Nearly two years since Covid-19 first hit the United States, our healthcare system continues to grapple with staffing shortages, sicker patients, and an aging population that has overburdened the workforce. Fortunately, technologies such as machine learning algorithms (MLA) and artificial intelligence continue to evolve and support care teams in making important clinical decisions that lead to optimal levels of care and better patient outcomes.
In the coming year, we’ll see more health systems rely heavily on these tools to fill staffing gaps and streamline workflows. Here are three predictions for innovation in clinical decision support (CDS) in healthcare that will have a major impact on how our healthcare workers deliver the best level of care in 2022.
MLA in the form of a CDS tool will be seen as an integral part of care management to improve clinician workflow.
With the increasing availability of personal health data from electronic health records, genetic information, and continuous monitoring devices, CDS tools can enable personalized healthcare for individuals. MLA can provide patients with risk prediction and early detection of complex diseases such as sepsis, diabetes or alcohol use disorder. In addition, they can assist clinicians by classifying patients into disease risk groups and categorizing disease severity in individuals.
CDS tools are already used in many different care settings, and their potential impact is enormous.By using large datasets and computing power cloud-driven technology, these tools can integrate and analyze multiple data input variables, resulting in countless benefits for healthcare systems. Both MLA and CDS tools are currently used in a variety of applications, including advising patient care, optimizing workflow procedures, and assisting healthcare providers with care setting challenges.other applications, such as using Improving priorities for diagnostic imaging, is also booming. Strong support from radiology and imaging providers attests to the clinical utility and expected market growth of such tools.
looming reality 500,000 nurses to retire by 2022 Continue to exacerbate staffing shortages in an already strained healthcare system.Resource-constrained suppliers should leverage CDS tools to address critical issues Patient-Centered Emergencies Such as patient status, potential risk of exacerbation, patient status and probability of readmission or death. In addition, the focus can provide solutions to problems in the healthcare setting; optimize patient flow and predict expected discharges or bed shortages; and assist with medical records and workflow, among others. Importantly, as dynamic healthcare situations evolve, such as the Covid-19 pandemic, CDS tools can also be rapidly implemented to address new emergency patient indications and care management practices.
MLA developers will demonstrate the effectiveness of these tools and increase their adoption through clinical validation, provide transparent and well-understood tools, and produce unbiased and externally validated products.
The past few years have seen changes in understanding the benefits of CDS tools and areas for improvement; comprehensive Comment in peer-reviewed journals, and criticism and give back Products and services from consumers and stakeholders have been launched.These discovered tools for patient safety and Reduce bias in healthcare delivery in certain settings.
However, areas for improvement were also identified: the risk of biased algorithm development data and subsequent results; the lack of clinical validation in prospective studies; and the real danger of alert fatigue, which clinicians ignore due to a flood of unnecessary or clinically irrelevant alerts CDS alert.With the first wave of feedback and demands for more regulations, MLA developers have Think critically about how to solve these problems. Several solutions have been discussed to provide transparent information on algorithm design, data training and development, feature input, and potentially biased results. Armed with this information, healthcare providers and stakeholders can more effectively understand and select CDS tools that fit their needs.
MLA developers are at a critical juncture in communicating with healthcare professionals, clinicians, patients and hospital stakeholders.
Regulatory pathways currently available for MLA in healthcare settings include FDA 510(k) or de novo regulatory approval for Software as Medical Device (SaMD) products, or CDS tools overseen by the Office of the National Health IT Coordinator (ONC). Seek FDA approval as a SaMD product may “lock in” algorithm design, inputs, alert delivery systems and intended patient populations, among other capabilities. Additionally, producing CDS tools overseen by the ONC would give developers the flexibility to adapt MLAs to respond to dynamic health conditions (eg, Covid-19). The ability to update algorithms helps optimize the performance of CDS tools for specific hospital data, patient populations, rapidly changing clinical care management practices, and specific healthcare professional preferences for alert notifications. This affects the clinical relevance of the tool, clinician adoption and product availability, and ultimately patient benefits and outcomes. While FDA approval is not required to deploy CDS tools in healthcare settings, physicians and clinicians rely on such regulatory green signals as markers of safety and efficacy. It is therefore the developer’s responsibility to provide an interpretable MLA that clinicians can understand and rely on based on validated results and external validation.
Medical knowledge is growing rapidly, as are care management protocols and dynamic patient conditions. We will need to augment our human skills with machine learning to provide the best possible care. A Fundamental Theorem of Biomedical Informatics It is “better” for people who work with information resources than those who don’t. Consistent with this, we can optimize clinician workflow and patient outcomes through human-machine interaction and support tools. MLA developers combine machine learning data-driven analytics Bring practitioners with clinical expertise – together they are better than one.
Photo: Getty Images, Andrei Popov
