International Journal of Healthcare Technology and Management (15 papers in press)
Fuzzy soft set approach for classifying malignant and benign breast tumours
by S. Sreedevi, Elizabeth Sherly
Abstract: Breast cancer is one of the most common health problems faced by women all over the world and mammography is an effective technique used for its early detection. This work is concentrated on developing machine learning algorithms combined with a mathematical model for classifying malignant or benign images in digital mammograms. The mathematical concept of the fuzzy soft set theory is advocated here, which is an extension of crisp and fuzzy with parameterization. Even though fuzzy and other soft computing techniques have made great progress in solving complex systems that involve uncertainties, imprecision and vagueness, the theory of soft sets open up a new way for managing uncertain data with parameterization. The classification is performed by using fuzzy soft aggregation operator to identify the abnormality in a mammogram image as malignant or benign. This work is a fully automated computer aided detection method which involves automated noise removal, pectoral muscles removal, segmentation of ROI, identification of micro-calcification clusters, feature extraction and feature selection followed by classification. The experiment is performed on images from MIAS dataset resulted in 95.12% accuracy.
Keywords: digital mammography; computer-aided diagnosis; fuzzy soft set theory; fuzzy c-means; NL-means; fuzzy soft aggregation operator.
An intelligent model for diagnosis of breast cancer
by Raj Kamal Kaur Grewal, Babita Pandey
Abstract: Breast cancer, the most common disease in India in comparison to the United States and China, is not easily diagnosed in its initial stage. The early diagnosis of breast cancer can save lives, therefore it is very important to diagnose it at the initial stage. The development of an effective diagnosis model is an important issue in breast cancer treatment. This study accordingly employs J48 classification algorithm and case-based reasoning to construct an intelligent integrated diagnosis model aiming to provide a comprehensive analytic framework to raise the accuracy of breast cancer diagnosis at two levels. The dataset used in the diagnosis is based on the advice and assistance of doctors and medical specialists of breast cancer. At the first level, J48 algorithm is deployed for classifying the breast cancer dataset into malignant and benign cancer types. At the second level, malignant cases are further classified as Ductal carcinoma in situ, Lobular carcinoma in situ, Invasive ductal carcinoma, Invasive lobular carcinoma, and Mucinous carcinoma using case-based reasoning. The result specifies that the J48 accuracy rate is 90%. In case-based reasoning at a second level, the new case is supported by a similar ratio, and the case-based reasoning diagnostic accuracy rate is 98.25%. The implemented result shows that the intelligent integrated diagnosis model is able to examine the breast cancer with considerable accuracy. This model can be helpful for making the decision regarding breast cancer diagnosis.
Keywords: breast cancer; data mining; case-based reasoning; J48.
Accuracy comparison of the data mining classification techniques for diabetic disease prediction
by Rakesh Garg
Abstract: In the present scenario, the speedy use of the data mining (DM) techniques is observed for predicting and categorizing symptoms in large medical datasets. Classification is one major DM technique that is widely used for classifying various unnoticed information from various diagnostic data. In a popular country like India, Diabetes is characterized as a dangerous disease which has affected the majority of the population. The present research emphasizes on the accuracy comparison of the various classifiers such as J48, Random Forest, Sequential Minimal Optimization (SMO), Stochastic Gradient Descent (SGD), Naive Bayes, Logistic Regression, Random Tree, Decision Stump, Simple Logistic, Hoeffding Tree, Adaboost, and Bagging, when applied on a diabetic data.
Keywords: data mining; diabetes; classification; Weka.
A new deep learning structure to improve detection of P300 signals using supervised neural networks as convolutional kernel of CNN
by Seyed Vahab Shojaedini, Sajedeh Morabbi, MohammadReza Keyvanpour
Abstract: Brain-Computer Interface (BCI) systems provide a safe and reliable interface between brain and outer world and detecting P300 signal plays a vital role in these systems. In recent years Convolutional Neural Networks (CNNs) have made a vast and rapid development in P300 signal detection. In this paper, a novel structure for CNN is proposed to enhance separability of the selected features in its convolutional layer. In the proposed scheme an artificial neural network is applied in the above layer as nonlinear filter which extracts nonlinear features which lead to improve detecting of P300 signals. The performance of the proposed structure is assessed on EPFL BCI group dataset. Then the achieved results are compared with the basic structure for P300 detection. The obtained results demonstrate the improvement of True Positive Rate (TPR) of the proposed structure against its alternative by extent of 19.69%. Such improvements for false detections and accuracy are 1.97% and 10.87% which show the effectiveness of applying the proposed structure in detecting P300 signals.
Keywords: brain-computer interface; P300 signal detection; conventional neural network; convolutional kernel; nonlinear filter.
Diagnosis of diseases from retinal images using support vector machine
by C. Malathy, Sneha Das
Abstract: Retina is a thin membrane present at the back of our eye which allows us to see the world around us. Identifying retinal diseases at an early stage is very important, as it may lead to loss of vision. The problem arises in the recognition procedure of some patients. In this paper, the focus is made to automatically diagnose the retinal diseases from the retinal images using machine learning technique. The images are taken from the DRIVE database and also from retina_gallery.com website. The reason for selecting the DRIVE dataset and retina_gallery.com is that all the available retinal diseases are present in these datasets and it becomes easy to identify which disease the retina is facing. It is then pre-processed and the region of interest (ROI) is taken. Then by local binary pattern (LBP) the features of the images are extracted, image segmentation is done on the images and finally support vector machine (SVM) is applied to diagnose the diseases. The accuracy of the proposed method is compared with the neural network (NN) and the accuracy found by neural network (NN) is 93% whereas for support vector machine (SVM) it is 96%. So, for the diseases diagnose and for the accuracy support vector machine (SVM) is taken. SVM was taken for solving the problem because classification is done between the diseased eye and normal eye .The accuracy with SVM was found to be more compared to other algorithms of machine learning. They are robust, accurate and are effective even with small training sample .
Keywords: Local Binary Pattern (LBP); Support Vector Machine(SVM).
Patient perception of interactive mobile healthcare apps: a predictive model
by Mattie Milner, Scott Winter, Rian Mehta, Stephen Rice, Matthew Pierce, Emily Anania, Karla Candelaria-Oquendo, Diego Garcia, Nathan Walters
Abstract: Many professionals connect with consumers through mobile app technology. It is then no surprise that healthcare providers have begun exploring this technology as a tool to reach their patients. Despite increasing accessibility, the willingness to use mobile technology, such as healthcare apps, can be affected by several different factors. This study aims to determine what factors predict a persons willingness to use this type of technology. Four hundred and five participants completed the study over two stages, which included a hypothetical scenario using a mobile healthcare app and a survey identifying their willingness to use, knowledge of, and privacy concerns regarding mobile healthcare applications. A backward stepwise regression analysis revealed two significant predictors: privacy concerns and likelihood to use the internet. There was good model fit, highlighting the predictive power of the regression model on a new dataset. As technology becomes more prevalent among consumer e-health options, this research may help key stakeholder groups, such as healthcare providers, doctors, and patients, better understand patients willingness to use mobile health applications.
Keywords: e-health; willingness to use; privacy; health care; mobile apps; regression.
Examining Taiwan's national health insurance website quality and customers' loyalty
by Jengchung Victor Chen, Timothy McBush Hiele, Mei-Tsui Lin
Abstract: The favourable impact on the assessment of website quality and response from customers can explain and justify the manifestation and significance of an online healthcare platform. This study applies the information systems (IS) success model and technology acceptance model (TAM) to assess Taiwans bureau of national health insurance (BNHI) online platform. In particular, this study presents and empirically tests the research framework by using a SmartPLS path analysis through the use and validation of Taiwans BNHI customers feedback. Overall, this study sheds light on the value and importance of a medical institution and/or organisation to offer better and quality information technology (IT) platform that can serve and retain its loyal customers.
Keywords: customer loyalty; e-government; IS success model; Taiwan national health insurance; website quality.
General analytics limitations with coronavirus healthcare big data
by Kenneth David Strang
Abstract: The goal of this study was to reveal factual big data statistical general analytics issues in the healthcare industry using COVID-19 coronavirus as an empirical example. Search engines and the SPSS Python R extension were used to analyse healthcare big data information stored on the internet. The research question was focused on what were the significant limitations of statistical techniques when analysing the effect of publicly available healthcare big data, using the coronavirus as an example. The sample was a manageable subset of dynamic information from the internet time-stamped to midnight of April 14, 2020 with a filter set for coronavirus confirmed cases or deaths in Wuhan in Hubei province in China, New York State in USA and New South Wales, Australia. There were surprising results, indicating using general analytics that the healthcare big data were not reliable. Nevertheless, interesting relationships were detected when linking foreign property ownership to the two Australian cities of Sydney and Melbourne experiencing the largest coronavirus related fatalities. During this study several useful and practical general analytics effect size equations were shown and proven to help detect reliability limitations when examining healthcare big data.
Keywords: healthcare big data problems; privacy; security; systems thinking action research.
Application of GIS and SPSS for prostate cancer and health disparity detection in Texas
by Jose Huerta, Gayle Prybutok, Victor Prybutok
Abstract: This study uses a geographic information system to create and analyse choropleth maps determining the distribution of prostate cancer in Texas and uses SPSS software to analyse social determinants of health that may explain prostate cancer mortality. The data, collected for period 19992009, was furnished by the Texas Health Rankings and VitalWeb. The dataset was for 19992004 and 20042009. It comprised age-adjusted data specific to the 2000 US Standard Population data, based on an age-distributed and -weighted methodology to create age adjustments for statistical purposes. The study found there was a statistically significant (P < .05) percentage of African Americans with age-adjusted prostate cancer mortality, but no statistically significant correlations were found in other races. The study indicates a number of ways in which medical communities and public health agencies can employ GIS and SPSS to screen for and treat prostate cancer more effectively.
Keywords: GIS; prostate cancer; social determinants of health; spatial patterns; SPSS; choropleth maps; geographic information systems.
The Stent for Life Initiative in Portugal: a critical realist perspective
by Maria Major
Abstract: Drawing on the morphogenetic approach proposed by Archer (1995), this study explains how the Stent for Life (SFL) initiative emerged and developed in Portugal and how it was embraced as a means to reduce mortality following acute myocardial infarction. A qualitative research strategy based upon a case study was adopted. Only by conducting qualitative research was it possible to join context with explanation and produce understandings for the 'who', 'when', 'why', 'how' and 'what' of SFL in Portugal (Mason, 2018; Ackroyd and Karlsson, 2014; Yin, 2011, 2015). By making analytical dualism explicit, Archers morphogenetic approach proved very useful in accounting for the existence of prior material structures and cultural structures that affected (but did not determine) agents' actions. The investigation revealed that there was morphogenesis (i.e. the former structures were transformed) and that new structures emerged from the intended consequences of agents' activities.
Keywords: Stent for Life in Portugal; Archer’s morphogenetic approach; critical realism; ST-segment elevation myocardial infarction.
The impact of healthcare legislation on the relationship between intangible investments and investor returns in the USA Pharmaceutical Industry
by Richard A. Heiens, Robert Leach
Abstract: The current study examines the relationship between intangibles and investor returns in the pharmaceutical industry over two separate time frames: before the Patient Protection and Affordable Care Act (ACA) was initially enacted in 2010, and post-ACA. Our results show that pharmaceutical firms in the post-ACA sample significantly outperformed those firms in the pre-ACA sample. Secondly, pharmaceutical firms investing little in R&D offered higher investor returns in comparison with those that invested more in R&D. Also, the negative relationship between R&D and holding-period returns was more pronounced in the post-ACA sample than in the pre-ACA sample. Finally, in the post-ACA period, investor holding-period returns were above average for firms that included goodwill and other intangible assets on their balance sheets, and negative for firms that did not.
Keywords: Affordable Care Act; intangibles; R&D; holding period returns; pharmaceutical industry.
A qualitative analysis of Lean implementation in Brazilian surgical centres: a multiple-case study
by Paulo R. De Sousa, Marcelo Werneck Barbosa, Gerson Tontini, Aloisio Rosado Filho, Maria Madalena Ferreira
Abstract: Different studies have been published in recent years describing the experiences of adopting Lean thinking in healthcare institutions. However, the analysis of Lean implementations in hospitals located in emerging economies is lacking. Besides, Lean implementations are usually reported based on a single case study. This study analyses Lean adoption in four Brazilian surgical centres, using a multiple-case approach with semi-structured interviews. Results suggest that Lean implementation reduces the duration of operations and the time the patient stays in the hospital, improves the punctuality of procedures, reduces losses, errors, materials' usage, and preparation time, and increases collaborators' satisfaction. Primary challenges to implementing Lean include changes to organisational culture and involvement of multidisciplinary teams. By using a multiple-case design, it is possible to compare different outcomes and more confidently generalise these results to other surgical centres. We expect this study to stimulate the implementation and analysis of Lean adoption in other emerging economies.
Keywords: surgical centre; operating rooms; Lean healthcare.
Investigation of ECG signal suitability for key generation to secure body area sensor networks
by Rajagopal Rekha
Abstract: Remote Patient Monitoring (RPM) systems help to continuously monitor the health condition of people affected by chronic diseases. The proposed research work analyses the suitability of ECG signals for 128-bit cryptographic key generation in RPM systems. The randomness and distinctiveness of keys generated from the ECG signal using the following three methods are evaluated: (a) the Inter Pulse Interval (IPI) of the ECG signal captured from the patient's body is converted into a 128-bit binary sequence using modular encoding; (b) Short-Time Fourier Transform (STFT) is applied over the ECG signal and the key is generated using the features extracted from ECG signal; (c) the binary sequence obtained with the help of STFT features of ECG signal is partitioned into subkeys and transformed into a 128-bit binary sequence using Tiny Encryption Algorithm (TEA). The results illustrate that the ECG signal can generate a highly random and distinctive 128-bit cryptographic key.
Keywords: cryptographic key; body area sensor network; short-time Fourier transform; tiny encryption algorithm; inter pulse interval; randomness; electrocardiogram.
The efficiency of healthcare spending in lower-middle-income countries: an empirical investigation using a two-stage data envelopment analysis approach
by Israa A. El Husseiny
Abstract: This study examines the technical efficiency of healthcare systems and its determinants in a sample of 49 lower-middle-income countries using a two-stage Data Envelopment Analysis (DEA) approach. The first stage of analysis indicates that the overall efficiency potential gains range from 2% under the output-oriented approach to 38.7% under the input-oriented approach. Among the examined countries, Algeria, Bangladesh, Laos PDR, Sri Lanka, Tunisia, and Vietnam have fully efficient healthcare systems under both orientations of efficiency. Sri Lanka and Bangladesh are the most referenced countries under the output- and input-oriented approaches, respectively. In contrary, Nigeria and Micronesia have the least efficient healthcare systems, under the output- and input-oriented approaches, respectively. The second stage of analysis indicates that, while GDP per capita, government health expenditure as percentage of total health expenditure, and population density tend to have a positive impact on health efficiency, unemployment seems to have a negative one.
Keywords: technical efficiency; healthcare systems; lower-middle-income countries; data envelopment analysis; health resources; health outputs; Tobit.
Value and organisational implications of big data and advanced analytics in healthcare: a case survey analysis
by Nishita Mehta, Sharvari Shukla
Abstract: Despite the possibilities of big data and advanced analytics in healthcare, very few organisations have been able to exploit its potential. One of the key reasons is inadequate practical evidence on the clinical and business impact of advanced analytics. To address the issue, this study proposes an advanced analytics healthcare value creation model. The model suggests that well-defined objectives for using the technology and readiness of big data analytics capability (BDAC) are rudimentary for an organisation's performance enhancement. It establishes four dimensions of robust BDAC - BDA infrastructure, personnel, management and relational BDA capability. In order to test the theoretical model, a case survey method was employed. Data from 254 case studies was used to verify its validity. The findings from case synthesis confirm the need for strategic intent and strong BDAC for effective implementation. The results also illuminate the implications of BDA in terms of improved clinical, research and business value.
Keywords: big data analytics; advanced analytics; clinical value; research value; business value; healthcare; case survey method; meta-analysis.