If we could save upwards of $30 billion a year on avoidable healthcare costs, why wouldn’t we? This is what the CDC estimates that hospital-acquired infections cost annually. 1 in 25 patients will suffer a hospital-acquired infection—many of these result in actual harm to the patient. For example, a central line or urinary catheter left in place too long causes an infection. Frustratingly, hospitals can also have these infections identified that may not be true infections, but fall into the NHSN criteria. These are costly in terms of unnecessary testing, financial penalties for hospitals, and lower publicly reported scores.

AI would alter the approach with an infection prediction with increased diagnostic accuracy. It could help discern acute inflammation from infection.

AI is poised perfectly to help us predict the patients that will get these infections through the use of predictive analytics scanning vast amounts of data combined with real time monitoring of the patient’s vital signs and other data points. This would allow us to mitigate those risks by removing or replacing the problematic lines. It can also predict multidrug-resistant organisms that could put a patient at risk. Multidisciplinary teams all have their roles in preventing these infections and AI suggestions and recommendations could be targeted towards the members of these teams.

AI would alter the approach with an infection prediction with increased diagnostic accuracy. It could help discern acute inflammation from infection.

While the risk of hospital-acquired infections depends on the hospital’s infection control practices, and those steps taken to reduce the risk, patient factors opposing these efforts are also at play which include immune status, recent antibiotic use, frequent visits to healthcare facilities, length of stay (LOS), major procedures, age, ventilatory support and intensive care stays. It seems quite plausible that artificial intelligence (AI) could identify risk factors and generate a score. Steps could be suggested and taken to mitigate infections by keeping devices out as much as possible and guiding clinician care decisions.

Cleveland Clinic investigators recently presented that AI could very accurately predict multi drug-resistant organisms days prior than a culture is available. It is exciting to think that we can use AI to predict and tailor antibiotics and isolation precautions towards these days ahead of a final culture. Taking antimicrobial stewardship to the next level to get patients appropriately treated earlier will save lives, time and money.

AI has recently been used to model new designs of urinary catheters to block the migration of bacteria towards the bladder. Catheters were made consistent with these designs, creating an obstacle course of geometric designs inside the catheter that blocked the migration of bacteria upstream. The design was optimized for E. coli, and testing showed that after 24 hours the bacterial burden was 1/100 of that of traditional Foley catheter design. This is exciting that we can use AI technology to predict the behavior of microbes and design ways to inhibit their growth and migration.

Machine learning (ML) algorithms have demonstrated value in predicting clostridium difficile infection with just 6 hours of data. With almost 30,000 deaths per year related to c. difficile infections, early diagnostics to treat, identify those at risk and isolate to prevent spread would be an incredible advance to saving lives.

CLABSI (central line-associated bloodstream infection) could be predicted allowing physicians to remove the lines prone to infection and avoid those consequences. Suggestions of treatment based on probability and risk would help discern true CLABSI from blood culture contamination.

The support for antimicrobial stewardship that AI could provide would adjust the approach toward treating infections. Currently, the physician evaluates data for the likelihood of an infection. Cultures are taken, the results of which will not be available for several days, and empiric antibiotics are started. When cultures and sensitivities are available, antibiotics are sometimes changed based on those results, or de-escalated. Have we then given a patient a toxic or broad-spectrum antibiotic for a few days that was unnecessary? Have we bred more resistance? AI would alter the approach with an infection prediction with increased diagnostic accuracy. It could help discern acute inflammation from infection. (Is this sepsis or something else?) The correct antibiotic could be chosen which would eliminate the need to de-escalate or change, and the best duration of therapy would be suggested.

Enhanced cleaning and sterilization practices could be suggested by algorithms to identify and mitigate risks with equipment and high touch areas in the healthcare setting.

Patient and caregiver education and engagement could be enhanced through AI based on their medical conditions or procedures to provide targeted and relevant material about their care and infection prevention practices. This would certainly foster a collaborative culture of safety, and mitigate the spread of infections.

I dream of a day when a patient is admitted to the hospital, using AI tools, we are able to get predictive scores on the likelihood of hospital-acquired infections or other complications. More informed decision-making can be made based on the probability instead of blindly pan-culturing when not needed or leaving devices in patients at high risk. Imagine that when a patient comes to the hospital, your risk of CAUTI, CLABSI or hospital-acquired pressure ulcer is present and available to the admitting team so that decision-making to reduce or eliminate this outcome altogether can be made. A most likely diagnosis and risks are presented along with the likely infection cause/ Is it MDRO or not? The right antibiotics are given and a short LOS gets the patient home safely and efficiently. Lives will be spared, and billions of dollars potentially saved.

The post AI to Combat Hospital-Acquired Infections – A Revolution for Patient Safety appeared first on HealthTech Magazines.

A gifted few have a memory of an elephant, smoothly juggling large amounts of information, but many of us rely on reminders. Similar to healthcare, it is impossible to remember every aspect of patient care. Medicine is quite complex, and memory alone cannot keep up with changing best practices, standards of care, and medical technology growth while simultaneously taking care of multiple patients.

The relentless pursuit of safety in aviation and the chase for patient safety and quality in healthcare are often studied and compared. One of the widely touted safety successes in aviation is using checklists. In the insightful book The Checklist Manifesto, we discover the transformative power of a simple checklist. Translated to current healthcare, clinical decision support systems (CDSS) have become the dynamic checklist of modern medicine. When embedded in the electronic health record (EHR) workflow, it will ensure standards of care that will standardize diagnostics and treatment plans.

Understanding CDSSs

CDSSs provide recommendations, alerts, or other methods designed to assist clinicians with making decisions with patient care, ideally at the point of care. However, what does that entail during patient care? It depends on the type of CDSS deployed. Primarily, there are two types: knowledge-based and non-knowledge-based CDSS. Knowledge-based CDSS leverages a comprehensive knowledge base comprising medical data and rules to provide informed recommendations and alerts to clinicians. Non-knowledge-based CDSS deploy machine learning (ML) algorithms and artificial intelligence (AI) to analyze vast data volumes, identifying patterns and trends that may not be immediately discernible.

Additionally, CDSSs are characterized as active or passive, meaning a passive system requires the user to do something to receive advice, like opening a dialogue box. In contrast, an active system automatically provides prompts based on information received. CDSS are potent tools that manifest in different forms, such as alerts, clinical diagnostics, drug contraindications, or order sets. We have seen evidence of effectiveness in medication safety with consistent CDSS usage. My organization incorporated drug-to-drug alerts that warn of possible allergy contraindications and suggest alternatives. Although pharmacists also review drug interactions, a CDSS provides an initial check that saves pharmacist a 45% reduction in time spent on verifying. The next step is to incorporate a decision support system that can adhere to clinical practice guidelines (CPGs).

While the adoption of CDSS in healthcare settings is growing, there are still substantial challenges to overcome. As we advance, the focus should be on addressing these barriers to further increase the adoption and optimization of these systems.

Clinical Practice Guidelines

Clinical practice guidelines are a consortium of recommendations to aid with diagnosing and treating a concern or condition. “The principal benefit of guidelines is to improve the quality of care received by promoting interventions of proven benefit and discouraging ineffective or potentially harmful interventions.” Additionally, CPGs can assist with diagnosis variance and treatment plans by establishing standards, including physician biases, which adversely affect patient care. “Bias in clinical practice, in particular in relation to race and gender, is a persistent cause of healthcare disparities.”

With the advancement of technology, the EHR can incorporate continuous CPGs. Unlike traditional provider methods of solely learning new techniques via continuing education credits or conference attendance, CPGs are updated as new facts and acceptances are established in real-time, which allows for the utmost patient care. “Clinical practice guidelines are the most important document for incorporating scientific evidence into healthcare decision-making recommendations.” When the CDSS is active within the workflow, it can reduce the biases for treatment.

Addressing Misconceptions About CDSS

Oftentimes, physicians’ reluctance to use CDSS is because of the notion that their expertise is discounted. However, CDSS is a tool that complements a clinician’s knowledge and experience while mitigating implicit bias. Additionally, it provides clinicians with additional information and resources, assisting them in making better-informed decisions. The ultimate decision will always be with the physician.

There’s a concern that clinicians might become overly reliant on CDSS, leading them to overlook their clinical judgment or become less skilled over time. CDSS is an aid, not a substitute, for thorough clinical assessment and sound clinical judgment. It’s there to support and enhance decision-making, not to take it over. Lastly, CDSS is not infallible. Various interest organizations have their “own perspective, goals, and intended uses.” Therefore, CPGs incorporated in a CDSS must be transparent, undergo a rigorous review process, and remain independent from payers. Additionally, it’s crucial for clinicians to critically evaluate the recommendations provided by CDSS, considering the entire clinical context and their knowledge.

Need for Widespread Adoption

Paying homage to the movie Field of Dreams alludes to the idea that “If you build it, he will come.” This is not the case for CDSS. Adoption by users remains low even though the evidence shows an increase in patient safety and quality. Suggestions or alerts are ignored for “reasons attributed to end-users’ negative attitudes, evasion or skepticism regarding the system.” Other barriers include a lack of integration within the EHR and trust in the system due to data quality issues. Due to common misconceptions, healthcare culture and resistance to change remain a significant barrier. While the adoption of CDSS in healthcare settings is growing, there are still substantial challenges to overcome. As we advance, the focus should be on addressing these barriers to further increase the adoption and optimization of these systems.

In conclusion, CDSSs are not merely technological advancements but are dynamic checklists poised to reform modern healthcare from its very core. Integral to EHR workflows, CDSS ensures that standards of care are met and consistently exceeded, standardizing diagnostics and treatment plans across the board. The result is an enhanced safety profile with tangible evidence in medication error reductions and a streamlined approach to healthcare delivery. I urge healthcare professionals, policymakers, EHR vendors, and experts to advocate for integrating and optimizing CDSS in healthcare. Our collective effort will ensure that these systems are not only adopted but also to ensure high-quality patient care. The future implications are clear: CPGs, intertwined with CDSS and embedded into clinical workflows, are set to become critical tools in directing evidence-based medicine into actionable healthcare decisions.

The post The Role of Clinical Decision Support System in Achieving Patient Care and Enhancing Patient Safety appeared first on HealthTech Magazines.