Impact of Multidisciplinary Care Coordination on Healthcare Utilization among Patients Coinfected with Human Immunodeficiency Virus and Hepatitis C

Abstract

People coinfected with HIV and hepatitis C (HCV) have complex needs that include pharmaceutical therapy, psychiatric care, and social work. Multidisciplinary care coordination (MCC) is the concerted effort to address health using a diverse team of primary care physicians, mental health specialists, and social workers. To determine the effect of MCC in reducing avoidable hospital use, we examined longitudinal data from an academic equity-focused program in Upper Manhattan (CHP). In 2015, CHP implemented MCC, which integrated team members, panel management, and open access (walk-in) activities to better link, engage, and treat those who are coinfected. We compared healthcare utilization trends before 2015 (“pre-intervention”) to after 2015 (“post-intervention”). We found that the implementation of MCC was associated with a decrease in per-year hospital visits among those coinfected. To assist in future upscaling, we hope to identify strategies that facilitated the implementation of MCC in a setting like CHP.

Introduction:

Globally, about 2.3 million people living with human immunodeficiency virus (HIV) have had or are living with hepatitis C (HCV).¹ Individuals coinfected with HIV and HCV have worse physical and mental health outcomes compared to those of the general population.^2–4 Additionally, the coinfected population may be at higher risk of housing instability, excess policing, social marginalization, and other factors in their socioecological environment that can result in worse health.⁵ Consequently, coinfected individuals have a two times the hospitalization rate compared to those monoinfected with HIV.²

One potentially effective solution to reduce avoidable health complications is multidisciplinary care coordination (MCC), which is defined as the active coordination between multiple care providers such as primary care physicians, behavioral health specialists, social workers, and care coordinators.⁶ This approach to healthcare can be effective in addressing multi-level risk factors affecting patients with comorbidities and complex needs.⁷ In one meta-analysis among patients with chronic conditions like cardiovascular disease and diabetes, researchers found that those who had an MCC team had a 19% reduction in avoidable healthcare use compared to those who had a traditional care team.⁸

However, it is currently unknown whether MCC can specifically improve the overall health of those coinfected with HIV and HCV. Therefore, we conducted an interrupted time series analysis to compare healthcare utilization before and after the implementation of a Practice Transformation with enriched MCC. In quantifying these changes, we hope to inform future interventions involving enhanced MCC that aim to improve overall health and reduce health disparities.

Methods:

Study Setting

The Comprehensive Health Program (CHP) is located within an academic urban setting in Upper Manhattan. CHP is funded by Ryan White and provides equity-focused care to those living with HIV and other infectious diseases in New York City. CHP primarily caters to Black, Indigenous, and people of color, men who have sex with men, transgender individuals, people who use drugs, people with severe mental illness, and those who experience housing instability. The program delivers care in both English and Spanish, and more than 100 staff provide HIV antiretroviral treatment, HIV counseling, and on-site social work.⁹

The Intervention

In 2015, CHP was awarded funds from the Delivery System Reform Incentive Payment Program and U.S. Department of Health and Human Services to implement Practice Transformation with the goal of achieving various quality of care outcomes among those living with HIV and coinfected with HCV.¹⁰ Chief among them was the development of Clinical Care Teams (CCT) that used a MCC framework. The CCTs shared a panel of patients and were comprised of primary providers, care coordinators, patient navigators, social workers, and psychiatric providers. CHP was divided into 5 CCTs and each CCT met weekly to review patient panels based on key indicators. To facilitate communication among team members, Practice Transformation entailed implementation of a Health Information Technology dashboard that could be accessed by all members of the team and allowed teams to identify patients who met criteria for customized HIV quality indicators (e.g. HIV viremia, untreated HCV, Lost to Follow-up, Recent Hospitalization, etc.).

Participants

Those living with HIV who accessed care at CHP were eligible for this analysis. We used laboratory results (ribonucleic qualitative and antibody tests) to determine the HIV statuses of our patients. We focused on those coinfected with HIV and HCV for the study. We ascertained yearly HCV status also using laboratory results. If a patient had a detectable HCV RNA or reactive HCV antibody test, they were categorized as coinfected for the given calendar year. This study was reviewed and approved by Institutional Review Board at Columbia University Irving Medical Center (IRB-AAAS1227), and the procedures followed were in accordance with the Helsinki Declaration as revised in 2013. The patients/participants provided their written informed consent to participate in this study.

Statistical Analysis

To contextualize our results, we created descriptive statistics of our study population. We calculated emergency visit rates and inpatient visit rates by first totaling the number of visits and dividing it by the observed person-years; we calculated this for those monoinfected and those coinfected. To assess changes in healthcare utilization among people coinfected with HIV and HCV, we conducted an interrupted time series analysis and compared trends in hospital utilization rates before and after implementation of the intervention. For each year, we calculated the total number of visits among those coinfected cohorts and divided the sum by the number of people. We multiplied this value by 1 000 to get the inpatient hospital and emergency visit rate per 1 000 patients per year. We adjusted this rate by baseline age to get the age-adjusted hospitalization rate. We then computed the rate of change slope for these rates before 2015 (pre-intervention) and after 2015 (post-intervention). For all statistical analyses, we used RStudio (Version 2023.03.1+446).

Results:

We included 2 954 CHP patients living with HIV in our study. Throughout the course of the study period, we identified 392 (13%) ever-coinfected patients. The yearly coinfected caseloads ranged from 84 to 140. Table 1 shows the characteristics and healthcare utilization of the study population. We found that those coinfected tended to be older but were similar on other characteristics. In terms of emergency visits, the coinfected cohort had higher rates (1 401 visits per 1 000 person-years) than the monoinfected (794 per 1 000). The coinfected population also had higher rates of inpatient visits (663 per 1 000) compared to the monoinfected population (255 per 1 000).

Table 1. Characteristics and Healthcare Utilization of the Study Population (n=2 954).

We describe the study population.

Characteristics	N (%)
Ever HIV/HCV Coinfection Status
Yes	392 (13%)
No	2 562 (87%)
Gender
Males	1 999 (68%)
Females	955 (32%)
Self-Reported Race
Black or African American	1 221 (41%)
White or European American	519 (18%)
American Indian or Alaska Native Asian, Native Hawaiian or Other Pacific Islander, Asian/Asian American, or Other Race	690 (23%)
Missing	524 (18%)
Self-Reported Ethnicity
Hispanic	1 167 (40%)
Non-Hispanic	1 243 (42%)
Missing	544 (18%)
Baseline Age
< 30	793 (27%)
30 – 39	526 (18%)
40 – 49	776 (26%)
≥ 50	859 (29%)

HIV: Human Immunodeficiency Virus

HCV: Hepatitis C Virus

During the pre-intervention period, the per-year change in emergency visit rate among the coinfected was positive, increasing about 133 additional visits per 1 000 per year. This decreased to −167 per 1 000 per year during the post-intervention period. Prior to the intervention, the inpatient visit rate was nearly flat (−3 per 1 000 per year) but decreased to −65 per 1 000 per year after CHP implemented Practice Transformation. A figure of the interrupted-time series can be found in the Supplemental Digital Content.

Discussion

Implementation of MCC was associated with a reduction in avoidable healthcare utilization among those coinfected with HIV and HCV. At baseline, those coinfected had 1.7 times the emergency visit rate and more than twice the inpatient visit rate compared to those of the monoinfected. Among those coinfected, the rate of change for emergency and inpatient visits began to decline when MCC was implemented. Broadly speaking, MCC has historically had mixed effectiveness; one meta-analysis that included 16 MCC effectiveness studies found that the intervention reduced hospitalization and emergency visits in four of the studies, while the others had unclear effectiveness results.¹¹ This may be largely due to the diverse contexts of these studies. Additionally, the way in which researchers implemented MCC was diverse; researchers used different implementation strategies, from provider training, patient coaching, and even role-playing. While all of the studies examined MCC, how it was implemented and the context it was implemented in were sundry, which could explain the observed heterogenous effectiveness.

We believe that the intervention was potentially effective because it addressed multiple needs of the coinfected population: the primary care providers administer the medication for HIV and HCV treatment; social workers and psychiatric providers managed substance use and mental health disorders; care coordinators/social workers offered care coordination and addressed SDOH such as housing insecurity or food insecurity, conducting critical outreach in between visits to keep them engaged in care. Furthermore, we believe that the implementation strategies that include the weekly CCT meetings and the shared CCT-focused dashboard facilitated the effectiveness of the intervention in the given context. In the future, we plan to quantify the outcomes of these implementation strategies to identify ways to improve MCC.

There are several limitations in this study. Because our study uses an ecological design, there may be potential confounding with time-varying factors. The composition of those who were coinfected may have changed before and after the implementation of HCV direct-acting antivirals. Additionally, some might consider using yearly HCV coinfection status to be crude, and that we would have more observations and therefore more power had we used monthly data. However, we risk having more misclassification of HCV status if we use monthly data– people who are HCV reactive do not get retested every month after the initial positive diagnosis. Retesting is not recommended for RNA and antibody tests because there may be false positives; a patient will continue to have HCV antibodies even after they have successfully treated it.¹² Therefore, we used a coarser time frame to avoid underestimating HCV prevalence. We hope to revisit this analysis in the future when we have more time points, so we can do a statistical analysis of the coefficients with enough power.

We found that the implementation of MCC was associated with a decrease in per-year hospital visits among those coinfected. To assist in future upscaling, we hope to identify strategies that facilitated the implementation of MCC in a setting like CHP. We believe that interventions like MCC have the potential to improve overall health and reduce health disparities for those with complex health needs.

Implications for Policy and Practice.

• Those coinfected with HIV and HCV have complex and dynamic health needs. They require coordinated care between primary providers, care coordinators, patient navigators, social workers, and psychiatric providers.

• Practice Transformation with MCC was associated with improved health among those coinfected with HIV and HCV in this cohort.

• There is potential to scale up coordinated care in both quantity and quality, and to also implement MCC for other patients with complex health needs. This, however, requires more research on implementation strategies to optimize effectiveness.