Big Data for Health
Introduction
For recent years, big data has turned out to be a buzzword due to its common usage doubling every year as per the standard search engines[1]. Even though the popularity of big data for the prior period is current, existing challenges have lived with the people for a more extended period. The researchers are pursuing many difficulties to get the solutions to those problems. In healthcare, big data has been a great concern having datasets which meaningful [1]. The datasets are too complicated, big and fast for the healthcare providers to do their processing as well as give the interpretation of such existing tools. This is steered by constant attempts of making healthcare to be more efficient and sustainable provided the demand for these tools are expanding the population constantly with an inverted age pyramid. Besides, there is a shift in the paradigm recurrent physiological characteristics in humans. Contrary, environmental features are presenting another set of variables which can be captured through continuous sensing, which are essential to the health of a population.
Nonetheless, the size is not giving the appropriate qualification of big data. Some challenges incorporate heterogeneity, speed, as well as different data available in healthcare. When it comes to the management of influenza pandemic, for instance, information notably the heterogynous ones-form managed, or unmanaged sources might be mined, processed then converted to actions that can assist in controlling the pandemic.
Within the healthcare facilities, variety and data heterogeneity come as a result of associating the diverse ranges of available biomedical data sources. The sources might be quantitative, for instance, sensor data, gene arrays, laboratory tests and free test and the qualitative test like demographics and free texts. The goal is underlying data challenge certainly is to support observational evidence answers to various clinical questions. These would never be solved through studies anchored directly on random trials. Besides, the act of generalizing outcomes based on a narrow spectrum of participants might be addressed by taking advantage of big prospective data in support of the use of longitudinal studies. The three Vs.-variety, volume and velocity form the key definitions in the characteristic of big data. Some other factors are also considered are veracity, variability and value.
Veracity is essential for big data. For instance, personal health record might at times have typographical mistakes, cryptic notes and abbreviations. The ambulatory measurements are deemed to be less accurate within an uncontrolled environment as matched to the clinical data gathered by practitioners who are trained. The utilization of unmanaged data spontaneously like attained from social media might bring about wrong data forecast, which cannot be easily known. What is more, most of the sources are biased towards young persons who are internet savvy and have the capacity of expressing themselves through various online platforms. The genuine value of healthcare systems and patients can be realized when the challenges of analyzing big data are addressed appropriately in a coherent manner [1].
It is imperative to note that multiple underlying theories are often explored within research communities. In this article, the author discussed various activities together with future opportunities which are associated with the big data within healthcare systems. Notably, the author discussed the values of imaging informatics, sensor informatics, translational bioinformatics, medical and health informatics.
Activities associated with big data in healthcare
Translational bioinformatics
This is a field which came to existence after initial mapping of the genome for human beings. It is focusing on bridging the gaps in statistical genetic, biostatistics, molecular biology with clinical informatics. The area is developing at a quicker pace with several related fields being proposed by the scientists. For instance, pharmacogenomics is an arm of genomics which deals with drug variation and responses in an individual as a result of differences in genetics. This field is essential when it comes to designing particular medicine for managing various ailments in future. Currently, the discoveries from the project of the human genome are being applied in developing improved diagnostics, therapies and prognosis of multifaceted diseases. To be specific, the cost of sequencing for every genome has been marked to be reduced for the past decades. The studies on bioinformatics have widened from genome sequencing of a person towards quantifying epigenomic data that incorporate processes which change the expression of genes apart from transformations in the first sequencing of DNA [1]. Information technology is being employed in support of analyzing and acquiring biological molecules apart from the genome. It is this possible for one to design precision of medicines for the specific patients using their profiles of genomics [1]. Pharmacogenomics is seen beyond studying a particular response of the response attained from the drug anchored on genome features. This now includes additional transcriptions as well as metabolic characteristics like gene expressions taking into considerations elements which influence drug concentration attaining targets together with associated factors on the drug tests. The profiles on gene expressions vary considerably within the process of prolonged culture within various conditions used in the drug response predictions within patients. This is an issue which has been controversial through research ethics.
Translation genomics has a higher prospective of genetic stratification when it comes to screening patients. For example, there can be arising factors such as DNA mutations, intolerance of gluten and milk, mucoviscidosis and genotyping issues. The provided information on by the HER on genetics and phenotype might assist in offering more significant insights into lower penetrant alleles.
OMICS and large-scale database have been launched to promote research of progression and mechanisms of diseases, especially at the system level. The other example which has mass spectrometry is the Proteomics DB. It contains drug databases and genes which are acquired from the human proteome gotten from human tissues, body fluids and cell lines. Contrary to these creations, the human metabolome database is made up of 40, 0000 entries of annotated metabolic studies which have been compiled from 2013. The database offers experimental metabolic concentrations as well as analyses which have been done via mass spectrometry as well as NMR spectrometry. Mostly, the databases are known to promote translations of information into an acquaintance in support of changing various clinical practices. The accessibility of proteomics, metabolic sans genomic databases permits excellent comprehension of the development of multifaceted ailments like cancer [1]. Such databases allow searches of new biomarkers by the use of various techniques of mining and clustering. The clusters might be either hierarchical or partitioned. All these approaches are facilitated through the use of GPU, multicore CPU and arrays of field-programmable gates having techniques of parallel processing.
Health informatics
In the capacity to handle larger volumes of data, which are both unstructured and structured from various sources, most of the great analytical tools are holding good promises of getting good outcomes when it comes to longitudinal researches. In multiple instances, unique chances lie with the integration of conventional medical informatics with social and mobile health [3]. These are employed to address both chronic and acute ailments in a manner which have never been possible in times back
In most cases, electronic health records (EHRs) which give details on the treatments and results of patients have plenty of information which at times is never used appropriately. Conventional health data centres take and store the massive amount of data in the structured form about various details [2]. These structured data might include laboratory tests, diagnostics, medications and ancillary clinical trial data. In the patient reports, the use of natural language in report compilation assists in the systematic analysis of the information. Data mining through the use of EHR tools improves clinical acquaintance as well as support concerning clinical studies. However, data mining of local knowledge with the use of EHR data has been active when it comes to supporting various healthcare challenges. For instance, the support of disease management, pharmacovigilance as well as building models which can be used in predicting risk assessment within the healthcare.
Still on EHR, telemedicine assists in promoting social health amongst the patients. It is used in linking doctors with the patients beyond clinical set-ups. With the use of communication techniques, social networks have increased, promoting more significant levels of social interactions. Such a new feature has created new possibilities of communication between patient to patient paradigms. For instance, patients having a chronic illness like cancer, heart conditions and cancer currently are in a position of using social networks to share their experiences with other ailing persons. On top of vital information, social applications and geo-location offer an additional feature which can assist in comprehending social, behaviours and demographic characteristics of a patient. The use of social media, as well as internal searches, can be used in providing data about the trend in various diseases [2].
In predicting the personal health of patients, the climatological data like heat-stress and mortality due to cold can be attained from the use of technology. Currently, the use of technologies on remote sensing and geographic information system enables the collection of global data on the climatic conditions of areas are interpolated to be having specific impacts on patient health.
Sensor informatics
For the past years, advancement on sensor hardware has gone higher and the trend of no indication of slowing down soon. The prices of MEMS sensors have reduced, which is driven by many people using internet applications. This has promoted development in the subsequent generations in the use of sequencing technologies as well as biomolecule detections. The patients who have diabetes have multiple sensors which are both implanted and wearable. These are providing continuous monitoring as well as offering corresponding responses timely on the levels of glucose for patients. This has assisted in ensuring that patients get proper treatment through the use of technology.
Through the use of physiological sensing, which is done by smart devices, interpretation of clinical data is made possible. For instance, contemporary clinical practice has the potentiality of defining hypertension using the measurements which have been compiled from the frequent hospital visits by the patients. The mobile devices offer complete BP related profiles of a person. The data can be made accessible to everyone. However, interpretation of these data is not significant since the data might be equivalent to the clinical readings of BP, which are presently used in various healthcare facilities. The signals which are being portrayed are that when the data is processed correctly or managed, then uncontrolled hypertension can be comprehended in depth. Smart implants might bear reactive roles through delivering neurological ailments which incorporate chronic pains and activations of brain stimulators in support of various neurological diseases. Thus, intelligent implants act as a resource for data collections as well as an integral component which assists in the early stages of disease intervention.
Again, mobile health has been enabled through the use of smartphones. The data generated through the use of smartphones offer continues and descriptive information concerning any disease in real-time. Most of the new generation smartphones come with health apps having standardized protocols which are used in connecting sensors installed by various firms. Smartphones can be used potentially in serving the platforms of centralized data for health. Mostly, the earlier versions of mobile phones have motion sensors which do not require the readings to be done from external sensors. When these sensors are used properly, then they have the capability of offering valuable information concerning the health of the patients to the management in the long-run. However, hardware levels, for instance, PET/CT and MRI/PET are opening up broader opportunities, especially the targeted therapy and oncological imaging.
Imaging informatics
Through the use of imaging informatics, there has been an increased increase in real-time data regarding medical imaging. There are many modalities which are coming up to be used in imaging processes [4]. Most of the techniques of imaging are geared towards in vivo apps, which make multimodality imaging to be more comfortable. Various MRI techniques are presently being used as a healing and interventional aids. There have been advancements of CT scans as well as other modalities of imaging. Approaches like MRIs and functional diffusion tension imaging offer flexible information within the form of microstructural and macro-structural matrices.
Discussions
The application of big data is essential in both physical and biological sciences. However, it is comparatively significant in soft sciences such as social and behavioural sciences. In most cases, it is known that human behaviours are the significant drivers of some environmental factors such as air pollutions, medical issues, and climatic changes, among others. With the use of advanced technology studied in this paper, all the human behaviours like observable emotions, physical actions, temperament, personalities and the patterns of social interactions can be monitored easily. Health informatics assists in the generation of datasets which are complex and provide interpretation for effective decision-making process [6]. Big data tools assist in processing complex data and offer useful analysis for the users.
Conclusion
In summary, the use of big data boosts applicability of clinical researches into scenarios of real-world. Again, the use of big data offers the users with chances of enabling appropriate precisions in medicine through doing stratification of patients. Hence, advancements in significant data processes such as sensing, imaging, bioinformatics, health informatics shall have more significant inputs in future clinical studies.
References
[1] Andreu-Perez, J., Poon, C.C., Merrifield, R.D., Wong, S.T. and Yang, G.Z., 2015. Big data for health. IEEE Journal of biomedical and health informatics, 19(4), pp.1193-1208. https://ieeexplore.ieee.org/abstract/document/7154395/
[2] Cottle, M., Hoover, W., Kanwal, S., Kohn, M., Strome, T. and Treister, NW, 2013. Transforming health care through big data. Institute for Health Technology Transformation, Washington DC, USA.
[3 ] Kovats, R.S. and Hajat, S., 2008. Heat stress and public health: a critical review. Annu. Rev. Public Health, 29, pp.41-55.
[5] Moltchanov, S., Levy, I., Etzion, Y., Lerner, U., Broday, D.M. and Fishbain, B., 2015. On the feasibility of measuring urban air pollution by wireless distributed sensor networks. Science of The Total Environment, 502, pp.537-547.
[6] Ram, S., Zhang, W., Williams, M. and Pengetnze, Y., 2015. Predicting asthma-related emergency department visits using big data. IEEE Journal of biomedical and health informatics, 19(4), pp.1216-1223.