How should clinicians proceed when a patient presents with symptoms that may or may not point to prostate cancer?
In this scenario, the first step is usually a prostate-specific antigen (PSA) test, but high levels could point to a number of other issues and the test alone is not always a good indicator of cancer. Other tests to determine the likelihood of prostate cancer are both costly and invasive, since diagnosis requires a biopsy.
As the cost of healthcare continues to increase, providers are feeling the pressure to lower costs, while at the same time, improve patient outcomes, particularly in hard-to-diagnose conditions like prostate cancer. This balancing act requires clinicians and other stakeholders to make use of software applications with integrated clinical decision support so they can learn more about the patient’s history, family history, and conditions, and to guide their care plan. Often, however, the information they need is locked in unstructured notes in the patient record.
But consider this: Natural-language programming (NLP) and machine learning – both forms of artificial intelligence (AI) – can be a true differentiator in clinical decision support. After combining data on family members and conditions from reference data, such as SNOMED CT, and dismissing irrelevant data, such as non-related family members and non-relevant conditions, using NLP, machine learning can be used to create lexical rules to link family members to conditions and better understand the patient’s risk. If the patient has a family history of prostate cancer, the physician can use that information to make more-informed decisions on further tests and potential treatment options.
Leveraging AI, busy clinicians have access to the information they need, when they need it, and to provide the best care and spend the most time with the patient. This is crucial if healthcare is to achieve the quadruple aim of better care, improving population health outcomes, reducing costs, and improving clinician satisfaction.
Improving the clinician workflow
Streamlining the clinician workflow is a priority for many healthcare organizations, as a poll demonstrated during a recent Wolters Kluwer webinar on applying AI in healthcare. Nearly half (48%) of the attendees reported this was a priority area when considering AI.
It’s a focus area because so much clinician time is spent creating reports and data from patient encounters, resulting in complex, inefficient workflows and wasted time, all contributing to physician burnout. AI has the potential to alleviate or eliminate those complex workflows. The physician can document the patient encounter and allow technology to extract the data for use in clinical decision support and analytics – instead focusing on taking care of the patient.
AI is also being leveraged for predictive analytics to improve population health outcomes. According to the Wolters Kluwer webinar poll on AI applications, 83% would like an intelligent, automated way to handle predictive analytics. To truly move the needle on population health outcomes, healthcare organizations need access to good, clean data from multiple sources. AI and machine learning can more easily address issues that are a barrier to accessing that data and navigate common challenges such as interoperability, data quality, data mapping, and data normalization.
