AI in Healthcare _ ___________________________________________________________________________
[24]. In a bibliometric analysis of 100 of the most commonly cited articles about AI in medicine, researcher found that the majority were published after 2000 and that oncology appears to be at the forefront in implementing AI in practice, with cardiovascular medicine lagging [25].
APPLICATIONS OF AI IN HEALTH AND MENTAL HEALTH CARE In 2020, 90% of healthcare organizations indicated they use AI for automation; this has likely increased in the years since [9]. As noted, AI has been employed to improve diagnosis, facilitate and expedite the discovery of new drugs, organize and manage healthcare data, and perform surgery [3]. The consensus is that AI applications have potential to improve health and mental healthcare delivery, reduce costs, and enhance efficiency and possibly efficacy from the diagnosis to treatment/intervention stages. This section will explore a few examples. When specific AI applications are mentioned, these examples are merely used as illustrative purposes and does not indicate endorsement. ASSESSING, DIAGNOSING, AND INTERPRETING IMAGES AI has been used to assess, diagnose, and interpret health and mental health conditions via a variety of methods, but perhaps the most obvious is through the analysis of imaging. AI applications can analyze images such as x-rays, computed tomography (CT) scans, and magnetic resonance imaging (MRI) to detect abnormal patterns. This approach can decrease human error and make more accurate diagnoses. For example, in a dataset of mammograms, an AI system was able to interpret the mammograms and make diagnoses of breast cancer, with a reduction in false positives of 5.7% and of false negatives of 9.4% [14]. In this case, the AI system was more sensitive and accurate compared to the radiologists (90% vs. 78% accuracy rates, respectively) [14]. In a systematic review and meta- analysis, researchers concluded that AI demonstrated a high level of accuracy in detecting lung cancer, which has positive implications for early diagnosis [30]. AI platforms can also assist with the differential diagnosis of certain disorders that present similar clinical symptoms, such as types of dementia or depression. AI systems can help to differentiate between these disorders by examining brain imaging and structural MRI scans [31]. Ultimately, accurate diagnosis is the lynchpin to effective treatment planning, and the use of AI technology to improve diagnosis is an exciting advancement on this front. Voice data can be analyzed using AI technology to examine for volume, tone, breathing patterns, and vocal cord vibrations to assess and screen for type 2 diabetes, stroke, Parkinson disease, depression, post-traumatic stress disorder (PTSD), schizophre- nia, heart conditions, larynx cancer, speech disorders, and autism [32]. In these cases, the human voice can be a digital biomarker to detect disorders that can affect the voice. In one systematic literature review involving 145 studies, Parkinson disease was the most common condition that used voice data as a digital biomarker [33]. However, the researchers found that the studies included in their review had limited and unbal- anced data sets and focused primarily on diagnostic detection versus longitudinal monitoring [33].
BENEFITS OF AI IN HEALTH AND MENTAL HEALTH CARE
Many have posited that AI will revolutionize the health and mental health fields, offering multiple benefits. Its potential benefits are briefly summarized here; however, specific applica- tions will be discussed in the next section, which will further illustrate AI’s practical uses. INCREASE SERVICE ACCESS AI-based health and mental health delivery of services can promote equity and justice, particularly for those underserved and marginalized populations that may not have easy access to services. These technologies could enhance remote monitoring and the provision of telehealth and mental health services [26]. Some have argued that AI could play a role in democratizing health and mental health disparities, reducing long wait lists, costs, and other systemic barriers [27]. ENHANCED DIAGNOSTIC ACCURACY AI has been found to increase the accuracy of detection and identification of various conditions, which can then facilitate improved prevention or early intervention efforts [23]. PERSONALIZED TREATMENT AND SUPPORT AI can tailor treatments and interventions to individual patients’/clients’ needs based on the synthesis of clinical data, history, and demographics [14]. Individuals can also receive personalized reminders and support to assist in treatment adherence. PROMOTE PUBLIC HEALTH SURVEILLANCE AI can rapidly aggregate, synthesize, and analyze large public health datasets to track and predict health and mental health trends. This can be used to guide decisions related to public health policy and prevention efforts [14]. INCREASE EFFICIENCY AND PRODUCTIVITY AI platforms can automate and streamline clinical day-to-day administrative tasks, reducing costs and time burdens [28]. IMPROVED DATA ANALYSIS AI platforms can also rapidly aggregate large volumes of data from multiple datasets. This can be used to identify patterns and obtain the most up-to-date best evidence recommendations quickly and implement these approaches into practice [4; 29].
38
EliteLearning.com/Physician
Powered by FlippingBook