females are more likely to receive maintenance therapies, likely related to more respiratory symptoms among females, than males. 38Other studies report that males are more likely to receive guidelineconcordant inhaler therapies. 39In our study, females who were hospitalized for COPD exacerbation were more likely to be prescribed guideline-discordant inhaler regimens.Reasons for this are unclear, though potential explanations could include delay in or under diagnosis of COPD in females and differences in prescribing habits of providers. 40ur study has several important limitations.First, the inhaler data we used relied on VA pharmacy data.
We did not have data on inhalers prescribed by non-VA providers.Therefore, we only included patients who were actively using VA inhaler prescriptions prior to and after hospitalization.Findings were similar in sensitivity analyses that included patients who were not receiving inhaler prescriptions for the VA before and/or after hospitalization (Supplementary Table S5).Second, we were not able to ascertain patient's inhaler technique and actual use of these inhalers at home.Although this paper focused on inhaler regimens, other components of COPD care including tobacco cessation, immunizations, and pulmonary rehabilitation are also essential in COPD management.Future interventions to improve delivery of COPD care should also address access to these essential components of COPD care.Third, we excluded patients with concurrent diagnosis of asthma; however, we did not have peripheral blood eosinophil counts to further aid in determining COPD patients who may be appropriately treated with ICS-containing regimens.Fourth, we did not have data regarding inhospital pulmonary consultation, which may be associated with prescription of guideline-concordant inhaler regimens at discharge.We also did not have data on inhaler dosing.Fifth, our cohort was predominantly male and White which may limit generalizability of the findings.Sixth, the correlations among the exposure variables were among the larger correlations observed in the dataset and the largest present in the table.Rurality and drive time are naturally associated and we anticipated that fragmented care would be associated with the other two exposure variables.Finally, both of our geographic variables have limitations.Our drive time data is limited to the closest VA facility that provides pulmonary specialty care; we did not have data on drive time to services in non-VA settings.Our measure of rurality was broad, and future research is needed to better understand differences in risk within rural populations.

Conclusion
Living in a rural area, longer drive time to pulmonary specialty care, and fragmented care were associated with higher odds of receiving guideline-discordant inhaler prescriptions after COPD hospitalization.Our findings suggest the need for development of innovative programs to improve delivery of guideline-concordant COPD care, especially in high-risk COPD patients with geographic barriers to care and fragmented care.
Take responsibility for the integrity of the data and the accuracy of the data analysis: All.

Data sharing statement
Data sets underlying publications will be shared in electronic format through a de-identified, anonymized dataset.Researchers interested in access to final study data that is not otherwise publicly available would need to prepare a brief (i.e., 3-5 page) research proposal, including the specific sample types and data they wish to obtain.Data sharing will occur under a written agreement that adheres to any applicable informed consent provisions and prohibits the recipient from identifying or re-identifying any individual whose data are included in the dataset.A signed data distribution agreement will be required for each interested researcher, as well as IRB approval, in order to ensure proper procedures of subject protection.project.Data sets will not be shared without first consulting with VA Privacy Officer for approval.

Declaration of interests
AKB reported grants from the National Institutes of Health.CHS reported grants from the Health Resources and Services Administration, Federal Office of Rural Health Policy, the National Institutes of Health and the CDC; payment or honoraria from Iowa Primary Care Association, Southern Gerontological Society, and the University of Michigan; and leadership or fiduciary role in the National Rural Health Association Board of Trustees.KMK reported personal fees from Nuvaira and Organicell (Data and Safety Monitoring Boards) outside of this work; consulting fees from Allergan/AbbVie and leadership or fiduciary role in the American Thoracic Society Proposal Review Committee, the American Thoracic Society Clinical Problems, and the Assembly Planning Committee.

Fig. 1 :
Fig. 1: Flow diagram of patients hospitalized for COPD exacerbation included in the analyses.