Internal Software Research Articles

Warfarin versus direct oral anticoagulants for patients needing distal deep vein thrombosis treatment

ObjectiveGreat interest exists in standardizing the anticoagulant choice for patients requiring treatment of distal deep vein thrombosis (DDVT). In the present multicenter, retrospective cohort study, we evaluated the outcomes of patients with DDVT who had been treated with warfarin vs direct oral anticoagulants (DOACs; ie, rivaroxaban, apixaban, edoxaban, dabigatran). MethodsQueries were built for the TriNetX database (TriNetX LLC, Cambridge, Mass), a federated network of healthcare organizations across the United States that provides de-identified patient data through aggregated counts and statistical summaries. International Classification of Diseases, 10th revision, diagnostic codes were used to identify eligible patients. Data from January 1, 2013 to January 1, 2020 were reviewed. Statistical analyses, including propensity matching, were performed using TriNetX’s internal software. The inclusion criterion was treatment with either warfarin or a DOAC started within the first 24 hours of diagnosis of an isolated thrombosis of the following veins: anterior tibial, posterior tibial, peroneal, or calf muscular veins. The exclusion criteria were a history of an adverse reaction to anticoagulant agents, SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2) infection, thrombophilia, mechanical heart valve, chronic proximal DVT (PDVT) and/or DDVT, and 6-month history of the following: acute PDVT, pulmonary embolism (PE), or anticoagulant usage. The outcomes measured included the incidence of mortality, PE, PDVT, stroke, myocardial infarction, and major bleeding within 6 months after initiating anticoagulation therapy. ResultsIn a cohort of 6509 patients, 1570 were treated with warfarin and 4939 were treated with a DOAC drug. After propensity matching for age, sex, ethnicity, and comorbidities, the DOAC cohort had a significantly lower incidence of PE (1.795% vs 3.590%; P = .0020) and major bleeding (7.949% vs 10.513%; P = .0134). Differences in the incidence of mortality, PDVT, myocardial infarction, and stroke were not statistically significant. ConclusionsBefore the present study, no strong evidence was available to suggest an optimal treatment modality for DDVT requiring anticoagulation therapy. The data from the present study suggest that patients receiving DOACs for the treatment of DDVT will have significantly lower rates of progression to PE and a lower incidence of major bleeding compared with patients receiving warfarin. This suggests that DOACs are superior to warfarin for treatment of DDVT.

Methodology for Neural Network-Based Material Card Calibration Using LS-DYNA MAT_187_SAMP-1 Considering Failure with GISSMO.

A neural network (NN)-based method is presented in this paper which allows the identification of parameters for material cards used in Finite Element simulations. Contrary to the conventionally used computationally intensive material parameter identification (MPI) by numerical optimization with internal or commercial software, a machine learning (ML)-based method is time saving when used repeatedly. Within this article, a self-developed ML-based Python framework is presented, which offers advantages, especially in the development of structural components in early development phases. In this procedure, different machine learning methods are used and adapted to the specific MPI problem considered herein. Using the developed NN-based and the common optimization-based method with LS-OPT, the material parameters of the LS-DYNA material card MAT_187_SAMP-1 and the failure model GISSMO were exemplarily calibrated for a virtually generated test dataset. Parameters for the description of elasticity, plasticity, tension–compression asymmetry, variable plastic Poisson’s ratio (VPPR), strain rate dependency and failure were taken into account. The focus of this paper is on performing a comparative study of the two different MPI methods with varying settings (algorithms, hyperparameters, etc.). Furthermore, the applicability of the NN-based procedure for the specific usage of both material cards was investigated. The studies reveal the general applicability for the calibration of a complex material card by the example of the used MAT_187_SAMP-1.

Mapping of choriocapillaris vascular density in young and aged healthy subjects.

To compare macular vascular density (VD) of the choriocapillaris (CC) between young and aged healthy individuals. A cross-sectional study was performed enrolling young and senior healthy individuals of Caucasian race and an axial length (AL) lower than 26 mm, and without systemic or ophthalmological diseases. CC VD was imaged with DRI Triton OCTA using a 6 × 6 mm macular analysis. Internal software delimited CC boundaries and gave colour pictures, which were analysed and codified into numbers, and a grid of 30 × 30 VD values was obtained. Two-dimension (2D) and three-dimension (3D) representations were created. 53 eyes of 53 young healthy individuals and 30 eyes of 30 senior healthy individuals were enrolled. Mean age was 27.17 ± 3.90 years, and 67.00 ± 7.41 years, respectively (p < 0.001). There were no differences in intraocular pressure (IOP) or AL (23.73 ± 0.79 mm, 23.18 ± 0.80 mm, respectively, p = 0.24). There were differences in foveal VD and in temporal perifoveal macula, but not in nasal perifoveal macula. Foveal VD was the highest in both groups. Foveal CC VD has been found to be considerably high with this method, and it is the area which most decreases with age. Nasal perifoveal VD is not reduced in older individuals. These outcomes are opposite to other studies using different methods but they are in line with previous histological findings.

A longitudinal study of the impact of refactoring in android applications

Mobile applications have to continuously evolve in order to meet new user requirements and technological changes. Addressing these constraints may lead to poor implementation and design choices, known code smells. Code refactoring is a key practice that is employed to ensure that the intent of a code change is properly achieved without compromising internal software quality. While previous studies have investigated the impact of refactoring on traditional code smells in desktop applications, little attention has been paid to the impact of refactoring activities in mobile application development.

Aims and criteria for advancing technology management research at the Journal of Operations Management

Aims and criteria for advancing technology management research at the <scp>Journal of Operations Management</scp>

To the problem of using an automated workplace by people with disabilities

Introduction. Employees of the banking sector with health restrictions have negative experience of using internal software to interact with customers and perform their official duties. Many employees, for example, with hearing problems, would like to work in call centers, but do not have this opportunity due to the outdated software. The research objective is to analyze the priority tasks for the further development of software products, taking into account the existing health problems of employees.Materials and Methods. One of the subsystems of the automated workplace (hereinafter referred to as the AWP) was selected the software, which allows the employee to interact directly with the clients of the given organization. The analysis used the method of expert evaluation by T. L. Saati with the assistance of one of the experts in the development of software for people with disabilities.Results. Using the fundamental preference scale and expert opinion in the field of software development for people with disabilities, a priority matrix was built for each of the criteria (subtitles, simplified fonts, voice guidance, simplified and remote management) and platforms (IOS, Android, Windows OS), as well as a global priority matrix for all criteria and platforms.Discussions and Conclusions. An expert assessment of several characteristics of the software of a commercial banking organization of the Russian Federation was carried out to identify the disadvantages of using the software by employees with disabilities. During the analysis, intermediate conclusions were made: the most demanded criterion for people with hearing problems is “Subtitle”; for people without the ability to leave the house — “Remote control”; for people with amputations or irreversible limb injuries — “Simplified control”. The other parameters are not recommended for implementation.

Selection and Validation of Candidate Reference Genes for Gene Expression Analysis by RT-qPCR in Rubus.

Due to the lack of effective and stable reference genes, studies on functional genes in Rubus, a genus of economically important small berry crops, have been greatly limited. To select the best internal reference genes of different types, we selected four representative cultivars of blackberry and raspberry (red raspberry, yellow raspberry, and black raspberry) as the research material and used RT-qPCR technology combined with three internal stability analysis software programs (geNorm, NormFinder, and BestKeeper) to analyze 12 candidate reference genes for the stability of their expression. The number of most suitable internal reference genes for different cultivars, tissues, and fruit developmental stages of Rubus was calculated by geNorm software to be two. Based on the results obtained with the three software programs, the most stable genes in the different cultivars were RuEEF1A and Ru18S. Finally, to validate the reliability of selected reference genes, the expression pattern of the RuCYP73A gene was analyzed, and the results highlighted the importance of appropriate reference gene selection. RuEEF1A and Ru18S were screened as reference genes for their relatively stable expression, providing a reference for the further study of key functional genes in blackberry and raspberry and an effective tool for the analysis of differential gene expression.

The Agile Success Model

Organizations are increasingly adopting Agile frameworks for their internal software development. Cost reduction, rapid deployment, requirements and mental model alignment are typical reasons for an Agile transformation. This article presents an in-depth field study of a large-scale Agile transformation in a mission-critical environment, where stakeholders’ commitment was a critical success factor. The goal of such a transformation was to implement mission-oriented features, reducing costs and time to operate in critical scenarios. The project lasted several years and involved over 40 professionals. We report how a hierarchical and plan-driven organization exploited Agile methods to develop a Command &amp; Control (C2) system. Accordingly, we first abstract our experience, inducing a success model of general use for other comparable organizations by performing a post-mortem study. The goal of the inductive research process was to identify critical success factors and their relations. Finally, we validated and generalized our model through Partial Least Squares - Structural Equation Modelling, surveying 200 software engineers involved in similar projects. We conclude the article with data-driven recommendations concerning the management of Agile projects.

Back to the OR of the Future: How Do We Make It a Good One?

Back to the OR of the Future: How Do We Make It a Good One?

COMPARISON OF HIGH-LEVEL ATHLETE TRAINING BETWEEN CHINESE AND AMERICAN UNIVERSITIES BASED ON THE INTERNET OF THINGS

ABSTRACT In order to improve the training quality of high-level athletes in Chinese and American universities, an athlete’s sports information transmission model is designed based on the Internet of Things (IoT). The communication protocol between Ayla module and the main motion control board, Ayla module and client APP or cloud platform, and APP and cloud platform in the system is designed in detail. For the Ayla module, the most important hardware part of the system, the internal composition and software design are described. In the mobile phone client part that is closely related to the user, the MVC architecture is adopted, the singleton and agent design patterns are utilized, and the functional design of each part is elaborated, including APP interface animation, data transmission format, network communication, and database storage. The research results show that the system of this study can handle most of the athlete training information, and the prediction accuracy exceeds the traditional algorithm. This research study is of great significance for improving the training efficiency of high-level athletes and further expanding the scope of application of the IoT.

THEORETICAL ASPECTS OF THE FORMATION OF ACCOUNTANT’S PROFESSIONAL COMPETENCE IN WORKING ENVIRONMENT

The last few years have seen significant changes in the world economy and by extension in accounting. Modern accountants have an important role in global entrepreneurship. As a result, the majority of employers prefer candidates with advanced accounting or auditing experience, knowledge of IT, including accounting and internal audit software, and good foreign language skills. The article contains an analysis of the content and components of the professional competence of accountants working in Latvia. The theoretical review is based on the information gathered and analysed from open sources and scientific databases.

A hierarchical model for quantifying software security based on static analysis alerts and software metrics

Despite the acknowledged importance of quantitative security assessment in secure software development, current literature still lacks an efficient model for measuring internal software security risk. To this end, in this paper, we introduce a hierarchical security assessment model (SAM), able to assess the internal security level of software products based on low-level indicators, i.e., security-relevant static analysis alerts and software metrics. The model, following the guidelines of ISO/IEC 25010, and based on a set of thresholds and weights, systematically aggregates these low-level indicators in order to produce a high-level security score that reflects the internal security level of the analyzed software. The proposed model is practical, since it is fully automated and operationalized in the form of a standalone tool and as part of a broader Computer-Aided Software Engineering (CASE) platform. In order to enhance its reliability, the thresholds of the model were calibrated based on a repository of 100 popular software applications retrieved from Maven Repository. Furthermore, its weights were elicited in a way to chiefly reflect the knowledge expressed by the Common Weakness Enumeration (CWE), through a novel weights elicitation approach grounded on popular decision-making techniques. The proposed model was evaluated on a large repository of 150 open-source software applications retrieved from GitHub and 1200 classes retrieved from the OWASP Benchmark. The results of the experiments revealed the capacity of the proposed model to reliably assess internal security at both product level and class level of granularity, with sufficient discretion power. They also provide preliminary evidence for the ability of the model to be used as the basis for vulnerability prediction. To the best of our knowledge, this is the first fully automated, operationalized and sufficiently evaluated security assessment model in the modern literature.

Using Log Data to Measure Provider EHR Activity at a Cancer Center during Rapid Telemedicine Deployment.

Accurate metrics of provider activity within the electronic health record (EHR) are critical to understand workflow efficiency and target optimization initiatives. We utilized newly described, log-based core metrics at a tertiary cancer center during rapid escalation of telemedicine secondary to initial coronavirus disease-2019 (COVID-19) peak onset of social distancing restrictions at our medical center (COVID-19 peak). These metrics evaluate the impact on total EHR time, work outside of work, time on documentation, time on prescriptions, inbox time, teamwork for orders, and undivided attention patients receive during an encounter. Our study aims were to evaluate feasibility of implementing these metrics as an efficient tool to optimize provider workflow and to track impact on workflow to various provider groups, including physicians, advanced practice providers (APPs), and different medical divisions, during times of significant policy change in the treatment landscape. Data compilation and analysis was retrospectively performed in Tableau utilizing user and schedule data obtained from Cerner Millennium PowerChart and our internal scheduling software. We analyzed three distinct time periods: the 3 months prior to the initial COVID-19 peak, the 3 months during peak, and 3 months immediately post-peak. Application of early COVID-19 restrictions led to a significant increase of telemedicine encounters from baseline <1% up to 29.2% of all patient encounters. During initial peak period, there was a significant increase in total EHR time, work outside of work, time on documentation, and inbox time for providers. Overall APPs spent significantly more time in the EHR compared with physicians. All of the metrics returned to near baseline after the initial COVID-19 peak in our area. Our analysis showed that implementation of these core metrics is both feasible and can provide an accurate representation of provider EHR workflow adjustments during periods of change, while providing a basis for cross-vendor and cross-institutional analysis.

Monitoring of Erosive Tooth Wear with Intraoral Scanners In vitro

Intraoral scanners (IOS) have been used to quantify tooth wear, but so far they have not been systematically validated for monitoring of tissue loss. The aim of this in vitro study was to investigate whether progression of tissue loss can be detected with an IOS and whether IOS values agree with those obtained with noncontacting profilometry (PRO) serving as a standard method. Model jaws were mounted in a phantom head positioned in a dental chair. Flattened areas were prepared on the non-load-bearing cusps of the first molars (model teeth; n = 16) in order to fix flat enamel samples with an experimental area and a reference area. After baseline PRO and IOS, the experimental enamel area was stepwise etched with 35% H₃PO₄ gel (4 × 30 s and 4 × 15 s). After each etching, PRO and IOS was performed and the vertical tissue loss between the reference and experimental areas was measured, each at the same 3 measurement points. Furthermore, cupped cusps were simulated by stepwise preparation of the load-bearing cusps of the model teeth with a spherical diamond bur, and the maximum vertical depth after each preparation step was measured only by IOS. Trios3 (3Shape, Denmark), Carestream CS3600 (Carestream, USA) and an optical profilometer (MicroProf, Fries, Germany) were used to measure the flat areas of the enamel samples, whereas only IOS were used to measure curved surfaces on the load-bearing cupped cusps of the model teeth. The IOS data were analyzed with an external software (GOM Inspect, Germany) and with the respective internal IOS software. PRO revealed a mean (±SD) tissue loss of 17.1 ± 4.7 µm after 30-s etching steps and 10.1 ± 5.1 µm after the 15-s etching steps. IOS and software types were able to detect the progression of tissue loss after each etching step (p ≤ 0.001 each); Bland-Altmann plots revealed good agreement with PRO regardless of the order of tissue loss, and no systematic difference was found. Increasing cupped lesion depths were detected by all IOS, with no significant differences between IOS and analysis methods. IOS were able to detect small amounts of tissue loss under simulated clinical conditions and seem to be a promising tool for monitoring even initial erosive tooth wear.

On the use of textual feature extraction techniques to support the automated detection of refactoring documentation

Refactoring is the art of improving the internal structure of a program without altering its external behavior, and it is an important task when it comes to software maintainability. While existing studies have focused on the detection of refactoring operations by mining software repositories, little was done to understand how developers document their refactoring activities. Therefore, there is recent trend trying to detect developers documentation of refactoring, by manually analyzing their internal and external software documentation. However, these techniques are limited by their manual process, which hinders their scalability. Hence, in this study, we tackle the detection of refactoring documentation as binary classification problem. We focus on the automatic detection of refactoring activities in commit messages by relying on text-mining, natural language preprocessing, and supervised machine learning techniques. We design our tool to overcome the limitation of the manual process, previously proposed by existing studies, through exploring the transformation of commit messages into features that are used to train various models. For our evaluation, we use and compare five different binary classification algorithms, and we test the effectiveness of these models using an existing dataset of manually curated messages that are known to be documenting refactoring activities in the source code. The experiments are carried out with different data sizes and number of bits. As per our results, the combination of Chi-Squared with Bayes point machine and Fisher score with Bayes point machine could be the most efficient when it comes to automatically identifying refactoring text patterns in commit messages, with an accuracy of 0.96, and an F-score of 0.96.

Pharmacokinetics and Safety of Inhaled Oxytocin Compared with Intramuscular Oxytocin: The First Randomised Open-Label Study in Women in the Third Stage of Labour

Background: Intramuscular (IM) oxytocin for postpartum haemorrhage (PPH) prevention in resource-poor settings is limited by the need for cold-chain storage and skilled staff for safe injection. We compared the pharmacokinetics (PK) and safety of heat-stable oxytocin inhaled (IH) versus IM administration in women during third stage of labour (TSL). Methods: The phase 1, randomised, open-label clinical study (NCT02999100) was conducted across three centres in the UK and Australia. Participants (Group 1, women in TSL; Group 2, non-pregnant women of childbearing potential) were randomised 1:1 via validated GSK internal software: Group 1, oxytocin 240 IU (400 μg) IH or oxytocin 10 IU (17 μg) IM immediately after delivery; Group 2, intravenous (IV) oxytocin 5 IU (8·5 μg) and oxytocin 240 IU (400 μg) IH at two separate dosing sessions in a cross-over design. Primary endpoints included characterising oxytocin IH PK. The safety population included patients who received ≥1 dose; any PK sample obtained and analysed was included in the PK population. Additional investigations explored the validity of the PK concentration data. Findings: Participants were recruited between November 2016 and March 2019. In Group 1, 29 participants were randomised; 17 received IH (n=9) or IM (n=8) oxytocin. After IH and IM administration most plasma oxytocin concentrations were below quantification limits (2 pg/ml). In Group 2 (n=14), oxytocin IH concentrations remained quantifiable up to 3 hrs post dose. Adverse events were reported for: Group 1, IH n=3 (33%) and IM n=2 (25%); Group 2, n=14 (100%). No deaths were reported. Interpretation: Safety profiles of oxytocin IH and IM were similar; PK profiles could not be defined for oxytocin IH or IM in women in TSL, despite using a highly sensitive assay. Ex vivo investigations suggested oxytocin metabolism occurred in vivo and not during sample collection and processing. Trial Registration: Trial Registration Number, (ClinicalTrials.gov: NCT02999100). The study protocol can be found online (https://clinicaltrials.gov/ProvidedDocs/00/NCT02999100/Prot_000.pdf). Funding: GlaxoSmithKline (GSK 205920; NCT02999100) Declaration of Interest: KG-K, AS-C, KRP, PL, TN, JM, CG, CK and VLO have no conflicts to declare. SK, AC, RAG, SP, IS, AS, KH, MP, ME and SS are employees of and hold stocks in GSK. MPM is an inventor on the patent method and formulation for oxytocin inhalation (WO2013/016754 A1 [PCT/AU2011001430]). Ethical Approval: The study protocol was reviewed and approved by the Office for Research Ethics Committees of Northern Ireland (UK) and the Monash Health Human Research Ethics Committee (Australia) in accordance with Good Clinical Practice (GCP) guidelines.

Inhalation of Soluble Plutonium: 53-year Follow-up of Manhattan Project Worker.

This whole-body tissue donor to the United States Transuranium and Uranium Registries was occupationally exposed to plutonium nitrate-dioxide mixture via chronic inhalation. This individual was involved in the Manhattan Project operations and later participated in medical follow-up studies. Soft tissues and bones collected at autopsy were analyzed for 238Pu, 239+240Pu, and 241Am. Fifty-three years post-intake, 700±2 Bq of 239+240Pu were still retained in the skeleton, 661±11 Bq in the liver, and 282±3 Bq in the respiratory tract. Bioassay measurements and organ activities at the time of death were used to estimate the intake and radiation doses using the TAURUS internal dosimetry software. For this individual, an ICRP Publication 130 Human Respiratory Tract Model with case-specific particle size of 0.3 μm, ICRP Publication 100 Human Alimentary Tract Model, and ICRP Publication 141 Plutonium Systemic Model adequately described long-term plutonium retention and excretion. The total cumulative 239+240Pu intake of 31,716 Bq was estimated, of which 24,853 Bq (78.4%) were contributed by inhalation of plutonium nitrate and 6,863 Bq (21.6%) of plutonium dioxide. The committed equivalent doses to the red bone marrow, bone surface, liver, lungs, and brain were 0.71 Sv, 6.5 Sv, 8.3 Sv, 3.8 Sv, and 0.068 Sv, respectively. The committed effective dose was 1.22 Sv.

Cost Analysis of Routine Vitrectomy Surgery

To use electronic health record (EHR) time logs to calculate the complete cost profiles of routine pars plana vitrectomy surgery. Economic analysis. Patients undergoing elective vitrectomy procedures (Current Procedural Terminology codes 67040, 67041, and 67042) at Vanderbilt University Medical Center in fiscal year2019. Process flow mapping for routine vitrectomy surgery was used to define the operative episode. De-identified time logs were sourced from an internal perioperative data warehouse to calculate procedure-level durations. The costs of materials and overhead were calculated from internal financial management software. Costs per minute for space, equipment, and personnel were based on internal figures. These inputs were used for a time-driven activity-based costing (TDABC) analysis. Complete cost profile of routine pars plana vitrectomy surgery. Cost analysis of routine vitrectomy surgery resulted in a total cost of $7169.79 per patient, which was $2053.85 more than the maximum Medicare reimbursement for the equivalent episode, $5115.93. Vitrectomy cases do not break even unless the case duration is fewer than 26.81 minutes, overhead is reduced by 53.78%, or reimbursement is increased by 40.15%. Reimbursement does not compensate for variable costs alone for cases lasting longer than 55.09 minutes. In the cohort used here, 68% of cases are completely unprofitable, with increasing losses directly proportional to the length of the case. This analysis showed that true costs for routine vitrectomy procedures are significantly more than the maximum allowable Medicare reimbursement. Academic ophthalmology departments may benefit from more accurate costing approaches using existing EHR data. These approaches may be informative for policy discussion regarding appropriate reimbursement.

ENHANCED ALGORITHM OF ARTIFICIAL BEE COLONY (ABC) TO OPTIMIZE MODELS OF SYSTEM RELIABILITY

In recent times, computer software applications are increasingly becoming an essential basis in several multipurpose domains including medicine, engineering, transportation etc. Consequently, with such wide implementation of software, the imperative need of ensuring certain software quality physiognomies such as efficiency, reliability and stability has ascended. To measure such software quality features, we have to wait until the software is executed, tested and put to use for a certain period of time. Numerous software metrics are presented in this study to circumvent this long and expensive process, and they proved to be awesome method of estimating software reliability models. For this purpose, software reliability prediction models are built. These are used to establish a relationship between internal sub-characteristics such asinheritance, coupling, size, etc. and external software quality attributes such as maintainability, stability, etc. Usingsuchrelationships, one canbuildamodelinordertoestimatethereliabilityofnewsoftware system.Suchmodelsaremainlyconstructedbyeitherstatisticaltechniquessuchasregression,or machine learningtechniquessuchasC4.5andneuralnetworks.The prototype presented isinvigoratedemployingprocedures of machine learninginparticularrule-basedmodels.Thesehaveawhite-boxnaturewhich accordsthecataloguingandmakingthemgood-looktoexpertsinthedomain. In this paper, wesuggest a powerfulinnovative heuristic based on Artificial Bee Colony (ABC) to enhance rule-based software reliability prediction models. The presented approach is authenticated on data describing reliability of classes in an Object-Oriented system. We compare our models to others constructed using other well-established techniques such as C4.5, Genetic Algorithms (GA), Simulated Annealing (SA), Tabu Search (TS), multi-layer perceptron with back-propagation,multi-lay perceptron hybridized with ABC and the majority classifier. Results show that, in most cases, the propose technique out- performs the others in different aspects.

The Technical Debt Density Over Multiple Releases and the Refactoring Story

Do developers postpone fixing Technical Debt (TD) in software systems? TD is a metaphor that refers to short-term decisions in software development that may affect the cost of the software development life cycle. The bad smell is an imperfect solution in the software system that negatively impacts the internal software quality and maintainability. In this paper, we will study five open-source software projects (OSSPs) that have several releases and also estimate the numbers of architecture smells (ASs), design smells (DSs), and code smells (CSs) for every release. Designite will be used to detect smells. We describe a case study conducted to explore the following: (1) What is the average smells density for architecture, design, and code smells in an OSSP? (2) Does the density of each smell type increase over multiple releases? (3) What percentage of each smell-type density is eliminated by refactoring? We collected around 2 million LOC from five OSSPs that have multiple releases from the GitHub repository to statistically analyze the software concerning the smells as indicators of TD. We find 36% of Architecture Technical Debt (ATD) is Cyclic Dependency, while 33% of Design Debt (DD) is Cyclically-dependent Modularization. More than 70% of Code Debt (CD) is Magic Number. Even though the developers do refactoring between releases, the TD density in general increases. On average, by refactoring, developers remove around 48%, 16%, and 22% from the introduced ATD, DD, and CD from their next release, respectively.