UPMC Center for High Value Health Care

Machine learning algorithms can advance clinical care, including identifying mental health conditions. These algorithms are often developed without considering the perspectives of the affected populations. This study describes the process of incorporating end-user perspectives into the development and implementation planning of a prediction algorithm for new perinatal depression onset. A focus group (N = 12 providers) and four virtual community engagement studios (N = 21 patients) were conducted. The project team presented on the initial development of a novel prediction algorithm used to detect first time perinatal depression. Rapid qualitative analysis coded the prediction algorithm's completeness, interpretability, and acceptability to stakeholders, with the goal of informing clinical implementation of a patient-facing screener produced from the prediction algorithm. Providers and patients showed consensus on the interpretability of the prediction algorithm's variables and discussed additional variables believed to be predictive of depression to ensure its completeness. In terms of acceptability, patients expressed a desire to discuss predictive risk screening results with their provider, while providers voiced concerns about limited bandwidth for these discussions. Both groups identified the need for post-screening resource connection but raised concerns over the availability of depression prevention specific resources. Providers and patients reported positively about their engagement in the sessions. Qualitative findings were incorporated into iterative algorithm development and informed an implementation pilot plan. This study demonstrates how the expertise of the end-users of a risk prediction algorithm can be incorporated into its development, which may ultimately increase clinical adoption.

This cross-sectional study assesses the rate of prostate cancer screening in men with high-deductible vs traditional health plans.

This dissertation explores the barriers and success factors of VBP models that influence their design, implementation, and application within the context of NOC and transmural care in the Netherlands. In doing so, this dissertation is expected to further aid the VBHC reform in the country. This dissertation identifies elements that enhance or inhibit the design, implementation, and application/evaluation of VBP models. It also analyzes the impact of different types of models on outcomes and reviews what facilitating and inhibiting factors are associated to each model type. These facilitating and inhibiting elements are also studied considering stakeholders’ insights in a Dutch network. Furthermore, this dissertation explores the introduction of a VBP model monitoring method and shares the results of its application in the Netherlands. Lastly, this dissertation takes a step back from the VBP models and analyzes the complex elements that shape the contractual negotiations between hospitals and health insurers in the Netherlands, examining why VBHC, despite its advantages, has not received more attention in this negotiation process.

Many interventions implemented for multi-visit patients (MVP) have been developed to address patient-centric needs of these individuals and reduce unplanned care for ambulatory-sensitive conditions. More rigorous research is needed to better understand the impact of these interventions on changes in care utilization including unplanned care. To evaluate the impact of the Enhanced Care Program (ECP), a payer-provider collaborative model, on unplanned care use and cost of care. Using propensity methods, a comparison group was constructed using insurer membership files. Comparisons were performed using a difference-in-differences analysis. Patients enrolled in ECP through December 2019 were considered eligible for the study (n = 357). All patients had five or more ED visits in the past year or two or more inpatient hospitalizations in the past year prior to enrollment. ECP is a high-intensity outpatient intervention intended to reduce avoidable unplanned care such as ED visits and inpatient hospital stays through home visits, chronic/acute disease management, and intensive care coordination. The primary outcomes of interest were events per 100 members per year of ED use with return to home, unplanned inpatient and observational status admissions, and unplanned behavioral health inpatient admission, and cost of care per member per month. Overall total unplanned care encounters were significantly reduced with a difference-in-difference of 320 unplanned care encounters per 100 members per year in the intervention group (p < 0.05). The ECP group showed statistically significant decreases in costs of unplanned ED, unplanned observation admission, and unplanned inpatient behavioral medicine costs, but statistically significant increases in overall pharmacy costs and lab costs. Changes in total costs of care for the ECP group were not statistically different than the control group (p = 0.55). ECP showed significant reduction of unplanned care for MVP patients.

To accelerate healthcare transformation and advance health equity, scientists in learning health systems (LHSs) require ready access to integrated, comprehensive data that includes information on social determinants of health (SDOH). We describe how an integrated delivery and finance system leveraged its learning ecosystem to advance health equity through (a) a cross-sector initiative to integrate healthcare and human services data for better meeting clients' holistic needs and (b) a system-level initiative to collect and use patient-reported SDOH data for connecting patients to needed resources. Through these initiatives, we strengthened our health system's capacity to meet diverse patient needs, address health disparities, and improve health outcomes. By sharing and integrating healthcare and human services data, we identified 281 000 Shared Services Clients and enhanced care management for 100 adult Medicaid/Special Needs Plan members. Over a 1-year period, we screened 9173 (37%) patients across UPMC's Women's Health Services Line and connected over 700 individuals to social services and supports. Opportunities exist for LHSs to improve, expand, and sustain their innovative data practices. As learnings continue to emerge, LHSs will be well positioned to accelerate healthcare transformation and advance health equity.

Improving medication adherence remains an important goal to improve therapeutic outcomes and lower health care costs. Point-of-sale prescription costs and forgetfulness remain top reasons why patients do not adhere to medications. Programs using both text message-based reminders and financial incentives may encourage patients to refill their prescriptions on time by reducing copays through discounts at the point of sale. Sempre Health, the subject of our analysis, provides both text message refill reminders and a dynamic discount incentive program to improve medication adherence. To evaluate the impact of a financial incentive/refill reminder program on medication adherence and total cost of care for patients taking the antithrombotic agents ticagrelor, apixaban, or rivaroxaban in a large regional health plan. After propensity-score matching on demographics, socioeconomic status, baseline copay, prior pharmacy/medical spend, and morbidity, we compared-using a difference-in-differences analytic approach-adherence (measured by proportion of days covered), unplanned health care utilization, and costs (total cost of care, medical, and pharmacy cost) of health plan members who did and did not enroll in the financial incentive/refill reminder program between February 1, 2019, and October 31, 2021, over 1 and 2 years. Because of differences in patient characteristics, we analyzed patients on ticagrelor (the antiplatelet group), apixaban, and rivaroxaban (the anticoagulant group) separately. There were a total of 1,292 one-to-one program and control propensity-matched patients: 166 each for the antiplatelet group and 480 each for the anticoagulant group. The average age of the anticoagulant group was 62 years; more than 60% were male, and approximately 45% had no prior unplanned care events. In contrast, the average age of the antiplatelet group was 57 years; more than 70% were male, and approximately 21% had no prior unplanned care events. In the antiplatelet group, the proportions adherent (proportion of days covered ≥80%) were 63.3% vs 42.8% (P = 0.0002) for program vs controls. Similarly, in the anticoagulant group, the proportion adherent was 77.9% vs 60.2% (P < 0.0001) for program vs controls. Reflecting improved adherence, costs of apixaban and rivaroxaban increased by $79 per member per month (PMPM) (P < 0.0001), with no statistically significant differences in other costs. Similarly, the cost of ticagrelor increased by $77 PMPM (P = 0.0102) with no statistically significant differences in other costs. Finally, there was a 16% (P = 0.032) reduction in emergency department use for those in the program. The financial incentive and refill reminder program was associated with improved adherence to antithrombotic medications, reduced emergency department use, and increased medication costs, but not in total pharmacy, medical, or total cost of care in both subgroups.

Patients with slow-transit constipation have bowel movements less than 3 times per week, experience severe constipation symptoms, and have a reduced quality of life. For patients who do not respond sufficiently to usual treatment (such as medication or bowel irrigation), sacral neuromodulation (SNM) may offer a solution. SNM is a minimally invasive surgical treatment that stimulates the nerves controlling the bowel. This thesis examined the effectiveness and costs of SNM compared to usual treatment for patients with slow-transit constipation. In a comparative study, patients were followed for 6 months. After 6 months, there was a 49,9% difference in treatment success, favoring SNM. Additionally, SNM patients had less severe constipation, less fatigue, and an improved quality of life. However, SNM was more expensive, with an cost-difference of €14,000. Lastly, this thesis showed a high variation in the types of outcomes and measurement tools used in research on slow-transit constipation and SNM, making it difficult to compare study results. As a result, there is currently no clear evidence on the effectiveness and costs of SNM for slow-transit constipation.