Multisite Practice Research Articles

Accelerating access: The power of gold carding in reducing treatment delays.

114 Background: Delays in time to treatment initiation (TTI) for cancer patients have been shown to impact survival/mortality/outcomes, and cause patient distress. The process of obtaining prior authorizations (PA) often leads to significant treatment delays, affecting patient outcomes and posing administrative challenges for healthcare providers. PAs have become a ubiquitous part of modern healthcare, aimed at ensuring appropriate utilization of medical treatments and reducing costs. In recent years, the concept of gold carding (GC), a practice where payers waive PA on services and prescription drugs ordered by providers with a proven track record of PA approvals based on guideline concordance or pathway adherence, has gained attention as a potential solution to provide administrative relief to all stakeholders affected by the PA process. This abstract seeks to explore the impact of GC on TTI for a large multi-site community-based oncology practice in The US Oncology Network having a GC contract with a regional payer. Methods: Using administrative claims data from 2021 to 2023, we derived a population of cancer patients who initiated chemotherapy as their first treatment, calculated the TTI (time from the date of first consultation to the date of treatment initiation), and compared TTI between GC and non-GC patients. Results: The average TTI for 4,836 patients that met the inclusion criteria was 26.3 days. Patients initiating therapy under a GC eligible payer coverage started therapy 4.5 days earlier (22.13 days vs 26.68 days, p<0.00001, Kruskal–Wallis test). 80% of GC eligible patients started their treatment within 30 days of their initial consultation. Conclusions: The findings indicate that gold carding has the potential to expedite the time to treatment initiation for cancer patients. Further research is needed to quantify the cost savings, patient and provider satisfaction, and other intangible benefits from GC. Gold carding offers a viable approach to balance guideline concordant care with lowering administrative burden. We enthusiastically support legislative reforms that encourage gold carding and selectively implement prior authorization requirements based on stratification of health care providers’ performance and adherence to evidence-based medicine. Patient Count Average TTI (days) GC 401 22.13 Commercial 217 21.13 Managed Medicare 184 23.32 Non-GC 4435 26.68 Commercial 1610 25.41 Managed Medicare 1405 27.37 Medicare 959 26.48 Managed Medicaid 254 29.06 Medicaid 112 29.54 All Other 95 30.15

Clinician engagement strategies to increase palliative care access in a multi-site community oncology practice.

196 Background: Despite nearly 20 years since PC was recognized as a medical subspecialty, and ACS, ASCO and NCCN’s recommendations for early involvement, many oncologists continue to categorize PC synonymously with hospice and therefore refer to palliative medicine specialists very late in the disease trajectory, if at all. Each year, more oncology and PC researchers study and report the life changing impact of early PC for patients living with stage III and IV “advanced cancer” (AC). While mandating PC would have its challenges, creating workflows to encourage providers to change their behavior and attitude regarding PC, while empowering other clinical team members to directly refer, should result in patients and families receiving earlier PC services. Methods: The study included a pre-intervention period (9/1/2023 - 12/31/2023), and an intervention period (1/1/2024 - 4/30/2024). Pre-intervention referral data was collected from OncoEMR, OneAnalytics and Social Work reports. PC referrals were requested for both oncology and hematology patients, regardless of diagnosis date or stage. During the intervention period, 6 strategies were designed to increase PC referrals and were deployed equally in all 8 oncology/hematology offices. Strategies included 1) weekly MD/APP huddles with one of the goals to discuss the palliative care needs of all patients, 2) self-directed referrals from nursing, 3) self-directed referrals from social work, 4) APP referrals for advance care planning, 5) proactive chart reviews by PC APNs and direct PC referral suggestions to MD/APPs for AC patients, and 6) hardwiring timely scheduling of PC referrals. The primary end point is the percent of unique cancer (UC) patients who had a PC referral; the secondary end point is the percent of referred AC patients with early PC referral as defined by referral within 8 weeks of diagnosis. Results: During the pre-intervention period, of the 8,551 UC patients, 154 (1.8%) received PC referrals as compared to 227 (2.8%) of the 8,055 UC patients during the intervention period, yielding a 56% increase in the intervention period. In the pre-intervention period, 24 (14%) of the 177 newly diagnosed AC patients received early PC referrals compared to 32 (21%) of the 153 newly diagnosed AC patients in the intervention period, giving rise to a 50% increase. Conclusions: Strategies to increase PC referrals in a community oncology practice increased overall referrals as well as earlier referrals in newly diagnosed AC patients. As the interventions mature, the total number of PC referrals and a shift toward earlier referrals will likely be seen. Future goals will include assessing patient and family comfort with their treatment plan with PC versus without PC, as well as assessing providers' perceived value of PC over time. With attitude and behavior change, practice culture change is anticipated amid PC becoming the standard of care for all AC patients.

Identification of complex co-occurring symptom burden in persons with cancer through multisite implementation of health-related social needs assessment in community oncology settings.

306 Background: Health related social needs (HRSN) impact cancer outcomes and patient experience, but their integration into routine cancer care is highly variable. To expand the implementation-centered evidence base for effective HRSN assessments, we studied a scalable HRSN screening tool to identify distress and complex symptom burden in patients with cancer. Methods: We prospectively tracked HRSN screenings (using the National Comprehensive Cancer Network [NCCN] Distress Thermometer [DT]) of patients at varying points during their cancer care between 07/2023-12/2023 at seven multi-site community oncology practices within the US Oncology Network. DT responses were used to evaluate the positive finding rate, types of distress reported, and correlations between distress areas and concerns reported. Results: Overall, 219,720 unique patients with cancer received 269,666 HRSN screenings using the NCCN DT; 27% had a distress score ≥1 and 38.5% reported a concern (of whom, 58.2% reported >1). Co-occurring concerns that were most correlated (Pearson correlation coefficient) were emotional & practical concerns in 49% of screenings, physical & practical 44%, emotional & social 37% and social & practical concerns 36%. Emotional symptoms most frequently reported together were sadness/depression + worry/anxiety 47%, loss of interest + sadness/depression 44%, and loneliness & sadness/depression 38.5%. Regarding complex symptom burden, the physical and psychosocial symptoms that were most correlated were fatigue & sleep 43%, pain & fatigue 40%, pain & sleep 38%, anxiety/worry & fatigue 37%, anxiety/worry & sleep 36%. Fatigue and pain, the two most frequently reported concerns were strongly correlated with emotional and practical concerns like sleep, worry or anxiety, memory or concentration, loss or change of physical abilities, sadness or depression, loss of interest or enjoyment, changes in eating, ability to take care of self, feelings of worthlessness or being a burden, etc. Concerns with the strongest correlations are reported in the table. Conclusions: Distress and its components impact multiple cancer outcomes making their rapid identification and consequent action key to high-quality cancer care. The findings highlight the NCCN DT as a scalable approach to integrate HRSN screening and to identify complex symptom burden in the real-world practice. Co-occurring physical and psychosocial concerns. Concern 1 Concern 2 Correlation r Sadness or depression Worry or anxiety 47.3% 0.5 Loss of interest or enjoyment Sadness or depression 44.3% 0.4 Fatigue Sleep 42.5% 0.4 Pain Fatigue 40.0% 0.4 Loneliness Sadness or depression 38.5% 0.4 Pain Sleep 37.8% 0.4 Worry or anxiety Fatigue 36.7% 0.4 Worry or anxiety Sleep 36.1% 0.4 Taking care of myself Taking care of others 35.9% 0.4

Venous thromboembolism prevention program implementation in a community oncology practice: a cohort study

While national guidelines recommend Venous Thromboembolism (VTE) risk assessment in cancer outpatients and consideration of pharmacologic prophylaxis in high-risk patients, prophylaxis rates are low in community oncology practices. A successful model for guideline implementation (the Vermont Model, VM) is validated in an academic tertiary oncology setting. We undertook an implementation study to determine the success of this model in a multi-site community oncology practice. The study objectives were to: 1) adapt the VM to the community practice setting; 2) implement the adapted VM into practice; and 3) evaluate clinical and implementation outcomes. The study was carried out in three phases: (1) Pre-implementation, a multidisciplinary team addressed the need to adapt the VM to the local context including electronic medical record (EMR) optimisation and clinician education; (2) implementation of the strategies adapted to the local context, informed by VM and adapted based on stakeholder feedback; (3) prospective evaluation of clinical and implementation outcomes at six months after implementation. Following creation of a comprehensive initiation roadmap for the adaptation of VM program to the community practice, 302 cancer outpatients initiating new treatment met inclusion criteria over a 6 month implementation period. VTE risk education was provided to 100% of patients, and 98% (296) of patients received a VTE risk assessment. Of 52 patients (18%) who scored as high risk based on a modified Khorana (Protecht) score, 14 (27%) initiated prophylaxis. Barriers to program adaptation included EMR optimization challenges and practice-level responsibility assignment, time constraints, concern about potential drug interactions, and financial & insurance issues. Implementation of a multidisciplinary VTE prevention model in the community-based oncology setting successfully increased VTE education and risk assessment rates. AC prophylaxis rates were modestly increased, highlighting the need to understand and address barriers to anticoagulant prophylaxis prescribing in this setting. Northern New England Clinical Oncology Society Research Funding Program.

Utilization and outcomes of transcranial magnetic stimulation and usual care for MDD in a large group psychiatric practice

BackgroundGeneral psychiatrists’ practice standards vary regarding when to implement transcranial magnetic stimulation (TMS) for care of patients with major depressive disorder (MDD). Furthermore, few studies have examined real-world utilization and clinical outcomes of TMS. This study analyzed data from a large, multi-site psychiatric practice to evaluate utilization and outcomes of TMS as well as usual care (UC) for patients with MDD.MethodsDepression outcomes for TMS and UC among adult patients at a multi-site psychiatric group practice were examined in this retrospective cohort analysis. Patients with a primary diagnosis of MDD, PHQ-9 ≥ 10, and a visit in November 2020 with 6-month follow-up were included and categorized into the TMS or UC cohorts.ResultsOf 1,011 patients with qualifying PHQ-9 at the baseline visit, 9% (89) received a full course of TMS, and 583 patients receiving UC met study inclusion criteria (339 patients were excluded due to lacking a 6-month follow-up visit or receiving esketamine during the study period). The TMS cohort had higher baseline PHQ-9 than UC (17.9 vs. 15.5, p < .001) and had failed more medication trials (≥ 4 vs. 3.1, p < .001). Mean PHQ-9 decreased by 5.7 points (SD = 6.7, p < .001) in the TMS cohort and by 4.2 points (SD = 6.4, p < .001) in the UC cohort over the study period. Among patients who had failed four or more antidepressant medications, PHQ-9 decreased by 5.8 points in the TMS cohort (SD = 6.7, p < .001) and by 3.2 points in the UC cohort (SD = 6.3, p < .001).ConclusionsTMS utilization was low, despite TMS showing significant real-world clinical benefits. Future research should examine and address barriers to wider adoption of TMS into routine patient care for patients with treatment-resistant MDD. Wider adoption including routine use of TMS in less treatment-resistant patients will allow statistical comparisons of outcomes between TMS and UC populations that are difficult to do when TMS is underutilized.

Impact of an evidence-based bundle on incontinence-associated dermatitis prevalence in hospital patients: A quasi-experimental translational study.

The study aimed to evaluate the effect of an intervention on the prevalence and severity of incontinence-associated dermatitis (IAD) in six hospitals in one state in Australia. This quasi-experimental pre-and post-study,conducted in 18 wards, was part of a larger implementation science study on incontinence-associated dermatitis. Skin and incontinence assessments were conducted on patients during February and March 2020 (pre-intervention) and July and August 2021 (post-intervention). The intervention comprised continence assessment and management, an education brochure for patients, family and caregivers on IAD, the Ghent Global IAD Categorisation Tool (GLOBIAD) and a skin care regime with patient skin protection measures (three-in-one barrier cream cloths, minimisation of bed protection layers, use of appropriate continence aid). A total of 1897 patients were assessed (pre-intervention = 964, post-intervention = 933). A total of 343 (35.6%) pre-intervention patients and 351 (37.6%) post-intervention patients had incontinence. The prevalence of hospital-acquired IAD was 6.71% in the pre-intervention group and 4.27% in the post-intervention group; a reduction of 36.3% (p = 0.159) despite higher patient acuity, prevalence of double incontinence and the COVID-19 pandemic in the post-intervention group compared with the pre-intervention group. Our multisite best practice IAD prevention and treatment intervention was able to reduce the prevalence and severity of hospital-acquired IAD, suggesting enduring effectiveness of the intervention.

5-year multicenter analysis of clinical trial participation and total cost of care for older adults with cancer.

11013 Background: Clinical trial access particularly for older adults aged 65 years and above with cancer is impacted by multilevel barriers, including cost concerns to payers in the usual models of cancer care delivery. Alternative models recently studied include the Centers for Medicare and Medicaid Innovation’s Oncology Care Model (OCM), a value-based care pilot to enhance quality &amp; reduce costs. Here, we investigated the association between clinical trial participation (CTP) and the total cost of care (TCOC) within this non-traditional model. Methods: We identified prospective longitudinal observational cohort of patients cared for in 11 states within 323 clinics in 14 multi-site community practices participating in OCM, in The US Oncology Network and extracted data (episode claims, medical records, episodes) from OCM performance periods (PP) 3-11 between 7/1/17 and 6/30/22. Propensity score matching by PP, age, gender, date of death, cancer type identified the analysis cohort of matched episodes, then stratified by trial participation (CTP vs usual care) to compare TCOC and drug expenditures. Results: Over 5 years, 121,717 unique patients (94% of whom were aged 65y+) received 282,604 episodes of cancer care in the community, most common cancers being breast, lung, multiple myeloma. Propensity score matching identified 13,260 matched episodes (6630 usual care, 6630 CTP). The actual expenditure differed by $2,341 per episode between CTP &amp; usual care (CTP $43,890 vs usual care $41,548; p &lt; 0.0001), while trial participation had significant savings against the OCM CTP benchmark ($4816 saved per CTP episode), compared to usual care vs its benchmark ($826 saved per usual care episode, p &lt; 0.0001). Drug costs per episode did not differ between the two groups (CTP $29,516 vs usual care $30,553; p 0.54). Further breakdown shows that drug spending on anticancer agents (CTP $28,008 vs usual care $29,139) and other drugs including supportive care meds (CTP $1,511 vs usual care $1,414) did not differ significantly. Conclusions: Clinical trial participation saw relative cost savings against the benchmark compared to usual care, owing largely to the higher benchmark prices for CTP episodes in OCM’s risk-adjusted price prediction model, rather than an absolute cost savings to payers. Drug expenditures remain the largest contributor to total cost of care in this alternative model. The contributors to both drug expenditures as well as the absence of absolute cost savings between the two cohorts represent key areas for further investigation. CMMI Disclaimer: The statements contained in this document are solely those of the authors and do not necessarily reflect the views or policies of CMS. The authors assume responsibility for the accuracy and completeness of the information contained in this document.

Integration of Genetic Testing and Counseling in Patients With Breast Cancer in a Large, Multisite Community-Based Practice.

Despite data-driven consensus recommendations, there remains significant nonadherence to genetic screening and testing. More than 300,000 patients are diagnosed with breast cancer annually, with one third of these estimated to be eligible for homologous recombination deficiency (HRD)/BRCA testing following National Comprehensive Cancer Network (NCCN) guidelines. Only 35% of eligible patients are referred for genetic counseling. The goal of this project was to apply NCCN guidelines for germline genetic testing to all new patients with breast cancer within a large community oncology practice to improve HRD/BRCA testing. Plan-Do-Study-Act methodology was used, and cycles were built on a proven teaching infrastructure. In cycle 1, providers were educated and directed to use electronic health record (EHR) templates in the setting of an initial diagnosis visit and treatment planning. Discreet data fields were created in the EHR during cycle 2 to streamline and automate the process. Appropriate patients were referred to the genetics team for further evaluation, counseling, and testing. Adherence to the plan was maintained and measured using data analytic reports and chart audits. Of the 1,203 patients with breast cancer eligible for inclusion, 1,200 (99%) were screened according to NCCN guidelines. Of the screened patients, 631 (52.5%) met the referral/testing criteria. In total, 585 (92.7%) of the 631 were referred to a genetic specialist. Seven percent had previous referrals. A total of 449 (71%) patients were acceptable to genetics referral while 136 (21.5%) patients refused. The implemented methods of education, NCCN guidelines imbedded within provider notes, and discreet data fields in the EHR have proven to be highly effective in screening appropriate patients and ordering subsequent genetic referrals.

Patient Priorities–Aligned Care for Older Adults With Multiple Conditions

Older adults with multiple conditions receive health care that may be burdensome, of uncertain benefit, and not focused on what matters to them. Identifying and aligning care with patients' health priorities may improve outcomes. To assess the association of receiving patient priorities care (PPC) vs usual care (UC) with relevant clinical outcomes. In this nonrandomized controlled trial with propensity adjustment, enrollment occurred between August 21, 2020, and May 14, 2021, with follow-up continuing through February 26, 2022. Patients who were aged 65 years or older and with 3 or more chronic conditions were enrolled at 1 PPC and 1 UC site within the Cleveland Clinic primary care multisite practice. Data analysis was performed from March 2022 to August 2023. Health professionals at the PPC site guided patients through identification of values, health outcome goals, health care preferences, and top priority (ie, health problem they most wanted to focus on because it impeded their health outcome goal). Primary clinicians followed PPC decisional strategies (eg, use patients' health priorities as focus of communication and decision-making) to decide with patients what care to stop, start, or continue. Main outcomes included perceived treatment burden, Patient-Reported Outcomes Measurement Information System (PROMIS) social roles and activities, CollaboRATE survey scores, the number of nonhealthy days (based on healthy days at home), and shared prescribing decision quality measures. Follow-up was at 9 months for patient-reported outcomes and 365 days for nonhealthy days. A total of 264 individuals participated, 129 in the PPC group (mean [SD] age, 75.3 [6.1] years; 66 women [48.9%]) and 135 in the UC group (mean [SD] age, 75.6 [6.5] years; 55 women [42.6%]). Characteristics between sites were balanced after propensity score weighting. At follow-up, there was no statistically significant difference in perceived treatment burden score between groups in multivariate models (difference, -5.2 points; 95% CI, -10.9 to -0.50 points; P = .07). PPC participants were almost 2.5 times more likely than UC participants to endorse shared prescribing decision-making (adjusted odds ratio, 2.40; 95% CI, 0.90 to 6.40; P = .07), and participants in the PPC group experienced 4.6 fewer nonhealthy days (95% CI, -12.9 to -3.6 days; P = .27) compared with the UC participants. These differences were not statistically significant. CollaboRATE and PROMIS Social Roles and Activities scores were similar in the 2 groups at follow-up. This nonrandomized trial of priorities-aligned care showed no benefit for social roles or CollaboRATE. While the findings for perceived treatment burden and shared prescribing decision-making were not statistically significant, point estimates for the findings suggested that PPC may hold promise for improving these outcomes. Randomized trials with larger samples are needed to determine the effectiveness of priorities-aligned care. ClinicalTrials.gov Identifier: NCT04510948.

Self-scheduling in a Large Multispecialty and Multisite Clinic: A Retrospective, Longitudinal Examination of Multiple Self-Scheduled Visit Types.

Self-scheduling of medical visits is becoming available at many medical institutions. We aimed to examine the self-scheduled visit counts and rate of growth of self-scheduled visits in a multispecialty practice. For 85 weeks extending from January 1, 2022 through August 24, 2023, we examined self-scheduled visit counts for over 1500 self-scheduled visit types. We compared completed self-scheduled visit counts to all scheduled completed visit counts for the same visit types. We collected counts of the most frequently self-scheduled visit types for each week and examined the change over time. We also determined the proportion that each visit type was self-scheduled. There were 20,769 699 completed visits during the course of the study that met the criteria for inclusion. Self-scheduled visits accounted for 4.0% of all completed visits (838 592/20,769 699). Over the 85-week span, self-scheduled visits rose from 3.0% to 5.3% of the total. There were 1887 unique visit types that were associated with completed visits. There were just 6 appointment visit types of the total 1887 self-scheduled visit types that accounted for 50.7% of the total 838 592 self-scheduled visits. Those 6 visit types were a lab blood test visit (19.5%, 163 K visits), two Family Medicine office visit types (13.0%, 109 K visits), a screening mammogram visit type (6.6%, 55 K visits), a scheduled express care visit type (6%, 50 K visits) and a COVID immunization visit type (5.7%, 48 K visits). Twenty-one visit types that were self-scheduled accounted for 75% of the total self-scheduled visits. Four seasonal visits, accounting for 10.6% of the total self-scheduled visits, were responsible for almost all the non-linear change in self-scheduling. Self-scheduling accounted for a small but growing percent of all outpatient scheduled visits in a multispecialty, multisite practice. A wide range of visit types can be successfully self-scheduled.

Clinical Interventions Following Escalations from a Continuous Respiratory Monitoring Service in Patients With Chronic Obstructive Pulmonary Disease.

Continuous respiratory monitoring can support integrated care for chronic obstructive pulmonary disease (COPD) patients, by coupling them with remote clinical personnel who triage patients in coordination with their health care providers. When deploying such services, there remains uncertainty surrounding outcomes when at-risk patients are proactively identified and escalated for provider evaluation. This study presents findings from a service deployed in a real-world COPD cohort by analyzing the clinical interventions made during in-person and telehealth pulmonary outpatient visits following remote escalations. A single-center, retrospective, observational study of real-world COPD patients at a multi-site pulmonary practice was conducted. Patients who were enrolled in a continuous respiratory monitoring service for at least one year and were seen by a provider within seven days of an escalation by the service (N=168) were included. To evaluate the potential impact of these escalations on provider and patient burden, medical charts from outpatient visits were manually reviewed and grouped into six categories based on the clinical action(s) taken by the provider. A total of 245 outpatient visits occurred from 168 patients within seven days of escalation. Of the 245 visits, 206 (84.1%) resulted in clinical intervention and 163 (66.5%) resulted in treatment consistent with acute exacerbations of COPD (AECOPDs). 1.6% of the outpatient visits resulted in referral to the emergency room. Provider encounters occurring following the escalation of a patient from a continuous respiratory monitoring service consistently resulted in that provider administering a treatment to the escalated patient.

Using Cost Conversations to Address Financial Toxicity in Pediatric Asthma Care: Findings From a Survey of Caregivers

Asthma care teams are well-positioned to help caregivers address financial toxicity in pediatric asthma care, although discussing cost can be challenging. We sought to characterize cost conversations in pediatric asthma specialty care. We surveyed 45 caregivers of children aged 4-17 with asthma. Eligible caregivers reported costs concerns and had accompanied their child to a multisite asthma specialty practice in North Carolina. About one-third of caregivers reported a cost conversation (36%). Cost conversations were less common among caregivers whose child had public versus private health insurance (16% vs. 56%), who attended a telehealth versus in-person visit (6% vs. 52%), or who did not versus did want a conversation (19% vs. 77%, all p < .05). Common cost conversation topics were medications and equipment like spacers. Our findings suggest cost conversations may be relatively uncommon in pediatric asthma care, particularly for publicly insured patients and telehealth visits.

Impact of remote symptom monitoring (RSM) with electronic patient-reported outcomes (ePRO) on hospitalization, survival, and cost in community oncology practice: The Texas Two-Step study.

569 Background: There is rising interest to integrate electronic patient reported outcomes (ePRO) solutions for symptom monitoring to enhance the quality of care in oncology and to support value-based care (VBC) models. Step 1 of the Texas Two-Step Study demonstrated successful implementation of an ePRO system in &gt;200 sites of service of a large community oncology practice. We now report step 2 of our study which evaluates the impact of ePROs in reducing adverse outcomes and cost of care among cancer patients enrolled in CMS’ Oncology Care Model (OCM) VBC program. Methods: This observational study focused on patients with metastatic cancer enrolled in the OCM program at Texas Oncology between July 2020 to December 2020. Patients who enrolled and completed ≥1 survey via the ePRO tool defined the study arm. Patients were matched and balanced for confounders with propensity score matching (PSM). Data used for this analysis included data collected from OCM claims files and the ePRO vendor database. Adverse events (death (D)/hospitalizations (H)/Emergency Department (ED) visits), average H/ED visits per patient, average length of H, total Medicare cost (i.e. Medicare paid amounts for different services except Part D drugs) were study measures. Mann-Whitney U and Chi-Square tests compared continuous and categorical variables, respectively, and multivariable logistic regression was used to adjust the covariates (age, gender, cancer type, therapy line). Results: Out of 831 metastatic cancer patients, 458 matched patients (229 patients per group) were identified through PSM. In our matched cohort, ~52% of patients were male, with a mean age of 74 years. Mean total adverse events were significantly less (0.98 vs 1.41 p-value: 0.007) in the study group (SG) as compared to control (C). We observed decreased rates of H (20% vs 32.5%; p value: &lt;0.05), ED (38.4% vs 42.3%) and D (11.8% vs 16.6%) in the SC compared to C. After adjusting for covariates, the SG had decreased odds of H (aOR: 0.51; 95% CI 0.33, 0.8; p-value 0.002) compared to C. Similar, but not statistically significant, effects were observed in the case of D and ED visits. Average number of H (0.28 vs 0.52, p-value-0.003) and ED visits (0.51 vs 0.75, p value: 0.01) per patient were also lower in the SG as compared to C. In addition, mean H duration per patient (days) was significantly less in the SG as compared to C (1.9 vs 3.2, p value: 0.03). ePRO engagement also reduced the total cost of care (per member per month) by an average of $1146. Conclusions: RSM with ePROs reduced adverse outcomes and cost of care among patients with metastatic cancer in a large multi-site community oncology practice. Additional studies to include a larger patient population are warranted to strengthen these observations. This work suggests that ePRO use in the community setting does improve patient outcomes and enhance VBC.

Colonoscopy withdrawal times and adenoma detection rates in a large, multi-site colorectal surgery practice.

433 Background: Colonoscopy withdrawal times (CWT) and adenoma detection rates (ADR) have been extensively correlated in medical literature. The American College of Gastroenterology (ACG) 2021 guidelines recommend measuring quality indicators for screening colonoscopy that achieve minimum benchmarks for ADR (&gt;=25%) and CWT (&gt;=6 minutes). In October 2022, M Desai et al reported in a sample of 1142 patients (46% screening and 54% surveillance), ADR increased as withdrawal time increased from 6 to 13 minutes. Results demonstrated a 6% increase per minute in the likelihood of detecting an additional patient with at least 1 adenoma. Texas Colon and Rectal Specialists routinely reviews ADR’s and CWT’s. This retrospective descriptive study presents a concise review of the correlation between CWT’s and ADR’s in a homogenous set of screening and surveillance patients. Methods: We examined 7164 abstracted colonoscopy records from 31 colorectal surgeons. Diagnostic colonoscopies were excluded. We studied patients between 45 and 84 years of age at the time of the colonoscopy. Each patient was noted for the colonoscopy CWT as well as whether at least one adenomatous polyp was detected. These data were reported in a scatter plot of ADR over CWT as well as a bubble chart of ADR over CWT by each surgeon. A two-sample t-test confirmed a statistical difference between the mean CWT of the adenoma-positive patients and those that were not positive. Results: Consistent with the literature, ADR increased with CWT. The mean CWT was 14.2 minutes in the group where an adenoma was detected (n=3289) and 10.6 minutes in the group where an adenoma was not detected (n=3875), (p&lt;0.05 and CI of 95%). Mean CWT was 12.3 minutes, mode was 7 minutes, median was 10.25 minutes, and the mean ADR was 45.9%. The best ADR performance was 62.5% at a mean of 26.8 minutes CWT and the worst performance was 27.7% at 10 minutes CWT. A bubble chart demonstrates the relationship between CWT and ADR by surgeon with the bubbles representing individual colonoscopy volumes. Consistent with the literature, ADR increased with CWT. The mean CWT was 14.2 minutes in the group where an adenoma was detected (n=3289) and 10.6 minutes in the group where an adenoma was not detected (n=3875), (p&lt;0.05 and CI of 95%). Mean CWT was 12.3 minutes, mode was 7 minutes, median was 10.25 minutes, and the mean ADR was 45.9%. The best ADR performance was 62.5% at a mean of 26.8 minutes CWT and the worst performance was 27.7% at 10 minutes CWT. Conclusions: Our data indicate a significant relationship between CWT and ADR, and may suggest that CWT well beyond six minutes may yield better ADR’s. Furthermore, scope volume appears to play a role in the achievement of higher ADR’s, though this would require further investigation for optimal CWT, minimum expected ADR, optimal techniques and bowel prep, and emerging assistive technology.

Leveraging a Real-Time Machine Learning Platform to Personalize Quality Assurance at a Large Multi-Site Oncology Practice

Peer review is a critical component of safe and effective oncology care and provides care teams with an opportunity to improve patient outcomes with predictive analytics. We describe a paradigm for incorporating predictive analytics into daily prospective peer review at a large multisite oncology center. Previously, we deployed a department-wide proprietary oncology analytics platform connected to the Electronic Medical Record System (outpatient and inpatient), Oncology Information System, Tumor Registry, Treatment Planning System and Radiology PACS. Variables extracted and calculated by the analytics platform on a nightly basis include patient demographics, past medical history, detailed pathology data, recent laboratory data, vitals, prior systemic cancer therapies, prior surgical treatments, prior radiation treatments, recent hospitalization, vital status, and current treatment details. Using this platform, we trained machine learning models for critical patient events, such as unplanned hospitalization, mortality, occurrence of grade 3 or higher toxicities, and treatment interruption. Importantly, these models leverage so-called interpretable machine learning techniques, allowing us to interrogate the models and obtain explanations for why a patient is predicted to be at higher or lower risk of an event. All such models were previously evaluated for performance on standalone testing sets. In this abstract, we propose a paradigm to include predictive analytics in our daily prospective peer review process by creating a list of patients at high risk of critical events (hospitalization, mortality, treatment interruption, and serious adverse events) and reviewing those lists in conjunction with our daily peer review conference. We also propose a process to review any gaps in clinical care (e.g., abnormal labs, decreases in patient weight, or missed progress visits) to recommend suitable interventions to the care team to reduce patient risk. A total of 64,286 radiation treatment cases were included in our analytics platform and used to train and validate our predictive models. Moving forward, we aim to peer review 20 patient charts per day using this paradigm. While numerous studies in the field of radiation oncology have focused on the benefits of machine learning when applied to contouring and personalized radiation dose, few have focused on how machine learning can improve supportive care. In this abstract, we propose a framework to introduce machine learning into quality assurance, with the goal of improving patient outcomes.

Number of patients with interstitial cystitis/bladder pain syndrome seen before versus during the COVID-19 pandemic at an academic, urban, multisite urogynecology practice.

Little is known about the impact of the COVID-19 pandemic on interstitial cystitis/bladder pain syndrome (IC/BPS). We aim to compare the number of newly diagnosed IC/BPS cases and number of patients with flares prior to and during the pandemic. We conducted a retrospective cohort study of women ≥18 years who were diagnosed with or treated for IC/BPS between March 2019 and March 2021 at an academic, urban, multisite urogynecology practice. The primary outcome was the number of IC/BPS cases from March 1, 2019 to February 29, 2020 (pre-pandemic) compared with March 1, 2020 to February 28, 2021 (during pandemic). The secondary outcome was the number of patients with flares during those same two time periods. Demographic and clinical characteristics were compared using nonparametric tests and interrupted time series (ITS) was used to evaluate our outcomes of interest. p-Value <.05 was considered significant. Fifty-four women (4.87% of new patients) were diagnosed with IC/BPS during the pandemic compared with 40 women pre-pandemic (4.05% of new patients). The median age was 35.0. Seventy-two percent were premenopausal, 75% sexually active, and 31% had anxiety, and there were no significant differences between groups. Although the number of patients newly diagnosed with IC/BPS was higher during the pandemic, the diagnosis rates between time periods were not statistically different. Thirty-five patients experienced flares during the pandemic compared with 49 patients the year prior (p = .43). This difference was also not statistically significant on ITS analysis. Although more patients were diagnosed with IC/BPS during versus before the pandemic, the difference in diagnosis rates was not different between these periods.

The impact of receipt of systemic anticancer therapy (SACT) near the end of life (EOL) on cost among oncology practices participating in CMS' Oncology Care Model (OCM).

e18923 Background: Chemotherapy administered near the EOL is associated with higher costs due to downstream care that occurs when patients do not transition to hospice, such as hospitalizations and ICU use. Since 2015 rates of systemic therapy at the EOL have remained stable, while chemotherapy has been replaced with costly immunotherapy. Recently, we showed receipt of immunotherapy at the EOL is also associated with higher rates of acute care use than among patients who don’t receive by SACT near the EOL. Using data from practices participating in OCM, we evaluated the relationship between receipt of chemotherapy, immunotherapy, or no systemic therapy at the EOL and categories of cost. Methods: We used anonymized Integra Connect data for OCM episode claims attributed to 8 multi-site OCM practices from July 2016-December 2021 to identify decedents capable of making treatment decisions. We stratified patients by SACT type received within 30 days of death (immunotherapy (IO), chemotherapy (CT), or none) and used t-tests to compare the mean episode costs. We evaluated differences in overall costs as well as those billed under Medicare subgroups: Part A, inpatient acute care use and other services including (hospice, SNF, HHA), Part B, infusion drug administered in the ambulatory setting and medical and ED costs, and Part D, oral pharmacy costs. Results: Among 14,169 OCM decedents overall mean cost within 30 days of death was $22,215. Overall cost within 30 days of death was higher for patients receiving some form of SACT (IO: $26,072 or CT: $23,259) compared with no SACT $19,296 (table) all p &lt; 0.01). Patients who received no SACT at EOL had significantly lower Part B and D compared with both IO and CT respectively (all p &lt; 0.01). Additionally, patients who received CT had lower average part B costs $8,081 compared with patients who received IO within 30 days of death $13,951 p &lt; 0.01. Conclusions: Overall costs for patients who received IO or CT within 30 days of death were significantly higher than patients who received no SACT within 30 days of death. While IO patients had nearly twice as high mean Part B costs per episode compared with CT patients, CT patients had higher mean Part D costs per episode. Further research should explore whether these findings are generalizable to patients outside of OCM. [Table: see text]

The association between systemic anticancer therapy (SACT) at the end of life (EOL) and acute care use among patients treated at oncology practices participating in the Oncology Care Model (OCM).

6580 Background: Receipt of chemotherapy near EOL has been shown to harm patient and caregiver experience, increase hospitalizations, intensive care unit (ICU) and emergency department (ED) use and drive-up costs. As rates of chemotherapy at EOL have declined, we have seen increases in the rates of immunotherapy at EOL. It is unknown whether use of EOL immunotherapy is associated with the same pattern of higher downstream acute care use and cost. Using data from OCM, a Centers for Medicare and Medicaid Services (CMS) alternative payment model, we evaluated the relationship between SACT type and acute care utilization within 30 days of death. Methods: Using anonymized Integra Connect data for OCM episode claims attributed to 8 multi-site OCM practices between July 2016 and December 2021, we stratified patients by SACT type received within 30 days of death (immunotherapy (IO), chemotherapy (CT), or none) and used chi-square tests to compare the proportion of decedents who had inpatient admission, ICU stay, ED visits without inpatient admission, and hospice use within 30 days of death. Results: Among 14,169 OCM decedents, 2,227 (15.7%) received IO within 30 days of death, 6,627 (46.8%) received CT and 5,315 (37.5%) received no SACT (table). Patients who received no SACT at EOL were significantly more likely to start hospice within 30 days of death (76%) and to be in hospice for ≥3 days prior to death (61%) compared with IO (65% and 40%) and CT patients (59% and 37%) respectively (all p-values&lt;0.05). Patients who received IO or CT were significantly more likely to have an inpatient stay within 30 days of death (66% and 68% respectively) compared with patients who received no SACT at EOL (58%, p-values &lt;0.001). No difference in ED visits within 30 days of death between IO (18%), CT (20%), and no SACT (19%) p &lt; 0.270. Conclusions: Patients who received chemotherapy or IO therapy within 30 days of death had higher proportion of inpatient admissions, ICU stay and lower rates of hospice than patients who received no SACT at the EOL. Patients who did not have any SACT at EOL were more likely to utilize hospice services and for a longer time. Further research should explore whether these findings are generalizable to patients outside of OCM and the magnitude of costs associated with these acute care metrics. [Table: see text]

Closing the Testing Gap: Standardization of Comprehensive Biomarker Testing for Metastatic Non-Small-Cell Lung Cancer in a Large Community Oncology Practice.

Non-small-cell lung cancer (NSCLC), the leading cause of cancer death in the United States, accounts for 85% of all lung cancer cases. Biomarker testing is an integral part of the care of patients with NSCLC. Despite broad consensus recommendations that all patients with metastatic NSCLC (mNSCLC) undergo comprehensive biomarker testing (comprehensive genomic profiling and PD-L1), testing rates remain suboptimal. The primary goal of this project was to apply National Comprehensive Cancer Network (NCCN) guidelines for comprehensive biomarker testing to all new patients with mNSCLC within a large community practice. Plan-Do-Study-Act methodology was used, with cycle 1 focused on provider education and the creation of a mNSCLC initial consult Note (electronic health record template/McKesson iKnowMed G2) and accompanying order set. Staging, template/order set utilization, and comprehensive biomarker testing rates were recorded while workflow processes were monitored. Cycle 2 centered on improved cancer staging, data analytic reporting, auditing, and reeducation. The comprehensive biomarker testing rates increased from a historic rate of 68% to 92.7% during the 1-year intervention period. The template utilization rate was 71% with complete staging (TNM stage and relevant biomarkers) documented in 40%. Implementation and standardization of comprehensive biomarker testing of patients with mNSCLC in a large multisite community-based oncology practice is feasible and results in significant improvement in comprehensive biomarker testing and reporting. Establishing reliable and measurable tracking metrics to ensure that these new processes are used and maintained can assist in scaling these processes. Efforts to scale this best practice are planned across the US Oncology Network.

Multi-faceted intervention to improve management of antibiotics for children presenting to primary care with acute cough and respiratory tract infection (CHICO): efficient cluster randomised controlled trial

ObjectiveTo assess whether an easy-to-use multifaceted intervention for children presenting to primary care with respiratory tract infections would reduce antibiotic dispensing, without increasing hospital admissions for respiratory tract infection.DesignTwo arm...