You are sick. You walk to the hospital to see doctors and nurses overwhelmed and overworked. You wait hours for the clinician to finally see you, but they spend more time entering information into a computer than talking to you face to face. This is the reality for Americans living in the healthcare worker crisis. The number of sick patients is increasing faster than the number of newly trained healthcare workers in America. The hospital subsector’s workforce has decreased by nearly 90,000 people since March of 2020 [1]. The pandemic amplified how healthcare workers are not being accommodated and rather, are forced to put their personal lives at risk and sacrifice their safety and time with family (which more people are not willing to do). Implications of this shortage include longer wait times, slower diagnosis rates, and less personalized care. Ultimately, the current trend in healthcare workers cannot sustainably address the medical needs of the growing population. But why is there a healthcare shortage? Currently, resident physicians work up to 80 hours a week while being trained in high-stress environments. Furthermore, it is becoming increasingly difficult to delegate the tasks of a hospital effectively. Doctors spend their valuable time entering information into electronic medical records when they could be spending more time building meaningful relationships with their patients. One study found that clinicians spent approximately 5.9 hours of an 11.4-hour workday on electronic health record (EHR) data entry [2]. To combat this, artificial intelligence is able to minimize the damage from the growing health practitioner shortage.
The most pertinent problem that artificial intelligence could address is EHR data entry. There are software solutions that automatically send information from monitors to a patient’s chart which could eliminate errors when entering data like blood pressure, temperature, weight, and pulse oximetry [3]. Patients could be triaged (assigned priority of care) by completing a questionnaire that is then run through an algorithm. One prospective cohort study found that, compared to the traditional triage system’s 1.2% mis-triage rate, the mis-triage rate in a machine learning system assisted group was 0.9% [4]. The machine learning systems, along with real-time explanations for triage officers, were able to significantly lower the mis-triage rate of critically ill ED patients. Additionally, software that can read CT or MRI scans could assist radiologists with detecting anomalies on scans. These methods rely on predefined engineered feature algorithms with explicit parameters based on expert knowledge and large amounts of data. These technologies are designed to identify specific radiographic characteristics, such as the 3D shape of a tumor or the intratumoral texture and distribution of pixel intensities (histogram) [5]. These AI technologies could improve patient outcomes by reducing common human errors and lowering the current limitations of human organ supplies. Current organ transplant lists are long because demand highly outweighs supply. AI technologies can be used to determine the organ transplant priority and even help match donors to patients. Beyond the hospital, supercomputers can be used to speed up clinical trials for drugs.
This has the potential to decrease the risk of participants in clinical trials and quicken the process of delivering the most effective drugs to consumers. Even pharmacies are shifting to digital sorting methods and online deliveries that reduce the cases of wrong medication deliveries and decrease the time it takes for patients to receive their medication.
Ultimately, the applications of artificial intelligence in healthcare are nearly endless, and they offer a promising solution to multiple issues in the healthcare industry. However, the most pertinent issue that technology in healthcare is able to address is provider burnout with the end goal of improving patient outcomes and delivering quality care to all patients.
References
1. U.S. Bureau of Labor Statistics. (2022, March 4). Employment situation summary - 2022 M02 results. U.S. Bureau of Labor Statistics. Retrieved March 14, 2022, from https://www.bls.gov/news.release/empsit.nr0.htm
2. Monica, K. (2017, September 12). Primary care doctors spending 6 hours daily on EHR Data Entry. EHRIntelligence. Retrieved March 14, 2022, from https://ehrintelligence.com/news/primary-care-doctorsspending-6-hours-daily-on-ehr-dataentry#:~:text=Ultimately%2C%20researchers%20determined%20throu gh%20direct,1.5%20occurring%20during%20off%2Dhours.
3. Franklin, R. (2019, August 27). Top 5 tools to reduce data entry errors in ehrs. Mobius MD. Retrieved March 15, 2022, from https://mobius.md/2019/08/20/top-5-tools-to-reduce-data-entry
4. Liu, Y., Gao, J., Liu, J. et al. Development and validation of a practical machine-learning triage algorithm for the detection of patients in need of critical care in the emergency department. Sci Rep 11, 24044 (2021). https://doi.org/10.1038/s41598-021-03104-2-errors-in-ehrs/
5. Hosny, A., Parmar, C., Quackenbush, J., Schwartz, L. H., & Aerts, H. (2018). Artificial intelligence in radiology. Nature reviews. Cancer, 18(8), 500–510. https://doi.org/10.1038/s41568-018-0016-5
Kommentare