An infrastructure for precision medicine through analysis of big data

Abstract

BackgroundNowadays, the increasing availability of omics data, due to both the advancements in the acquisition of molecular biology results and in systems biology simulation technologies, provides the bases for precision medicine. Success in precision medicine depends on the access to healthcare and biomedical data. To this end, the digitization of all clinical exams and medical records is becoming a standard in hospitals. The digitization is essential to collect, share, and aggregate large volumes of heterogeneous data to support the discovery of hidden patterns with the aim to define predictive models for biomedical purposes. Patients’ data sharing is a critical process. In fact, it raises ethical, social, legal, and technological issues that must be properly addressed.ResultsIn this work, we present an infrastructure devised to deal with the integration of large volumes of heterogeneous biological data. The infrastructure was applied to the data collected between 2010–2016 in one of the major diagnostic analysis laboratories in Italy. Data from three different platforms were collected (i.e., laboratory exams, pathological anatomy exams, biopsy exams). The infrastructure has been designed to allow the extraction and aggregation of both unstructured and semi-structured data. Data are properly treated to ensure data security and privacy. Specialized algorithms have also been implemented to process the aggregated information with the aim to obtain a precise historical analysis of the clinical activities of one or more patients. Moreover, three Bayesian classifiers have been developed to analyze examinations reported as free text. Experimental results show that the classifiers exhibit a good accuracy when used to analyze sentences related to the sample location, diseases presence and status of the illnesses.ConclusionsThe infrastructure allows the integration of multiple and heterogeneous sources of anonymized data from the different clinical platforms. Both unstructured and semi-structured data are processed to obtain a precise historical analysis of the clinical activities of one or more patients. Data aggregation allows to perform a series of statistical assessments required to answer complex questions that can be used in a variety of fields, such as predictive and precision medicine. In particular, studying the clinical history of patients that have developed similar pathologies can help to predict or individuate markers able to allow an early diagnosis of possible illnesses.

Clinical Information Systems – Seen through the Ethics Lens

“…The minority of studies put a technical concept forward, e.g. , for data curation 20 or described an existing infrastructure 21 22 . Two studies demonstrated how EHRs could be augmented to accommodate ethically relevant information, i.e.…”

“…Ashton and Sullivan [55] Baldini et al [38] Boers et al [36] Bourla et al [45] Brill et al [49] Brisson et al [33] Carter et al [35] Davenport and Kalakota [25] De Riel et al [31] Duckett [29] Eberlin et al [52] Erikainen et al [28] Evans and Whicher [47] Galvin et al [48] Gensheimer et al [54] Gooding [46] Graham et al [26] Ho and Quick [57] Lenca et al [56] Kogetsu et al [21] Kuhnel [53] Laurie [58] Lehmann et al [23] Loftus et al [37] Macdonald et al [50] Mars et al [51] McBride et al [40] McWilliams et al [20] Meredith et al [30] Moscatelli et al [22] Musher et al [39] Natsiavas et al [43] Pathak and Chou [44] Rashidi et al [27] Robichaux et al [41] Sánchez et al [34] Sanelli-Russo et al [24] Stockdale et al [42] Wilburn [32] [56] provided an overview of ethical themes grouped as families and subfamilies. There was a preponderance of studies found on independence and safety considerations.…”

EULAR points to consider for the use of big data in rheumatic and musculoskeletal diseases

“…A generic and comprehensive definition of big data is based on the five V paradigm, ie, volume of data, variety of data, velocity of processing, veracity and value 77…”

Current status of use of big data and artificial intelligence in RMDs: a systematic literature review informing EULAR recommendations

“…A key finding from this review is that there is no consensual definition of big data. First defined as data sets too large or complex for traditional analysis methods,11 this concept has evolved and the ‘5 V’ paradigm (for volume, velocity, veracity, variety and value) is more and more used 26–28. The definition provided in recent EMA recommendations may be considered as a synthesis of all these notions 12.…”