A randomized, placebo-controlled crossover trial to assess the influence of body weight on aspirin-triggered specialized pro-resolving mediators: Protocol for the DISCOVER Study.

Background: Low-dose aspirin is ineffective for primary prevention of cardiovascular events in people with body weight greater than 70kg. While the prevalent explanation for this is reduced platelet cyclooxygenase-1 (COX-1) inhibition at higher body weights, supporting data are limited, thereby demanding further investigation of the reason(s) underlying this observation. We propose that aspirin-mediated cyclooxygenase-2 (COX-2) acetylation and the resulting synthesis of 15-epi-lipoxin A4, a specialized pro-resolving mediator, is suboptimal in higher weight individuals, which may contribute to the clinical trial findings. Methods: To test this hypothesis, we are conducting a double-blind, placebo-controlled, randomized, mechanistic crossover trial. Healthy men and women exhibiting a wide range of body weights take 81mg aspirin and 325mg aspirin for 3 weeks each, following 3-week placebo run-in and wash-out phases. Our target sample size is 90 subjects, with a minimum of 72 completing all visits estimated to be necessary to achieve power adequate to test our primary hypothesis. Results: Our primary endpoint is the difference in change in plasma 15-epi-lipoxin A4 occurring with each dose of aspirin. Secondary endpoints include lipid mediator profiles, serum bioactive lipid profiles, and other endpoints involved in the resolution of vascular inflammation. Conclusions: Study enrollment began in November 2021 and is ongoing. The results of this study will improve our understanding of the mechanisms underlying aspirin's role(s) in the prevention of adverse cardiovascular outcomes. They may also lead to additional studies with the potential to inform dosing strategies for patients based on body weight.

Bias correction models for electronic health records data in the presence of non-random sampling.

Electronic health records (EHRs) contain rich clinical information for millions of patients and are increasingly used for public health research. However, non-random inclusion of subjects in EHRs can result in selection bias, with factors such as demographics, socioeconomic status, healthcare referral patterns, and underlying health status playing a role. While this issue has been well documented, little work has been done to develop or apply bias-correction methods, often due to the fact that most of these factors are unavailable in EHRs. To address this gap, we propose a series of Heckman type bias correction methods by incorporating social determinants of health selection covariates to model the EHR non-random sampling probability. Through simulations under various settings, we demonstrate the effectiveness of our proposed method in correcting biases in both the association coefficient and the outcome mean. Our method augments the utility of EHRs for public health inferences, as we show by estimating the prevalence of cardiovascular disease and its correlation with risk factors in the New York City network of EHRs.

Quantitative multiplex immunohistochemistry reveals inter-patient lymphovascular and immune heterogeneity in primary cutaneous melanoma.

Introduction: Quantitative, multiplexed imaging is revealing complex spatial relationships between phenotypically diverse tumor infiltrating leukocyte populations and their prognostic implications. The underlying mechanisms and tissue structures that determine leukocyte distribution within and around tumor nests, however, remain poorly understood. While presumed players in metastatic dissemination, new preclinical data demonstrates that blood and lymphatic vessels (lymphovasculature) also dictate leukocyte trafficking within tumor microenvironments and thereby impact anti-tumor immunity. Here we interrogate these relationships in primary human cutaneous melanoma. Methods: We established a quantitative, multiplexed imaging platform to simultaneously detect immune infiltrates and tumor-associated vessels in formalin-fixed paraffin embedded patient samples. We performed a discovery, retrospective analysis of 28 treatment-naïve, primary cutaneous melanomas. Results: Here we find that the lymphvasculature and immune infiltrate is heterogenous across patients in treatment naïve, primary melanoma. We categorized five lymphovascular subtypes that differ by functionality and morphology and mapped their localization in and around primary tumors. Interestingly, the localization of specific vessel subtypes, but not overall vessel density, significantly associated with the presence of lymphoid aggregates, regional progression, and intratumoral T cell infiltrates. Discussion: We describe a quantitative platform to enable simultaneous lymphovascular and immune infiltrate analysis and map their spatial relationships in primary melanoma. Our data indicate that tumor-associated vessels exist in different states and that their localization may determine potential for metastasis or immune infiltration. This platform will support future efforts to map tumor-associated lymphovascular evolution across stage, assess its prognostic value, and stratify patients for adjuvant therapy.

Major Adverse Cardiovascular Events After Colchicine Administration Before Percutaneous Coronary Intervention: Follow-Up of the Colchicine-PCI Trial.

Periprocedural inflammation is associated with major adverse cardiovascular events in patients who undergo percutaneous coronary intervention (PCI). In the contemporary era, 5% to 10% of patients develop restenosis, and in the acute coronary syndrome cohort, there remains a 20% major adverse cardiovascular events rate at 3 years, half of which are culprit-lesion related. In patients at risk of restenosis, colchicine has been shown to reduce restenosis when started within 24 hours of PCI and continued for 6 months thereafter, compared with placebo. The Colchicine-PCI trial, which randomized patients to a 1-time loading dose of colchicine or placebo 1 to 2 hours before PCI, showed a dampening of the inflammatory response to PCI but no difference in postprocedural myocardial injury. On mean follow-up of 3.3 years, the incidence of major adverse cardiovascular events did not differ between colchicine and placebo groups (32.5% vs 34.9%; hazard ratio 0.95 [0.68 to 1.34]).

Adapting the Diabetes Prevention Program for Older Adults: Descriptive Study.

BACKGROUND: Prediabetes affects 26.4 million people aged 65 years or older (48.8%) in the United States. Although older adults respond well to the evidence-based Diabetes Prevention Program, they are a heterogeneous group with differing physiological, biomedical, and psychosocial needs who can benefit from additional support to accommodate age-related changes in sensory and motor function. OBJECTIVE: The purpose of this paper is to describe adaptations of the Centers for Disease Control and Prevention's Diabetes Prevention Program aimed at preventing diabetes among older adults (ages ≥65 years) and findings from a pilot of 2 virtual sessions of the adapted program that evaluated the acceptability of the content. METHODS: The research team adapted the program by incorporating additional resources necessary for older adults. A certified lifestyle coach delivered 2 sessions of the adapted content via videoconference to 189 older adults. RESULTS: The first session had a 34.9% (38/109) response rate to the survey, and the second had a 34% (30/88) response rate. Over three-quarters (50/59, 85%) of respondents agreed that they liked the virtual program, with 82% (45/55) agreeing that they would recommend it to a family member or a friend. CONCLUSIONS: This data will be used to inform intervention delivery in a randomized controlled trial comparing in-person versus virtual delivery of the adapted program.

Does Incident Cardiovascular Disease Lead to Greater Odds of Functional and Cognitive Impairment? Insights From the Health and Retirement Study.

BACKGROUND: Although studies to date have broadly shown that cardiovascular disease (CVD) increases cognitive and physical impairment risk, there is still limited understanding of the magnitude of this risk among relevant CVD subtypes or age cohorts. METHODS: We analyzed longitudinal data from 16 679 U.S. Health and Retirement Study participants who were aged ≥65 years at study entry. Primary endpoints were physical impairment (activities of daily living impairment) or cognitive impairment (Langa-Weir Classification of dementia). We compared these endpoints among participants who developed incident CVD versus those who were CVD free, both in the short term (<2-year postdiagnosis) and long term (>5 years), controlling for sociodemographic and health characteristics. We then analyzed the effects by CVD subtype (atrial fibrillation, congestive heart failure, ischemic heart disease, and stroke) and age-at-diagnosis (65-74, 75-84, and ≥85). RESULTS: Over a median follow-up of 10 years, 8 750 participants (52%) developed incident CVD. Incident CVD was associated with significantly higher adjusted odds (aOR) of short-term and long-term physical and cognitive impairment. The oldest (≥85) age-at-diagnosis subgroup had the highest risk of short-term physical (aOR 3.01, 95% confidence interval [CI]: 2.40-3.77) and cognitive impairment (aOR 1.96, 95% CI: 1.55-2.48), as well as long-term impairment. All CVD subtypes were associated with higher odds of physical and cognitive impairment, with the highest risk for patients with incident stroke. CONCLUSIONS: Incident CVD was associated with an increased risk of physical and cognitive impairment across CVD subtypes. Impairment risk after CVD was highest among the oldest patients (≥85 years) who should therefore remain a target for prevention efforts.

A joint fairness model with applications to risk predictions for underrepresented populations.

In data collection for predictive modeling, underrepresentation of certain groups, based on gender, race/ethnicity, or age, may yield less accurate predictions for these groups. Recently, this issue of fairness in predictions has attracted significant attention, as data-driven models are increasingly utilized to perform crucial decision-making tasks. Existing methods to achieve fairness in the machine learning literature typically build a single prediction model in a manner that encourages fair prediction performance for all groups. These approaches have two major limitations: (i) fairness is often achieved by compromising accuracy for some groups; (ii) the underlying relationship between dependent and independent variables may not be the same across groups. We propose a joint fairness model (JFM) approach for logistic regression models for binary outcomes that estimates group-specific classifiers using a joint modeling objective function that incorporates fairness criteria for prediction. We introduce an accelerated smoothing proximal gradient algorithm to solve the convex objective function, and present the key asymptotic properties of the JFM estimates. Through simulations, we demonstrate the efficacy of the JFM in achieving good prediction performance and across-group parity, in comparison with the single fairness model, group-separate model, and group-ignorant model, especially when the minority group's sample size is small. Finally, we demonstrate the utility of the JFM method in a real-world example to obtain fair risk predictions for underrepresented older patients diagnosed with coronavirus disease 2019 (COVID-19).

When More is Less: Incorporating Additional Datasets Can Hurt Performance By Introducing Spurious Correlations.

Survival analysis is a general framework for predicting the time until a specific event occurs, often in the presence of censoring. Although this framework is widely used in practice, few studies to date have considered fairness for time-to-event outcomes, despite recent significant advances in the algorithmic fairness literature more broadly. In this paper, we propose a framework to achieve demographic parity in survival analysis models by minimizing the mutual information between predicted time-to-event and sensitive attributes. We show that our approach effectively minimizes mutual information to encourage statistical independence of time-to-event predictions and sensitive attributes. Furthermore, we propose four types of disparity assessment metrics based on common survival analysis metrics. Through experiments on multiple benchmark datasets, we demonstrate that by minimizing the dependence between the prediction and the sensitive attributes, our method can systematically improve the fairness of survival predictions and is robust to censoring.

Kinase Insert Domain Receptor Q472H Pathogenic Germline Variant Impacts Melanoma Tumor Growth and Patient Treatment Outcomes.

BACKGROUND: We previously reported a higher incidence of a pathogenic germline variant in the kinase insert domain receptor (KDR) in melanoma patients compared to the general population. Here, we dissect the impact of this genotype on melanoma tumor growth kinetics, tumor phenotype, and response to treatment with immune checkpoint inhibitors (ICIs) or targeted therapy. METHODS: The KDR genotype was determined and the associations between the KDR Q472H variant (KDR-Var), angiogenesis, tumor immunophenotype, and response to MAPK inhibition or ICI treatment were examined. Melanoma B16 cell lines were transfected with KDR-Var or KDR wild type (KDR-WT), and the differences in tumor kinetics were evaluated. We also examined the impact of KDR-Var on the response of melanoma cells to a combination of VEGFR inhibition with MAPKi. RESULTS: We identified the KDR-Var genotype in 81/489 (37%) patients, and it was associated with a more angiogenic (p = 0.003) and immune-suppressive tumor phenotype. KDR-Var was also associated with decreased PFS to MAPKi (p = 0.022) and a trend with worse PFS to anti-PD1 therapy (p = 0.06). KDR-Var B16 murine models had increased average tumor volume (p = 0.0027) and decreased CD45 tumor-infiltrating lymphocytes (p = 0.0282). The anti-VEGFR treatment Lenvatinib reduced the tumor size of KDR-Var murine tumors (p = 0.0159), and KDR-Var cells showed synergistic cytotoxicity to the combination of dabrafenib and lenvatinib. CONCLUSIONS: Our data demonstrate a role of germline KDR-Var in modulating melanoma behavior, including response to treatment. Our data also suggest that anti-angiogenic therapy might be beneficial in patients harboring this genotype, which needs to be tested in clinical trials.

Baseline Serum Autoantibody Signatures Predict Recurrence and Toxicity in Melanoma Patients Receiving Adjuvant Immune Checkpoint Blockade.

PURPOSE: Adjuvant immunotherapy produces durable benefit for patients with resected melanoma, but many develop recurrence and/or immune-related adverse events (irAE). We investigated whether baseline serum autoantibody (autoAb) signatures predicted recurrence and severe toxicity in patients treated with adjuvant nivolumab, ipilimumab, or ipilimumab plus nivolumab. EXPERIMENTAL DESIGN: This study included 950 patients: 565 from CheckMate 238 (408 ipilimumab versus 157 nivolumab) and 385 from CheckMate 915 (190 nivolumab versus 195 ipilimumab plus nivolumab). Serum autoAbs were profiled using the HuProt Human Proteome Microarray v4.0 (CDI Laboratories, Mayaguez, PR). Analysis of baseline differentially expressed autoAbs was followed by recurrence and severe toxicity signature building for each regimen, testing of the signatures, and additional independent validation for nivolumab using patients from CheckMate 915. RESULTS: In the nivolumab independent validation cohort, high recurrence score predicted significantly worse recurrence-free survival [RFS; adjusted HR (aHR), 3.60; 95% confidence interval (CI), 1.98-6.55], and outperformed a model composed of clinical variables including PD-L1 expression (P < 0.001). Severe toxicity score was a significant predictor of severe irAEs (aHR, 13.53; 95% CI, 2.59-86.65). In the ipilimumab test cohort, high recurrence score was associated with significantly worse RFS (aHR, 3.21; 95% CI, 1.38-7.45) and severe toxicity score significantly predicted severe irAEs (aHR, 11.04; 95% CI, 3.84-37.25). In the ipilimumab plus nivolumab test cohort, high autoAb recurrence score was associated with significantly worse RFS (aHR, 6.45; 95% CI, 1.48-28.02), and high severe toxicity score was significantly associated with severe irAEs (aHR, 23.44; 95% CI, 4.10-212.50). CONCLUSIONS: Baseline serum autoAb signatures predicted recurrence and severe toxicity in patients treated with adjuvant immunotherapy. Prospective testing of the signatures that include datasets with longer follow-up and rare but more severe toxicities will help determine their generalizability and potential clinical utility. See related commentary by Hassel and Luke, p. 3914.

Associations of Age at Diagnosis and Duration of Diabetes With Morbidity and Mortality Among Older Adults.

Importance: Older adults vary widely in age at diagnosis and duration of type 2 diabetes, but treatment often ignores this heterogeneity. Objectives: To investigate the associations of diabetes vs no diabetes, age at diagnosis, and diabetes duration with negative health outcomes in people 50 years and older. Design, Setting, and Participants: This cohort study included participants in the 1995 through 2018 waves of the Health and Retirement Study (HRS), a population-based, biennial longitudinal health interview survey of older adults in the US. The study sample included adults 50 years or older (n = 36 060) without diabetes at entry. Data were analyzed from June 1, 2021, to July 31, 2022. Exposures: The presence of diabetes, specifically the age at diabetes diagnosis, was the main exposure of the study. Age at diagnosis was defined as the age when the respondent first reported diabetes. Adults who developed diabetes were classified into 3 age-at-diagnosis groups: 50 to 59 years, 60 to 69 years, and 70 years and older. Main Outcomes and Measures: For each diabetes age-at-diagnosis group, a propensity score-matched control group of respondents who never developed diabetes was constructed. The association of diabetes with the incidence of key outcomes-including heart disease, stroke, disability, cognitive impairment, and all-cause mortality-was estimated and the association of diabetes vs no diabetes among the age-at-diagnosis case and matched control groups was compared. Results: A total of 7739 HRS respondents developed diabetes and were included in the analysis (4267 women [55.1%]; mean [SD] age at diagnosis, 67.4 [9.9] years). The age-at-diagnosis groups included 1866 respondents at 50 to 59 years, 2834 at 60 to 69 years, and 3039 at 70 years or older; 28 321 HRS respondents never developed diabetes. Age at diagnosis of 50 to 59 years was significantly associated with incident heart disease (hazard ratio [HR], 1.66 [95% CI, 1.40-1.96]), stroke (HR, 1.64 [95% CI, 1.30-2.07]), disability (HR, 2.08 [95% CI, 1.59-2.72]), cognitive impairment (HR, 1.30 [95% CI, 1.05-1.61]), and mortality (HR, 1.49 [95% CI, 1.29-1.71]) compared with matched controls, even when accounting for diabetes duration. These associations significantly decreased with advancing age at diagnosis. Respondents with diabetes diagnosed at 70 years or older only showed a significant association with the outcome of elevated mortality (HR, 1.08 [95% CI, 1.01-1.17]). Conclusions and Relevance: The findings of this cohort study suggest that age at diabetes diagnosis was differentially associated with outcomes and that younger age groups were at elevated risk of heart disease, stroke, disability, cognitive impairment, and all-cause mortality. These findings reinforce the clinical heterogeneity of diabetes and highlight the importance of improving diabetes management in adults with earlier diagnosis.

Risk and tropism of central nervous system (CNS) metastases in patients with stage II and III cutaneous melanoma.

BACKGROUND: Recent data suggest that patients with stage III melanoma are at high risk for developing central nervous system (CNS) metastases. Because a subset of patients with stage II melanoma experiences worse survival outcomes than some patients with stage III disease, the authors investigated the risk of CNS metastasis in stage II melanoma to inform surveillance guidelines for this population. METHODS: The authors examined clinicopathologic data prospectively collected from 1054 patients who had cutaneous melanoma. The χ2 test, the cumulative incidence, and Cox multivariable regression analyses were performed to evaluate the association between baseline characteristics and the development of CNS metastases. RESULTS: Patients with stage III melanoma had a higher rate of developing brain metastases than those with stage II melanoma (100 of 468 patients [21.4%] vs. 82 of 586 patients [14.0%], respectively; p = .002). However, patients who had stage IIC melanoma had a significantly higher rate of isolated first recurrences in the CNS compared with those who had stage III disease (12.1% vs. 3.6%; p = .002). The risk of ever developing brain metastases was similarly elevated for patients who had stage IIC disease (hazard ratio [HR], 3.16; 95% CI, 1.77-5.66), stage IIIB disease (HR, 2.83; 95% CI, 1.63-4.91), and stage IIIC disease (HR, 2.93; 95% CI, 1.81-4.74), and the risk was highest in patients who had stage IIID disease (HR, 8.59; 95% CI: 4.11-17.97). CONCLUSIONS: Patients with stage IIC melanoma are at elevated risk for first recurrence in the CNS. Surveillance strategies that incorporate serial neuroimaging should be considered for these individuals until more accurate predictive markers can be identified.

Utilization of Palliative Care for Patients with COVID-19 and Acute Kidney Injury during a COVID-19 Surge.

BACKGROUND AND OBJECTIVES: AKI is a common complication of coronavirus disease 2019 (COVID-19) and is associated with high mortality. Palliative care, a specialty that supports patients with serious illness, is valuable for these patients but is historically underutilized in AKI. The objectives of this paper are to describe the use of palliative care in patients with AKI and COVID-19 and their subsequent health care utilization. DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS: We conducted a retrospective analysis of New York University Langone Health electronic health data of COVID-19 hospitalizations between March 2, 2020 and August 25, 2020. Regression models were used to examine characteristics associated with receiving a palliative care consult. RESULTS: Among patients with COVID-19 (n=4276; 40%), those with AKI (n=1310; 31%) were more likely than those without AKI (n=2966; 69%) to receive palliative care (AKI without KRT: adjusted odds ratio, 1.81; 95% confidence interval, 1.40 to 2.33; P<0.001; AKI with KRT: adjusted odds ratio, 2.45; 95% confidence interval, 1.52 to 3.97; P<0.001), even after controlling for markers of critical illness (admission to intensive care units, mechanical ventilation, or modified sequential organ failure assessment score); however, consults came significantly later (10 days from admission versus 5 days; P<0.001). Similarly, 66% of patients initiated on KRT received palliative care versus 37% (P<0.001) of those with AKI not receiving KRT, and timing was also later (12 days from admission versus 9 days; P=0.002). Despite greater use of palliative care, patients with AKI had a significantly longer length of stay, more intensive care unit admissions, and more use of mechanical ventilation. Those with AKI did have a higher frequency of discharges to inpatient hospice (6% versus 3%) and change in code status (34% versus 7%) than those without AKI. CONCLUSIONS: Palliative care was utilized more frequently for patients with AKI and COVID-19 than historically reported in AKI. Despite high mortality, consultation occurred late in the hospital course and was not associated with reduced initiation of life-sustaining interventions. PODCAST: This article contains a podcast at https://www.asn-online.org/media/podcast/CJASN/2022_02_24_CJN11030821.mp3.

Factors associated with hospital admission and severe outcomes for older patients with COVID-19.

BACKGROUND: Morbidity and death due to coronavirus disease 2019 (COVID-19) experienced by older adults in nursing homes have been well described, but COVID-19's impact on community-living older adults is less studied. Similarly, the previous ambulatory care experience of such patients has rarely been considered in studies of COVID-19 risks and outcomes. METHODS: To investigate the relationship of advanced age (65+), on risk factors associated with COVID-19 outcomes in community-living elders, we identified an electronic health records cohort of older patients aged 65+ with laboratory-confirmed COVID-19 with and without an ambulatory care visit in the past 24 months (n = 47,219) in the New York City (NYC) academic medical institutions and the NYC public hospital system from January 2020 to February 2021. The main outcomes are COVID-19 hospitalization; severe outcomes/Intensive care unit (ICU), intubation, dialysis, stroke, in-hospital death), and in-hospital death. The exposures include demographic characteristics, and those with ambulatory records, comorbidities, frailty, and laboratory results. RESULTS: The 31,770 patients with an ambulatory history had a median age of 74 years; were 47.4% male, 24.3% non-Hispanic white, 23.3% non-Hispanic black, and 18.4% Hispanic. With increasing age, the odds ratios and attributable fractions of sex, race-ethnicity, comorbidities, and biomarkers decreased except for dementia and frailty (Hospital Frailty Risk Score). Patients without ambulatory care histories, compared to those with, had significantly higher adjusted rates of COVID-19 hospitalization and severe outcomes, with strongest effect in the oldest group. CONCLUSIONS: In this cohort of community-dwelling older adults, we provided evidence of age-specific risk factors for COVID-19 hospitalization and severe outcomes. Future research should explore the impact of frailty and dementia in severe COVID-19 outcomes in community-living older adults, and the role of engagement in ambulatory care in mitigating severe disease.

mTOR Inhibition with Sirolimus in Multiple System Atrophy: A Randomized, Double-Blind, Placebo-Controlled Futility Trial and 1-Year Biomarker Longitudinal Analysis.

BACKGROUND: Multiple system atrophy (MSA) is a fatal neurodegenerative disease characterized by the aggregation of α-synuclein in glia and neurons. Sirolimus (rapamycin) is an mTOR inhibitor that promotes α-synuclein autophagy and reduces its associated neurotoxicity in preclinical models. OBJECTIVE: To investigate the efficacy and safety of sirolimus in patients with MSA using a futility design. We also analyzed 1-year biomarker trajectories in the trial participants. METHODS: Randomized, double-blind, parallel group, placebo-controlled clinical trial at the New York University of patients with probable MSA randomly assigned (3:1) to sirolimus (2-6 mg daily) for 48 weeks or placebo. Primary endpoint was change in the Unified MSA Rating Scale (UMSARS) total score from baseline to 48 weeks. (ClinicalTrials.gov NCT03589976). RESULTS: The trial was stopped after a pre-planned interim analysis met futility criteria. Between August 15, 2018 and November 15, 2020, 54 participants were screened, and 47 enrolled and randomly assigned (35 sirolimus, 12 placebo). Of those randomized, 34 were included in the intention-to-treat analysis. There was no difference in change from baseline to week 48 between the sirolimus and placebo in UMSARS total score (mean difference, 2.66; 95% CI, -7.35-6.91; P = 0.648). There was no difference in UMSARS-1 and UMSARS-2 scores either. UMSARS scores changes were similar to those reported in natural history studies. Neuroimaging and blood biomarker results were similar in the sirolimus and placebo groups. Adverse events were more frequent with sirolimus. Analysis of 1-year biomarker trajectories in all participants showed that increases in blood neurofilament light chain (NfL) and reductions in whole brain volume correlated best with UMSARS progression. CONCLUSIONS: Sirolimus for 48 weeks was futile to slow the progression of MSA and had no effect on biomarkers compared to placebo. One-year change in blood NfL and whole brain atrophy are promising biomarkers of disease progression for future clinical trials. © 2022 International Parkinson and Movement Disorder Society.

Fair Generalized Linear Models with a Convex Penalty.

Despite recent advances in algorithmic fairness, methodologies for achieving fairness with generalized linear models (GLMs) have yet to be explored in general, despite GLMs being widely used in practice. In this paper we introduce two fairness criteria for GLMs based on equalizing expected outcomes or log-likelihoods. We prove that for GLMs both criteria can be achieved via a convex penalty term based solely on the linear components of the GLM, thus permitting efficient optimization. We also derive theoretical properties for the resulting fair GLM estimator. To empirically demonstrate the efficacy of the proposed fair GLM, we compare it with other wellknown fair prediction methods on an extensive set of benchmark datasets for binary classification and regression. In addition, we demonstrate that the fair GLM can generate fair predictions for a range of response variables, other than binary and continuous outcomes.

Reinforcement learning assisted oxygen therapy for COVID-19 patients under intensive care.

BACKGROUND: Patients with severe Coronavirus disease 19 (COVID-19) typically require supplemental oxygen as an essential treatment. We developed a machine learning algorithm, based on deep Reinforcement Learning (RL), for continuous management of oxygen flow rate for critically ill patients under intensive care, which can identify the optimal personalized oxygen flow rate with strong potentials to reduce mortality rate relative to the current clinical practice. METHODS: We modeled the oxygen flow trajectory of COVID-19 patients and their health outcomes as a Markov decision process. Based on individual patient characteristics and health status, an optimal oxygen control policy is learned by using deep deterministic policy gradient (DDPG) and real-time recommends the oxygen flow rate to reduce the mortality rate. We assessed the performance of proposed methods through cross validation by using a retrospective cohort of 1372 critically ill patients with COVID-19 from New York University Langone Health ambulatory care with electronic health records from April 2020 to January 2021. RESULTS: The mean mortality rate under the RL algorithm is lower than the standard of care by 2.57% (95% CI: 2.08-3.06) reduction (P < 0.001) from 7.94% under the standard of care to 5.37% under our proposed algorithm. The averaged recommended oxygen flow rate is 1.28 L/min (95% CI: 1.14-1.42) lower than the rate delivered to patients. Thus, the RL algorithm could potentially lead to better intensive care treatment that can reduce the mortality rate, while saving the oxygen scarce resources. It can reduce the oxygen shortage issue and improve public health during the COVID-19 pandemic. CONCLUSIONS: A personalized reinforcement learning oxygen flow control algorithm for COVID-19 patients under intensive care showed a substantial reduction in 7-day mortality rate as compared to the standard of care. In the overall cross validation cohort independent of the training data, mortality was lowest in patients for whom intensivists' actual flow rate matched the RL decisions.

Preexisting immune-mediated inflammatory disease is associated with improved survival and increased toxicity in melanoma patients who receive immune checkpoint inhibitors.

BACKGROUND: Immune-related adverse events (irAEs) are common, clinically significant autoinflammatory toxicities observed with immune checkpoint inhibitors (ICI). Preexisting immune-mediated inflammatory disease (pre-IMID) is considered a relative contraindication to ICI due to the risk of inciting flares. Improved understanding of the risks and benefits of treating pre-IMID patients with ICI is needed. METHODS: We studied melanoma patients treated with ICI and enrolled in a prospective clinicopathological database. We compiled a list of 23 immune-mediated inflammatory diseases and evaluated their presence prior to ICI. We tested the associations between pre-IMID and progression-free survival (PFS), overall survival (OS), and irAEs. RESULTS: In total, 483 melanoma patients were included in the study; 74 had pre-IMID and 409 did not. In patients receiving ICI as a standard of care (SoC), pre-IMID was significantly associated with irAEs (p = 0.04) as well as improved PFS (p = 0.024) and OS (p = 0.007). There was no significant association between pre-IMID and irAEs (p = 0.54), PFS (p = 0.197), or OS (p = 0.746) in patients treated through a clinical trial. Pre-IMID was significantly associated with improved OS in females (p = 0.012), but not in males (p = 0.35). CONCLUSIONS: The dichotomy of the impact of pre-IMID on survival and irAEs in SoC versus clinical trial patients may reflect the inherit selection bias in patients accrued in clinical trials. Future mechanistic work is required to better understand the differences in outcomes between female and male pre-IMID patients. Our data challenge the notion that clinicians should avoid ICI in pre-IMID patients, although close monitoring and prospective clinical trials evaluating ICI in this population are warranted.

Clinical outcomes in cancer patients with COVID-19.

BACKGROUND: Early reports on cancer patients with coronavirus disease 2019 (COVID-19) corroborated speculation that cancer patients are at increased risk for becoming infected with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and developing severe COVID-19. However, cancer patients are a heterogeneous population and their corresponding risk may be different. AIM: To compare COVID-19 presentation in patients with active malignancy to those with a history of cancer to determine the impact of cancer status on COVID-19 outcomes in the two groups. METHODS AND RESULTS: Of the 6724 patients who were hospitalized at NYU Langone Health (3/16/20-7/31/20) and tested positive for SARS-CoV-2, 580 had either active cancer (n = 221) or a history of cancer (n = 359). We compared the baseline clinicodemographic characteristics and hospital courses of the two groups. We studied the relationship between cancer status and the rate of admission to the intensive care unit (ICU), use of invasive mechanical ventilation (IMV), and all-cause mortality. The two groups had similar laboratory results associated with COVID-19 infection, incidence of venous thromboembolism, and incidence of severe COVID-19. Active cancer status was not associated with the rate of ICU admission (p = .307) or use of IMV (p = .236), but was significantly associated with worse all-cause mortality in both univariate and multivariate analysis with odds ratios of 1.48 (95% confidence interval [CI]: 1.04-2.09; p = .028) and 1.71 (95% CI: 1.12-2.63; p = .014), respectively. CONCLUSION: Active cancer patients had worse survival outcomes compared to patients with a history of cancer despite similar COVID-19 disease characteristics in the two groups. Our data suggest that cancer care should continue with minimal interruptions during the pandemic to bring about response and remission as soon as possible.

A Joint Fairness Model with Applications to Risk Predictions for Under-represented Populations.

In data collection for predictive modeling, under-representation of certain groups, based on gender, race/ethnicity, or age, may yield less-accurate predictions for these groups. Recently, this issue of fairness in predictions has attracted significant attention, as data-driven models are increasingly utilized to perform crucial decision-making tasks. Existing methods to achieve fairness in the machine learning literature typically build a single prediction model in a manner that encourages fair prediction performance for all groups. These approaches have two major limitations: i) fairness is often achieved by compromising accuracy for some groups; ii) the underlying relationship between dependent and independent variables may not be the same across groups. We propose a Joint Fairness Model (JFM) approach for logistic regression models for binary outcomes that estimates group-specific classifiers using a joint modeling objective function that incorporates fairness criteria for prediction. We introduce an Accelerated Smoothing Proximal Gradient Algorithm to solve the convex objective function, and present the key asymptotic properties of the JFM estimates. Through simulations, we demonstrate the efficacy of the JFM in achieving good prediction performance and across-group parity, in comparison with the single fairness model, group-separate model, and group-ignorant model, especially when the minority group's sample size is small. Finally, we demonstrate the utility of the JFM method in a real-world example to obtain fair risk predictions for under-represented older patients diagnosed with coronavirus disease 2019 (COVID-19).