Exploring the critical utility of big data in boosting performance in healthcare (YEL2022)
Date: 17 November 2022
Exploring the critical utility of big data in boosting performance in healthcare (YEL2022)
Authors: Jessica Gannon (Australia), Rizza Jean Rivera (Philippines) J Laurence Arnfinsen (Norway), Abel Mwale (Zambia), Pritindira Kaur (India).
Reviewers, YEL Alumni: Sudha Pathak (UAE), Fatma Al Jahwari (Oman).
Introduction
In recent years the utilization of big data has become more prominent within the healthcare sector. The possibilities for utilizing big data stretch far beyond improving patient outcomes. The application of big data has also shown promise in improving staffing and management, reducing costs and inefficiencies, product development, and strategic planning (Kruse et al., 2016). Big data equips healthcare institutions with the ability to rapidly capture and analyze vast amounts of data into meaningful information.
Big data in healthcare
Over the years, the term Big Data has been coined to mean a massive amount of information on a given topic and includes information that is generated, stored, and analyzed on a vast scale (Tulane University et al., 2021). While there may have been delays in the collection and utilization of data in healthcare, Big Data analytics is embarking to make significant changes in the healthcare industry (Datapine et al., 2022).
Zooming in on big data in healthcare, the input data is just as important as the output. Therefore, collecting high-quality data requires optimization of data collection tools in health care and proper use of such tools by both the patients and the providers. Velocity, volume, and variety might be necessary aspects of defining big data. However, in healthcare, the utility of data is often not confined to the size of a data set but rather the ability to draw meaningful analysis from data that facilitate decision-making. The ultimate goal is to boost performance and patient outcomes. The sources of healthcare data may vary between different healthcare systems. The most common sources from which data is extracted include medical records, Surgical records, Behavioral data (i.e., patient’s diet), Biometrics (i.e., patient’s blood vital signs), and living conditions, among others.
The above data sources mainly contain patient-centered healthcare data. Other forms of traceable data in healthcare include:
Staffing schedules (i.e., determine how many medical workers to put on staff in a given time period to care for patients)
Process times (i.e., patient waiting time, room preparation time, medication delivery time in inpatient areas)
Insurance claim data
Medical referrals
Employee performance metrics (i.e., number of patients cared for per hour)
Supply chain metrics (i.e., for ordering the correct amounts of personal protective equipment)
Consumption data (for example, patient wise drugs, and consumable items)
The scope of big data in healthcare
While the size of the data may matter depending on the context and industry, the focus should not be on how “big” the data is, but on how smartly it is managed. With the help of the right technology, data can be extracted from the above-mentioned sources in a smart and fast way.
1 Electronic health records (EHRs)
The EHR is the most widespread application of big data in medicine. Every person has their own digital record which includes demographics, medical history, allergies, diagnostic results, monitoring parameters and treatment records. Records are shared via secure information systems and are available for providers from both the public and private sectors. Every record consists of one modifiable file, which means that doctors can implement changes over time with no paperwork and no danger of data replication.
Although EHR is a great idea, many countries still struggle to fully implement them. U.S. has made a major leap with 94% of hospitals adopting EHRs according to the Health Information Technology for Economic and Clinical Health (HITECH) research.
2 Utilizing big data in lowering waiting times
There is something more universal than standards of care in the healthcare setting and that is, waiting time.
The length of waiting time has become an important indicator of the efficiency of medical services and the quality of medical care. Lengthy waiting times for patients will inevitably affect their health-seeking behaviors and reduce satisfaction. For patients who are in urgent need of hospitalization, delayed admission often leads to exacerbation of the patient’s condition which may consequently threaten the patient’s life.
Waiting time is defined as the total time that a patient spends in a facility from arrival at the registration desk until the time she/he leaves the facility. (Biya, et. al, 2022)
In one of the studies on waiting time, the findings indicate that on average, patients wait for more than two hours from registration to getting the prescription slip, while the contact time with medical personnel is only on average 15 minutes. Employee surveys on factors contributing to the lengthy waiting time indicate employee attitude and work process, heavy workload, management and supervision problems, and inadequate facilities to be among the contributory factors to the waiting time problem. (Pillay DI, et al)
Healthcare facilities all over the world strive to find means to utilize big data in improving patient outcomes and decreasing waiting time. In Davao Oriental Provincial Medical Center, a 100-bed capacity hospital in the Southern Region of the Philippines in Southeast Asia, healthcare managers utilize big data in employing process improvement measures and ensuring client delightedness and satisfaction. Although health executives admit that they still got a long way to go especially in the advent of utilizing data in healthcare vis a vis the latest trends in healthcare technology, they continue to strive to improve waiting time by gathering data through pen and paper survey method for patient’s satisfaction and delightedness.
As the delivery of care continues to evolve, we continue to see innovations and strategies being set up to reduce waiting time in hospitals. Healthcare managers and executives continue to develop processes and strategies to lessen the waiting time. However, the first step in addressing such waiting time is to collate and make use of big data. The use of these extremely large data sets was analyzed computationally to reveal patterns, trends, and associations, especially relating to waiting time in the hospital.
It is important to find the key factors affecting waiting time for admission, to improve the status quo. The results of big data utilization and analysis allow health care key personnel to develop recommendations for hospital admission management, which can assist in patient management and improve work efficiency.
3 Utilizing big data to improve laboratory turnaround times
Laboratory turnaround times are key drivers in improving patient flow in hospitals and health services and in the provision of diagnostic care. Evidence suggests that early diagnosis and intervention are generally associated with better outcomes for patients. This phenomenon is particularly evident in Emergency Departments (ED) where the length of stay (LOS) correlates closely with a longer inpatient LOS, suggesting that delays in management and diagnosis in the ED continue to have an impact and repercussions on admitted patients. The bulk of early diagnostic data in a patient encounter is laboratory and imaging data. Consequently, improving laboratory turnaround times will have positive effects on the end-to-end patient journey.
Laboratory medicine has always been one of the medical disciplines with the highest degree of digitalization. Since its emergence, automation, electronic medical records, electronic transmission of results and electronic reporting have become prevalent in healthcare. Medical laboratories maintain extensive data bases of test results and various quality controls. As a result of increasing accessibility to high volumes of data, the expectations and promise related to the fields of Big Data and Artificial Intelligence (AI) are growing exponentially in the field of laboratory medicine (Blatter, T, U., Witte, H., Nakas, C, T., and Leichtle, A, B., 2022).
Several examples of AI applications during COVID-19 have been reported such as AI-enabled outbreak tracking apps and prognosis prediction tools There are other examples where Big Data and AI in laboratory medicine have been used for the management of non-communicable chronic diseases such as cardiovascular disease and cancer where they have been demonstrated to support decision making by improving both diagnostic and prognostic performance. One such example is, AI-powered Cardiovascular Disease Risk Score (AICVD Risk Score), developed by Apollo Hospitals, India, helps physicians to better predict the risk of cardiac disease in their patients. The tool is built on algorithms based on ten years of anonymized data relating to around 400,000 individuals across the country collected by the team at Apollo Hospitals. It delivers CVD risk score using algorithms and helps in developing standardized care regimens. The aim of the use of AI is to augment the effectiveness and accuracy of clinical decision making and improve the safety and quality of care delivered to patients (Blatter et al, 2022 and Molero et al, 2022).
In addition to the clinical impact, it is worth noting that Big Data and AI can contribute to laboratory efficiency through process improvements, recognising areas of waste, improving processes and rationalizing laboratory test ordering.
Laboratory processes are complex; however, it is recognised that data science has the potential to improve complex analytical tasks and flows within the laboratory domain. Data science could be used to examine data in real time and calculate the most efficient clinical and operational pathways. If these systems are applied with lean methodologies, there is potential for significant improvements in process flows (Schulte and Bohnet-Joschko, 2022).
A quality improvement initiative was implemented at the Rush University Cancer Centre (RUCC) Chicago to optimize pre-analytic laboratory processes in two practice settings to enhance efficiency, quality, safety and patient care delivery. The issues identified were a rise in patient wait times, laboratory errors, lengthy turnaround times and decrease in patient satisfaction (Wojciechowski et al, 2022). A total of four lean initiatives were chosen and implemented to standardize workflows and increase lab order adherence. The first initiative was an infusion chart prep plan, which detailed a multi-step check process to ensure information needed prior to the patient’s appointment was readily available. The second, team-specific lab panels were created for each oncology subspecialty. The third initiative; a preview screen in the EHR was developed, the aim of which was to decrease the amount of time spent on the lab pre-check. The final initiative entailed a modification to the appointment template in the EHR, to prevent the clinic overbooking and reducing efficiency. The intended outcomes of this process redesign project resulted in positive results from baseline to post implementation. Specifically, a decrease in lab turnaround times, patient wait times, safety events, and inefficient use of staff time was accomplished (Wojciechowski et al, 2022).
The use of big data, and data more broadly in enhancing laboratory turnaround times gives rise to countless opportunities for improvement. There is large scale potential for the increasing use of artificial intelligence to deliver more precise and data driven decision making for population health outcomes. Its application is also relevant in smaller organizations looking to enhance the patient experience, improve safety and efficiency.
4 Utilizing big data for patient safety improvement
Safety is an issue that all healthcare organizations must prioritize. Patients come to hospitals to seek treatment and not to suffer from preventable harm. Patient safety not only helps in improving patient outcomes, but also creates system efficiencies by reducing the number of errors.
Patient safety is supported through quality measures that include clinical, operational and cost data analytics for providing actionable insight and direction for improvement. Leveraging predictive analytics and machine learning can make safety an overarching cultural goal. Machine learning supports patient safety improvement with capabilities that are reactive, proactive, and fully integrated and helps to further enhance the embedment of the culture of safety (Health Analytics, Apr 2019)
The following are the methods through which big data can be used to improve patient safety:
Risk assessment
By implementing big data analytics, clinical risks for individual patients can be identified for certain chronic illnesses such as cardiovascular diseases, COPD, diabetes and other disease conditions. This helps in timely management and treatment of such patients thereby reducing individual risks.
Patient care process safety
Application of analytics for medication error data helps to improve the medication process by understanding the causal factors and mitigating the risk points. Analytics that are used for medication error analysis include- type of error eg. prescription, transcription, dispensing errors, documentation errors; Segregation of data- ward/ICU wise, drug type, physician wise, common errors. Improvement strategies, staff education are instituted as per the results obtained through the detailed analysis.
As an example, Apollo Hospitals, uses an algorithm called HIPAR (Healthcare India Pharmaceutical Registry). This module is a database of drugs used in the hospital and interacts with the HIS to generate alerts for the physicians and the nursing times (Apollo Hospitals et al, 2022)
Visual dashboards for measuring patient safety performance
Improved visual analytic capabilities and big data dashboards are invaluable for performance improvement. Visualizations make complex datasets clear by presenting information in intuitive and user-friendly ways, helping to dive deeply into existing data assets and identify opportunities for improvement. Dashboards show the progress being made and provide critical management information about effectiveness and efficiency (Healthcare Analytics,2017). Patient safety dashboards provide a visual overview of the information needed for patient safety performance and facilitate decision-making and improvement (JMIR, 2022). Examples include the Patient Safety Dashboard tool used by the Brigham and Women’s Hospital (BWH). The dashboard is used for the clinical staff in the Medicine, Oncology, and Neurology Care Units. Data is gathered from various siloed parts of the electronic health record and placed in one convenient display, prompting interdisciplinary conversations regarding patient safety concerns. Tailored and personalized for each service, the patient safety dashboard encourages a collective, multi-disciplinary focus on preventing avoidable harms (BWH, HsyE Northeastern University et al, 2015). The Apollo Quality Program (AQP), the patient safety dashboard used by Apollo Hospitals helps to monitor compliance of essential patient safety parameters across its various locations in the country. It has helped to strengthen the culture of patient safety and clinical governance across the Group Hospitals.
Diagnostic and therapeutic decision support for clinicians – enhancing clinical accuracy and minimizing diagnostic errors
For the purpose of handling diverse medical image data obtained from X-rays, CT-scan, MRI and in order to attain better insights for diagnosis, Big Data Analytics platforms are being leveraged to a great extent. Disease surveillance can be effectively improved using Big Data Analytics. Unstructured medical image data sets can be evaluated with great efficiency to create a better discernment about the disease and requisite prevention and curing methodologies, hence leading to much better critical decision making, clinical accuracy and minimizing diagnostic errors (Research Gate, 2021). For example, Apollo Hospitals partnered with Medtronic to integrate artificial intelligence (AI) for advanced stroke management. The AI software aims to provide automated analysis in less than two minutes as against currently accepted imaging practices for diagnosis of stroke that takes up to one hour to complete, thus enabling faster decision making in stroke where every second counts. The AI platform uses artificial intelligence to create high quality, advanced images from non-contrast CT, CT angiography, CT perfusion, and MRI diffusion and perfusion scans, helping hospitals to improve time-critical triage or transfer decisions and facilitate better patient outcomes (Apollo Hospitals, 2020)
Infection control and antimicrobial stewardship programmes
Big data analytics can provide patient safety teams with the insights that they need to reduce HAIs. Many of the current HAI surveillance strategies are labor intensive and subject to limitations (Healthcare Analytics, 2018). For reducing infections, one needs to analyze all the data — lab values, antibiotic resistance, patient diagnosis, and the primary treatment plan. The lab, the pathologist, the infection control team, the treating clinicians, the primary clinicians who are treating the condition are all part of a typical infection control protocol. With Big Data analytics doctors are able to see the pattern of the patient in question and that of similar patients within a few minutes. This is enabled through patient tracking for culture reports, device related infections, automated surveillance data, liasoning lab equipment with clinical dashboards and real time alerts. Big Data analytics aids Antibiotic Stewardship Program by tracking antibiotic usage patient wise, monitoring overall consumption patterns of antibiotics, generating anti-biograms and real time alert generation. As an example, Apollo Hospitals has implemented data analytics for presaging infection risks for better infection control. Its entire data in terms of the antibiotics, microbial systems, microbiology tests that go along with it and the diagnosis pattern of diseases has been analyzed for prescribing patterns to visually examine how to obviate and control hospital-acquired infections (ET CIO.com, 2020)
5 Key challenges
Insufficient data structures, issues surrounding privacy and data security, lack of interoperability of systems, management and governance and data reliability are some key elements that pose significant challenges within the realm of big data in healthcare (Kruse et al, 2016). The success of utilizing big data in the future development of the healthcare industry heavily relies on understanding its limitations and overcoming the myriad of challenges that lie ahead.
Data structure
The inherent structure of health data causes considerable challenges in drawing meaningful conclusions from big data. Due to technological progress and rapid adaptation of new digital tools there has been monumental growth in the volume, velocity and variety of data pertaining to health (Senthilkumar et al., 2018). Some estimates suggest that the global amount of available health data has increased by a fiftyfold during the past three years
Furthermore, the exponential increase in data is derived from a constantly expanding variety of sources which exceed far beyond traditional biomedical sources. Such sources range from new physiological trackers, wearable biosensors, a person’s digital persona and any other data source which affect social determinants of health. Many of these items and sources are constantly being updated in real time and yield huge amounts of health data. In order to sufficiently analyze and draw meaningful conclusions from big data it is crucial to have the ability to access data from a huge variety of sources which far exceed traditional biomedical sources. This is difficult to achieve considering that data from different sources are often siloed and are not easily rendered interoperable (Kruse et al., 2016). Moreover, the vast majority of health data is unstructured and is often fragmented, dispersed and rarely standardized which in turn is a significant concern when assessing the utility of big data.
Privacy and data security
The use of big data analytics in any industry entails extensive concerns regarding privacy and data security given the extensive and complex legislation by which institutions must comply. The concern of privacy and data security is especially true for healthcare institutions given their inherent disposition in handling sensitive patient data that has to be carefully protected. Big data in healthcare is, more often than not, derived from multiple sources which complicate the ability of healthcare institutions to comply with regulatory standards. Moreover, being able to collect, analyze and use data in real-time is an irrefutable requirement for utilizing big data in health care. Analyzing data in real-time often requires cloud computing technologies or other measures which ensure interoperability between systems. Such technologies pose significant threats to data security and handling data in accordance with regulatory standards.
To address these and other potential risks in handling healthcare data, the Health Insurance Portability and Accountability Act (HIPAA) Security Rule has come up with a list of safeguards for healthcare organizations storing Protected Health Information (PHI). These safeguards may involve encrypting sensitive data, enabling firewalls, implementing multi-factor authentication, and ensuring anti-virus software is up-to-date.
Some specific ways of ensuring the above objectives include practices such as:
Ensuring transmission security
Adopting authentication protocols
Managing controls over data access and integrity
Scheduling regular data security audits
Conclusion
Rapid advances in technology and digital transformation have spiked interest in utilizing big data for boosting performance in the healthcare industry. The utility of advanced big data analytics has paved the way for unprecedented opportunities and challenges in healthcare. In general, Big Data applications make it possible to aggregate vast amounts of information that is derived from multiple data sources, thereby enabling clinicians and healthcare leaders to make better decisions and enhance performance.
Big data has the potential to increase the quality of care by reducing waiting times in most healthcare facilities thus leading to better patient outcomes and client delightedness and satisfaction. Furthermore, the use of big data in enhancing laboratory turnaround times may significantly improve patient flow and clinical outcomes as clinicians are able to establish diagnoses more quickly and consequently initiate medical interventions. Big data may also contribute significantly to improving patient safety measures. Analytics has the ability to visualize complex data structures in an intuitive and user-friendly manner which in turn aid patient safety improvement measures and facilitate decision-making. Big data analytics may also boost clinical performance through various applications for assessing clinical risk of patients for certain chronic illnesses, disease surveillance, infection control and antibiotic stewardship.
Healthcare industries all over the world face a multitude of critical challenges. Healthcare workforce shortages, the possibility of future pandemics, antimicrobial resistance and rising costs. Although in its infancy, this paper demonstrates that Big Data may enable healthcare leaders and clinicians to gain additional insight that might lead to more purposeful conclusions, and enhance decision making and clinical outcomes. Even though big data has paved the way for unprecedented opportunities and innovation, the inherent limitation of big data surrounding data structure, privacy and data security warrant caution. Therefore, appropriate utilization of big data arguably relies on it being used in conjunction with established best practice in healthcare professions and robust governance processes. Consequently, big data may prove to be an important piece to a broader puzzle in order to manage and overcome present and future challenges facing healthcare service delivery.
In consonance with the International Hospital Federation’s thrust of encouraging global learning and propelling local actions to achieve sustainable healthcare, healthcare leaders should invest in utilizing big data as well as ascertain that technological and global innovations are welcomed. In this manner, the aim of harnessing big data in order to augment and extend innovation, clinical outcomes, and overall efficiency. The aim should be to utilize the power of big data to boost healthcare performance.
References
Belle, R. Thiagarajan, S. M. R. Soroushmehr, F. Navidi, D. A. Beard, K. Najarian, Big Data Analytics in Healthcare, (n. d.). doi:10.1155/2015/370194.
Blatter, T, U., Witte, H., Nakas, C, T., and Leichtle, A, B,. Big Data in Laboratory Medicine – FAIR Quality for AI? Diagnostics, 2022; 12 (1923), 1 – 13. https://doi.org/10.3390/diagnostics12081923
Biya, M., Gezahagn, M., Birhanu, B. et al. Waiting time and its associated factors in patients presenting to outpatient departments at Public Hospitals of Jimma Zone, Southwest Ethiopia. BMC Health Serv Res22, 107 (2022). https://doi.org/10.1186/s12913-022-07502-8
JMIR Publications, Advancing Digital Health & Open Science, Hilco J van Elten, Sandra Sülz, Erik M van Raaij, Rik Wehrens, 22.2.2022 in Vol 24, No 2 (2022): February, doi:10.2196/30201, https://www.jmir.org/2022/2/e30201/
Kruse CS, Goswamy R, Raval Y, Marawi S. Challenges and Opportunities of Big Data in Health Care: A Systematic Review. JMIR Med Inform. 2016 Nov 21;4(4):e38. doi: 10.2196/medinform.5359. PMID: 27872036; PMCID: PMC5138448.
Molero, A., Calabro, M., Vignes, M., Gouget, B., and Gruson, D. Sustainability in Healthcare: Perspectives and Reflections Regarding Laboratory Medicine. Annals of Laboratory Medicine, 2022; 41 (2), 139 – 144 https://doi.org/10.3343/alm.2021.41.2.139
Pillay DI, Ghazali RJ, Manaf NH, Abdullah AH, Bakar AA, Salikin F, Umapathy M, Ali R, Bidin N, Ismail WI. Hospital waiting time: the forgotten premise of healthcare service delivery? Int J Health Care Qual Assur. 2011;24(7):506-22. doi: 10.1108/09526861111160553. PMID: 22204085.
Schulte, T and Bohnet-Joschko, S. How can Big Data Analytics Support People-Centred and Integrated Health Services: A Scoping Review. International Journal of Integrated Care, 2022; 22(2): 23, 1–18. DOI: https://doi.org/10.5334/ijic.5543
Senthilkumar SA, Bharatendara K Rai, Amruta A Meshram, Angappa Gunasekaran, Chandrakumarmangalam S. Big Data in Healthcare Management: A Review of Literature. American Journal of Theoretical and Applied Business. Vol. 4, No. 2, 2018, pp. 57-69. doi: 10.11648/j.ajtab.20180402.14
Wojciechowski, M, and Rodts, M, F. Laboratory Process Improvement: A quality initiative in an outpatient oncology clinic. Clinical Journey of Oncology Nursing, 2022; 24 (5), 571 – 574.
