Big data has been referred to as data that are too complex and large that it cannot be processed and managed by traditional data processing tools. Gartner describes big data along three dimensions: variety, velocity and volume.1 ‘Variety’ refers to the fact that big data must be made up of many different types of data; ‘velocity’ addresses the fact that big data is about data that is transmitted and available in real-time, arrives in varying bursts rather than at a constant steady speed; and ‘volume’ refers to the fact that big data must be extremely large in size. A recent literature review defines healthcare big data as datasets with Log (n * p) ≥ 7, and have the properties of great variety and high velocity.2
Big data in the healthcare context refers to the routine collection, storage, assimilation and analysis of patient level data as they are collected in a real-time, live setting. Data collection can occur in a wide number of ways. Electronic Medical Records (EMRs), which exist in the hospital inpatient and outpatient settings, are a potential source of data, together with supplementary laboratory data as a further asset. Capturing attributes such as physical testing results (X-ray, MRI, CT results) as well as blood chemistry.
These are regularly collected and compiled and then fed back to providers for use in patient care within hours if not minutes after they are collected.3 But also, data such as from monitoring vital signs and wearable devices, social media posts, including Twitter feeds (so-called tweets), blogs, status updates on Facebook and other platforms, and web pages; and less patient-specific information, including emergency care data, news feeds, and articles in medical journals.4
Regarding hospital pharmacy, big healthcare data could become essential for understanding the effectiveness and safety of therapeutics, because big healthcare data are characterised by a large number of patients covered, a reflection of the variations in routine care practices, a lack of researcher-designed data capture (yielding inaccurate or missing information), and a lack of uniform data standards. Therefore, different potential uses are proposed for big data in healthcare settings, and of course in hospital pharmacy, in order to improve effectiveness, safety, efficiency and quality etc.:
- Population description and pattern exploration. Large insurance claims and electronic medical records (EMR) databases are very useful in understanding disease burden and medical need, as well as the underuse and guideline-recommended use of therapeutics, because they reflect care outside tightly controlled research environments. Conclusions drawn from EMR-based research may trigger care interventions to optimise the use of drugs, for example, improving adherence to chronic-use medications, and may be used to monitor the success of such interventions. Hypothesis-free pattern exploration with powerful visualisations may identify populations with particular utilisation and outcome patterns, stimulating new lines of inquiry. Data generated from outside the professional healthcare system, for example, through blogs, smartphone apps, or patient support groups, provide a different type of health information but are less straightforward to interpret for population-level insights, as they lack meaningful denominators and are subject to selected participation. Linkage between these novel data sources and structured healthcare information would create a very useful data asset.
- Genome-wide association studies (GWAS). Big genomics data are increasingly linked to big healthcare data. The latter includes phenotypic data and temporality preserving drug use and health outcomes data, allowing large-scale GWAS and genome–drug interaction studies. Even if they do not imply causal relationships, such association studies using big healthcare data can be useful when interpreted cautiously.
- Improve the understanding of causal relationships. Ultimately, providers and drug developers need to understand causal relationships between drugs and health outcomes. Understanding causality is arguably even more critical in medicine than in other data-rich fields, because healthcare professionals and regulators are responsible for making decisions about the wellbeing of patients. Big healthcare data have proven useful for assessing the safety of medications, drug–drug interactions (including the risk of unintended clinical events), and increasingly the comparative effectiveness of different drugs on health outcomes.5 Studies that conduct baseline randomisation and follow subjects using secondary healthcare data are of particular interest.6 However, in order to be useful for patient care, evidence needs to cross a quality threshold that allows interpreting associations as causal relationships. When analyses lead to causal interpretations of the effectiveness of therapeutics they become subject to more scientific scrutiny, in terms of transparency, auditability, reproducibility, and replicability.
- Clinical decision support systems. The objective is to deploy clinical decision support systems for enhancing the efficiency and quality of operations. These systems include computerised physician order-entry (CPOE) capabilities. The current generation of such systems analyses physician entries and compares them against medical guidelines to alert for potential errors such as adverse drug reactions or events. By deploying these systems, providers can reduce adverse reactions and lower treatment error rates and liability claims, especially those arising from clinical mistakes. In one particularly powerful study conducted at a paediatric critical care unit in a major US metropolitan area, a clinical decision support system tool cut adverse drug reactions and events by 40% in just two months.7
- Remote patient monitoring. Collecting data from remote patient monitoring for chronically ill patients and analysing the resulting data to monitor adherence (determining if patients are actually doing what was prescribed) and to improve future drug and treatment options. An estimated 150 million patients in the United States in 20108 were chronically ill with diseases such as diabetes, congestive heart failure, and hypertension, and they accounted for more than 80% of health system costs that year. Remote patient monitoring systems can be highly useful for treating such patients. The systems include devices that monitor heart conditions, send information about blood-sugar levels, transmit feedback from caregivers, and even include “chip-on-a-pill” technology – pharmaceuticals that act as instruments to report when they are ingested by a patient – that feeds data in near real-time to electronic medical record databases. Simply alerting a physician that a congestive heart failure patient is gaining weight because of water retention can prevent an emergency hospitalisation. More generally, the use of data from remote monitoring systems can reduce patient in-hospital bed-days, cut emergency department visits, improve the targeting of nursing home care and outpatient physician appointments, and reduce long-term health complications.
- Advanced analytics applied to patient profiles. For instance, applying advanced analytics to patient profiles (for example, segmentation and predictive modelling) to identify individuals who would benefit from proactive care or lifestyle changes. These approaches can help to identify patients who are at high risk of developing a specific disease (for example, diabetes) and would benefit from a preventive care programme. These approaches can also enable the better selection of patients with a pre-existing condition for inclusion in a disease management programme that best matches their needs. And, of course, patient data can provide an enhanced ability to measure the success of these programmes, an exercise that poses a major challenge for many current preventive care programs.
- Planning-based prediction. Particularly in integrated healthcare systems, it is now possible to program prediction algorithms for treatment effectiveness versus failure and feed the suggestion back to the provider. Because these are individual-level probabilistic predictions without any implication of causality, such prediction algorithms inform the provider at the point of care; however, they will not culminate in automated prescribing unless their performance improves substantially.9 Improvements are more likely to come from richer data than from new algorithms. For example, predicting the lack of adherence to a medication regimen is an area in which dynamic analyses of big data, including claims, EMR, in addition to consumer apps and electronic devices measuring behavioural factors, hold the promise of meaningful improvements in healthcare.10 The meaningful utilisation of big data is driving the personalisation of medicine, wherein diagnoses are made, treatment regimens designed, and prognoses forecast based on comprehensive analyses of individuals, not on population-based research. As the healthcare landscape shifts from fee-for-service (FFS) to value-based care, subspecialty medicine, and especially the surgical specialties, will be required to report and be held accountable to generic and non-specific quality measures in the absence of specialty-specific metrics. Existing generic measures not only have little relevance to specialty clinical practice, but also fail to meaningfully correlate with patient outcomes.11
In summary, there are immense opportunities for big data in hospital pharmacy settings to improve patient’s health and lower costs for the healthcare systems. However, the challenge lies in being able to effectively process and make sense of the big data that is available, and of course their quality with the best legal support. Intellectual property will become an even more important consideration for policy makers, who will set the rules that determine the legal rights associated with the ownership and use of data, and what constitutes “fair use” of data. Finally, one of the advantages of big data in healthcare settings is the ability to go beyond improving profits and cutting down on wasted overhead to predict epidemics, cure disease, improve quality of life and avoid preventable deaths.
- Big healthcare data could become essential for understanding the effectiveness and safety of therapeutics
- Evidence needs to cross a quality threshold that allows interpreting associations as causal relationships, collecting data from remote patient monitoring for chronically ill patients to monitor adherence.
- The meaningful utilisation of big data is driving the personalisation of medicine.
- Gartner. Gartner says solving ‘Big Data’ challenge involves more than just managing volumes of data. Stamford, Connecticutt, June 27, 2011. www.gartner.com/newsroom/id/1731916. Last accessed April 2016.
- Baro E et al. Toward a literature-driven definition of Big data in healthcare. Biomed Res Int 2015;2015:9.
- Alyass A, Turcotte M, Meyre D. From Big data to personalized medicine for all: Challenges and opportunities. Med Genom 2015;43:425–9.
- Raghupathi W, Raghupathi V. Big data analytics in healthcare: promise and potential. Health Info Sci Sys 2014;2:3.
- Schneeweiss S. Developments in post-marketing comparative effectiveness research. Clin Pharmacol Ther 2007;82(2):143–56.
- Tunis SR, Stryer DB, Clancy CM. Practical clinical trials: increasing the value of clinical research for decision making in clinical and health policy. JAMA 2003;290(12):1624–32.
- Potts AL et al. Computerized physician order entry and medication errors in a pediatric critical care unit. Pediatrics 2004;113(1):59–63.
- Big Data: The next frontier for innovation, competition, and productivity. McKinsey Global Institute. September 20; 2011.
- Pencina MJ, D’Agostino RB, Sr. Evaluating discrimination of risk prediction models: The C Statistic. JAMA 2015;314(10):1063–4.
- Shrank WH. A case for why health systems should partner with pharmacies. Harvard Business Review 2005.
- Marr B. How big data is changing healthcare. April 21, 2015. www.forbes.com/sites/bernardmarr/2015/04/21/how-big-data-ischanging-healthcare/print/. Last accessed April 2016.