Imagine being a nurse on the general ward of a hospital. You have nine patients, several with multiple diseases. You must keep a close eye on vital signs, log a long list of metrics and most importantly, ensure you’re watching for any early signs of deterioration. It’s impossible to be everywhere at once – where do you focus? As Harvard professor emeritus E.O. Wilson articulates, “We are drowning in data while starving for wisdom.”
Today’s clinicians face a serious challenge: they’re collecting and accessing more data than ever, but rather than translating into positive patient outcomes, they’re getting overwhelmed by the data. Which numbers are meaningful? It’s a difficult situation to navigate and one that contributes to the 440,000 adverse but preventable events leading to patient deaths in U.S. hospitals every year.
The troves of data clinicians collect are useless if they can’t be utilized to improve patients’ outcomes. Physicians and nurses agree: 73% of physicians and 79% of nurses said patient safety keeps them up at night, with incomplete data being a top threat. 
With their ability to identify patterns and critical data points, intelligent technologies can augment clinical decision-making helping nurses and doctors know whose care to prioritize, making them more efficient and able to improve patient outcomes. As these technologies offer more diagnostic and clinical decision support, they can help to deliver more appropriate treatments in a timelier manner and because of that, can reduce costs for both the provider and the patient.
At Philips we believe in adaptive intelligence: technology that adapts to the people using it, not the other way around. Adaptive intelligence combines artificial intelligence with knowledge of the clinical, operational, or personal context in which they are used. Here are a few examples of how we can create measurable impact for clinicians and their patients:
Diagnosis and Treatment – In patient imaging, AI can provide critical context around anatomy and patient history (including lab results, diseases, prior radiology reports) to allow the radiologist to more thoroughly understand what they’re seeing. For example, there are advanced diagnostic solutions that can feed in a patient overview and disease data gathered from thousands of other patients, enabling clinicians to compare data, determine abnormalities and decide the best treatment. Patients are also more equipped to understand the situation, and empowered to participate in their care plan.
Other solutions utilize AI to organize data and create important context and meaningful information. Certain genomics platforms can help pathologists and oncologists unify the whole patient picture – genomic profiling, imaging, patient history, anatomic pathology – as a way to more precisely match tailored treatments to a patient’s needs. These applications bypass large amounts of data to provide an oncologist with the exact information he needs, exactly when he needs it, to provide efficient care for each patient.
Acute Care – technologies that use adaptive intelligence are one of the best ways we can help the busy nurses I mentioned earlier. Early Warning Scoring is a capability that aggregates patient data and applies intelligence to detect signs of deterioration early (sometimes 6-8 hours prior to an event). The system then alerts nursing staff, helping to take the guesswork out of clinical decision-making and, in the case of Saratoga Hospital, reducing patient transfers to the ICU by 63%.
Learn more about Philips adaptive intelligence - applying AI in a meaningful way to improve people's lives.
AI can also be applied to remote patient monitoring scenarios to further improve efficiency and overall costs. After a hospital stay, many patients are discharged with no long-term monitoring, leaving them without proactive care while at home. Connected technology with predictive analytics – such as Philips CareSage – can help prevent avoidable hospital readmissions and emergency room visits by detecting changes in behavior that may signal a change in a patient’s wellbeing in the comfort of their home.
CareSage combines patient demographics and medical condition data with Philips Lifeline medical alert service data to score a patient’s risk of transport to the hospital in an upcoming 30-day period. A study by Partners Connected Health showed that within a population of 2,318 patients monitored by CareSage, 224 hospital admissions can potentially be avoided each year, equal to a 40% reduction or $2.2 million in potential net savings.
A promising path ahead
With a clear understanding of patient and clinician needs, adaptive intelligence has a realistic chance to truly disrupt healthcare by making sense of the troves of data clinicians create every day to improve patient outcomes. The growth of AI allows providers to now answer many critical questions, but introduces many more. How do we ensure physicians do not become overly dependent on AI and maintain safe care delivery and patient relationships? Where is the line between the type of procedure an AI-powered assistant can lead and a procedure that is more appropriate for a physician to handle without AI? These are questions we are exploring at Philips and with our partners around the world. I’m eager to share more with you as we continue to uncover new trends, collaborate and innovate.
 James, J. A New, Evidence-based Estimate of Patient Harms Associated with Hospital Care, Journal of Patient Safety 2013; 19:3, 122-128. 78.https://journals.lww.com/journalpatientsafety/fulltext/2013/09000/A_New,_Evidence_based_Estimate_of_Patient_Harms.2.aspx
 Patient Safety in Critical Condition, Philips 2017. https://www.usa.philips.com/c-dam/b2bhc/master/feature-details/pm-deepdive/pm-group-page/PCMS_survey_project_report_UPDATED.pdf
 Saratoga Hospital Partners with Philips to Improve Patient Care and Safety, 2018: https://www.usa.philips.com/a-w/about/news/archive/standard/news/press/2018/20180307-saratoga-hospital-partners-with-philips-to-improve-patient-care-and-safety.html
 Golas, S.B., et. al. (2016, April) Retrospective Evaluation of Philips Lifeline CareSage Predictive Model on Patients of Partners Healthcare at Home. Poster session presented at the American Telemedicine Association, Minneapolis, MN