scholarly journals Interplay between upsampling and regularization for provider fairness in recommender systems

2021 ◽  
Author(s):  
Ludovico Boratto ◽  
Gianni Fenu ◽  
Mirko Marras
Keyword(s):  
Recommender Systems ◽  
Real World ◽  
Minority Groups ◽  
Minority Group ◽  
Prior Work ◽  
Real World Data ◽  
Online Platforms ◽  
Key Stakeholders ◽  
Sensitive Attribute ◽  
The Impact

AbstractConsidering the impact of recommendations on item providers is one of the duties of multi-sided recommender systems. Item providers are key stakeholders in online platforms, and their earnings and plans are influenced by the exposure their items receive in recommended lists. Prior work showed that certain minority groups of providers, characterized by a common sensitive attribute (e.g., gender or race), are being disproportionately affected by indirect and unintentional discrimination. Our study in this paper handles a situation where (i) the same provider is associated with multiple items of a list suggested to a user, (ii) an item is created by more than one provider jointly, and (iii) predicted user–item relevance scores are biasedly estimated for items of provider groups. Under this scenario, we assess disparities in relevance, visibility, and exposure, by simulating diverse representations of the minority group in the catalog and the interactions. Based on emerged unfair outcomes, we devise a treatment that combines observation upsampling and loss regularization, while learning user–item relevance scores. Experiments on real-world data demonstrate that our treatment leads to lower disparate relevance. The resulting recommended lists show fairer visibility and exposure, higher minority item coverage, and negligible loss in recommendation utility.

scholarly journals NCDB Analysis of Melanoma 2004–2015: Epidemiology and Outcomes by Subtype, Sociodemographic Factors Impacting Clinical Presentation, and Real-World Survival Benefit of Immunotherapy Approval

Cancers ◽  
2021 ◽  
Vol 13 (6) ◽  
pp. 1455
Author(s):  
Sunny R. K. Singh ◽  
Sindhu J. Malapati ◽  
Rohit Kumar ◽  
Christopher Willner ◽  
Ding Wang
Keyword(s):  
Metastatic Melanoma ◽  
Metastatic Disease ◽  
Real World ◽  
Care Delivery ◽  
Sociodemographic Factors ◽  
P Value ◽  
Real World Data ◽  
Factors Associated ◽  
Improvement In Survival ◽  
The Impact

Background: The incidence of invasive melanoma is rising, and approval for the first immune checkpoint inhibitor (ICI) to treat metastatic melanoma occurred in 2011. We aim to describe the epidemiology and outcomes in recent years, sociodemographic factors associated with the presence of metastasis at diagnosis, and the real‐world impact of ICI approval on survival based on melanoma subtype and race. Methods: This is a retrospective analysis of the National Cancer Database (NCDB) from the years 2004–2015. The primary outcome was the overall survival of metastatic melanoma by subtype. Secondary outcomes included sociodemographic factors associated with the presence of metastasis at diagnosis and the impact of treatment facility type and ICI approval on the survival of metastatic melanoma. Results: Of the 419,773 invasive melanoma cases, 93.80% were cutaneous, and 4.92% were metastatic at presentation. The odds of presenting with metastatic disease were higher in African Americans (AA) compared to Caucasians (OR 2.37; 95% CI 2.11–2.66, p < 0.001). Treatment of metastatic melanoma at an academic/research facility was associated with lower mortality versus community cancer programs (OR 0.75, 95 % CI 0.69–0.81, p-value<0.001). Improvement in survival of metastatic melanoma was noted for Caucasians after the introduction of ICI (adjusted HR 0.80, 95% CI 0.78–0.83, p < 0.001); however, this was not statistically significant for AA (adjusted HR 0.80, 95% CI 0.62–1.02, p‐value = 0.073) or ocular cases (HR 1.03, 95% CI 0.81–1.31, p‐value 0.797). Conclusion: Real‐world data suggest a 20% improvement in survival of metastatic melanoma since the introduction of ICI. The disproportionately high odds of metastatic disease at presentation in AA patients with melanoma suggest the need for a better understanding of the disease and improvement in care delivery.

scholarly journals 979. Disparities in PrEP uptake and adherence among cisgender women using a pharmacologic measure

2020 ◽  
Vol 7 (Supplement_1) ◽  
pp. S517-S518
Author(s):  
Shane Hebel ◽  
Elijah Kahn-Woods ◽  
Casper Enghuus ◽  
Helen Koenig ◽  
Linden Lalley-Chareczko ◽  
...  
Keyword(s):  
Real World ◽  
Minority Groups ◽  
Hiv Infections ◽  
Urine Samples ◽  
Real World Data ◽  
Women And Gender ◽  
Gender Minority ◽  
Adherence Data ◽  
And Gender ◽  
Exposure Prophylaxis

Abstract Background HIV pre-exposure prophylaxis (PrEP) is 99% effective at preventing new HIV infections if taken daily. To be successful, PrEP requires concurrent efforts to optimize uptake, persistence, and adherence. In 2018, cisgender (cis) women accounted for 19% of new HIV infections in the US but comprised only 7% of all PrEP users. Studies show poor PrEP adherence amongst cis women, but there is a paucity of real-world clinical data describing PrEP adherence among cis women and gender minority people. Methods An adherence test that measures the concentration of tenofovir in urine samples using a liquid chromatography mass spectrometry (LC-MS/MS) was used to assess recent PrEP adherence at 8 clinics. Urine samples were collected during routine visits and analyzed using the LC-MS/MS assay. Test results were retrospectively paired with gender data, when available, and sex assigned at birth (SAAB) data. Adherence data were aggregated and analyzed to assess non-adherence proportions by sub-population. Results Gender data were available from 1,461 patients at 5 clinics, 1,344 (92%) of whom were cis males (Figure 1). From the 5 clinics where gender and SAAB data were available, 3,835 tests were conducted and 517 (13.5%) indicated non-adherence (Figure 2). 3 additional clinics conduct routine adherence testing and collect SAAB data (gender data not available). At these 8 clinics, SAAB data were available for 2,773 PrEP patients, totaling 5,602 urine tests (Figure 3). Among these 5,602 adherence tests, 813 (14.5%) indicated non-adherence (Figure 4). SAAB females demonstrated significantly higher non-adherence than SAAB males (22% vs 14%, p&lt; 0.001). Across clinics, 89%-98% of PrEP patients are SAAB male (Figure 5). Within these 8 clinics, SAAB female demonstrated consistently higher non-adherence (17%-44%, vs 12%-17% for SAAB males) (Figure 6). Figures 1 and 2 Figures 3 and 4 Figures 5 and 6 Conclusion Real-world data align with nationwide trends in PrEP utilization and show that the majority of PrEP patients are cis men. When initiated on PrEP, cis women exhibit higher rates of non-adherence than cis men. These data underscore the need to collect gender-identity data to monitor PrEP disparities and suggest that greater efforts are needed to target PrEP access, utilization, and accompanying support services to cis women and gender minority groups. Disclosures All Authors: No reported disclosures

PD-0776 A Bayesian approach to evaluate the impact of change in IGRT protocol using real world data

2021 ◽  
Vol 161 ◽  
pp. S608
Author(s):  
I. Fornacon-Wood ◽  
H. Mistry ◽  
C. Johnson-Hart ◽  
J.P.B. O’Connor ◽  
C. Faivre-Finn ◽  
...  
Keyword(s):  
Bayesian Approach ◽  
Real World ◽  
Real World Data ◽  
World Data ◽  
The Impact

Assessment and design of real world driving cycles targeted to the calibration of vehicles with electrified powertrain

2021 ◽  
pp. 146808742110387
Author(s):  
Stylianos Doulgeris ◽  
Zisimos Toumasatos ◽  
Maria Vittoria Prati ◽  
Carlo Beatrice ◽  
Zissis Samaras
Keyword(s):  
Real World ◽  
Simulation Models ◽  
Cost Effective ◽  
Operational Parameters ◽  
Virtual Design ◽  
Passenger Cars ◽  
Real World Data ◽  
Vehicle Simulation ◽  
Driving Cycles ◽  
The Impact

Vehicles’ powertrain electrification is one of the key measures adopted by manufacturers in order to develop low emissions vehicles and reduce the CO2 emissions from passenger cars. High complexity of electrified powertrains increases the demand of cost-effective tools that can be used during the design of such powertrain architectures. Objective of the study is the proposal of a series of real-world velocity profiles that can be used during virtual design. To that aim, using three state of the art plug-in hybrid vehicles, a combined experimental, and simulation approach is followed to derive generic real-world cycles that can be used for the evaluation of the overall energy efficiency of electrified powertrains. The vehicles were tested under standard real driving emissions routes, real-world routes with reversed order (compared to a standard real driving emissions route) of urban, rural, motorway, and routes with high slope variation. To enhance the experimental activities, additional virtual mission profiles simulated using vehicle simulation models. Outcome of the study consists of specific driving cycles, designed based on standard real-world route, and a methodology for real-world data analysis and evaluation, along with the results from the assessment of the impact of different operational parameters on the total electrified powertrain.

VP08 Real-World Data Use In Health Technology Assessments: A Comparison Of Five Health Technology Assessment Agencies

2017 ◽  
Vol 33 (S1) ◽  
pp. 149-150
Author(s):  
Amr Makady ◽  
Ard van Veelen ◽  
Anthonius de Boer ◽  
Hans Hillege ◽  
Olaf Klunger ◽  
...  
Keyword(s):  
Health Technology Assessment ◽  
Technology Assessment ◽  
Real World ◽  
Relative Effectiveness ◽  
Internal Validity ◽  
Health Technology ◽  
Real World Data ◽  
World Data ◽  
The Impact

INTRODUCTION:Reimbursement decisions are usually based on evidence from randomized controlled trials (RCT) with high internal validity but lower external validity. Real-World Data (RWD) may provide complimentary evidence for relative effectiveness assessments (REA's) and cost-effectiveness assessments (CEA's) of treatments. This study explores to which extent RWD is incorporated in REA's and CEA's of drugs used to treat metastatic melanoma (MM) by five Health Technology Assessment (HTA) agencies.METHODS:Dossiers for MM drugs published between 1 January 2011 and 31 December 2016 were retrieved for HTA agencies in five countries: the United Kingdom (NICE), Scotland (SMC), France (HAS), Germany (IQWiG) and the Netherlands (ZIN). A standardized data-extraction form was used to extract data on RWD mentioned in the assessment and its impact on appraisal (for example, positive, negative, neutral or unknown) for both REA and CEA.RESULTS:In total, fourty-nine dossiers were retrieved: NICE = 10, SMC = 13, IQWiG = 16, HAS = 8 and ZIN = 2. Nine dossiers (18.4 percent) included RWD in REA's for several parameters: to describe effectiveness (n = 5) and/or the safety (n = 2) of the drug, and/or the prevalence of MM (n = 4). CEA's were included in 25/49 dossiers (IQWiG and HAS did not perform CEA's). Of the twenty-five CEA's, twenty (80 percent) included RWD to extrapolate long-term effectiveness (n = 19), and/or identify costs associated with treatments (n = 7). When RWD was included in REA's (n = 9), its impact on the appraisal was negative (n = 4), neutral (n = 2), unknown (n = 1) or was not discussed in the appraisal (n = 2). When RWD was included in CEA's (n = 11), its impact on the appraisal varied between positive (n = 2), negative (n = 5) and unknown (n = 4).CONCLUSIONS:Generally, RWD is more often included in CEA's than REA's (80 percent versus 18.4 percent, respectively). When included, RWD was mostly used to describe the effectiveness of the drug (REA) or to predict long-term effectiveness (CEA). The impact of RWD on the appraisal varied greatly within both REA's and CEA's.

550 Real World Treatment Trends for Patients TBSA ≤ 20%: Evaluating Important Shifts in Care and Budget Impact of ASCS Adoption

2020 ◽  
Vol 41 (Supplement_1) ◽  
pp. S115-S115
Author(s):  
James H Holmes ◽  
Stacey Kowal ◽  
Cheryl P Ferrufino
Keyword(s):  
Survey Data ◽  
Resource Use ◽  
Real World ◽  
Budget Impact ◽  
Patient Characteristics ◽  
Real World Data ◽  
Treatment Pathways ◽  
World Data ◽  
Burn Care ◽  
The Impact

Abstract Introduction Treatment pathways in burn care are typically determined based on burn center (BC) and patient characteristics, although decisions may be influenced by anecdotal experience, personal preference, and hospital policies/purchasing decisions. Health economic (HE) evaluations can support improved decision-making, identifying the most cost-effective interventions for tailored care. A novel burn care model (BEACON) was developed with burn surgeons over several years and validated through numerous publications, including an assessment of the HE impacts of autologous skin cell suspension (ASCS) use for definitive burn closure. To ensure that BEACON accurately represents the current state of care, it is vital to update data that underpins model projections. This study collected real world data on practice patterns and patient outcomes for the most commonly seen burns (TBSA ≤ 20%) to update the current understanding of standard of care (SOC) costs and outcomes and to refine estimates on the impact of ASCS use in TBSA ≤ 20% patients. Methods Data was collected from a 10% sample of BCs, including: BC and patient characteristics, resource use, inpatient costs, and length of stay (LOS). NBR based inputs in BEACON were updated to reflect survey data for patients with TBSA ≤ 20%, with the ability to view data as a national aggregate sample and across BC characteristics. BEACON estimates patient and BC costs and outcomes across a spectrum of patient profiles (age, gender, inhalation injury, comorbidity status, burn depth, TBSA) and combines information on each patient profile to understand annual budget impact. Key outcomes were compared across the survey sample and published NBR trends. Using the updated BEACON, the BC budget impact of ASCS in burns TBSA ≤ 20% was assessed. Results The survey was collected from 16+ BCs, focusing on inpatient encounters in 2018. LOS was lower than NBR estimates, with some centers reporting LOS per %TBSA far below 1 d/%TBSA. Using the detailed bottom-up estimation of cost from BEACON with survey data, trends suggest total hospital costs for SOC are lower than published NBR charges given shorter LOS and updated cost and resource use assumption. Conclusions Compared to NBR 8.0, contemporary data suggests that fewer small TBSA burns are being treated in the inpatient setting; those treated have a LOS below NBR estimates. When using real world data, the impact of ASCS use in burns TBSA ≤ 20% was still calculated to be cost saving to a BC overall, given reductions in LOS and number of definitive closure procedures. Incorporating ASCS into appropriate TBSA ≤ 20% procedures can still result in a positive financial impact for BCs. Applicability of Research to Practice

scholarly journals Economic interventions to ameliorate the impact of COVID-19 on the economy and health: an international comparison

2020 ◽  
Author(s):  
Shaun Danielli ◽  
Raman Patria ◽  
Patrice Donnelly ◽  
Hutan Ashrafian ◽  
Ara Darzi
Keyword(s):  
Life Expectancy ◽  
Gross Domestic Product ◽  
International Comparison ◽  
Real World ◽  
Domestic Product ◽  
Real World Data ◽  
Economic Collapse ◽  
Available Information ◽  
The Impact ◽  
Second Wave

Abstract Background The COVID-19 pandemic continues to challenge governments and policymakers worldwide. They have rightfully prioritised reducing the spread of the virus through social distancing interventions. However, shuttered business and widespread restrictions on travel and mobility have led to an economic collapse with increasing uncertainty of how quickly recovery will be achieved. Methods The authors carried out a review of publicly available information on the economic intervention’s countries have put in place to ameliorate the impact of COVID-19. Results The strategies and scale of economic interventions have been broad, ranging from 2.5% to a reported 50% of Gross Domestic Product. Conclusions Numerous countries are beginning to ease lockdown restrictions and restart economies in different ways. There is therefore evolving, real-world data that should be used dynamically by governments and policymakers. The strategies on restarting the economy must be balanced against the uncertainty of a possible second wave of COVID-19. A nuanced approach to easing restrictions needs to take into account not only immediate risk to life but longer-term risks of widening inequalities and falling life expectancy.

Sign in / Sign up

Export Citation Format

Share Document