Deriving information from external Big Databases and Big Data analytics: all that glitters is not gold

Abstract

@ERSpublicationsBig Data analytics are a further step on the road to the ideal model of personalised medicine http://ow.ly/XWr1CTrue genius resides in the capacity for evaluation of uncertain, hazardous, and conflicting information Winston Churchill It seems more and more obvious that we are now living in the era of technology and information. Technological advances have allowed us to generate vast amounts of information of all kinds extremely quickly and, more importantly, store it so that it remains available for subsequent analysis. Accordingly, many large companies and institutions centre a large part of their strategy on the analysis of enormous databases on their customers. These provide information that determines future decisions and optimises resources.For a long time, medical insurance companies have been handling huge quantities of information about millions of customers worldwide, stored in well-structured databases so that it can be analysed for business purposes. Such an enormous depository of information is also potentially useful for other purposes, such as the attainment of valuable scientific knowledge that would be very difficult to achieve by other means [1,2]. This process does have a number of limitations, however, because these databases were not specifically created with scientific objectives in mind. The information they contain may not therefore be representative of the population under analysis. Alternatively, other conditioning factors may emerge, such as ethical problems derived from the use of personal data, a lack of key scientific information and insufficient quality control of the stored data [3][4][5].

Predictors of positive airway pressure therapy termination in the first year: analysis of big data from a German homecare provider

“…Analysis of big data allows inclusion of a very large population of patients and means that subgroup analyses include adequate numbers of patients to allow statistically meaningful comparisons. However, there are also some limitations associated with conducting scientific research using databases that were created for administrative, rather than scientific, purposes [ 26 ]. Such databases, including the one used in this analysis, include limited baseline clinical and demographic data, and no information about the severity of sleep-disordered breathing, method of diagnosis and comorbidities.…”

“…Although a number of studies have investigated compliance with positive airway pressure (PAP) therapy, the results have not always been consistent [ 18 – 24 ]. In addition, while the pattern of compliance within first weeks of CPAP therapy appears to be predictive of longer term compliance [ 25 ] – highlighting the importance of achieving good compliance early to ensure adequate long-term device use – there is a general lack of robust data from big data analyses [ 26 ] on specific predictors of PAP therapy persistence or termination. Identifying patients who are at risk of stopping CPAP therapy and the time course of when this might occur could help to optimize and target the provision of support strategies designed to increasing compliance [ 17 , 19 , 27 , 28 ].…”

Patología respiratoria en la era del big data

“…con 1.704.905 pacientes y un grupo control, concluyendo que la HTA, la diabetes mellitus, la cardiopatía isquémica, la depresión, la obesidad mórbida, las arritmias, la insuficiencia cardiaca congestiva y el IAM son comorbilidades más frecuentes en el SAHS que en la población general. Dichas asociaciones se conocían previamente, pero este estudio confirma la relación con mayor precisión y aporta, en contraposición con otras cohortes tradicionales, nuevas asociaciones específicas en subgrupos de pacientes poco estudiados hasta la fecha: mujeres y diferentes rangos de edad 20 . En Europa, a partir de la European Sleep Apnea Database (ESADA), también se trabaja en la identificación de fenotipos clínicamente relevantes y en su relación con las comorbilidades, destacando la caracterización de pacientes en función de los patrones de sueño (hipersomnolencia diurna vs. síntomas de insomnio) 21 .…”

α1-Antitrypsin deficiency: count me in please!

“…Reassuringly, the well-known association of liver disease with AATD was confirmed in this study by GREULICH et al [7], giving support to the methodology utilised. The novelty of this "real-world" AATD analysis, recently also applied to obstructive sleep apnoea [10,11], can provide us with hints of the population distribution and trends of respiratory conditions as seen in primary care and chest clinics. Despite the continuing controversy over population spirometry for COPD screening [12], recent case-finding strategies in COPD proved successful [13].…”