PCORI Integrated Care (IC) Models for Patient-Centered Outcomes (IC)

Leveraging Integrated Models of Care to Improve Patient-Centered Outcomes for Publicly-Insured Adults With Complex Health Care Needs

Multiple chronic conditions (MCC) are widely recognized as the U.S. public health challenge of the 21st century. These physical and behavioral health conditions take a large toll on those suffering from the diseases, including many who are publicly insured, as well as caregivers and society. While evidence-based integrated care models can improve outcomes for individuals with MCC, such models have not yet been widely implemented. Insurance providers/payers have innovative system features that can be used to deploy these models; however, the investigators do not yet know which of these features can best help to improve outcomes for individuals with MCC in general or high-need subgroups in particular. As a result, patients lack information to make important decisions about their health and health care, and system-level decision makers face ongoing challenges in effectively and efficiently supporting those with MCC. This real-world study will provide useful information about available options for supporting individuals with MCC. Building on existing integrated care efforts, the investigators will enroll N=1,927 (N=265 Phase I and N=1,662 Phase II) adults with MCC at risk for repeated hospitalizations and assess the impact of three payer-led options (e.g. High-Touch, High-Tech, Usual Care) on patient-centered outcomes, namely patient activation in health care, health status, and subsequent re-hospitalization. The investigators will also determine which option works best for whom under what circumstances by gathering information directly from individuals with MCC through self-report questionnaires, health care use data, and interviews.

Study aims. Given the documented need for valuable information about system-level features that can be used to effectively and efficiently support adults in living well with MCC, this study is designed to achieve the following aims: Aim 1: Compare the effectiveness of High-Touch, High-Tech, and Usual Care on primary outcomes including hospital readmission, health status, and patient activation, and on several secondary outcomes including functional status, quality of life, care satisfaction, emergent care use, engagement in primary, specialty, and mental health care, and gaps in care. Aim 2: Examine the differential effects of the interventions for patient subgroups, based on age, race, illness complexity, and comorbid behavioral health conditions to evaluate heterogeneity of treatment effects (HTE) and determine for whom and in what circumstances the interventions are most effective. Aim 3: Examine perceived barriers and facilitators to efficient and effective implementation of High-Touch and High-Tech interventions for delivering evidence-based integrated care. An individual-level randomized design along with a pragmatic, mixed-methods approach to compare system-level features for delivering evidence-based components of integrated care for Medicaid or dual-eligible adult members with MCC who reside in in Western, Central, or Eastern PA and are at high risk for rehospitalization has been selected for this study. This design, based on significant input from patient stakeholders and Drs. Kevin Kraemer (Scientific Co-I; health services researcher) and Doug Landsittel (Co-I; biostatistician/CER expert), accords fully with the PCORI Methodology Standards. Intervention effectiveness will be determined by examining the differential impact on outcomes that are most meaningful to patients in our target population and those delivering their care. The scope and duration of the study interventions and evaluation are sufficient to measure change in patient-centered outcomes. High-Touch, High-Tech, and Usual Care will serve as the comparators for this study. Usual Care follows standardized procedures for patient engagement including when a patient is either hospitalized or transitioning from the hospital setting into ambulatory care for follow-up and condition management. Due to resources and other limitations, not all patients who are eligible for High-Touch/High-Tech enroll in these programs. Thus, the addition of the Usual Care arm will allow for a less intensive model to be examined and targeted to appropriate patient populations. For Aims 1 and 2, an individual, stratified randomized trial design was selected to randomly assign each enrollee to one of the three interventions arms, minimizing and balancing for confounding variables. Individual-level randomization was selected as opposed to cluster randomization at a system level (e.g. practice-, hospital-level) because the interventions are delivered by a single payer and are not subject to within-practice contamination. Based on valuable system-level stakeholder feedback, an unequal randomization ratio of 2:2:1 for High-Touch, High-Tech, and Usual Care, respectively, was utilized. While the less resource intensive Usual Care may, in fact, improve meaningful outcomes for certain patient subgroups, the health care system has invested heavily in High-Touch and High-Tech as evidence-based solutions for chronic disease care. Additionally, stakeholders have indicated that they would like as many participants as possible to have a fully integrated care experience offered by High-Touch/High-Tech and would like to limit enrollment into Usual Care. The investigators will use a mixed-methods approach that incorporates both qualitative and quantitative data. The addition of qualitative data collection and analyses in Aim 3 will permit more comprehensive understanding of patient and staff experiences with the interventions and results will aide in dissemination of study findings in a manner that is most consistent with patient and other stakeholder perceptions and experiences. The overall, four-year study timeline includes three phases: Pre-Intervention (months 1-6), Intervention and Data Collection (months 7-40), and Data Analysis and Reporting (months 41-48). The study population includes Medicaid or dual-eligible (Medicare-Medicaid) adults age 21 years and older with MCC, including at least one physical health condition (e.g., cardiovascular disease, hypertension, COPD, diabetes) and at least one additional physical or behavioral health condition (e.g., depression, serious mental illness, substance abuse disorder) and at least one hospital discharge in the previous 30 days. These individuals will reside in PA and will be insured through physical and/or behavioral health payers within the UPMC Insurance Services Division (ISD). In addition, these individuals will have several comorbidities, will have been prescribed several medications, and/or will be predicted future high health care utilizers. Assuming a 75% enrollment rate, 1,662 individuals will be randomized to either High-Tech (n=667), High-Touch (n=667) or Usual Care (n=328). The study will use web-based randomization to one of the three interventions for those individuals who consent to participate in the study. Once a member of the Community Team (CT), multidisciplinary community-based team of nurses, licensed social workers, and licensed professional counselors, determines eligibility, CT personnel will enter key identification information, and the system will then generate a Study ID (numeric identification number) along with assignment to an intervention arm. Randomization will be stratified by gender, type of insurance (Medicaid or Medicare-Medicaid), and technology/digital literacy, which will be assessed at time of enrollment and before randomization, to ensure that intervention arms are balanced with respect to these important variables. Within each stratum, random block sizes of 5 and 10 will be used to maximize balance between intervention groups while minimizing the ability to unmask investigators to the next treatment assignment, triggering an automated alert to CT staff regarding which intervention to implement for each participant and documented accordingly in HealthPlaNET, UPMC ISD's integrated health management software program. If a participant is unwilling to be randomized, they will be excluded from the study. Each patient is assigned a care manager (CM) who provides comprehensive services for the duration of intervention implementation. Bilingual CMs will be available to support native Spanish speaking participants. CMs are currently employed to develop and implement care plans with patients, coordinate healthcare services, work with the pharmacist to manage patient's medications, make home visits, and deliver telehealth care and remote monitoring. Patients in both High-Touch and High-Tech will experience similar procedures at the start of their participation. A CM engages patients in a face-to-face assessment in-home or telephonically to dialogue about the social determinants affecting continued hospital readmissions and emergency department use. At the completion of the assessment, the study is presented to the member and if agreeable informed consent occurs. Individuals randomized to Usual Care will be provided with the transitional care services. High-Touch and High-Tech interventions are provided for approximately four months following hospitalization, and Usual Care participants are transitioned to Health Plan telephonic care management with in 14 days.

Diabetes, Arthritis, Asthma, Chronic Obstructive Pulmonary Disease, Hypertension, Anxiety, Atrial Fibrillation, Congestive Heart Failure, Depression, Bipolar Disorder, Schizophrenia

We chose an individual, stratified randomized trial design to randomly assign each enrollee to one of the three interventions arms, minimizing and balancing confounding variables. We will utilize an unequal randomization ratio of 2:2:1 for High-Touch, High-Tech, and Usual Care, respectively. While the less resource intensive Usual Care may, in fact, improve meaningful outcomes for certain patient subgroups, our health care system has invested heavily in High-Touch and High-Tech as evidence-based solutions for chronic disease care. We will use a mixed-methods approach that incorporates both qualitative and quantitative data.

Delivered primarily via face-to-face interactions, with telephonic interactions and information sharing that does not require access to mobile devices or the Internet. In-person support and/or telephonic interactions to occur at least four times over at least a four-month period.

Delivered via a remote care management platform and digital health tools. Remote care support interactions to occur for at least a four-month period.

Delivered via Health Plan support and resources within 14 days of an initial home or telephonic visit.

Patient activation will be assessed using the Patient Activation Measure (PAM), a 13-item scale that gauges individual knowledge, skills, and confidence essential to managing one's own health. The PAM identifies where an individual falls within four different levels of activation. Level 1 - Disengaged and overwhelmed Level 2 - Becoming aware, but still struggling Level 3 - Taking action Level 4 - Maintaining behaviors and pushing further

Health Status will be measured using the 36-Item Short Form Survey (SF-36). The SF-36 is a set of 36 health status and quality-of-life measures that are patient self-reported and measure functional health and well-being within eight domains, including physical functioning, physical ability to complete tasks, bodily pain, general health, vitality, social functioning, emotional ability to complete tasks, and mental health.. Patients are asked to reply to questions based on how they have felt over the previous week. There are eight scaled scores and each scale is translated into a 0-100 scale with each question carrying equal weight. Lower scores represent more disability and higher scores less disability. The items use Likert-type scales, some with 5 or 6 points and others with 2 or 3 points.

Rehospitalization will be measured using an all-cause readmission rate existing in UPMC ISD inpatient claims for physical and behavioral health service use within 30 and 90 days following the index inpatient admission that triggered enrollment in the study.

Functional status will be measured using the PROMIS Physical Function - Short Form 6b which is a brief self-report tool that consists of 6 self-reported physical function measures that assess current function including the functioning of upper extremities, lower extremities, central regions, as well as activities of daily living. Each question has five response options (a 5 point Likert scale) ranging in value from one to five. The instrument is scored using item-level calibrations using responses to each item for each participant. A single Physical Function capability score is generated from the short form.

Quality of Life will be assessed using the Quality of Life Enjoyment and Satisfaction Questionnaire - Short Form (Q-LES-Q-SF): which is a self-report measure consisting of 16 questions designed to enable investigators to easily obtain sensitive measures of the degree of enjoyment and satisfaction experienced by subjects in various areas of daily functioning during the past week. The scoring of the Q-LES-Q-SF involves summing only the first 14 items to yield a raw total score. The last two items are not included in the total score but are standalone items. The raw total score ranges from 14 to 70. The raw total score is transformed into a percentage maximum possible score using the following formula: (raw total score - minimum score)/(maximum possible raw score - minimum score) The minimum raw score on the Q-LES-Q-SF is 14, and the maximum score is 70. Thus the formula for % maximum can also be written as (raw score - 14)/56.

Care satisfaction will be assessed using the Patient Assessment of Care for Chronic Conditions (PACIC) Survey. The PACIC Survey consists of 20-items that measures specific actions or qualities of care that patients report they have experienced in the care of their chronic conditions over the past 6 months. There is an additional item that is not part of the PACIC Survey and is used to track the patient's hopefulness over time. The PACIC Survey is grouped into five scales where each scale is scored by averaging the items completed within the scale, and the overall PACIC is cored by averaging scores across all 20 items. The five scales are: Patient Activation (items 1-3) Delivery System Design/Decision Support (items 4-6) Goal Setting (items 7-11) Problem-solving/Contextual Counseling (items 12-15) Follow-up/Coordination (items 16-20)

Emergent care use will be measured using existing behavioral and physical health claims data from the UPMC Insurance Services Division (ISD) data warehouse to determine the frequency of ED/inpatient admits in the 12 months following randomization.

Engagement in primary, specialty, mental health care will be measured using existing behavioral and physical health claims data from the UPMC ISD data warehouse to determining participant frequency of non-acute visits for participants in the 12 months following randomization.

Gaps in care will be assessed using HEDIS quality metrics that will be compiled using existing behavioral and physical health claims data from the UPMC ISD data warehouse. Gaps in care will be assessed for six of the most common chronic diseases present among our target population, including: asthma, chronic obstructive pulmonary disease, congestive heart failure, coronary artery disease, diabetes, and depression.

Inclusion Criteria: Enrollment Phase I Inclusion Criteria: Medicaid or dual-eligible (Medicare-Medicaid) adults age 21 years and older with Multiple Chronic Conditions (MCC). Have at least one physical health condition (e.g., cardiovascular disease, hypertension, COPD, diabetes). Have at least one additional physical or behavioral health condition (e.g., depression, serious mental illness, substance abuse disorder). Reside in Western, Central, or Eastern Pennsylvania. Be insured through physical and/or behavioral health payers within the UPMC ISD. Are identified as members with consistently high or rising resource utilization (which takes into account avoidable hospital readmissions, emergency department usage, and high cost, among other utilization-related variables). Enrollment Phase II Inclusion Criteria: Medicaid or dual-eligible (Medicare-Medicaid) adults age 21 years and older with Multiple Chronic Conditions (MCC). Have at least one physical health condition (e.g., cardiovascular disease, hypertension, COPD, diabetes). Have at least one additional physical or behavioral health condition (e.g., depression, serious mental illness, substance abuse disorder). Reside in Western, Central, or Eastern Pennsylvania. Be insured through physical and/or behavioral health payers within the UPMC ISD. ) Individuals will have several comorbidities, will have been prescribed several medications, and/or will be predicted future high health care utilizers. Must have at least one hospital discharge within 30 days of enrollment. Exclusion Criteria: Individuals who are pregnant. Individuals in skilled nursing facilities. Individuals on hemodialysis for kidney disease. Individuals who are in active cancer treatment. Individuals who have participated in High-Touch or High-Tech within the previous 12 months. Individuals who are unable to operate a smart phone due to reasons of literacy, vision or dexterity.

Due to the sensitive nature of the questions asked in this study, survey respondents were assured raw data would remain confidential and would not be shared.

Williams K, Markwardt S, Kearney SM, Karp JF, Kraemer KL, Park MJ, Freund P, Watson A, Schuster J, Beckjord E. Addressing Implementation Challenges to Digital Care Delivery for Adults With Multiple Chronic Conditions: Stakeholder Feedback in a Randomized Controlled Trial. JMIR Mhealth Uhealth. 2021 Feb 1;9(2):e23498. doi: 10.2196/23498. Erratum In: JMIR Mhealth Uhealth. 2021 Feb 26;9(2):e27996.

Kearney SM, Williams K, Nikolajski C, Park MJ, Kraemer KL, Landsittel D, Kang C, Malito A, Schuster J. Stakeholder impact on the implementation of integrated care: Opportunities to consider for patient-centered outcomes research. Contemp Clin Trials. 2021 Feb;101:106256. doi: 10.1016/j.cct.2020.106256. Epub 2020 Dec 29.