Effect of an Artificial Intelligence Decision Support Tool on Palliative Care Referral in Hospitalized Patients: A Randomized Clinical Trial.

CONTEXT: Palliative care services are commonly provided to hospitalized patients, but accurately predicting who needs them remains a challenge. OBJECTIVES: To assess the effectiveness on clinical outcomes of an artificial intelligence (AI)/machine learning (ML) decision support tool for predicting patient need for palliative care services in the hospital. METHODS: The study design was a pragmatic, cluster-randomized, stepped-wedge clinical trial in 12 nursing units at two hospitals over a 15-month period between August 19, 2019, and November 17, 2020. Eligible patients were randomly assigned to either a medical service consultation recommendation triggered by an AI/ML tool predicting the need for palliative care services or usual care. The primary outcome was palliative care consultation note. Secondary outcomes included: hospital readmissions, length of stay, transfer to intensive care and palliative care consultation note by unit. RESULTS: A total of 3183 patient hospitalizations were enrolled. Of eligible patients, A total of 2544 patients were randomized to the decision support tool (1212; 48%) and usual care (1332; 52%). Of these, 1717 patients (67%) were retained for analyses. Patients randomized to the intervention had a statistically significant higher incidence rate of palliative care consultation compared to the control group (IRR, 1.44 [95% CI, 1.11-1.92]). Exploratory evidence suggested that the decision support tool group reduced 60-day and 90-day hospital readmissions (OR, 0.75 [95% CI, 0.57, 0.97]) and (OR, 0.72 [95% CI, 0.55-0.93]) respectively. CONCLUSION: A decision support tool integrated into palliative care practice and leveraging AI/ML demonstrated an increased palliative care consultation rate among hospitalized patients and reductions in hospitalizations.

Cannabis in Cancer Survivors Who Report High Impact Chronic Pain: Findings from a 1500+ Patient Survey.

BACKGROUND: Chronic pain in cancer survivors negatively impacts quality of life. This study sought to investigate the relationship between high-impact chronic pain (HICP) -- defined as chronic pain that limits life or work activities on most days or every day in the past 3 months -- and cannabis in cancer survivors. METHODS: An electronic survey was developed in conjunction with the National Cancer Institute Comprehensive Cancer Centers in the United States. This survey was distributed to cancer survivors within a multi-site, single institution setting. RESULTS: The survey response rate was 23.0% (2304/10,000); 72.7% of these patients (1676/2304) did in fact have a confirmed cancer diagnosis. Among these cancer survivors, 16.5% (unweighted 278/1676) had HICP, and 12.4% (208/1676) reported cannabis use since their cancer diagnosis. The prevalence of past 30-day cannabis use was 12.3% (206/1676). Compared to cancer survivors without pain, those with HICP were more likely to believe in the benefits of cannabis (unweighted 92.1% vs. 74.7%; age-adjusted odds ratio [OR] = 3.1; 95% CI: 1.9-5.1) and less likely to believe in its risks (unweighted 48.2% vs. 58.4%; age-adjusted OR = 0.6; 95% CI: 0.4-0.7). CONCLUSIONS: Cancer survivors with HICP have a higher prevalence of cannabis use compared to those patients without pain. More research is needed to advance pain and symptom management among cancer survivors and to identify clinical scenarios in which benefit is greater than potential harm.

Improving time to palliative care review with predictive modeling in an inpatient adult population: study protocol for a stepped-wedge, pragmatic randomized controlled trial.

BACKGROUND: Palliative care is a medical specialty centered on improving the quality of life (QOL) of patients with complex or life-threatening illnesses. The need for palliative care is increasing and with that the rigorous testing of triage tools that can be used quickly and reliably to identify patients that may benefit from palliative care. METHODS: To that aim, we will conduct a two-armed stepped-wedge cluster randomized trial rolled out to two inpatient hospitals to evaluate whether a machine learning algorithm accurately identifies patients who may benefit from a comprehensive review by a palliative care specialist and decreases time to receiving a palliative care consult in hospital. This is a single-center study which will be conducted from August 2019 to November 2020 at Saint Mary's Hospital & Methodist Hospital both within Mayo Clinic Rochester in Minnesota. Clusters will be nursing units which will be chosen to be a mix of complex patients from Cardiology, Critical Care, and Oncology and had previously established relationships with palliative medicine. The stepped wedge design will have 12 units allocated to a design matrix of 5 treatment wedges. Each wedge will last 75 days resulting in a study period of 12 months of recruitment unless otherwise specified. Data will be analyzed with Bayesian hierarchical models with credible intervals denoting statistical significance. DISCUSSION: This intervention offers a pragmatic approach to delivering specialty palliative care to hospital patients in need using machine learning, thereby leading to high value care and improved outcomes. It is not enough for AI to be utilized by simply publishing research showing predictive performance; clinical trials demonstrating better outcomes are critically needed. Furthermore, the deployment of an AI algorithm is a complex process that requires multiple teams with varying skill sets. To evaluate a deployed AI, a pragmatic clinical trial can accommodate the difficulties of clinical practice while retaining scientific rigor. TRIAL REGISTRATION: ClinicalTrials.gov NCT03976297 . Registered on 6 June 2019, prior to trial start.

Central Sensitization Symptom Severity and Patient-Provider Relationships in a Community Setting.

INTRODUCTION: Central sensitization syndromes (CSS) comprise an overlapping group of clinical conditions with the core feature of "pain arising as a direct consequence of a lesion or disease affecting the somatosensory system." Patients with CSS are known to have challenging interactions with healthcare providers contributing to psychological distress and increased healthcare utilization. CSS symptom severity has been associated with psychologic comorbidities, but little is known about how symptom severity relates to provider interactions. METHODS: We performed a cross-sectional survey among patients with CSS in our primary care practices to examine the relationship between CSS symptom severity and experiences with doctors. RESULTS: A total of 775 respondents completed the survey (775/5000; 15.5%) with 72% reporting high CSS symptom severity. About 44% of respondents had a prior diagnosis of fibromyalgia, 72% had migraines, and 28% had IBS. Patients with high CSS symptom severity were more likely to report that doctor(s) had often/always told them that they don't need treatment when they feel like they do (OR = 3.6, 95% CI 1.9-7.5), that doctor(s) often/always don't understand them (OR = 3.1, 95% CI 1.9-5.4), and that doctor(s) often/always seem annoyed with them when compared with respondents with low-moderate CSS symptom severity (OR = 4.8, 95% CI 2.2-12.5). Patients with high CSS symptom severity were at greater than 5 times odds of reporting being told that their symptoms were "all in their head" when compared to patients with low-moderate symptom severity (OR = 5.4, 95% CI 3.3-9.0). CONCLUSION: Patients with CSS spectrum disorders experience frequent pain and decreased quality of life. A high degree of CSS symptom severity is associated with negative experiences with healthcare providers, which deters the establishment of a positive provider-patient relationship. Further research is needed to help understand symptom severity in CSS and harness the power of the therapeutic alliance as a treatment modality.

Effect of Integrated Gastroenterology Specialists in a Primary Care Setting: a Retrospective Cohort Study.

BACKGROUND: Gastrointestinal (GI) complaints are common in primary care practices. The patient-centered medical home (PCMH) may improve coordination and collaboration by facilitating coordination across healthcare settings and within the community, enhancing communication between providers, and focusing on quality of care delivery. OBJECTIVE: To investigate the effect of integrated community gastroenterology specialists (ICS-GI) model within a large primary care practice. DESIGN: Retrospective cohort with propensity-matched historic controls. PATIENTS: We identified 265 patients who had a visit with one of our ICS-GI specialists and matched them (1:2) to 530 similar patients seen prior to the implementation of the ICS-GI model. MAIN MEASURES: Frequency of diagnostic testing for GI indications, visits to our outpatient GI referral practice, emergency department and hospital utilization, and time to access of specialty care for the whole population and by GI condition group. KEY RESULTS: Patients seen in our ICS-GI model had similar outpatient care utilization (OR = 1.0, 95% CI 0.7-1.4, p = 0.90), were more likely to have visits in primary care (OR OR=1.5, 95% CI 1.1-2.2, p = 0.02), and were less likely to have visits to our GI outpatient referral practice (OR = 0.3, 95% CI 0.2-0.7, p < 0.0001). Condition-specific analyses show that all GI conditions experienced decreased visits to the outpatient GI referral practice outside of patients with GI neoplasm. Populations did not differ in emergency department, hospital, or diagnostic utilization. CONCLUSIONS: We observed that an embedded specialist in primary care model is associated with improved care coordination without compromising patient safety. The PCMH could be extended to include subspecialty care.

High rates of cancer screening among dialysis patients seen in primary care a cohort study.

Routine preventive cancer screening is not recommended for patients with end-stage renal disease (ESRD) due to their limited life expectancy. The current extent of cancer screening in this population is unknown. Primary care (PC) reminder systems or performance incentives may encourage indiscriminate cancer screening. We compared rates of cancer screening in patients with ESRD, with and without PC visits. This is a retrospective cohort study using United States Renal Data System (USRDS) billing data and electronic medical record data. Patients aged ≥18 years starting dialysis from 2001 to 2008, Midwest regional dialysis network were categorized with or without a PC visit (defined as an office visit in family practice, internal medicine, pediatrics, geriatrics or preventive medicine during the first two years of dialysis). Cancer screening was based on Current Procedural Terminology codes in USRDS. We identified 2512 incident dialysis patients (60% men, median age 65y). Cancer screening rates were more frequent among those seen in PC: 38% vs 19% (P = 0.0002), for breast; 18% vs 10% (P = 0.047) for cervical; 13% versus 8% (P = 0.024) for prostate; and 18% vs 9% (P = 0.0002) for colon cancer. Multivariable analyses found that those with PC were more likely to be screened after adjusting for age, sex, and comorbidities. In our practice, cancer screening rates among chronic dialysis patients are lower than those previously reported for our general population (64% for breast cancer). However, a sizeable proportion of our ESRD population does receive cancer screening, especially those still seen in primary care.

Controlled Substance Agreements for Opioids in a Primary Care Practice.

BACKGROUND: Opioids are widely prescribed for chronic non cancer pain (CNCP). Controlled substance agreements (CSAs) are intended to increase adherence and mitigate risk with opioid prescribing. We evaluated the demographic characteristics of and opioid dosing for patients with CNCP enrolled in CSAs in a primary care practice. METHODS: We conducted a retrospective cohort study of 1066 patients enrolled in CSAs between May 9, 2013 and August 15, 2016 for CNCP in a Midwest primary care practice. RESULTS: Patients were prescribed an average of 40.8 (SD ± 57.0) morphine milligram equivalents per day (MME/day), and 21.5% of patients were receiving ≥50 MME/day and 9.7% were receiving ≥90 MME/day. Patients who were younger in age (≥ 65 vs. < 65 years, P < 0.0001), male gender (P = 0.0001), and used tobacco (P = 0.0002) received significantly higher MME/day. Patients with more co-morbidities (Charlson Comorbidity Index, CCI) received higher MME/day (CCI > 3 vs. CCI ≤ 3, P = 0.03), and reported higher average pain (CCI > 3 mean 5.8 [SD ± 2.1] vs. CCI ≤ 3 mean 5.3 [SD ± 2.0], P = 0.0011). Patients on an identified tapering plan (6.9%) had higher MME/day than patients not on a tapering plan (P = 0.0002). CONCLUSIONS: CSAs present an opportunity to engage patients taking higher doses of opioids in discussions about opioid safety, appropriate dosing and tapering. CSAs could be leveraged to develop a population health management approach to the care of patients with CNCP.

Inherited variants in regulatory T cell genes and outcome of ovarian cancer.

Although ovarian cancer is the most lethal of gynecologic malignancies, wide variation in outcome following conventional therapy continues to exist. The presence of tumor-infiltrating regulatory T cells (Tregs) has a role in outcome of this disease, and a growing body of data supports the existence of inherited prognostic factors. However, the role of inherited variants in genes encoding Treg-related immune molecules has not been fully explored. We analyzed expression quantitative trait loci (eQTL) and sequence-based tagging single nucleotide polymorphisms (tagSNPs) for 54 genes associated with Tregs in 3,662 invasive ovarian cancer cases. With adjustment for known prognostic factors, suggestive results were observed among rarer histological subtypes; poorer survival was associated with minor alleles at SNPs in RGS1 (clear cell, rs10921202, p=2.7×10(-5)), LRRC32 and TNFRSF18/TNFRSF4 (mucinous, rs3781699, p=4.5×10(-4), and rs3753348, p=9.0×10(-4), respectively), and CD80 (endometrioid, rs13071247, p=8.0×10(-4)). Fo0r the latter, correlative data support a CD80 rs13071247 genotype association with CD80 tumor RNA expression (p=0.006). An additional eQTL SNP in CD80 was associated with shorter survival (rs7804190, p=8.1×10(-4)) among all cases combined. As the products of these genes are known to affect induction, trafficking, or immunosuppressive function of Tregs, these results suggest the need for follow-up phenotypic studies.