An Outpatient Critical Care Transition Clinic Model Reduces Admissions/Readmissions in Medically Complex Patients

, ,
The American Journal of Managed Care, September 2021, Volume 27, Issue 9

Critical care transition clinic patients with chronic conditions had a 31% reduction in relative risk for inpatient admissions, and the clinic reduced cost by more than $1 million.

ABSTRACT

Objectives: Reducing hospital admissions in patients with multiple complex chronic conditions is both a quality indicator and cost-effective to health care systems. This study assesses and compares utilization rates and cost of encounters between patients referred and seen in an outpatient critical care transition clinic (Healthy Transitions Clinic [HTC]) and those referred and not seen.

Study Design: Retrospective cohorts.

Methods: Patients with complex chronic conditions discharged from a tertiary/quaternary acute care hospital or emergency department (March 1, 2015, to February 29, 2016) were referred to an outpatient critical care transition clinic. Comparative cohorts were those patients who attended this transition clinic and those who did not. Pre– and post–HTC referral visits, with health care utilization evaluations including admissions/readmissions, attention to social determinants of health, and cost assessments, were compared among the cohorts.

Results: Insurance coverage differed significantly in its distribution between the groups (χ2 = 22.99; P < .001); therefore, an adjusted relative risk model was used. Inpatient admissions significantly increased, by 31%, in the non-HTC cohort (P = .03); a significant increase in the rate of 30-day readmissions (69%) occurred in the HTC group (P < .001) at 6 months post index admission. Length of stay did not differ pre– and post HTC visit. Although not statistically significant, visits to the HTC reduced median all-cost and HTC cohort cost by more than $1 million.

Conclusions: In patients with complex chronic medical conditions with recent hospital admissions, the HTC model appears to reduce both admissions and encounter costs. Further community/regional studies are needed to better define this observation on a longitudinal basis.

Am J Manag Care. 2021;27(9):e301-e307. https://doi.org/10.37765/ajmc.2021.88742

_____

Takeaway Points

Reducing admissions and readmissions in patients with complex chronic conditions indicates quality and cost-effective care. Patients with complex chronic conditions discharged from an acute tertiary/quaternary care hospital or emergency department were referred to an outpatient critical care transition clinic. Pre-/postclinic referral visits, readmission rates, and cost assessments were compared. Utilizing an adjusted relative risk model, inpatient admissions were significantly lower, by 31%, in the clinic cohort. Thirty-day readmissions did show a significant increase among the clinic cohort. Commercial and Medicaid insurance had a significant effect on rates of inpatient admission. Visits to the clinic reduced median cost by more than $1 million.

  • The Healthy Transitions Clinic model reduces inpatient admissions and encounter costs.
  • Clinics should consider financial limitations upon patients referred to transition services.
  • Clinic administrators should customize transition care to location-specific variables.

_____

Chronic health conditions lead to frequent utilization of health care services.1 The fundamental goals of the Ascension St. Vincent Medical Group Healthy Transitions Clinic (HTC) are to help patients manage their chronic diseases effectively and reduce encounters at emergency departments (EDs) or admissions. Transitional care services reduce ED encounters and inpatient stays.2 The benefits to patients, hospital systems, and payers from reduced hospital encounters include improved health, disease management, and financial savings. As the US health care system changes from fee for service to value based, efficiently managing medically complex chronic conditions is increasingly important.

Health care cost reduction is a major focus of the Affordable Care Act (ACA). Reports suggest that readmissions cost between $12 billion and $44 billion per year.3 Thus, a 10% reduction could save at least $1.2 billion. Avoidable hospital readmissions suggest inadequate transitional care.3 Readmissions are linked to misunderstood discharge instructions and lack of help for patients transitioning from hospital to home.4,5 To reduce the cost of readmission, reports suggest timely follow-up care for high-risk patients.6-8

Hospitals have seen progress in reducing readmissions using interventional strategies. Strategies identify at-risk patients, strengthen the discharge process, schedule follow-up appointments, improve communication with community providers, involve nurse-directed follow-up calls, provide necessary medications/resources, and enhance patient education.2,3,9 Slow progress and widely variable readmission rates exist because of social determinants of health (SDOH),10 the availability of local health care resources, and the socioeconomic status of the community.7,9 Critical care transition programs (CCTPs) are one strategy to reduce patient admissions and readmissions. Rosa et al11 have reported that “although heterogeneously reported in the literature, CCTPs are often focused on follow-up by health care professionals to support the primary hospital [inpatient] ward and prompt care in case of deterioration of patient’s condition.…”

Ascension St. Vincent Hospital – Indianapolis (SVH) is an 822-bed tertiary/quaternary teaching institution. SVH is a medical campus that includes primary care providers (PCPs), outpatient and specialty clinics, and other health care facilities. The HTC is a separate outpatient CCTP clinic, proximal to the hospital, designed to lower unplanned readmissions among referred medically complex hospital-discharged patients. Hospitalized or ED patients, often with 2 or more chronic conditions, were referred to the HTC for prompt access to critical care medical providers within 3 business days of discharge based on a referral algorithm and criteria (eAppendix Figure 1 [eAppendix available at ajmc.com]) that include (but are not limited to) patients without or with delayed (> 7 days) access to a PCP, with frequent ED visits or hospitalizations over the past year, with complex chronic conditions, or who are prescribed more than 6 medications. Occasionally, the hospital case managers could deviate from the referral algorithm if they deemed the patient high risk. The clinic provider worked with the patient’s PCP and specialist provider(s) to coordinate follow-up care, with the HTC often overseeing follow-up visits at the request of the patient’s PCP or if a patient did not have a PCP. Table 1 lists available services at the HTC. The HTC combined outside methodologies12-14 and developed a unique and specialized approach to reduce hospital readmissions. The HTC clinic is outpatient but accepts referrals involving identified high-risk patients visiting the ED or recently discharged inpatients. These recently discharged patients include individuals from critical care units. The clinic is staffed by experienced critical care physicians and advanced practice nurses with at least 5 years of critical care or ED experience. Uniquely, compared with published reports, the HTC may see patients on a frequent/daily basis (eg, frequent infusions or other therapies/procedures).

Grover and Joshi completed a systematic literature review in 2015 describing the 6 elements of the chronic care model, which include a health system or a health organization; clinical information systems; decision support; delivery system design; self-management support; and community, including organizations and resources for patients.15 The HTC incorporates all 6 of these elements with our HTC clinic model being a part of a large integrated health care system. The clinical information system includes an integrated inpatient-outpatient health information system with remote care monitoring. This information is reviewed by the providers in the HTC and forms the basis for decision-making processes with patients based on accepted clinical practice guidelines. The HTC provides patients with extensive information and education and with monitoring devices when appropriate, such as blood pressure cuffs, pulse oximeters, scales, and glucose monitors. The HTC is integrated into our parent health care system, which includes remote care monitoring, home health care, durable medical equipment, and various therapies (eg, physical therapy, occupational therapy, speech therapy), which give regular written and oral input and assist in the monitoring of each patient’s clinical course. Each provider engages in patient education and in helping to improve patients’ self-management skills. The HTC interacts with PCPs and specialty clinicians as a team when treating patients. Depending on patients’ sites of primary care, community resources assist in fulfilling various needs, based on an individual’s SDOH assessment.

The HTC functions as a hybrid of the chronic care model and the transitional care model. In the latter, patients are assessed for interplays among disease severity, medical care needs, and SDOH.14 Case managers classify patients deemed at high risk for 30-day hospital readmission prior to discharge and refer them to specialized clinics. The HTC operates in concert with the hospital to transition patients to providers who manage chronic diseases.

The objectives of this study were to determine and describe the differences in outcomes between referred populations who utilized the HTC and those who did not. We sought to understand if HTC utilization resulted in fewer inpatient admissions and 30-day readmissions and affected future length of stay (LOS). A secondary objective was to conduct a cost analysis to describe the effect that HTC services had on resource utilization in the SVH health care system.

METHODS

Patient Selection

We retrospectively reviewed data from 787 referred patients (SVH institutional review board approval: R2017-0008). Data included patients referred from SVH during the time period of March 1, 2015, to February 29, 2016. The 3 pathways of referral to the HTC are (1) admitted to the hospital for observation after an ED encounter, (2) treated and released for follow-up after an ED encounter, or (3) admitted directly for an inpatient stay (eAppendix Figure 1).

SVH Case Management, which coordinates hospital discharges and manages posthospitalization appointments, utilized the algorithm (eAppendix Figure 1) to identify patients who could benefit from a referral to the HTC. We used the HTC referral appointment date as the index date for the 2 cohorts in this study: (1) patients who attended their HTC appointments (HTC) and (2) patients who did not attend these appointments (non-HTC). For HTC patients, if a patient rescheduled their appointment, this rescheduled appointment was considered the index date.

Data Acquisition

Data were retrospectively collected 6 months prior to and after the index date (eAppendix Figure 2). Number of hospital admissions and LOS were compared from the preindex period to a similar postindex period following utilization/nonutilization of the HTC.

The HTC’s referral log was used to search the electronic health record for inpatient and financial data. Regenstrief Institute (RI), a health care research organization, provided statewide hospital utilization data from the Indiana Health Information Exchange. These data were obtained to examine the effect the HTC has on the health care network within Indiana, not solely SVH. One health care institution in the Indianapolis metro area declined to release deidentified data and to participate (9% of study referrals).

Patients were excluded from analyses if discharge or care setting data (dates, location) were unavailable or if encounter care setting or source were not inpatient (eAppendix Figure 3). Inpatient stays overlapping the patient’s index referral date were included in the data set as a preindex encounter.

Inpatient encounters were categorized into inpatient admissions and 30-day readmissions and examined in the pre- and postindex time periods. An inpatient stay included any ED encounter that resulted in an admission to any hospital agreeing to share data with RI (for study purposes). A 30-day readmission included inpatient stays that occurred within 30 days of the previous discharge from a reference hospital. Inpatient stays did not include readmissions. Although all inpatient encounters were examined together, inpatient admissions and 30-day readmissions were also examined separately. These designations were made to examine the utilization of inpatient services and the frequency of 30-day readmissions between the cohorts. ED visits were not examined in this study due to the inability to capture these visits in the data among regional hospitals.

Analysis

Inpatient LOS was evaluated between the HTC and non-HTC population cohorts. Frequency and rate of hospital admission were compared among both cohorts. Two-way Fisher exact tests were completed to see if any differences emerged in the distribution of inpatient stays. The relative risk (RR) and 95% CI for inpatient admissions were calculated. An adjusted RR (ARR) model was used to account for significant covariates for demographic variables found to be statistically different between groups. Variables in the ARR model included patient type (HTC vs non-HTC), visit type (inpatient admission vs 30-day readmission), and insurance type/status at the encounter. Inpatient encounter or utilization data were compared using ARR for 1-month and 6-month intervals surrounding the index date. These time periods were chosen to explore acute and long-term effects. Additionally, univariate subanalysis was conducted among the study cohorts’ super-users. Super-users have multiple ED visits and/or are frequently admitted to the hospital.1 The definition of super-user (or “super-utilizer”) in the literature is heterogeneous, with differing definitions used by researchers and CMS. CMS defines super-users as “patients who accumulate large numbers of [ED] visits and hospital admissions which might have been prevented by relatively inexpensive early interventions and primary care.”16 To examine the HTC effect on super-users, we utilized the definition of 2 or more readmissions within a 30-day period at any point during the study period.

The cost for inpatient encounters and the median cost per encounter were evaluated. The median cost per encounter was analyzed comparing cost savings of the HTC based on the reduction in the number of encounters observed.

Because costs at other hospitals are not provided to RI, total costs evaluated are representative only of SVH. Results of this total cost calculation between HTC and non-HTC patients were assumed to be the same to make relative comparisons.

Data aggregation was performed in R version 3.5 (R Foundation for Statistical Computing). Statistical analyses were performed using SPSS version 24 (IBM). Statistical significance was set with α = .05, and 95% CIs were used for RR and ARR models.

RESULTS

Demographics

The study included 283 non-HTC patients (36%) and 504 HTC patients (64%). No difference in age (P = .69) existed between non-HTC (median [interquartile range (IQR)] age, 53 [29] years) and HTC (54 [26] years) cohorts. A statistical difference in the distribution of insurance types (P = .005) was found. Of non-HTC patients, 41.0% had Medicare, 29.3% had commercial insurance, 22.3% had Medicaid, and the remaining 7.4% were self-pay or charity. For HTC patients, 41.3% had commercial insurance, 34.3% utilized Medicare, 20.0% had Medicaid, and 4.4% were self-pay or charity. This difference could have an effect on whether referred patients attend HTC. Inpatient utilization was compared using the associated insurance with each encounter. A significant difference was found among the insurance types between the groups (χ2 = 22.99; P < .001). Of non-HTC encounters, 45.6% had Medicare, 29.8% had commercial insurance, 23.0% had Medicaid, and the remaining 1.6% were self-pay or charity. For HTC encounters, 20.3% had commercial insurance, 47.4% utilized Medicare, 28.6% had Medicaid, and 3.7% were self-pay or charity. These rate changes reflect the possibility that patient insurance could change from one encounter to another and some patients would be represented more often. Because we observed a statistical difference in insurance type on a patient encounter level, an ARR model was used to account for these differences. Because insurance status could have an effect on LOS, a univariate generalized linear model with 2-factor analysis of variance was used to compare LOS between the HTC and non-HTC groups.

Inpatient Encounters: 6 Months Pre– and Post Index Date

Non-HTC LOS did not differ significantly from the pre- to the postindex period (F = 1.03; P = .38). A similar result was observed among HTC patients (F = 1.57; P = .19). No statistically significant differences were found in preindex (F = 0.88; P = .45) or postindex (F = 1.70; P = .17) LOS when comparing the cohorts (Table 2).

The distribution of inpatient admissions, 30-day inpatient readmissions, and all admissions were explored for HTC and non-HTC patients (Table 3). A statistical difference was found in the distribution of inpatient admissions and all admissions between HTC and non-HTC patients (P = .001) from the pre- to post-HTC periods. No statistical difference was found (P = .17) in the distribution of 30-day inpatient readmissions from pre-HTC and post-HTC. The ARR model used HTC patients with inpatient admissions and Medicaid as the reference intercept exploring for statistical differences between the study cohorts (Table 4). This intercept combination was statistically significant for admission (P < .001), 30-day readmissions (P < .001), having commercial insurance (P < .001), and being a non-HTC patient (P = .03). Utilizing the ARR, non-HTC patients had a 31% increase in inpatient admissions compared with HTC patients (ARR, 1.31; 95% CI, 1.02-1.66). Insurance did have a statistically significant effect on inpatient admissions if the HTC patient had Medicaid or commercial insurance. Compared with Medicaid, patients with commercial insurance had a 52% decreased chance of readmission (ARR, 0.48; 95% CI, 0.33-0.69). The model also revealed a 69% increase in the rate of 30-day readmissions among HTC patients (ARR, 1.69; 95% CI, 1.02-1.66).

Inpatient Encounters: 1 Month Pre– and Post Index Date

Non-HTC patients’ median (IQR) LOS was 4 (3-6) days preindex and 3 (2-6) days post index when evaluating 30 days pre- and post index (F = 0.45; P = .64) (Table 2). HTC patients had a median (IQR) LOS of 4 (2-6.25) days and 3 (2-6) days during the pre- and postindex periods, respectively (nonsignificant; F = 0.71; P = .55). When comparing the LOS among non-HTC and HTC patients during the 30-day preindex or 30-day postindex periods, no significant differences were found (pre-: F = 0.98; P = .40; post: F = 0.19; P = .83).

Similar distribution rates of admissions between HTC and non-HTC patients in the 1-month period relative to the index date were found for all admissions, inpatient admissions, and 30-day readmissions (Table 3). The ARR model for the 1-month period revealed statistically significant patient admissions among the reference intercept (HTC inpatient admissions with Medicaid) (P < .001), commercial insurance (P = .001), and non-HTC patients. Non-HTC patients had a 95% increase in inpatient admissions compared with HTC patients (ARR, 1.95; 95% CI, 1.33-2.88) (Table 4). Results from the 1-month study period were similar to those from the 6-month analysis in regard to insurance. Medicaid (P < .001) and commercial (P = .001) insurance were statistically significant for inpatient admission in the model. Similarly, HTC patients with commercial insurance had a 61% decrease in the rate of inpatient admission (ARR, 0.39; 95% CI, 0.22-0.69). Readmissions were not statistically significant when considered in the first month post index date for HTC patients (P = .92).

Super-users

Super-users are frequent users of the health care system.1 Seventeen non-HTC patients and 14 HTC patients were super-users. We found no effect of the HTC on decreasing super-user utilization. Super-users averaged 5 admissions per patient for both the HTC and non-HTC groups during the study period. Non–super-users averaged 1 admission per patient for both groups. Of these super-user admissions, 91.9% of the HTC and 82.1% of non-HTC super-users had either Medicaid or Medicare insurance.

Aggregated Total Costs per Patient Encounter

Total cost for encounters during pre- and postindex periods are presented in Table 5. Median costs of preindex encounters were higher for HTC patients but not significantly different. Pre- to postindex median cost savings occurred for HTC patients (–$1326) but not for non-HTC patients (+$211); pre- and postindex differences were not statistically different. Median costs of encounters during the 30 days preindex were higher for HTC patients but not significantly different. Thirty-day pre- to postindex median cost savings were more than 10 times greater for HTC patients (–$2839) than non-HTC patients (–$227) (HTC total median cost savings, $2612); differences were not statistically different due to cost variability.

Super-users saw benefit from using the HTC, as median costs declined by approximately $3000 per encounter compared with an increase in costs for non-HTC users when comparing pre- and post index (data not statistically significant). When comparing 30 days pre- and post index, the HTC kept costs stable; non-HTC patients saw median cost increase by $2600 per encounter.

DISCUSSION

The main goals of the HTC are to help high-risk patients better manage their chronic diseases and improve the cost-effectiveness of health care services. Common conditions included congestive heart failure, chronic obstructive pulmonary disease, high blood pressure, diabetes mellitus and complications, acute myocardial infarction, substance abuse, and renal and liver disease. Identification of other patient characteristics that may have differentiated users of the HTC, or those most assisted by it, was not included in the design of the study. This study demonstrates a reduction of HTC patients’ utilization of inpatient services and associated costs. The study revealed a statistically significantly lower rate of hospital inpatient visits among HTC patients compared with non-HTC patients (Tables 3 and 4). Looking at the 30-day postindex period, our data revealed a median per-patient encounter cost reduction of $2612 among HTC patients compared with non-HTC patients. The 6-month postindex time period showed a cost reduction of $1537 among HTC patients compared with non-HTC patients. Applying the 30-day postindex period cost savings to the 409 non-HTC inpatient encounters results in potential savings of approximately $1,070,000. These savings do not include any cost reduction of ancillary services (eg, imaging and laboratory services). In terms of costs to the HTC clinic, patient data for this study were pulled from the HTC’s initial year of operation (2015-2016), and transitional care management (TCM) billing codes were not initially used. The HTC clinic has grown in the number of patients seen on an annual basis and in revenue. In 2019, the HTC treated 1360 patients and used TCM billing codes. This generated a 50% increase in revenue per TCM patient. No significant increases have occurred in the costs to operate the HTC.

A February 2019 statistical brief notes that the mean cost of index admissions and 30-day all-cause readmissions were $12,500 and $14,400, respectively (CMS data from 2010 to 2016).17 Because our study did not separate cost information between index patient admissions and 30-day readmissions, the data from CMS suggest that our cost savings may, in fact, be greater.

Study data suggest that the HTC achieved its goal of lowering inpatient admissions among high-risk patients. However, this decrease in utilization among HTC patients was limited to our classification of inpatient visits, as HTC patients had a 69% increase in the ARR of 30-day readmissions (Table 4). Additionally, super-user patients attending the HTC did not see a lowered number of admissions compared with non-HTC patients. We semiquantitatively reviewed super-user differences among non-HTC and HTC patients (data not shown) and found that one major difference was the frequency of patients with polysubstance/alcohol misuse disorders (6 of 17 in non-HTC vs 1 of 14 in HTC). It appears that super-users will use health care services at high frequency1 and the current health system infrastructure does not yet effectively address these patients to manage their health conditions and reduce inpatient encounters.

Literature reports have discussed the cost benefits of various interventions aimed at reducing readmissions. Telephone support for Medicare patients during transition to home reduced ED encounters, 30-day readmissions, and physician office encounters by more than 1, 2, and 4 percentage points, respectively.3 Overall savings per person were nearly $300. Similar findings in a chronic disease population illustrated that each prevented readmission saved an estimated $15,000 (adjusted to 2016 US$).18 Another report suggests the cost of preventable readmission to be closer to $33,000 (2016 US$).9 Our data support a major 30-day reduction of inpatient admissions in HTC patients compared with non-HTC patients and a drop in per-patient encounter cost of approximately $2600. The methodologies and admission types among our study and these cited investigations differ; however, the trend shows that these types of transitional programs are clinically and financially effective.

The cost-effectiveness of the HTC program goes beyond what we measured. Several other factors exist because the HTC is part of SVH, a large integrated health care system. Initial patient appointments lasted a mean of 75 minutes, of which at least 40 minutes were spent on education about the various medical issues, assessments of the individual’s resources, and attendant SDOH, a significant factor in the success of the HTC. On a health care system basis, this clinic design permits more time for appointments for the individual’s PCP to see other patients. Another factor is that the SVH health care system is geographically diverse, with offices located in multiple locations throughout the state. We found that this consistent process, which focused on education with patients and family, achieved clinically improved quality of care and patient satisfaction compared with the non-HTC cohort. Although likely substantial, the resultant total financial impact to the system was unable to be accurately measured.

Key drivers of increased 30-day readmissions are important for health care providers and administrators to understand because of the implications the ACA has on penalties for 30-day unplanned readmissions. Our data showed that inpatient admissions decreased after patients visited the HTC. The cohort differences in insurance coverage may imply that socioeconomic factors and other SDOH potentially affected our results. Although not statistically significant with the univariate models, the ARRs did reveal a significant increase in the rate of 30-day readmissions among HTC patients with Medicaid coverage. Potential disease severity differences among our patients may explain why decreases in inpatient admissions were found but an increase in 30-day readmissions was revealed in the adjusted model. This study did not account for disease severity, and this variable should be considered in the future. After adjusting for insurance coverage differences, we still noticed a statistical impact of HTC services on utilization. Availability of local health care resources and the overall socioeconomic status of the community influence readmissions.9 Social and environmental factors account for 20% of the risk of premature death,19 indicating that comprehensive and multifaceted approaches are necessary to improve readmission rates. Predictive tools that can identify patients at high risk for readmission could greatly improve an integrated health care system’s ability to provide appropriate services and reduce readmissions. This study was focused on utilization outcomes among patients choosing to visit the HTC and those electing not to attend HTC appointments. Because the chronic care model and the transitional care model were used as blueprints in the design of the HTC, future research will compare outcomes of HTC patients with those of patients in these and other critical care transition models.

Limitations

A possible confounder is that the population who do not use the HTC may inherently utilize health care services differently (at a lower rate) than those who do, especially taking into account the effect of insurance status. Our data, however, suggest that non-HTC patients utilized services at a similar rate to HTC patients prior to the index date and had similar LOS, both pre– and post index date. This indicates that the HTC was able to reduce health care costs and affect patients who are willing to engage in services. Another limitation to our study is the lack of race/ethnicity demographics for our cohorts. The authors recognize that the addition of this SDOH variable would have strengthened the confidence of the differences found between the cohorts.

Patients may have died or relocated during the study period, which would appear as nonutilization of services post index. Data that we obtained did not include information related to patient death or relocation, thus estimates would be speculation. Additionally, patients were not randomly assigned to study cohort groups. Instead, this study followed the design of a natural retrospective experiment in which the patient’s choice in attending HTC appointments was their own.

Despite the study’s limitations, we conducted post hoc power analysis and adjusted for a number of variables. Focusing solely on the comparison of inpatient admissions and 30-day readmissions among the HTC and non-HTC patients at 6 months, the calculated power in our study, using Medicaid patients as the index, was 78.1% for Medicare patients, 75.2% for commercial insurance patients, and 76.1% for self-pay patients. These power calculations lend even more confidence to our results.

CONCLUSIONS

The HTC decreased posthospital inpatient admissions and costs in a large group of patients with medically complex chronic conditions. We suggest that other health care systems could benefit from setting up similar programs to help these patients manage their medically complex chronic conditions and reduce health care costs nationwide.

Acknowledgments

The authors wish to acknowledge the following individuals in their assistance in the development or completion of this project: Tammy Johnson, Melanie Holt-Fauth, Merrilee Freese, Victoria Andrews, Denise Stachke, Denise Bradford, Sarah Hoover, Anna Roberts, Adam Locke, Maxwell Aiken, and Jessica Leaman. Special thanks are extended to Cynthia Adams and Bruce Bethancourt for thoughtful review of the manuscript, and to Melissa Merrill and Stacey Agee for their clinical work and dedication.

Author Affiliations: Indiana Hemophilia & Thrombosis Center (IAJ), Indianapolis, IN; Office of Research & Clinical Trials (TLF) and Healthy Transitions Clinic (MRG), Ascension St. Vincent Hospital – Indianapolis, Indianapolis, IN.

Source of Funding: None.

Author Disclosures: The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (IAJ, TLF, MRG); acquisition of data (TLF, MRG); analysis and interpretation of data (IAJ, TLF, MRG); drafting of the manuscript (IAJ, TLF, MRG); critical revision of the manuscript for important intellectual content (IAJ, TLF, MRG); statistical analysis (IAJ, TLF); provision of patients or study materials (TLF, MRG); administrative, technical, or logistic support (TLF); and supervision (TLF, MRG).

Address Correspondence to: Todd L. Foster, PhD, Ascension St. Vincent Hospital – Indianapolis, 2001 W 86th St, Indianapolis, IN 46260. Email: todd.foster@ascension.org.

REFERENCES

1. Harris LJ, Graetz I, Podila PSB, Wan J, Waters TM, Bailey JE. Characteristics of hospital and emergency care super-utilizers with multiple chronic conditions. J Emerg Med. 2016;50(4):e203-e214. doi:10.1016/j.jemermed.2015.09.002

2. Takahashi PY, Naessens JM, Peterson SM, et al. Short-term and long-term effectiveness of a post-hospital care transitions program in an older, medically complex population. Healthc (Amst). 2016;4(1):30-35. doi:10.1016/j.hjdsi.2015.06.006

3. Costantino ME, Frey B, Hall B, Painter P. The influence of a postdischarge intervention on reducing hospital readmissions in a Medicare population. Popul Health Manag. 2013;16(5):310-316. doi:10.1089/pop.2012.0084

4. Kripalani S, Jackson AT, Schnipper JL, Coleman EA. Promoting effective transitions of care at hospital discharge: a review of key issues for hospitalists. J Hosp Med. 2007;2(5):314-323. doi:10.1002/jhm.228

5. Chugh A, Williams MV, Grigsby J, Coleman EA. Better transitions: improving comprehension of discharge instructions. Front Health Serv Manage. 2009;25(3):11-32.

6. Jackson C, Shahsahebi M, Wedlake T, DuBard CA. Timeliness of outpatient follow-up: an evidence-based approach for planning after hospital discharge. Ann Fam Med. 2015;13(2):115-122. doi:10.1370/afm.1753

7. Kripalani S, Theobald CN, Anctil B, Vasilevskis EE. Reducing hospital readmission rates: current strategies and future directions. Annu Rev Med. 2014;65:471-485. doi:10.1146/annurev-med-022613-090415

8. Pourat N, Chen X, Wu SH, Davis AC. Timely outpatient follow-up is associated with fewer hospital readmissions among patients with behavioral health conditions. J Am Board Fam Med. 2019;32(3):353-361. doi:10.3122/jabfm.2019.03.180244

9. Danner M. Taking aim at reducing hospital readmission rates.Lexington Insurance. February 2016. Accessed October 16, 2019. https://www.lexingtoninsurance.com/content/dam/aig/america-canada/us/documents/brochure/taking-aim-at-reducing-hospital-readmission-rates.pdf

10. Byyny RL. Social determinants of health: reforming education and public health to improve health in the United States. The Pharos. Autumn 2017. Accessed September 25, 2019. https://alphaomegaalpha.org/pharos/2017/Autumn/2017-4-Byyny.pdf

11. Rosa RG, Maccari JG, Cremonese RV, et al. The impact of critical care transition programs on outcomes after intensive care unit (ICU) discharge: can we get there from here? J Thorac Dis. 2016;8(7):1374-1376. doi:10.21037/jtd.2016.05.30

12. Glasgow RE, Orleans CT, Wagner EH. Does the chronic care model serve also as a template for improving prevention? Milbank Q. 2001;79(4):579-612, iv-v. doi:10.1111/1468-0009.00222

13. Coleman EA. Falling through the cracks: challenges and opportunities for improving transitional care for persons with continuous complex care needs. J Am Geriatr Soc. 2003;51(4):549-555. doi:10.1046/j.1532-5415.2003.51185.x

14. Roper KL, Ballard J, Rankin W, Cardarelli R. Systematic review of ambulatory transitional care management (TCM) visits on hospital 30-day readmission rates. Am J Med Qual. 2017;32(1):19-26. doi:10.1177/1062860615615426

15. Grover A, Joshi A. An overview of chronic disease models: a systematic literature review. Glob J Health Sci. 2015;7(2):210-227. doi:10.5539/gjhs.v7n2p210

16. Johnson TL, Rinehart DJ, Durfee J, et al. For many patients who use large amounts of health care services, the need is intense yet temporary. Health Aff (Millwood). 2015;34(8):1312-1319. doi:10.1377/hlthaff.2014.1186

17. Characteristics of 30-day all-cause hospital readmissions, 2010-2016. Healthcare Cost and Utilization Project statistical brief No. 248. February 2019. Accessed August 12, 2019. https://hcup-us.ahrq.gov/reports/statbriefs/sb248-Hospital-Readmissions-2010-2016.jsp

18. Harrison PL, Hara PA, Pope JE, Young MC, Rula EY. The impact of postdischarge telephonic follow-up on hospital readmissions. Popul Health Manag. 2011;14(1):27-32. doi:10.1089/pop.2009.0076

19. Schroeder SA. We can do better—improving the health of the American people. N Engl J Med. 2007;357(12):1221-1228. doi:10.1056/NEJMsa073350