Multisystem Healthcare Use among U.S. Veterans with Pulmonary Hypertension

To the Editor:

Multisystem healthcare use in the United States is common, owing to complicated payment plans, provider specialization, and isolated public and private networks. Although delivery of health care across multiple organizations can have positive effects, such as increased access to care, it may also lead to fragmented care, which can threaten the quality of care and increase costs. For patients with pulmonary hypertension (PH), a serious cardiopulmonary condition with high morbidity and mortality, highly coordinated care is vital to mitigate rates of missed or delayed diagnosis (1, 2) and to optimize outcomes. Clinical practice guidelines therefore recommend the diagnosis, classification, and management of PH be implemented through a multidisciplinary approach (3, 4).

The extent to which this integration of care for PH is being realized on a national level is currently unknown. We sought to characterize the prevalence of multisystem use among Veterans with PH and determine the location of diagnostic testing versus treatment among dual users as a marker of potentially fragmented care. We hypothesized that 1) multisystem healthcare use is common among Veterans with PH, 2) Veterans are more likely to receive diagnostic testing in the community because of financial incentives to perform procedures, and 3) treatment is more common in Veterans Health Administration (VA) because of lower patient copayments for medications.

Methods

To conduct this retrospective cohort analysis, we first linked national patient-level data maintained by VA National Data Systems and the Centers for Medicare and Medicaid Services from January 1, 2006, to December 31, 2015. We then constructed a cohort of all Veterans with PH defined by at least two visits linked to an International Classification of Diseases, 9th Revision or International Classification of Diseases, 10th Revision diagnosis code for PH (416.xx or I27.0, I27.2x). We restricted our sample to Medicare-eligible Veterans, defined by age ≥65 years or indication of eligibility in the Medicare denominator file. The Edith Nourse Rogers Memorial VA Hospital Institutional Review Board approved this study.

We identified the source of PH-associated care, defined as outpatient clinic visits linked to a PH diagnosis code, which we categorized as “VA-only care” (all PH-associated visits in VA), “Medicare-only care” (no PH-associated visits in VA), or “dual PH care” (PH-associated visits in both VA and Medicare). Our primary outcomes were receipt of any pulmonary vasodilator within 6 months before the first appearance of the PH diagnosis code or any time after diagnosis and receipt of diagnostic testing including right heart catheterization (RHC) and echocardiogram before initial diagnosis.

We assessed baseline patient characteristics including age, sex, race, geographic location, comorbidities, disease severity, and healthcare utilization across VA-only, Medicare-only, and dual-care groups. We calculated proportions of patients in each group who had received diagnostic testing or therapy and compared differences in rates of diagnostic tests and treatment by source of care using chi-square test of independence. Among the dual-care group, we identified the location of diagnostic testing and treatment.

Results

Of 225,175 Veterans with PH in our cohort, 97% were male, 82% were white, and the mean age was 76.8 years. Patients in the VA-only group were younger and less likely to be white than those in the Medicare-only group. Most Veterans received at least some PH care in the community, with 161,474 (71.4%) having PH visits in Medicare only and 30,378 (13.5%) having PH visits in both VA and Medicare (Table 1).

Table 1.

Demographic characteristics among U.S. Veterans by source of care

Baseline Patient Characteristics	VA Only (n = 33,323)	Medicare Only (n = 161,474)	Dual Care (n = 30,378)
Age, yr, mean (SD)	73.4 (9.9)	78.1 (9.4)	74.0 (9.9)
Male sex	32,302 (96.9)	157,560 (97.6)	29,330 (96.6)
Race
White	25,189 (75.6)	136,432 (84.4)	23,817 (78.4)
Black	5,157 (15.5)	12,442 (7.7)	4,351 (14.3)
Hispanic	1,268 (3.8)	2,174 (1.3)	624 (2.1)
Geographic location
North Atlantic	8,625 (25.9)	44,388 (27.5)	7,763 (25.6)
Southeast	6,189 (18.6)	28,993 (18.0)	5,437 (17.9)
Midwest	7,695 (23.1)	42,461 (26.3)	7,788 (25.6)
Continental	4,902 (14.7)	21,932 (13.6)	4,284 (14.1)
Pacific	5,912 (17.7)	18,733 (11.6)	5,106 (16.8)
Comorbidities, mean (SD)	9.8 (5.2)	10.9 (5.6)	10.2 (5.4)
Comorbid conditions
Neurological disorder	1,401 (4.2)	13,870 (8.6)	1,191 (3.9)
Mental health disorder	1,473 (4.4)	4,209 (2.6)	1,180 (3.9)
Substance abuse	14,581 (43.8)	13,833 (8.6)	9,141 (30.1)
Malignancy	6,214 (18.6)	35,485 (22.0)	5,264 (17.3)
HIV or HCV infection	2,057 (6.2)	11,813 (7.3)	1,947 (6.4)
Chronic liver disease	1,845 (5.5)	5,426 (3.4)	1,434 (4.7)
Chronic kidney disease	18,717 (56.2)	82,715 (51.2)	16,456 (54.2)
COPD	14,539 (43.6)	73,190 (45.3)	14,391 (47.4)
Heart failure	13,603 (40.8)	81,916 (50.7)	13,943 (45.9)
Obstructive sleep apnea	3,816 (11.5)	20,294 (12.6)	3,704 (12.2)
Disease severity
Acute right heart failure*	11,396 (34.2)	47,720 (29.6)	10,335 (34.0)
Acute respiratory failure*	5,064 (15.2)	21,074 (13.1)	4,983 (16.4)
Healthcare utilization,† mean (SD)
Inpatient hospitalizations	2.0 (3.5)	2.1 (3.3)	2.2 (3.7)
ED visits	1.5 (1.1)	1.2 (1.1)	1.4 (1.2)
Outpatient visits	45.4 (45.8)	44.1 (36.5)	44.9 (42.8)

Definition of abbreviations: COPD = chronic obstructive pulmonary disease; ED = emergency department; HCV = hepatitis C virus; HIV = human immunodeficiency virus; SD = standard deviation; VA = Veterans Health Administration.

Data presented as n (%) unless otherwise noted.

^*Within 90 days before diagnosis.

^†In the year before diagnosis.

Use of echocardiography was high across all three groups, ranging from 88.3% in the Medicare-only group to 92.4% in the VA-only group (Table 2). Conversely, use of RHC was low across all groups: 19.4% in the VA-only group, 22.9% in the Medicare-only group, and 30.4% in the dual-care group. Rates of PH treatment were also low across all groups: 3.5% VA-only patients, 1.8% Medicare-only patients, and 9.7% dual-care patients, with phosphodiesterase-5-inhibitors most commonly prescribed (94.3%). Among dual-care patients, most RHCs (56.7%) were performed in the community; however, most treatment (68.3%) was received through the VA (Figure 1). Of those who underwent both RHC and treatment (n = 1,771), 585 (33%) received diagnostic and therapeutic care in discordant locations.

Table 2.

Patterns of pulmonary hypertension evaluation and treatment among U.S. Veterans by source of care

	VA Only (n = 33,323)	Medicare Only (n = 161,474)	Dual Care (n = 30,378)	P Value
PH Evaluation
Echocardiogram	30,797 (92.4)	142,530 (88.3)	27,556 (90.7)	<0.01
Risk ratio (95% CI)	Reference	0.96 (0.95–0.96)	0.98 (0.98–0.99)
RHC	6,456 (19.4)	36,995 (22.9)	9,241 (30.4)	<0.01
Risk ratio (95% CI)	Reference	1.18 (1.15–1.21)	1.57 (1.53–1.61)
$\stackrel{.}{\mathrm{V}}$/$\stackrel{.}{\mathrm{Q}}$ scan	9,005 (27.0)	21,146 (13.1)	6,812 (22.4)	<0.01
Risk ratio (95% CI)	Reference	0.48 (0.47–0.50)	0.83 (0.81–0.85)
PH Treatment
Pulmonary vasodilators	1,176 (3.5)	2,708 (1.7)	2,932 (9.7)	<0.01
Risk ratio (95% CI)	Reference	0.48 (0.44–0.51)	2.73 (2.56–2.92)
Receipt of treatment without confirmatory RHC	747 (2.2)	2,052 (1.3)	1,468 (4.8)	<0.01

Definition of abbreviations: CI = confidence interval; PH = pulmonary hypertension; RHC = right heart catheterization; $\stackrel{.}{\mathrm{V}}$/$\stackrel{.}{\mathrm{Q}}$ = ventilation-perfusion; VA = Veterans Health Administration.

Data presented as n (%) unless otherwise noted.

Figure 1.

Location of diagnostic and therapeutic care among Veterans with dual pulmonary hypertension (PH) care (n = 30,378). RHC = right heart catheterization; VA = Veterans Health Administration.

Discussion

We found that most Veterans in our study received some portion of their PH care in the community, with a considerable proportion receiving care through both VA and Medicare. This is consistent with prior studies showing high rates of multisystem use among Veterans, a trend that is increasingly common since the implementation of several policies expanding Veterans’ access to non-VA health care (5, 6).

We found that among Veterans with PH visits in both VA and the community, most received diagnostic testing in the community, but those who received treatment were more likely to get it through the VA. A third of dual-use Veterans with both a diagnostic procedure and treatment received diagnostic versus therapeutic care in discordant locations. In this highly complex disease, this discordance in location of care may reflect the appropriate referral of patients with PH to expert referral centers. However, if systems are not in place to ensure coordination of care across the two systems, it could increase the potential for suboptimal or redundant care. Indeed, Veterans who participate in multisystem care often experience worse outcomes, including higher rates of readmissions, unsafe prescribing, and mortality (6–9), likely reflecting the challenges in coordinating care and sharing management plans across healthcare systems (10).

This study has several limitations. As our cohort was limited to Veterans older than 65 years of age, it may not be generalizable to all patients with PH. However, as recent cohort studies suggest that the general PH population is aging (11, 12), determining how PH care is delivered to older patients is valuable. Without additional physiologic data, we were unable to identify the World Health Organization PH classification. This limits our ability to interpret the appropriateness of PH treatment or sharing of care across healthcare systems. For example, guidelines recommend PH expert involvement in management of patients with groups 1 and 4 PH, which may require referral outside the VA.

This national analysis reveals a discordance in the location of diagnostic testing and treatment among Veterans with PH who participate in multisystem care. Further studies are warranted to determine the impact of multisystem use on clinical outcomes, including hospitalizations and mortality. To ensure high-quality care for patients with PH, organizational efforts must prioritize care coordination across health systems.