Abstract
Background
Healthcare fragmentation may lead to adverse consequences and may be amplified among older, sicker patients with mental health (MH) conditions.
Objective
To determine whether older Veterans with MH conditions have more fragmented outpatient non-MH care, compared with older Veterans with no MH conditions.
Design
Retrospective cohort study using FY2014 Veterans Health Administration (VHA) administrative data linked to Medicare data.
Participants
125,481 VHA patients ≥ 65 years old who were continuously enrolled in Medicare Fee-for-Service Parts A and B and were at high risk for hospitalization.
Main Outcome and Measures
The main outcome was non-MH care fragmentation as measured by (1) non-MH provider count and (2) Usual Provider of Care (UPC), the proportion of care with the most frequently seen non-MH provider. We tested the association between no vs. any MH conditions and outcomes using Poisson regression and fractional regression with logit link, respectively. We also compared Veterans with no MH condition with each MH condition and combinations of MH conditions, adjusting for sociodemographics, comorbidities, and drive-time to VHA specialty care.
Key Results
In total, 47.3% had at least one MH condition. Compared to those without MH conditions, Veterans with MH conditions had less fragmented care, with fewer non-MH providers (IRR = 0.96; 95% CI: 0.96–0.96) and more concentrated care with their usual provider (OR = 1.08 for a higher UPC; 95% CI: 1.07, 1.09) in adjusted models. Secondary analyses showed that those with individual MH conditions (e.g., depression) had fewer non-MH providers (IRR range: 0.86–0.98) and more concentrated care (OR range: 1.04–1.20). A similar pattern was observed when examining combinations of MH conditions (IRR range: 0.80–0.90; OR range: 1.16–1.30).
Conclusions
Contrary to expectations, having a MH condition was associated with less fragmented non-MH care among older, high-risk Veterans. Further research will determine if this is due to different needs, underuse, or appropriate use of healthcare.
Supplementary Information
The online version contains supplementary material available at 10.1007/s11606-022-07705-z.
INTRODUCTION
Advances in medical knowledge, rapidly changing diagnostics and therapeutics, and higher life expectancies have transformed medical care such that care is provided by multiple professionals with specialized knowledge from diverse disciplines.1 This approach is medically necessary to manage care of increasingly complex and high-risk patients. Among adults over 65 years old in the USA, 56% have two or more chronic conditions2 and may require a multidisciplinary approach that involves multiple healthcare providers and an increase in clinic visits. Such dispersion of care across healthcare providers and visits is referred to as care fragmentation. In the absence of care coordination among professionals, care fragmentation introduces inefficiencies,3 deviations from clinical guidelines,4 increases in preventable hospitalizations,5 higher mortality,6 greater healthcare spending,4 greater prescription drug use,3 more encounters with specialists, and fewer encounters with a primary care team.3
We hypothesize that older Veterans with mental health (MH) conditions have greater outpatient care fragmentation than those without MH conditions. MH conditions are present among 15% of adults 60 years or older globally.7 In the US Veterans Health Administration (VHA), one in four Veterans seen in primary care has at least one MH condition.8 Similar to care fragmentation, the presence of MH conditions is generally associated with worse outcomes, notably developing chronic medical conditions,9–11 exacerbating existing medical conditions,12,13 increasing the risk for preventable ED visits and hospitalizations,12,14,15 and premature mortality especially when MH conditions are comorbid.16 While literature linking MH conditions to outpatient care fragmentation is limited, there is evidence that individuals with MH conditions fare worse on the related idea of care coordination. Benzer et al.17 used a patient-reported survey to assess patient-centered care coordination among nearly 6000 Veterans. Consistent with their hypothesis, they found that individuals with MH conditions had consistently poorer care coordination than non-MH conditions. This was true across seven dimensions of care coordination, including knowledge fragmentation and disruptions in the flow of information. Care coordination was especially poor among high-severity MH conditions such as PTSD and schizophrenia. Others have shown that individuals with MH conditions may prefer to be actively involved in decision-making around psychiatric care but defer to their providers regarding non-psychiatric care.18,19 Care coordination and lack of engagement may make Veterans with MH conditions especially vulnerable to outpatient care fragmentation.
Older Veterans may experience care fragmentation because of the number of medical conditions, and logistical barriers to timely access (e.g., geographical distance from healthcare facilities). Unfortunately, the associations between MH conditions and outpatient care fragmentation have not been well explored within or outside of the VHA. A few non-VHA studies have found intriguing associations with inpatient care fragmentation.5,6,20,21 Schrag et al.5 examined fragmentation in hospital-based care among frequently hospitalized individuals living in New York City, and defined high care fragmentation as the number of hospitals visited relative to the number of hospitalizations over a 3-year period. The greatest predictor of care fragmentation was a diagnosis of psychosis and substance use disorder. More recently, Hempstead et al. conducted a similar study with residents of New Jersey who had multiple chronic conditions and were high utilizers of inpatient care. They defined care fragmentation as having received care at multiple hospitals during a 1-year period. Similar to Schrag et al., they found that hospital-based care fragmentation was highest among individuals with a MH condition.
Building on these studies, we sought to address the lack of information about associations between MH conditions and outpatient care fragmentation across non-MH providers and visits. The objectives of this study were to characterize non-MH care fragmentation among high-risk Veterans ages 65+ with no vs. any MH conditions, and to examine non-MH care fragmentation for Veterans with specific MH conditions as well as MH condition combinations previously shown to be associated with high mortality rates.8,16 Our primary hypothesis was that Veterans with MH conditions would experience greater non-MH outpatient care fragmentation compared to those who did not have a MH condition.
METHODS
Study Cohort
This is a secondary analysis of a retrospective cohort study that examined care fragmentation among VHA patients at high risk for hospitalization.22 Here, we conducted a retrospective cohort analysis of non-MH care fragmentation in a cohort of Veterans 65 years and older, all of whom were dually enrolled in VHA and fee-for-service (FFS) Medicare in federal fiscal year 2014 (FY14). We used two measures of care fragmentation as our outcomes: provider count and Usual Provider of Care (UPC). In our primary analysis, we compared care fragmentation among Veterans with any vs. no MH conditions. In secondary analyses, we measured care fragmentation among Veterans with specific MH conditions or combinations of MH conditions. All study procedures were approved by the local scientific review committee and the Institutional Review Board. This study adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.23 Participants in the main cohort study were VHA patients at high risk for hospitalization as determined by the Care Assessment Need (CAN) score, a VHA-based algorithm that predicts a patient’s risk of hospitalization over the next year.24,25 Veterans were included if their last recorded CAN score in FY14 was ≥ 90th percentile and if there was no recorded death by the end of FY14. To be eligible for the current study, they had to be 65 years of age or older and have four or more VHA, VHA-purchased Community Care, or Medicare FFS non-MH care encounters between October 1, 2014, and September 30, 2015 (FY14). Of the 541,382 Veterans in the parent study, 251,679 were 65 years of age or older and 143,074 patients had continuous Medicare FFS. At the time of this analysis, Medicare Advantage (managed care) encounter-level data were not available. Of these, 17,603 patients were excluded if they had fewer than four non-MH outpatient visits due to minimal variation in care fragmentation.26–28 The final cohort included 125,469 Veterans 65+ enrolled in VHA and Medicare, whose risk of hospitalization or death was in the top 10% and who had four or more non-MH visits (Fig. 1).
Figure 1.
Defining the cohort using Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. Notes: PSSG, Planning Systems Support Group maintains a dataset with locations of all VHA facilities and the distance from each zip code to the nearest VHA inpatient and outpatient facility.
Data Sources
VHA administrative data were abstracted from the VA Corporate Data Warehouse (CDW) and Medicare claims data from the VA Information Resource Center (VIReC) for FY14-FY15. Drive time to VHA specialty care, defined as “patient’s nearest VA facility providing ambulatory and inpatient care,” was abstracted from the VA Planning Systems Support Group (PSSG). Chronic physical conditions were identified using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes from a list of 47 chronic conditions reported previously in Yoon et al. (see Appendix).29 MH conditions were categorized from ICD-9 codes according to the Agency for Healthcare Research and Quality Clinical Classifications Software framework,30 VHA’s Health Economics Resource Center, and VA’s Women’s Health Evaluation Initiative (WHEI), and were linked to VHA administrative data.
Measures
Determination of MH Conditions
Veterans were designated as having a MH condition if they had two or more outpatient encounters or one inpatient encounter during FY14 for the following MH conditions defined by ICD-9 criteria: depression, post-traumatic stress disorder (PTSD), anxiety, substance use disorders (SUD), opioid use disorder (OUD) (OUD was included within SUD but also examined separately), serious mental illness (SMI), personality disorders, and other psychosis. We defined SMI as having a diagnosis of bipolar disorder and/or schizophrenia.31 These categories are commonly used in defining MH conditions within the VA due to their prevalence,8 increased risk for poor outcomes,16,32,33 or both. Two outpatient encounters were used to enhance the specificity of identifying Veterans with MH conditions, recognizing that this may result in a loss of power. Specific codes used are provided in Appendix.
Outcome Variables
Care fragmentation in non-MH care was determined using two measures:
Provider count: This was defined as the number of unique non-MH providers across all non-MH outpatient encounters observed in FY14 for each Veteran.
Usual Provider of Care (UPC):34 This was defined as the proportion of outpatient encounters with the Veteran’s most frequently seen non-MH provider, in other words, their care concentration with a primary provider.35
We excluded all MH encounters and MH care providers from the outcome measures, including those associated with VHA’s Primary Care Mental Health Integration (PCMHI), an integrated care program situated within VHA Primary Care in which 5% of the study population participated. This was done because many Veterans with a MH condition are likely to have more fragmented care from different MH specialties (e.g., psychologists, psychiatrists) and different modalities (e.g., individual therapy, group therapy, medication management). We focused on non-MH care to facilitate comparisons of the two groups. Encounters were defined as any Current Procedural Terminology (CPT) code pertaining to a clinic or home-based visit conducted by a physician, nurse practitioner, physician’s assistant, or medical resident. Outpatient encounters included any primary care, specialty care, or surgical care visit. In VHA data, stop codes were used to define these care categories while in Medicare and VHA-purchased Community Care, provider type and E&M codes were used. The Appendix describes additional information about the identification of encounters and providers used in constructing our care fragmentation measures.
Demographic and Other Health Characteristics
All demographic and health characteristic variables were based on the last reported quarter of FY14 except for age, which was based on the individual’s age at the beginning of FY14. Demographic variables included age (used as a continuous variable in analyses), sex (male/female), marital status (married/not married), race (White/Black/Other), ethnicity (Hispanic/not Hispanic), Veterans’ priority group enrollment, and drive time to VHA specialty care. Health variables included the Charlson Comorbidity Index (CCI)36 and number of chronic physical conditions. The CCI includes a list of 17 conditions, coded as present if they occurred in any encounter in FY14. Our rationale for including both measures was that CCI captures the burden of illness better than a count of conditions but count of conditions could better reflect non-MH care fragmentation.
Veteran priority status was used as a proxy measure for likelihood of using VHA over Medicare. Veteran priority status is based on a combination of factors, including military service history, disability rating, income level, Medicaid qualification, and whether the Veteran is receiving other benefits (e.g., VHA pension benefits). A higher priority status is indicated by a lower number, with 1 being the highest priority. Veterans with higher priority status have lower copayments and medication costs than lower priority groups.
Data Analytic Plan
The primary aim was to compare care fragmentation among Veterans with any vs. no MH conditions. Secondary analyses compared care fragmentation within individual MH condition groups—depression, PTSD, anxiety, SUD, OUD, SMI, personality disorders, and other psychosis with no MH conditions. We also compared Veterans with no MH conditions with prespecified MH condition combinations: depression + SUD + SMI; SUD + SMI; depression + SMI; and depression + SUD because these are associated with poor outcomes, namely, highest 10-year mortality rates among Veterans seen within the VHA.16 We used “no MH conditions” as the comparator group across all analyses for consistency.
We used Poisson regressions to model the odds of provider count and fractional logistic regressions to model the odds of UPC. We first calculated unadjusted models, which accounted only for our MH conditions and the two outcome variables. We then calculated adjusted models, controlling for demographic variables and health characteristics. To account for the interdependence of the data due to clustering of Veterans within VHA care locations, all models included fixed effects for the Veteran’s nearest VHA facility providing ambulatory and inpatient care.
RESULTS
Table 1 displays the demographic and health characteristics and care fragmentation measures for the total sample, as well as for Veterans with any and no MH conditions. As is typical among older Veterans, nearly all Veterans in the sample were male (97.7%) and over three-quarters (81.9%) identified as White. Over half (50.5%) of Veterans were between the ages of 65 and 74 years old. Most Veterans with a MH condition (54.2%) were in priority group 1 (service-connected disability > 50%) while most Veterans without a MH condition (40.9%) were in priority groups 4–6 (eligibility through aid and attendance, housebound, VA pension benefits, or having low incomes qualified for Medicaid). Veterans, on average, had eight chronic non-MH conditions and a CCI of 4.7, indicating high mortality risk.10,37
Table 1.
Patient Demographic and Health Characteristics by MH Conditions Category
| Total sample | Any MH conditions | No MH conditions | |
|---|---|---|---|
| N | 125,469 | 59,288 | 66,181 |
| Age, % (n) | |||
| 65–74 | 63.1 (79,205) | 71.7 (42,483) | 55.5 (36,722) |
| 75–84 | 24.2 (30,351) | 18.7 (11,093) | 29.1 (19,258) |
| 85–94 | 12.2 (15,313) | 9.3 (5519) | 14.8 (9794) |
| 95+ | 0.5 (600) | 0.3 (193) | 0.6 (407) |
| Mean (SD, range) | 73.9 (8.0, 65–104) | 72.4 (7.5, 65–103) | 75.2 (8.1, 65–104) |
| Sex, male, % (n) | 97.7 (122,572) | 97.4 (57,718) | 98.0 (64,854) |
| Married, % (n) | 47.2 (59,146) | 47.6 (28,204) | 46.8 (30,942) |
| Race, % (n) | |||
| White | 81.9 (98,973) | 82.4 (47,091) | 81.5 (51,882) |
| Black | 16.2 (19,550) | 15.6 (8927) | 16.7 (10,623) |
| Other | 1.9 (2275) | 1.9 (1101) | 1.8 (1174) |
| Ethnicity, % (n) | |||
| Hispanic/Latino | 3.8 (4615) | 4.1 (2361) | 3.5 (2254) |
| Drive time | |||
| > 60 min to specialty care | 26.0 (32,597) | 26.0 (15,409) | 26.0 (17,187) |
| Priority groups, % (n) | |||
| 1 (Service-connected disability > 50%) | 43.9 (55,047) | 54.2 (32,144) | 34.6 (22,903) |
| 2 (Service-connected disability 30–40%) | 4.7 (5953) | 4.5 (2646) | 5.0 (3307) |
| 3 (Service-connected disability 10–20%, POW, Purple Heart) | 7.0 (8732) | 5.9 (3490) | 7.9 (5242) |
| 4–6 (Eligibility through aid and attend, housebound, VA pension benefits, or having low incomes qualified for Medicaid | 35.5 (44,533) | 29.4 (17,439) | 40.9 (27,094) |
| 7–8 (Income above VA means test limits) | 8.9 (11,204) | 6.0 (3569) | 11.5 (7635) |
| Charlson Comorbidity Index (CCI) | |||
| Mean (SD, range) | 4.7 (3.0, 0–23) | 4.7 (3.1, 0–23) | 4.8 (2.9, 0–21) |
| Median (IQR*) | 4 (3–6) | 4 (2–6) | 4 (3–6) |
| Number of non-MH chronic conditions | |||
| Mean (SD, range) | 8.2 (3.7, 0–26) | 8.8 (3.9, 0–26) | 7.7 (3.4, 0–24) |
| Median (IQR) | 8 (6–10) | 8 (6–11) | 7 (5–10) |
| Number of non-MH visits | |||
| Mean (SD, range) | 11.5 (7.0, 4–136) | 11.8 (7.4, 4–136) | 11.3 (6.7, 4–121) |
| Median (IQR) | 10 (7–14) | 10 (7–15) | 10 (7–14) |
| Healthcare fragmentation measures | |||
| UPC | |||
| Mean (SD, range) | 0.4 (0.2, 0.05–1) | 0.4 (0.2, 0.05–1) | 0.4 (0.2, 0.05–1) |
| Median (IQR) | 0.3 (0.3–0.5) | 0.3 (0.3–0.5) | 0.3 (0.3–0.5) |
| Non-MH providers | |||
| Mean (SD, range) | 6.6 (3.5, 1–41) | 6.7 (3.7, 1–41) | 6.5 (3.4, 1–40) |
| Median (IQR) | 6 (4–8) | 6 (4–9) | 6 (4–8) |
| All providers | |||
| Mean (SD, range) | 7.2 (3.7, 1–42) | 7.8 (3.9, 1–42) | 6.6 (3.4, 1–40) |
| Median (IQR) | 6 (5–9) | 7 (5–10) | 6 (4–8) |
Note. Missing values possible across demographic variables. *IQR, interquartile range
Table 2 describes the MH conditions in the total sample and among those with any MH conditions. Nearly half (47.3%) of high-risk Veterans had at least one MH condition, which included depression (30%), PTSD (17.5%), and anxiety (13.5%). Among those with a MH condition, nearly two-thirds (61%) had depression suggesting a significant overlap with other MH conditions in this group. Among the combinations of MH conditions that we examined, 11.9% of Veterans had depression and SUD, 4.4% had depression and SMI, and 1.4% had depression, SUD, and SMI. We examined the number of providers with and without MH providers among Veterans with a MH condition.
Table 2.
Patient Mental Health (MH) Characteristics
| MH conditions | n | % total sample (N = 125,469) |
% any MH conditions (n = 59,288) |
|---|---|---|---|
| Depression | 37,611 | 30.0 | 63.4 |
| PTSD | 21,918 | 17.5 | 37.0 |
| Anxiety | 16,988 | 13.5 | 28.7 |
| SUD | 13,893 | 11.1 | 23.4 |
| OUD | 1469 | 1.2 | 2.5 |
| SMI | 5912 | 4.7 | 10.0 |
| Personality disorder | 1232 | 1.0 | 2.1 |
| Other psychosis | 5059 | 4.0 | 8.5 |
| Depression + SUD + SMI | 836 | 0.7 | 1.4 |
| SUD + SMI | 1476 | 1.2 | 2.5 |
| Depression + SMI | 2583 | 2.1 | 4.4 |
| Depression + SUD | 7038 | 5.6 | 11.9 |
| Number of MH conditions | |||
| 0 conditions | 66,181 | 52.8 | -- |
| 1 condition | 29,748 | 23.7 | 50.2 |
| 2 conditions | 18,220 | 14.5 | 30.7 |
| 3+ conditions | 11,320 | 3.1 | 6.7 |
| Mean (SD, range) | - | 0.8 (1.1, 0–10) | 1.8 (1.0, 1–10) |
| Median (IQR) | - | 0 (0–1) | 1 (1–2) |
Tables 3 and 4 provide detailed results regarding our unadjusted and adjusted models, including odds ratios, incident rate ratios, p values, and 95% CIs. Most of the unadjusted and adjusted models had identical patterns; the smaller sample size in Table 4 reflects missing data in the covariates. All adjusted analyses were significant at p <.001. In adjusted models, compared to Veterans without MH conditions, those with MH conditions saw 4% fewer non-MH providers (IRR = 0.96) and had a significantly higher UPC (more concentrated care; AOR = 1.08). When we compared Veterans with no MH conditions to Veterans with individual MH conditions, we found that having a MH condition was associated with lower provider count ranging from 2% (depression) to 16% fewer (psychosis) providers. Veterans with PTSD and anxiety saw 2% fewer non-MH providers, while those with personality disorders saw 6% fewer non-MH providers. Veterans with SUD in general and OUD in particular saw 10% and 7% fewer providers. This number was even lower among Veterans with SMI and psychosis, who saw 14% and 16% fewer non-MH providers respectively.
Table 3.
Unadjusted Associations Between MH Conditions or Combinations of MH Conditions and MH Fragmentation Outcomes
| Variable | n | Provider count | UPC* | ||||
| IRR† | 95% CI‡ | p value | OR§ | 95% CI | p value | ||
| Any vs. no MH conditions | 125,469 | 1.02 | 1.01–1.02 | <.001 | 1.03 | 1.02–1.03 | <.001 |
| Depression | 103,792 | 1.04 | 1.04–1.05 | <.001 | 1.01 | 1.00–1.02 | .03 |
| PTSD‖ | 88,099 | 1.0 | 1.03–1.04 | <.001 | 0.98 | 0.97–0.99 | .002 |
| Anxiety | 83,169 | 1.05 | 1.05–1.06 | <.001 | 1.03 | 1.01–1.04 | <.001 |
| SUD¶ | 80,074 | 0.94 | 0.94–0.95 | <.001 | 1.09 | 1.08–1.11 | <.001 |
| OUD# | 67,650 | 1.02 | 1.00–1.04 | .07 | 1.12 | 1.08–1.17 | <.001 |
| SMI** | 72,093 | 0.90 | 0.89–0.91 | <.001 | 1.18 | 1.15–1.20 | <.001 |
| Personality disorder | 67,413 | 1.01 | 0.98–1.03 | .57 | 1.05 | 1.00–1.09 | .04 |
| Other psychosis | 71,240 | 0.95 | 0.94–0.96 | <.001 | 1.15 | 1.12–1.17 | <.001 |
| Depression + SUD+ SMI | 67,017 | 0.90 | 0.88–0.93 | <.001 | 1.16 | 1.10–1.23 | <.001 |
| SUD + SMI | 67,657 | 0.86 | 0.84–0.87 | <.001 | 1.23 | 1.18–1.28 | <.001 |
| Depression + SMI | 68,764 | 0.95 | 0.93–0.96 | <.001 | 1.13 | 1.10–1.17 | <.001 |
| Depression + SUD | 73,219 | 0.97 | 0.96–0.98 | <.001 | 1.08 | 1.06–1.10 | <.001 |
Note. *UPC, Usual Provider of Care; †IRR, incidence rate ratio; ‡CI, confidence interval; §OR, odds ratio; ‖PTSD, post-traumatic stress disorder; ¶SUD, substance use disorder; #OUD, opiate use disorder; **SMI, serious mental illness
Table 4.
Adjusted Associations Between MH Conditions or Combinations of MH Conditions and MH Fragmentation Outcomes*
| Variable | n | Provider count | UPC | ||||
|---|---|---|---|---|---|---|---|
| IRR | 95% CI | p value | AOR† | 95% CI | p value | ||
| Any vs. no MH conditions | 118,474 | 0.96 | 0.96–0.96 | <.001 | 1.08 | 1.07–1.09 | <.001 |
| Depression | 98,038 | 0.97 | 0.97–0.98 | <.001 | 1.07 | 1.06–1.08 | <.001 |
| PTSD | 83,179 | 0.98 | 0.97–0.98 | <.001 | 1.06 | 1.05–1.08 | <.001 |
| Anxiety | 78,519 | 0.98 | 0.97–0.98 | <.001 | 1.09 | 1.07–1.10 | <.001 |
| SUD | 75,601 | 0.90 | 0.89–0.91 | <.001 | 1.15 | 1.13–1.16 | <.001 |
| OUD | 63,832 | 0.93 | 0.91–0.95 | <.001 | 1.22 | 1.17–1.27 | <.001 |
| SMI | 68,043 | 0.86 | 0.85–0.87 | <.001 | 1.23 | 1.20–1.25 | <.001 |
| Personality disorder | 63,622 | 0.94 | 0.92–0.97 | <.001 | 1.11 | 1.06–1.16 | <.001 |
| Other psychosis | 67,173 | 0.84 | 0.83–0.85 | <.001 | 1.23 | 1.20–1.26 | <.001 |
| Depression + SUD + SMI | 63,241 | 0.81 | 0.79–0.83 | <.001 | 1.27 | 1.21–1.35 | <.001 |
| SUD + SMI | 63,851 | 0.80 | 0.79–0.82 | <.001 | 1.30 | 1.25–1.36 | <.001 |
| Depression + SMI | 64,894 | 0.86 | 0.84–0.87 | <.001 | 1.23 | 1.19–1.27 | <.001 |
| Depression + SUD | 69,133 | 0.90 | 0.90–0.91 | <.001 | 1.16 | 1.14–1.19 | <.001 |
Note. *Analyses were adjusted for age, priority status, gender, race, ethnicity, marital status, drive time to VA specialty care, CCI, number of chronic physical conditions, and closest VA secondary care site
†AOR, adjusted odds ratio
We found that having a MH condition was also associated with more concentrated care for Veterans. Veterans with depression, PTSD, and anxiety had 1.07, 1.06, and 1.09 greater odds of having more concentrated care than those without MH conditions. Once again, the more serious conditions such as SUD, OUD, and SMI had more concentrated care (1.15, 1.22, 1.23 respectively). Veterans with a personality disorder had 1.11 greater odds of having more concentrated care while Veterans with psychosis had 1.23 greater odds of having more concentrated care.
Finally, we examined the combinations of MH conditions that are associated with greater morbidity and mortality. Here, too, we found the same pattern of results where each combination was associated with less care fragmentation as measured by both non-MH provider count and UPC. Veterans with depression, SUD, and SMI saw 19% fewer non-MH providers. Similarly, Veterans with SUD and SMI saw 20% fewer non-MH (IRR = 0.80). Other combinations also followed this pattern, with Veterans with depression and SMI seeing 14% fewer and those with depression and SUD seeing 10% fewer non-MH providers. Concentration of care was also higher among these groups as can be seen in Table 4, with the odds of having more concentrated care ranging from 1.16 to 1.30. Differences in effects sizes were observed with the MH condition combinations and specific MH conditions (i.e., SUD, OUD, SMI, personality disorders, other psychosis) demonstrating the largest effect sizes (Table 4).
DISCUSSION
In this study of older VHA patients at high risk for hospitalization, we found that having a MH condition is not associated with more fragmented outpatient care for non-MH conditions. This finding contradicted to our expectations, since we had hypothesized that those with MH conditions in our group would experience more fragmentation once we adjusted for patients’ number and type of conditions.
Care fragmentation is typically viewed as a negative side effect of multidisciplinary care among complex patients, and if so, less care fragmentation is more desirable. If true, our study findings suggest that the non-MH care received by high-risk Veterans with MH conditions is superior to those without MH conditions in that they are receiving less fragmented care. However, the converse is possible and high-risk Veterans with MH conditions may not be accessing appropriate specialty care. For instance, the providers could be focused on managing the serious or care-intensive MH conditions leaving little time to devote to non-MH conditions. This latter interpretation is supported by our finding that those with the most serious or care-intensive MH conditions such as SUD, OUD, and psychosis had the least fragmentation. Unfortunately, our study does not lend itself to determining whether care fragmentation is desirable or not.
Furthermore, this may be a more complex problem, since we are unable to ascertain Veterans’ tolerance for care fragmentation. It is possible that Veterans are only able to commit to a certain number of healthcare appointments due to limitations in time, resources, and MH. If so, those with MH conditions may prioritize their MH-related appointments. Future studies may consider developing a metric to determine how much fragmentation is optimal for complex and high-risk individuals and incorporate the patients’ perspectives on care fragmentation. Despite this limitation of our data, we believe our findings likely demonstrate a positive and desirable decrease in care fragmentation among Veterans with MH conditions given the aforementioned VHA initiatives (i.e., patient-centered medical home and primary care-MH integration), which have been successful in improving care coordination for Veterans with MH conditions.38
Another surprising observation was that the MH conditions that are associated with the highest utilization or most intensive MH treatment (e.g., inpatient care, residential treatment) and worst outcomes—SUD, OUD, SMI, personality disorders, other psychosis, and combinations—were associated with the least care fragmentation for non-MH conditions.31,39–41 We consider two possible explanations. One possibility is that VHA initiatives such as the patient-centered medical home and integrated primary care have been successful in keeping Veterans with MH conditions engaged with VHA care. Studies have indicated that the patients with MH conditions experience challenges in adhering to care regimens and coordinating their own care,39 and it is possible that recent VHA initiatives have helped these Veterans overcome these barriers. Indeed, Browne et al.31 compared quality of care among Veterans with and without MH conditions seen in the VHA’s patient-centered medical home (called Patient-Aligned Care Teams, or PACT). They found that the quality of care received by both groups was nearly identical across 33 quality indicators examined. Outside of the VHA, Daumit and colleagues have conducted studies examining the integration of primary care in specialty MH settings.40,41 In their recent review,41 they noted that integration of primary care into specialty MH settings resulted in improved primary care access, and improved screening and monitoring of cardiovascular disease among individuals with SMI. Similar trends could be examined among non-MH care providers as well and could show that comorbid MH and non-MH conditions could result in better care. Daumit and colleagues found that psychiatrists were likely to refer patients to preventive services if they had also had comorbid chronic health conditions.40 Similar trends could be examined among non-MH care providers as well and could show that comorbid MH and non-MH conditions could result in better care. A second possibility is that those with the most serious or care-intensive MH conditions have lower rates of referral for specialty care or are less likely to access primary and specialty care services. While this might be because of appropriate prioritization of their MH conditions (e.g., when patients are receiving intensive MH care), it is also possible that these Veterans are less likely to be referred to and receive specialty care. This issue warrants further investigation in future studies.
Strengths of our study included the large, national sample size; inclusion of VHA and Medicare enrolled Veterans; and rigorous selection of covariates. A common limitation of studies conducted within the VHA is that findings may not generalize to the broader population because Veterans seen within the VHA are predominantly male and sicker than the general population.42 Specific to these analyses, a key limitation was our inability to assess the quality of care received because we relied exclusively on administrative and claims data. Another limitation is the potential confounding of our results, since it is possible that Veterans who are at a high risk for hospitalization are also at an increased risk for outpatient fragmentation regardless of their MH condition. In addition, we did not distinguish between care that was dispersed within VHA versus care that was dispersed across VHA and other systems. Previous studies have found that Veterans who receive Medicare-covered non-VA care have greater outpatient care fragmentation,22 and that fragmentation across VHA and other health systems (e.g., dual use of VA and Medicare) is associated with poor clinical outcomes for a range of conditions.43–50 Our objective in this study was to examine measures of outpatient fragmentation across all non-MH visits, but future studies could consider hierarchical models to determine whether there are distinct patterns between vs. within systems. We also lacked information on what was discussed in clinical visits, and whether the complex care was appropriately managed. As mentioned earlier, we neither had insights from the Veterans on their willingness to seek care from multiple providers, nor how this may vary by MH conditions. Another limitation was that we did not have insights into efforts made by VHA care teams to engage those with MH conditions. Data regarding these efforts would help determine whether the low care fragmentation we observed among Veterans with MH conditions was a positive outcome of extensive outreach by the care teams. Finally, we also recognize that non-MH providers may provide MH care during their visits, and this would be coded as non-MH care in our data because we relied on clinic codes. Future studies may build on these findings to examine clinical outcomes, utilization of services, and address these limitations.
The VHA continues to be a pioneer in developing initiatives that support the care of Veterans with MH conditions, such as the SMI PACT51 which is a specialized model that combines aspects of both VHA’s patient-centered medical home and integrated primary care to improve care for Veterans with serious mental illness seen within VHA. This study provides a new direction for understanding how Veterans with MH conditions engage with healthcare systems. Future work is needed to understand why Veterans with MH conditions had lower levels of fragmentation, and to better understand if this indicates high-quality care, or underuse of available non-MH care. Future studies should also study how non-MH care fragmentation is associated with clinical outcomes and utilization among Veterans with MH conditions, as well as study care fragmentation within MH care. These insights will ultimately improve patient care in healthcare systems that are managing increasingly complex patients.
Supplementary Information
(DOCX 60 kb)
Funding
This study was supported by the Department of Veterans Affairs (5I01HX002127-02). Dr. Megan Vanneman was further supported by a career development award from VA HSR&D (CDA 15-259). Dr. Rossi was supported by the VA Office of Academics Affairs Advanced Fellowship in Medical Informatics. Dr. Javier was supported by the VA Office of Academics Affairs Advanced Fellowship in Health Services Research. The opinions expressed are those of the authors and not necessarily those of the Department of Veterans Affairs, the US Government, Stanford University, or the University of Utah.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Ouwens M, Wollersheim H, Hermens R, Hulscher M, Grol R. Integrated care programmes for chronically ill patients: a review of systematic reviews. Int J Qual Health Care. 2005;17(2):141-146. 10.1093/intqhc/mzi016 [DOI] [PubMed]
- 2.National Center for Health Statistics. Percent of U.S. Adults 55 and Over with Chronic Conditions. Centers for Disease Control and Prevention; 2009:2. Accessed May 24, 2021. https://www.cdc.gov/nchs/data/health_policy/adult_chronic_conditions.pdf
- 3.Agha L, Frandsen B, Rebitzer JB. Causes and Consequences of Fragmented Care Delivery: Theory, Evidence, and Public Policy. Natl Bur Econ Res Work Pap Ser. Published online 2017:50.
- 4.Frandsen BR, Joynt KE, Rebitzer JB, Jha AK. Care fragmentation, quality, and costs among chronically ill patients. Am J Manag Care. 2015;21(5):355-362. [PubMed]
- 5.Schrag D, Xu F, Hanger M, Elkin E, Bickell NA, Bach PB. Fragmentation of Care for Frequently Hospitalized Urban Residents. Med Care. 2006;44(6):560-567. [DOI] [PubMed]
- 6.McCrum ML, Cannon AR, Allen CM, Presson AP, Huang LC, Brooke BS. Contributors to Increased Mortality Associated With Care Fragmentation After Emergency General Surgery. JAMA Surg. 2020;155(9):841-848. 10.1001/jamasurg.2020.2348 [DOI] [PMC free article] [PubMed]
- 7.World Health Organization. Mental health of older adults. Mental health of older adults. Published 2017. Accessed May 24, 2021. https://www.who.int/news-room/fact-sheets/detail/mental-health-of-older-adults
- 8.Trivedi RB, Post EP, Sun H, et al. Prevalence, Comorbidity, and Prognosis of Mental Health Among US Veterans. Am J Public Health. 2015;105(12):2564-2569. 10.2105/AJPH.2015.302836 [DOI] [PMC free article] [PubMed]
- 9.Roy T, Lloyd CE. Epidemiology of depression and diabetes: A systematic review. J Affect Disord. 2012;142:S8-S21. 10.1016/S0165-0327(12)70004-6 [DOI] [PubMed]
- 10.Leung YW, Flora DB, Gravely S, Irvine J, Carney RM, Grace SL. The impact of pre-morbid and post-morbid depression onset on mortality and cardiac morbidity among coronary heart disease patients: A meta-analysis. Psychosom Med. 2012;74(8):786-801. 10.1097/PSY.0b013e31826ddbed [DOI] [PMC free article] [PubMed]
- 11.Gan Y, Gong Y, Tong X, et al. Depression and the risk of coronary heart disease: a meta-analysis of prospective cohort studies. BMC Psychiatry 2014;14:371. 10.1186/s12888-014-0371-z [DOI] [PMC free article] [PubMed]
- 12.Helmer DA, Dwibedi N, Rowneki M, et al. Mental Health Conditions and Hospitalizations for Ambulatory Care Sensitive Conditions Among Veterans with Diabetes. Am Health Drug Benefits 2020;13(2):61-71. [PMC free article] [PubMed]
- 13.Outcalt SD, Kroenke K, Krebs EE, et al. Chronic pain and comorbid mental health conditions: independent associations of posttraumatic stress disorder and depression with pain, disability, and quality of life. J Behav Med. 2015;38(3):535-543. 10.1007/s10865-015-9628-3 [DOI] [PubMed]
- 14.Lin WC, Bharel M, Zhang J, O’Connell E, Clark RE. Frequent Emergency Department Visits and Hospitalizations Among Homeless People With Medicaid: Implications for Medicaid Expansion. Am J Public Health. 2015;105(S5):S716-S722. 10.2105/AJPH.2015.302693 [DOI] [PMC free article] [PubMed]
- 15.Painter JM, Malte CA, Rubinsky AD, et al. High inpatient utilization among Veterans Health Administration patients with substance-use disorders and co-occurring mental health conditions. Am J Drug Alcohol Abuse. 2018;44(3):386-394. 10.1080/00952990.2017.1381701 [DOI] [PMC free article] [PubMed]
- 16.Trivedi RB, Post EP, Piegari R, et al. Mortality Among Veterans with Major Mental Illnesses Seen in Primary Care: Results of a National Study of Veteran Deaths. J Gen Intern Med. 2020;35(1):112-118. 10.1007/s11606-019-05307-w [DOI] [PMC free article] [PubMed]
- 17.Benzer JK, Singer SJ, Mohr DC, et al. Survey of Patient-Centered Coordination of Care for Diabetes with Cardiovascular and Mental Health Comorbidities in the Department of Veterans Affairs. J Gen Intern Med. 2019;34(S1):43-49. 10.1007/s11606-019-04979-8 [DOI] [PMC free article] [PubMed]
- 18.O’Neal EL, Adams JR, McHugo GJ, Van Citters AD, Drake RE, Bartels SJ. Preferences of Older and Younger Adults With Serious Mental Illness for Involvement in Decision-Making in Medical and Psychiatric Settings. Am J Geriatr Psychiatry. 2008;16(10):826-833. 10.1097/JGP.0b013e318181f992 [DOI] [PubMed]
- 19.Moise N, Ye S, Alcántara C, Davidson KW, Kronish I. Depressive symptoms and decision-making preferences in patients with comorbid illnesses. J Psychosom Res. 2017;92:63-66. 10.1016/j.jpsychores.2015.12.001 [DOI] [PMC free article] [PubMed]
- 20.Hempstead K, DeLia D, Cantor JC, Nguyen T, Brenner J. The Fragmentation of Hospital Use Among a Cohort of High Utilizers: Implications for Emerging Care Coordination Strategies for Patients With Multiple Chronic Conditions. Med Care. 2014;52(3):S67-S74. [DOI] [PubMed]
- 21.Kaltenborn Z, Paul K, Kirsch JD, et al. Super fragmented: a nationally representative cross-sectional study exploring the fragmentation of inpatient care among super-utilizers. BMC Health Serv Res. 2021;21(1):338. 10.1186/s12913-021-06323-5 [DOI] [PMC free article] [PubMed]
- 22.Zulman DM, Greene L, Slightam C, et al. Outpatient care fragmentation in Veterans Affairs patients at high-risk for hospitalization. Health Serv Res. Published online March 11, 2022:1475-6773.13956. 10.1111/1475-6773.13956 [DOI] [PMC free article] [PubMed]
- 23.von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: Guidelines for Reporting Observational Studies. Ann Intern Med. 2007;147(8):573. 10.7326/0003-4819-147-8-200710160-00010 [DOI] [PubMed]
- 24.Wang L, Porter B, Maynard C, et al. Predicting Risk of Hospitalization or Death Among Patients Receiving Primary Care in the Veterans Health Administration. Med Care. 2013;51(4):368-373. 10.1097/MLR.0b013e31827da95a [DOI] [PubMed]
- 25.Nelson KM, Chang ET, Zulman DM, Rubenstein LV, Kirkland FD, Fihn SD. Using Predictive Analytics to Guide Patient Care and Research in a National Health System. J Gen Intern Med. 2019;34(8):1379-1380. 10.1007/s11606-019-04961-4 [DOI] [PMC free article] [PubMed]
- 26.Kern LM, Rajan M, Pincus HA, Casalino LP, Stuard SS. Health Care Fragmentation in Medicaid Managed Care vs. Fee for Service. Popul Health Manag. 2019;23(1):53-58. 10.1089/pop.2019.0017 [DOI] [PubMed]
- 27.Nyweide DJ, Anthony DL, Bynum JPW, et al. Continuity of Care and the Risk of Preventable Hospitalization in Older Adults. JAMA Intern Med. 2013;173(20):1879. 10.1001/jamainternmed.2013.10059 [DOI] [PMC free article] [PubMed]
- 28.Rosenberg NA, Zulman DM. Measures of care fragmentation: Mathematical insights from population genetics. Health Serv Res. 2020;55(2):318-327. 10.1111/1475-6773.13263 [DOI] [PMC free article] [PubMed]
- 29.Yoon J, Chee CP, Su P, Almenoff P, Zulman DM, Wagner TH. Persistence of High Health Care Costs among VA Patients. Health Serv Res. 2018;53(5):3898-3916. 10.1111/1475-6773.12989 [DOI] [PMC free article] [PubMed]
- 30.Elixhauser A, Steiner C, Palmer L. Clinical Classifications Software (CCS). U.S. Agency for Healthcare Research and Quality; 2011. Accessed May 24, 2021. http://www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs.jsp.
- 31.Browne KC, Hoerster KD, Piegari R, et al. Clinical Care Quality Among Veterans Health Administration Patients With Mental Illness Following Medical Home Implementation. Psychiatr Serv. 2019;70(9):816-823. 10.1176/appi.ps.201800474 [DOI] [PubMed]
- 32.Dixon-Gordon KL, Whalen DJ, Layden BK, Chapman AL. A systematic review of personality disorders and health outcomes. Can Psychol Can. 2015;56(2):168-190. 10.1037/cap0000024 [DOI] [PMC free article] [PubMed]
- 33.Quirk SE, Williams LJ, Chanen AM, Berk M. Personality disorder and population mental health. Lancet Psychiatry. 2015;2(3):201-202. 10.1016/S2215-0366(14)00144-8 [DOI] [PubMed]
- 34.Breslau N, Haug MR. Service delivery structure and continuity of care: a case study of a pediatric practice in process of reorganization. J Health Soc Behav. 1976;17(4):339-352. [PubMed]
- 35.Pollack CE, Hussey PS, Rudin RS, Fox DS, Lai J, Schneider EC. Measuring Care Continuity: A Comparison of Claims-based Methods. Med Care. 2016;54(5):e30-e34. 10.1097/MLR.0000000000000018 [DOI] [PMC free article] [PubMed]
- 36.Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis. 1987;40(5):373-383. 10.1016/0021-9681(87)90171-8 [DOI] [PubMed]
- 37.Huang YJ, Chen JS, Luo SF, Kuo CF. Comparison of Indexes to Measure Comorbidity Burden and Predict All-Cause Mortality in Rheumatoid Arthritis. J Clin Med. 2021;10(22):5460. 10.3390/jcm10225460 [DOI] [PMC free article] [PubMed]
- 38.Leung LB, Yoon J, Rubenstein LV, et al. Changing Patterns of Mental Health Care Use: The Role of Integrated Mental Health Services in Veteran Affairs Primary Care. J Am Board Fam Med JABFM. 2018;31(1):38-48. 10.3122/jabfm.2018.01.170157 [DOI] [PubMed]
- 39.Grenard JL, Munjas BA, Adams JL, et al. Depression and Medication Adherence in the Treatment of Chronic Diseases in the United States: A Meta-Analysis. J Gen Intern Med. 2011;26(10):1175-1182. 10.1007/s11606-011-1704-y [DOI] [PMC free article] [PubMed]
- 40.Daumit GL, Crum RM, Guallar E, Ford DE. Receipt of Preventive Medical Services at Psychiatric Visits by Patients With Severe Mental Illness. Psychiatr Serv. 2002;53(7):884-887. 10.1176/appi.ps.53.7.884 [DOI] [PubMed]
- 41.Murphy KA, Daumit GL, Stone E, McGinty EE. Physical health outcomes and implementation of behavioural health homes: a comprehensive review. Int Rev Psychiatry. 2018;30(6):224-241. 10.1080/09540261.2018.1555153 [DOI] [PMC free article] [PubMed]
- 42.Schult TM, Schmunk SK, Marzolf JR, Mohr DC. The Health Status of Veteran Employees Compared to Civilian Employees in Veterans Health Administration. Mil Med. 2019;184(7-8):e218-e224. 10.1093/milmed/usy410 [DOI] [PubMed]
- 43.Jia H, Zheng Y, Reker DM, et al. Multiple System Utilization and Mortality for Veterans With Stroke. Stroke. 2007;38(2):355-360. 10.1161/01.STR.0000254457.38901.fb [DOI] [PubMed]
- 44.Tarlov E, Lee TA, Weichle TW, et al. Reduced Overall and Event-Free Survival among Colon Cancer Patients Using Dual System Care. Cancer Epidemiol Biomarkers Prev. 2012;21(12):2231-2241. 10.1158/1055-9965.EPI-12-0548 [DOI] [PubMed]
- 45.Axon RN, Gebregziabher M, Everett CJ, Heidenreich P, Hunt KJ. Dual health care system use is associated with higher rates of hospitalization and hospital readmission among veterans with heart failure. Am Heart J. 2016;174:157-163. 10.1016/j.ahj.2015.09.023 [DOI] [PubMed]
- 46.Maciejewski ML, Wang V, Burgess JF, Bryson CL, Perkins M, Liu CF. The Continuity and Quality of Primary Care. Med Care Res Rev. 2013;70(5):497-513. 10.1177/1077558713495454 [DOI] [PubMed]
- 47.Cooper AL, Jiang L, Yoon J, et al. Dual-System Use and Intermediate Health Outcomes among Veterans Enrolled in Medicare Advantage Plans. Health Serv Res. 2015;50(6):1868-1890. 10.1111/1475-6773.12303 [DOI] [PMC free article] [PubMed]
- 48.Pizer SD, Gardner JA. Is Fragmented Financing Bad for Your Health? Inq J Health Care Organ Provis Financ. 2011;48(2):109-122. 10.5034/inquiryjrnl_48.02.02 [DOI] [PubMed]
- 49.Ajmera M, Wilkins TL, Sambamoorthi U. Dual Medicare and Veteran Health Administration Use and Ambulatory Care Sensitive Hospitalizations. J Gen Intern Med. 2011;26(S2):669-675. 10.1007/s11606-011-1788-4 [DOI] [PMC free article] [PubMed]
- 50.Wolinsky FD, Miller TR, An H, Brezinski PR, Vaughn TE, Rosenthal GE. Dual use of Medicare and the Veterans Health Administration: are there adverse health outcomes? BMC Health Serv Res. 2006;6(1):131. 10.1186/1472-6963-6-131 [DOI] [PMC free article] [PubMed]
- 51.Young AS, Cohen AN, Chang ET, et al. A clustered controlled trial of the implementation and effectiveness of a medical home to improve health care of people with serious mental illness: study protocol. BMC Health Serv Res. 2018;18(1):428. 10.1186/s12913-018-3237-0 [DOI] [PMC free article] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
(DOCX 60 kb)