Research Paper in Public Health: Use of Big Data Analytics to Prevent Medication Errors

Medication errors are becoming rampant in hospital settings. Even though these errors are preventable, most of them are as a result of adverse drug effects (ADEs). ADEs are defined as any patient response to a drug, which is unintended, noxious, and which occurs at doses that are typically used in people for diagnosis, prophylaxis, and even therapy for diseases (Oren, Shaffer, & Guglielmo, 2003). As the authors articulate, medication errors that are attributed to illegible handwritten prescriptions, drug interactions, incorrect dosages, and overlooked allergies usually result in ADEs. Oren et al.s (2003) assertions are supported by research findings of Agrawal (2009), who highlighted the high frequency of medication errors in hospital settings along with the resultant patient cost and harm. For instance, Agrawal (2009) pointed out that medication errors are approximated to harm more than 1.5 million patients on a yearly basis, with about 400,000 preventable ADES. Agrawal (2009) established that indeed the prevention of medical errors should be made a priority for health systems. As Agrawal (2009) revealed, there is increased evidence that systems that capitalize on the use of information technology, including automated dispensing cabinets, computerized physician order entry, electronic medication reconciliation, as well as bedside barcoded medication administration, are key strategic considerations for the prevention of medical errors. Besides, use of the IT systems has the capability of allowing healthcare facilities to save approximately $88 billion in the next ten years in costs in the US with increasing adoption. Additionally, the use of IT systems in hospitals automated order entry, clinical decision support, and note and records have reported lower costs, lower mortality rates, as well as fewer complications (Agrawal, 2009).

Therefore, owing to the numerous medical errors, technology-based interventions have been recommended as a mechanism to reduce the likelihood of ADEs and medication errors. One of these technologies includes big data analytics. In essence, according to Ayers (2016), big data allows medical experts to analyze massive data sets in the healthcare sector, and thus, providing assistance and insights in solving healthcares most significant challenges, including rectification and mitigation of medical errors. Big data is a widely defined term, and therefore, according to Ward and Barker (2013), the term has become ubiquitous, owing to its shared origin within the media, academic, and industry proportions. As such, different stakeholders have provided variant definitions. Big data is composed of huge datasets whose capacity exceeds the current capability of conventional database and software tools in capturing, storing, managing, and analyzing these vast datasets (Demchenko, 2013). For this reason, todays Big Data may not be used in the future, which provides evidence of the fact that big data is ever growing (Franks, 2013). As Lohr (2013) purports, interest in Big Data has gained enormous attention, and its interest since 2011 has increased exponentially. Big Data is becoming related to all aspects of human activities, from mere events recording to design, research, digital services, production, as well as product delivery from manufacturers to the final consumer (Demchenko et al., n.d). Current technologies, including geographic information systems (GIS) and cloud computing, as well as the Internet, have allowed the automation of all data handling processes, including data collection, data storage, processing, and data visualization (Grimmer, 2015). GIS allows for real-time mapping of the medical errors in various departments. These can help reduce medication errors in the hospital setting. The purpose of this paper, therefore, is to highlight the importance of big data in the hospital settings, specifically in reducing medical errors.

Medical Errors in Healthcare Settings

According to Ayers (2016), prescription errors are increasingly becoming a significant problem, and big data is one of the technological tools that can be used in preventing prescriber mistakes. According to Stetson et al. (2002), medical errors are attributed to system failures that predispose to adverse clinical events. The failures are due to lack of appropriate and accurate information during clinical care. Also, problems with clinical personnel communication are commonly attributed to the medical errors (Stetson et al., 2002). Besides, the authors highlighted that many of the errors that lead to adverse drug effects (ADEs) and adverse effects (AEs) are preventable. 56% of the errors occurred at the stage of ordering, 34% in administration, 6% in transcription, and 4% during the dispensing process. These errors are categorized as cognitive errors as opposed to errors due to accidents, for example, a slip of a scalpel. The cognitive errors are classified as mistakes or slips, as shown in Figure 1.

Figure 1: Conceptual schema for errors and information needs. Source: Stetson et al. (2002).

Latent errors as shown in Figure 1 usually happen as a result of poor system design, and in most instances, they are not preventable by hospital personnel. The system failures consequently predispose to mistakes and slips. The most common sources of system failure errors are inadequate dissemination of drug knowledge and the unavailability of adequate information about the patient (Stetson et al., 2002). Therefore frequent errors are commonly caused by impaired access to information. This is further aggravated by the reliance on patient information with other healthcare workers compared to the use of paper-based resources that may have accurate information about the patients. However, it should be noted that omissions may result in inadequate information while making records on papers. As such, this calls for information automation, and capitalizing on big data analytics ensures that accurate patient information is available. Necessarily, the high mobility of the physicians coupled with the interrupt-driven hospital environment can institute synchronous-bias among the workers leading to most of the medical errors presented in Figure 1. To eliminate these errors, Stetson et al. (2002) recommended the use of technology to address the mobility issue that involves incorporation of a message board while also acknowledging tasks while also keeping a role-based database that can flag urgent tasks requests while also improving collaboration among the team members.

Use of Big Data to Prevent Medical Errors in Healthcare Settings

The concept of big data is characterized by veracity, velocity, variety, and volume, and usually goes beyond data types and includes data analysis including hypothesis generation instead of testing the hypothesis. It mainly focusses on association and when used in the healthcare settings can be used as a predictive tool, clinical decision support, surveillance of safety and diseases, research, as well as public health (Lee & Yoon, 2017). According to Grimmer (2015), this can be facilitated by geographic information systems (GIS) and cloud computing, as well as the Internet. In essence, using GIS, it is possible to increase hospital settings surveillance, and thus, help decrease the number of medical errors committed. Also, the Internet can be used in conducting research purposes while cloud computing can be used in predictive modeling, which when coupled together yields a reduction in medical errors. Mostly, big data analytics usually exploits the analytical methods that are developed specifically for data mining, such as regression, clustering, and classification. Therefore, big data analytics can be used in creating knowledge-based expert systems that can be used in avoiding medical errors.

Ayers (2016) articulated that some companies are currently capitalizing on big data, for example, MedAware, to reduce the number of prescription errors. The healthcare facility scans the patient data and consequently flags any prescriptions that do not match up with the patients file records. Once an error is detected, then the drug is not ordered until the physician responsible confirms the order, rectifies it or prescribes the correct drug. Also, as Ayers (2016) opines, even though current alert systems can catch the patients drug interactions and high drug doses, they are unable to determine wrongly prescribed orders. As such, there is a need for healthcare facilities to adopt big data, which will consequently increase patient safety by eliminating medication errors. Besides, reducing malpractice is another way that big data is revolutionizing the healthcare industry, which can be advantageous as it allows hospitals to hire only competent staff and not incompetent personnel who would otherwise lead to medication errors. Additionally, Ayers points out that big data is the backbone of artificial intelligence technology, which will eventually be used to scan cloud-based data to help prevent misdiagnosis and transcription errors. Even now, some hospitals are moving to cloud systems to help reduce errors, and integrate patient records more efficiently.

According to Whinters-Miner (2004), there are a variety of ways in which predictive analytics, a big data analytics tool can be used in improving healthcare, and thus, lead to a reduction in the number of medical errors. As the author articulates, predictive analytic utilizes statistical methods and technology in searching through massive amounts of information and predicting the outcomes for individual patients. In essence, the data can be from past treatment outcomes and the latest medical research. According to Whinters-Miner (2004), the big data analytics tool can be used in increasing the accuracy of the diagnoses, thereby reducing ant instance of medical errors. In addition, one of the major reasons for medical errors that is that the physicians may inaccurately prescribe the medication, and according to Whinters-Miner (2004), the predictive analytics provides the physicians with specific answers for the individual patients thereby reducing the likelihood of making the medical errors. Also, most hospitals also use handwritten scripts that are mostly paper-based, which increases the likelihood of committing errors. The use of digital entries, which will be facilitated through the use of big data will, therefore, reduce prescription errors, which are mostly paper-based (Raghupathi & Raghupathi, 2014).

Besides, as Raghupathi and Raghupathi (2014) point out big data analytics have allowed for increased veracity, volume, velocity, and variety of data, which when coupled together can lead to the reduction of medical errors. Big data allows for the collection of huge amounts of data sets in the hospital setting, including personal records, human genetics, and population data genomic sequences. The increased information volume allows hospitals to capitalize on the big data through the advancement of data management, cloud computing, and virtualization, which are platforms that allow for effective data capture, storage, as well as manipulation of large volumes of data, including genomics, 3D imaging, as well as biometric sensor readings, which allow for accurate medical processes that leave no room for errors (Raghupathi & Raghupathi, 2014), primarily because most of the medical errors are preventive. With the increase in accuracy of medical processes, less medical errors will occur. Besides, the increased velocity and constant flow of data will allow the practitioners to access information about patients, eliminating chances that they must talk to other practitioners or cross-check in a paper-based platform that can incur medical errors.

In fact, most of the hea...