Commercial Dataset Research Articles

Economic Burden Associated with Pulmonary Arterial Hypertension in the United States.

Pulmonary arterial hypertension (PAH) is a progressive disease characterized by elevated pressure in the pulmonary arteries, commonly resulting in right heart failure. PAH is associated with a high economic burden throughout the duration of the disease. This retrospective cohort study of the Milliman Contributor Health Source Data, the Medicare 100% Research Identifiable Files, and the Merative Marketscan® Commercial dataset between 2018 and 2020 identified adult patients with prevalent PAH based on the earliest qualifying diagnosis date or medication date ('index date') between January 1, 2019 and November 30, 2020. Outcomes were assessed using patient data from index date through the earliest of end of enrollment, end of data, or death (Medicare fee-for-service [FFS] only). All-cause and PAH-related medical and pharmacy costs per-patient per-month (PPPM) and healthcare resource utilization per 1000 patients were summarized. The study included 11,670 Medicare FFS, 1021 Medicare Advantage, 274 Medicaid, and 1174 commercially insured patients in the US. The annual national burden to payers was estimated to be US$3.1 billion. The PPPM payer costs ranged from US$6500 to US$14,742; out-of-pocket (OOP) costs ranged from US$341 to US$907 PPPM. Inpatient utilization rate ranged from 435 to 770 per 1000 patients for all-cause admissions and from 15 to 58 per 1000 patients for PAH-related admissions. This study demonstrates that PAH continues to be associated with a high economic burden and healthcare resource utilization across all payer types within the US healthcare system.

The art of time-bending: Data augmentation and early prediction for efficient traffic classification

The accurate identification of internet traffic is crucial for network management. However, the use of encryption techniques and constant changes in network protocols make it difficult to extract useful features for traffic classification. Additionally, there may be limited data availability and a lack of diversity within the dataset, which poses further challenges. To address these issues, our research proposes a novel solution that uses an innovative data augmentation technique. This approach leverages the capabilities of LSTM networks to create synthetic data points that closely resemble real traffic data. By doing so, we can significantly enrich the dataset used for training and improve classification efficiency. We conducted thorough experiments to validate our approach and found that combining LSTM-generated data with actual traffic data leads to notable improvements in classification efficiency. We demonstrated the effectiveness of our methodology using academic and commercial datasets. Our classifier, trained on the generated data, showed a performance boost of 6%. Moreover, when classifying with only half of the time, thus utilizing half of the signal, our approach achieved a notable 4% improvement compared to the original classifier. The inclusion of augmented samples within the training set led to a noticeable improvement in both accuracy and F1-score. These findings compellingly demonstrate our data augmentation strategy’s practical utility and efficiency in earlier prediction with improved performance for encrypted traffic classification systems.

Imaging Pleistocene volcanic edifices along the Egyptian Red Sea margin: Insights from reflection seismics and 3D constrained inversion of gravity and magnetic data

Several volcanic buildups have been documented using gravity and magnetic data at specific locations in the Northern Red Sea (NRS). Most of these volcanoes were never sampled, and only a few were imaged by seismic or bathymetry data. Furthermore, the confidentiality of commercial datasets does not allow adequate knowledge of these structures and their morphology and physical properties. In this study, 2D and 3D seismic reflection data and 3D constrained inversion of magnetic and gravity data have been used to infer densities and magnetic susceptibilities of volcanic features in the central part of the NRS and to evaluate the distribution and geometry. This study indicates that the results of the 3D inverted models show a good fit with the volcanic structures inferred from seismic data and, thus, that the inversion technique adopted can be a reliable tool for detecting other edifices where seismic data are missing. The 3D inverted density and magnetic susceptibility models suggest that at least eight isolated structures with similar gravity and magnetic properties exist in the study area. They are associated with high densities and magnetic susceptibility anomalies with average values of 2700 kg/m3 and 0.025 SI, respectively. These magmatic structures are deeply rooted in the sedimentary cover and emerge at elevations of 100 to >650 m relative to the seabed. They developed in a sector interpreted as containing thinned continental crust, accompanied by the ascension of magma through prominent rift-related normal faults. Magmatic activity likely started in the Pleistocene and continues to the present day.

Relationships Between Alcohol Policies and Infant Morbidities and Injuries

Previous research has found that policies specifically focused on pregnant people's alcohol use are largely ineffective. Therefore, the purpose of this study is to analyze the relationships between general population policies regulating alcohol physical availability and outcomes related to pregnant people's alcohol use, specifically infant morbidities and injuries. Outcome data were obtained from Merative MarketScan, a longitudinal commercial insurance claims data set. Policy data were obtained from the National Institute on Alcohol Abuse and Alcoholism's Alcohol Policy Information System, the National Alcohol Beverage Control Association, and Liquor Handbooks and merged using policies in effect during the estimated year of conception. Relationships between state-level policies regulating sites, days/hours, and government monopoly of liquor sales and infant morbidities and injuries were examined. Analyses used logistic regression with individual controls, fixed effects for state and year, state-specific time trends, and SEs clustered by state. The study analysis was conducted from 2021 to 2023. The analytic sample included 1,432,979 infant-birthing person pairs, specifically people aged 25-50 years who gave birth to a singleton between 2006 and 2019. A total of 3.1% of infants had a morbidity and 2.1% of infants had an injury. State government monopoly on liquor sales was associated with reduced odds of infant morbidities and injuries, whereas gas station liquor sales were associated with increased odds of infant morbidities and injuries. Allowing liquor sales after 10PM was associated with increased odds for infant injuries. No effect was found for allowing liquor sales in grocery stores or on Sundays. Findings suggest that limiting alcohol availability for the general population may help reduce adverse infant outcomes related to pregnant people's alcohol use.

Monitoring health and nutrition claims on food labels in Brazil.

The monitoring of nutrition and health claims on food and beverage labels has been proposed by international and national organizations because it can collaborate with the development of public policies to regulate food labeling and marketing strategies. One way of carrying out this monitoring is by using data collected by private companies. To compare information on nutrition and health claims available in a commercial database of a private company that monitors the launch of new foods and beverages in Brazilian food retailers with information on those same claims manually coded by trained research assistants. This is a cross-sectional observational study using a data sample of newly launched food and beverages available at a commercial database from 2018 to 2021. We compared the information on health and nutrition claims available on the commercial dataset with reliable information on the same nutrition and health claims manually coded by trained research assistants using a tested taxonomy to classify such claims. We used Gwet's Kappa AC1 with 95% CI and percentage agreement to compare both data sources and calculated sensitivity and specificity of the compared data. A total of 6,722 foods and beverages were analyzed. Mintel-GNPD presented 36.28% (n = 2,439) of nutrition claims, while in the trained researchers' coding, it was 33.73% (n = 2,267). We found a prevalence of 5.4% (n = 362) for health claims in Mintel-GNPD and 10.8% (n = 723) in the researchers' coding. All subcategories of nutrition and health claims showed high agreement (Kappa >0.81). Health claims presented kappa = 0.89 with 33.7% sensitivity and 98.0% specificity while nutrition claims showed kappa = 0.86 with 92.9% sensitivity and 92.5% specificity. Nutrition and health claims showed high agreement, with great results in nutrition claims, indicating that Mintel-GPND is suitable for monitoring such claims on food and beverage packaging in Brazil. Additionally, our findings show a high prevalence of nutrition and health claims on food packages launched in the Brazilian food retail, highlighting the need to monitor these to develop public policies to regulate food marketing on packaging in Brazil.

Open-Source, Step-Counting Algorithm for Smartphone Data Collected in Clinical and Nonclinical Settings: Algorithm Development and Validation Study.

Step counts are increasingly used in public health and clinical research to assess well-being, lifestyle, and health status. However, estimating step counts using commercial activity trackers has several limitations, including a lack of reproducibility, generalizability, and scalability. Smartphones are a potentially promising alternative, but their step-counting algorithms require robust validation that accounts for temporal sensor body location, individual gait characteristics, and heterogeneous health states. Our goal was to evaluate an open-source, step-counting method for smartphones under various measurement conditions against step counts estimated from data collected simultaneously from different body locations ("cross-body" validation), manually ascertained ground truth ("visually assessed" validation), and step counts from a commercial activity tracker (Fitbit Charge 2) in patients with advanced cancer ("commercial wearable" validation). We used 8 independent data sets collected in controlled, semicontrolled, and free-living environments with different devices (primarily Android smartphones and wearable accelerometers) carried at typical body locations. A total of 5 data sets (n=103) were used for cross-body validation, 2 data sets (n=107) for visually assessed validation, and 1 data set (n=45) was used for commercial wearable validation. In each scenario, step counts were estimated using a previously published step-counting method for smartphones that uses raw subsecond-level accelerometer data. We calculated the mean bias and limits of agreement (LoA) between step count estimates and validation criteria using Bland-Altman analysis. In the cross-body validation data sets, participants performed 751.7 (SD 581.2) steps, and the mean bias was -7.2 (LoA -47.6, 33.3) steps, or -0.5%. In the visually assessed validation data sets, the ground truth step count was 367.4 (SD 359.4) steps, while the mean bias was -0.4 (LoA -75.2, 74.3) steps, or 0.1%. In the commercial wearable validation data set, Fitbit devices indicated mean step counts of 1931.2 (SD 2338.4), while the calculated bias was equal to -67.1 (LoA -603.8, 469.7) steps, or a difference of 3.4%. This study demonstrates that our open-source, step-counting method for smartphone data provides reliable step counts across sensor locations, measurement scenarios, and populations, including healthy adults and patients with cancer.

Development of an Algorithm to Evaluate the Quality of Geolocated Addresses in Urban Areas

The spatial and semantic data of geographic addresses are extremely important for citizens, governments, and companies. The addresses can georeference environmental, economic, security, health, and demographic parameters in urban areas. Additionally, address components can be used by users to locate any point of interest (POI) with location-based systems (LBSs). For this reason, errors in address data can affect the geographic location of events, map representations, and spatial analyses. Thus, this paper presents the development of an algorithm for evaluating the quality of semantic and geographic information in any geospatial address dataset. The reference datasets are accessible using open data platforms or spatial data infrastructure (SDI) and volunteered geographic information (VGI), and both have been compared with commercial datasets using geocoding web services. Address quality analysis was developed using several open-source data science code libraries combined with spatial databases and geographic information systems. In addition, the quality of geographic addresses was evaluated by carrying out normalized tests in accordance with International Geospatial Standards (ISO 19157). Finally, this methodology assesses the quality of authorized and VGI address datasets that can be used for geocoding any relevant information in specific urban areas.

Obtaining reference values for nutrients in vineyard soils through boundary line approach using Bayesian segmented quantile regression on commercial farm data

Inefficient use of fertilizers in agriculture is a major issue in global food production as it hampers economic viability of farms and offers great risk of environmental pollution. Fertilizing recommendations are guided by plant nutritional assessments, which are compared to reference values so that decisions can be made regarding which nutrient and how much is to be supplied. Traditionally, these values are obtained by field trials that are particularly long and expensive for perennial crops such as grapevines. This research evaluated the use of boundary line (BL) approach and Bayesian segmented quantile regression (BSQR) in determining reference values for nutrients in vineyard soils using commercial farm data. Critical levels (CL) and sufficiency ranges (SR) of phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), sulfur (S), copper (Cu), zinc (Zn), boron (B) and manganese (Mn) were estimated using datasets from a winery in the Campanha Gaúcha region, Rio Grande do Sul, Brazil. Results indicated lower P and higher K, S, Zn and Mn levels, as compared to current state recommendations. Great inconsistencies were found in CL estimates for Cu and, to a lesser extend, Ca and Mg. Model assessments indicated that BSQR estimates were highly sensitive to quantile selection but poorly affected by changes in the prior. Data simulated using fitted models were consistently over-dispersed. Allowing for variable slopes and intercepts may partition data variability more properly, leading to better fit, higher parameter precision and greater accuracy of estimates. Despite current limitations, the BSQR model provides a flexible framework for BL analysis, enabling the use of commercial datasets in establishing reference values to support group- and site-specific recommendations, a valuable contribution to increase efficiency and sustainability of grapevine fertilization.

Nationwide Demand Modeling for an Urban Air Mobility Commuting Mission

In this paper, we present a comprehensive and reproducible urban air mobility (UAM) demand model centered around publicly available data and open source tools capable of demand estimation at the national level. A discrete mode-choice demand model is developed using longitudinal origin–destination employment statistics flow data, American community survey economic data, and the Open Source Routing Machine (OSRM) to identify the utility of a UAM commuter service relative to other modes of transportation. Using the implemented model, we identify New York City, San Francisco, and Los Angeles as cities with the highest potential commuter demand, and Seattle as the city most resilient to increases in delay time. A sensitivity study of demand is performed and shows that strong demand exists for short trips with low total delay times and for longer trips with a low ticket price per kilometer, with the former showing resilience to increases in operational costs and the latter showing resilience to increases in delays. The demand model is supported by a speed-flow model, which fuses highway performance monitoring system data with OpenStreetMap data to provide traffic-adjusted road segment speeds to OSRM. The speed-flow model has the capability of providing congestion data for road segments across the United States without the use of commercial data sets or routing services and is shown to improve routing duration accuracy in congested regions.

A DYNAMIC APPROACH FOR LIFE CYCLE ASSESSMENT. THE CASE OF DOMESTIC REFRIGERATORS

AbstractThe Life Cycle Assessment is a well-stated methodology whose application has recently spread over a multitude of sectors. Thus the need for very accurate and reliable analysis. The present work investigates how to achieve reliable and faithful results while still maintaining a micro-systemic approach and how to handle the evolution of the real cases through commercial solutions available. The works present an innovative dynamic approach that aims at filling the discrepancy between the attributional Life Cycle Assessment which is focused on the product at the point to appear short-sighted and isolated from the surrounding evolving system and the consequential, which is willing to include the consequences of the evolution of the surrounding system, with increased complexity. The approach is applied to the case of a domestic refrigerator; the application reveals a discrepancy of 16% between the results of the dynamic and attributional analysis and registered doubled environmental impacts than the consequential, carried out with the support of commercial datasets. The approach respects the 5 main criteria for methods in environmental systems analysis, namely feasibility, accuracy, easiness in communication, inspiration, robustness.

The Relationship Between the Timing of Knee Osteoarthritis Diagnoses and Arthroscopic Partial Meniscectomy.

There is ongoing debate regarding the efficacy of arthroscopic partial meniscectomy (APM) for meniscus tears in patients with knee osteoarthritis (OA). Some insurance payers will not authorize APM in patients with knee OA. The purpose of this study was to assess the timing of knee OA diagnoses in patients undergoing APM. A large commercial national claims data set containing deidentified information from October 2016 to December 2020 was used to identify patients undergoing arthroscopic partial meniscectomy. Data were analyzed to determine whether patients in this group had a diagnosis of knee OA within 12 months before surgery and for the presence of a new diagnosis of knee OA at 3, 6, and 12 months after APM. Five lakhs thousand nine hundred twenty-two patients with a mean age of 54.0 ± 8.52 years, with the majority female (52.0%), were included. A total of 197,871 patients underwent APM without a diagnosis of knee OA at the time of the procedure. Of these patients, 109,427 (55.3%) had a previous diagnosis of knee OA within 12 months preceding surgery, and 24,536 (12.4%), 15,596 (7.9%), and 13,301 (6.7%) patients were diagnosed with knee OA at 3, 6, and 12 months after surgery, respectively. Despite evidence against APM in patients with knee OA, more than half of the patients (55.3%) had a previous diagnosis of OA within 12 months of surgery and 27.0% received a new diagnosis of knee OA within one year of surgery. A notable number of patients had a diagnosis of knee OA either before or shortly after APM.

Simulation‐data‐driven load disaggregation based on multi‐channel neural network for industrial and commercial users

AbstractThe immeasurable energy‐saving potential of industrial parks and commercial buildings with unprecedented growth in scale and quantity can be well exploited by load disaggregation, a promising technique for demand‐side refinement management. However, traditional load disaggregation studies mainly focus on residential users and neglect the collection and analysis of industrial and commercial loads. This study employs a simulated response model of versatile load categories with voltage‐dependent load functions to overcome the unavailability of industrial and commercial datasets. Specially, a weighting factors allocation method considering probability statistics of power consumption distribution characteristics is proposed to construct a typical daily multi‐type load aggregation model. Furthermore, an intelligent load disaggregation strategy is developed based on a multi‐channel neural network to constructively improve multi‐feature tracking and convergence efficiency. Finally, the simulation results of the detailed study case confirm the tractability and effectiveness of the proposed approach to carry out load identification among multi‐type users, with an accuracy of 94.9% and errors of individual categories controlled between 3.32% and 10.686%. Moreover, additional interference tests demonstrate that the developed model has superiority over common load disaggregation structures, with better tolerance to exceptional voltage fluctuation and noise.

Optimal Knowledge Component Extracting Model for Knowledge-Concept Graph Completion in Education

As people have become accustomed to non-face-to-face education because of the COVID-19 pandemic, adaptive and personalized learning is being emphasized in the field of education. Learning paths suitable for each student may differ from those normally provided by teachers. To support coaching based on the concept of adaptive learning, the first step is to discover the relationships among the concepts in the curriculum provided in the form of a knowledge graph. In this study, feature reduction for the target knowledge-concept was first performed using Elastic Net and Random Forest algorithms, which are known to have the best performance in machine learning. Deep knowledge tracing (DKT) in the form of a dual-net, which is more efficient because of the already slimmer data, was then applied to increase the accuracy of feature selection. The new approach, termed the optimal knowledge component extracting (OKCE) model, was proven to be superior to a feature reduction approach using only Elastic Net and Random Forest using both open and commercial datasets. Finally, the OKCE model showed a meaningful knowledge-concept graph that could help teachers in adaptive and personalized learning.

Proposed variable sampling interval maximum EWMA and distance EWMA charts with unknown process parameters

The variable sampling interval (VSI) exponentially weighted moving average (EWMA) chart which varies the chart's sampling interval according to the value of the current plotting statistic increases the speed of the standard EWMA chart in detecting shifts. Joint monitoring schemes use a single combined statistic for the mean and variance in process monitoring. To simultaneously monitor the mean and variance of a process from the normal distribution, two VSI EWMA schemes with unknown process parameters, based on (i) Maximum (Max) and (ii) Distance (Dis) type combining functions, are proposed in this paper. Each of these schemes uses a single plotting statistic. The effects of parameter estimation on the performance of the proposed VSI Max EWMA and VSI Dis EWMA schemes, in terms of the average time to signal, standard deviation of the time to signal, expected average time to signal and median time to signal criteria, are studied using Monte Carlo simulation. The results show that the proposed schemes can identify process shifts quicker than the existing Max/Dis Shewhart (SH), Max/Dis cumulative sum (CUSUM) and Max/Dis EWMA schemes. The implementation of the proposed schemes is demonstrated using a commercial dataset.

The Effect of Intra-articular Hyaluronic Acid Injections and Payer Coverage on Total Knee Arthroplasty Procedures: Evidence From Large US Claims Database

BackgroundThere is debate regarding the efficacy of intra-articular (IA) hyaluronic acid (HA) injections for the management of knee osteoarthritis (OA). This study aimed to determine if IA HA utilization and payer coverage of viscosupplementation affected the prevalence of total knee arthroplasty (TKA) procedures and the age of TKA patients. MethodsWe performed a retrospective analysis from 2014 to 2020 using a large national commercial claims data set. We analyzed the number of TKA procedures and the age of the patients in states that covered IA HA vs those with limited coverage. Mixed random effects and slopes models were used to identify the impact of the IA HA injections. ResultsOf 7,335,301 patients with knee OA, 440,606 (6.0%) received a TKA procedure at an average age of 59 years. The rate of TKA procedures increased by 0.56% per year (95% confidence interval [CI] 0.46-0.66; P < .001). Payer coverage of IA HA injections had no effect on TKA prevalence (P = .926). The age of surgical patients increased yearly by 0.15 years (95% CI 0.12-0.18; P < .001), regardless of IA HA injections (P = .990). After controlling for demographics and comorbidities, patients that received an IA HA injection had a higher probability of receiving a subsequent TKA (odds ratio = 2.83; 95% CI 2.80-2.87; P < .001); this finding was conditional of patients' age at the first diagnosis of knee OA. ConclusionsAdditional clinical trials should be employed to identify the role of HA injections in the treatment armamentarium for knee OA.

Comparative Analysis of Incomplete Business Data Clustering

Incomplete values can significantly reduce the accuracy and usability of missing data. In particular, in analyzing commercial data sets, missing values often lead to the dilemma of data selection. It means that a common way to deal with missing data is to delete the sample that contains the missing attribute. However, this can lead to biased and invalidated conclusions, as some data are too critical to be omitted. Therefore, we should use some method to fill the data set rather than delete the data with missing values. The filling of missing data is divided into supervised learning and unsupervised learning. This paper compares six benchmark business datasets by adopting several different data imputation methods and supplementing the missing data with a clustering approach (unsupervised learning). The results are guided to dealing with incomplete business data.

Ownership diversity and fragmentation: A barrier to urban centre resilience

Fragmentation of ownership has long been a recognised constraint to UK city centre development, a complexity that is growing in significance as centres try to manage the decline in physical retailing and transform obsolete retail units. Yet, our understanding of the structure of ownership and how that might be facilitating or inhibiting urban change remains weak. In this paper, the objective is to address this gap by examining the structure and diversity of land ownership in five retailing centres - Edinburgh, Glasgow, Hull, Liverpool, and Nottingham – between 2000–2017 using original databases created by linking administrative and commercial property data sets. Overall, the analysis finds property ownership to be spatially complex with ownership richness and diversity generally rising over the study period. The study also reveals that ownership structure has been shifting away from financial institutions towards overseas investors, private individuals and unlisted property companies, implying greater fragmentation of ownership. While the greater diversity in ownership should stimulate competition and innovation in property market practices, the shift in balance from equity-rich larger investors towards smaller and sometimes unknown investors makes urban centre management harder to manage. This suggests policymakers need to rethink the urban governance model to find a better way to galvanise the actions of this increasing disparate group of stakeholders if their visions of more resilient, mixed use city centres are to be realised.

Real-world treatment (tx) patterns and economic burden of patients (pts) with marginal zone lymphoma (MZL).

e18730 Background: MZL remains understudied with limited real-world evidence on how current tx conform to clinical guidelines. This study aimed to assess real-world tx patterns, costs, and healthcare resource utilization (HRU) in US MZL pts. Methods: A retrospective, observational study was conducted using the IBM MarketScan® commercial and Medicare supplemental claims dataset (2017-2020). Newly diagnosed adult MZL pts (≥18 yr) continuously enrolled 6 month pre- and 3 month post-index date, defined as the first diagnosis date, were included. Tx regimens were identified by line of therapy and mutually exclusively categorized as rituximab monotherapy (R-mono), bendamustine + rituximab (BR), rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP), ibrutinib (ibr), or other. Descriptive analyses were conducted to assess pt sociodemographic/clinical characteristics and tx utilization patterns (frequency, duration, discontinuation). Costs and HRU assessed included inpatient, outpatient, and pharmacy visits per-patient-per-month (PPPM). Multivariable logistic regression was conducted to examine predictors of costs and HRU. Results: In 2491 MZL pts (median age = 63; 52% male), 59% were commercially insured (median age = 57) and 41% in Medicare (median age = 76). Common comorbidities were hypertension (44%), diabetes (17%), atrial fibrillation (AF; 16%), and gastroesophageal reflux disease (15%). Mean time from diagnosis to tx initiation was 223 days. A total of 1,781 (72%) pts received first-line (1L), 518 (29%) pts received second-line (2L) and 239 (13%) pts received third-line (3L) therapies. R-mono was the most common regimen across both commercial and Medicare pts and all tx lines (Table). R-CHOP and BR were the second most used regimen in 1L with decreased use in 2L+. Ibr was used more in 2L+ setting but had the lowest 1L PPPM cost (median $2958.9) than other regimens. Overall MZL pts had PPPM 4.6 outpatient visits, 0.5 hospitalization, and mean length of stay of 2.6 days. Total PPPM healthcare cost was $19,895.8. Multivariable regression showed that baseline comorbidities (AF, renal disease, neutropenia) and tx discontinuation were significant predictors of higher costs and HRU. Conclusions: This real-world data suggested that US MZL real-world tx pattern across lines of therapy follows the regimen recommendations by the National Comprehensive Cancer Network clinical practice guidelines and that MZL pts incur high economic burden. Future studies are needed to evaluate long-term outcomes and the impact of heterogenous MZL subtypes.[Table: see text]

From triple-class exposed (TCE) to penta-exposed and triple-class refractory (PE-TCR): Real-world outcomes across a spectrum of advanced relapsed or refractory multiple myeloma (RRMM) patients (Pts) in the United States.

e20011 Background: RRMM treatment (Tx) has been transformed by novel Txs including anti-CD38 monoclonal antibodies (mAbs) and 2nd/3rd generation immunomodulatory drugs (IMiDs) and proteasome inhibitors (PIs), resulting in an increasing population of TCE pts (i.e., prior exposure to ≥1 each of an IMID, a PI, and an anti-CD38 mAb). Many such pts are exposed to ≥5 drugs in these classes (i.e., “penta-exposed” PE: ≥2 IMIDs, ≥2 PIs, and an anti-CD38 mAb). Some are refractory to ≥1 drug in each class (“triple-class refractory” TCR) and some are both PE and TCR (PE-TCR). Real-world outcomes data for the spectrum of these advanced RRMM pts are limited. Methods: A US commercial insurance claims dataset (IQVIA PharMetrics Plus®) spanning 2007-20 was used to examine outcomes in pts with RRMM (≥1 prior line of therapy [LOT]) who were TCE, TCR, PE, and PE-TCR. The 1st LOT after becoming TCE or PE was defined as the “index LOT” (start date of index LOT = “index”). Subsets of TCE and PE pts who were also TCR at index were identified. To ensure complete documentation of treatment history, continuous enrollment for ≥6-mos before 1st claim with MM diagnosis was required. Outcomes assessed included Txs received, time to next Tx (TTNT), % with 2nd primary malignancy (SPM), healthcare resource use (HRU), and costs. Results: 500 TCE and 217 PE pts were identified including 228 TCR and 142 PE-TCR. Mean age across groups was 62-63 y; 39-42% were female, 40% (TCR) to 46% (PE) had prior autologous stem cell transplant; mean number of prior LOTs ranged from 3.2 (TCE ) to 4.3 (PE); 5.7% (TCR) to 8.8% (PE) had extramedullary disease at index. The no. of new TCE pts increased by ~6X from 25 in 2016 to 145 in 2020; similar increases were seen for TCR, PE, and PE-TCR. Mean post-index follow-up (mos) was 11.6 for TCE, 11.7 for TCR, 10.7 for PE, and 10.3 for PE-TCR. There was no standard Tx across cohorts; the most frequent index LOT regimens were pomalidomide (POM)+daratumumab (11% TCE), POM+carfilzomib (15% TCR), and POM+elotuzumab (7% PE and 9% PE-TCR). Kaplan-Meier median TTNT was short, ranging from 6.5 (TCR) to 11.3 mos. (TCE). TCE pts had high HRU with mean of 0.08 emergency department visit, 2.37 office/outpatient visits, 5.10 prescriptions, 0.14hospitalizations, and 1.2 days in hospital per month. Mean monthly all-cause healthcare costs (2021 $US) for TCE pts were $31,127, including $18,033 (58%) for MM medications and $10,297 (33%) for other MM-related care. Monthly HRU and costs for TCR, PE, and PE-TCR pts were similar. In TCE pts with no SPM before index date (N=291), 29 (10%) developed SPM post-index. Conclusions: Despite increasing numbers of TCE, TCR, PE, and PE-TCR pts, there is no apparent standard Tx. With current Tx, TTNT is short, HRU and costs are high, and risk of SPM is 10% (TCE). These data underscore the unmet need for new therapies in this growing population.

Short‐term commercial load forecasting based on peak‐valley features with the TSA‐ELM model

AbstractCommercial buildings are consuming an increasing amount of energy, and accurate load demand forecasting is critical for the reliable operation of power systems and the efficient use of resources. Therefore, in this paper, a short‐term commercial load forecasting model based on tunicate swarm algorithm (TSA) combined with an extreme learning machine (ELM) under peak‐valley features is proposed as a research case for a shopping mall in Romania. This paper's overall structure is divided into two steps. In the first step, the 24‐h day is divided into six periods by analyzing the daily load characteristics of the training set, and the peak and valley loads are obtained. The ELM optimized by TSA (TSA‐ELM) algorithm is then used to forecast the peak and valley values of the test set one day ahead. In the second step, the actual load, peak, and valley for the previous week of historical load are chosen using the maximum information coefficient (MIC). Following that, the MIC ≥ 0.8 features are added to the TSA‐ELM to achieve short‐term commercial electricity load forecasting. The results show that the PV (Peak &amp; Valley)‐TSA‐ELM model proposed in this paper has higher prediction accuracy compared with other models. Taking ELM as an example, compared with the traditional ELM, the mean absolute error, root mean square error, and mean absolute percentage error of PV‐TSA‐ELM are reduced by 20.59%, 20.13%, and 19.19% on average in the three commercial data sets. The proposed model is validated with an industrial data set, and ideal results are obtained, which verifies the effectiveness and superiority of the proposed method.