Introduction
The development of clinical drugs did not significantly change during the last 30 years. Some of the reasons for this are strict regulations, risks, and conservative views on new technologies. However, alternative learning, other than the human brain, will be potentially widely used in the future. The most obvious means for ‘alternate thinking’ in the present day is computer technology, which is now being rapidly developed. As computers and machines have a place in every sphere of life, it is obvious that it is the safest route for proposing further changes in clinical research and practice. In these areas, computers are useful via machine learning, which will cause a significant impact on clinical healthcare.
Discussion
Biomedical research grows in complexity, including technologies such as electronic health records (EHRs) and sensor data. These technologies allow studying cardiovascular disease, its symptoms, and its outcomes (Stevens et al., 2020). Next-generation sequencing has allowed for a better understanding of disease mechanisms in more patients and the creation of specialized care for them (Shah et al., 2019). Machine learning is, therefore, a potentially powerful tool that can be used for accounting for complex interactions, automatically collecting a lot of data. As machine learning can calculate many different factors which would otherwise remain unregistered, clinical investigators use it to deal with hypothesis-driven cases (Stevens et al., 2020). Thus, machine learning helps to determine the ailment in cases where humans doubt, studying the most dangerous diseases and helping determine the correct diagnosis.
There are many other roles in which machine learning can be utilized. For example, it may direct clinical trial research via its applications. It may help to identify candidates for clinical experiments especially, drawing the ones with the most interesting combinations of health issues that could be studied along with an experimental drug. Other options for choosing candidates for clinical trials include people whose tests will go quicker, safer, and more successfully, and finding the best sample sizes for every patient. Machine learning may cause interest from the media and lead doctors to the most endangered patients.
It is not just experiments that machine learning may be utilized for. For example, it is useful in remote monitoring endangered patients, drawing doctors’ attention to them in case of danger (Faggella, 2020). IBM’s research has allowed its machines to predict therapeutic indications and the side effects of drugs. One of their invention’s capabilities included finding out that mood disorders may be a reason for weight loss (IBM Research Editorial Staff, 2019). Weissler et al. (2021) agree that machine learning has the potential to improve clinical research’s efficiency, with all stakeholders benefiting from it and standard clinical trials being greatly improved upon. Thus, authors of scientific research and inventors value their machines’ usefulness, which is far from surprising.
Thus, every example demonstrated above shows that machine learning has promising opportunities in assisting medical professionals. Even in the present situation, when certain areas like drug development remain unchanged, machine learning finds many roles in which it can be utilized. Thus, it seems reasonable to suggest that machine learning should be implemented on a wider scale in the nearest future. The possible benefits of such an implementation are numerous, greatly increasing patient safety. Even simple machines assist in keeping patients alive by measuring their heartbeat. Machine learning can go much further and develop clinical drugs which will allow patients to treat their conditions at home while remotely tracking their illnesses. Other possibilities are determining conditions correctly in situations where clinical professionals fail to formulate a correct diagnosis.
Such an ambitious change in the basis of medicine’s functionality will require broad and effective means of implementation. Healthcare providers will be sponsored to acquire the hardware necessary for machine learning to work. Medical staff and pharmacists will be educated to work with the machines, with their purpose explained in detail. Every health provider working with machines will submit monthly reports on the rate of successful treatment in the organization, which will be compared to previous data, if available, to assess the impact of the change.
However, not all studies agree that machine learning is a direct benefit. Some researchers warn against human factors such as haste and lack of attention to detail which may interfere with the plans to utilize machine learning. For example, Futoma et al. (2020) doubt making machine learning’s universality the priority, claiming that aiming for it will decrease performance in single sites. They attest that machine learning must be reviewed on when and in what case it may be useful for both doctors and patients. Only then, as the authors claim, will machine learning be effective while trying to employ machines in several fields simultaneously will cause them to fail. Mateen et al. (2020) have a similar opinion, claiming that only after many tests it becomes safe to use machines in real-life circumstances rather than tests. To this end, as the authors say, researchers and users must work together to achieve a full understanding of their machines’ advantages and limitations. Thus, using machine learning in a clinical situation should be a distant goal for researchers, as the authors claim.
While the authors mentioned above may seem conservative, they are right that a successful implementation of machine learning requires established communication between the staff. Even with the benefits of machine learning in mind, it is necessary to teach staff to communicate with each other so that researchers and users can determine in what cases machine learning is best applied. However, some researchers, doctors, and scholars may be resistant to change and still claim that humans are more effective. For example, Futoma et al. (2020), while criticizing machine learning, make an example of trained clinicians, who are, as they claim, better than machines due to their adaptability. While it is beyond doubt that human clinicians are better than machines in some cases, especially at the current stage of machine learning development, it is conservative to note that machines will never surpass clinicians.
From a broader perspective, machine learning will inevitably face the same resistance as machines in other spheres, as humans will not want to have worse job opportunities because of machines. Some representatives of the older generation of professional doctors may find it uncomfortable to work with machines rather than people. Finally, it is still worth noting that unprepared utilization of machine learning on a broader scale than it was tested for will cause significant failures which will, in turn, lead to the abandonment of the technology.
Conclusion
Thus, machine learning will significantly impact patient safety, mainly by assisting in developing new drugs that will be able to treat patients with hazardous diseases. Other areas of its implementation are safer and more accurate tests during medicine trials, monitoring patients’ health, and predicting outcomes. The further development of machine learning promises great opportunities such as continuing to invent new drugs, tracking patients’ conditions even before they develop into dangerous stages, and improving the effectiveness of clinical trials. However, its implementation must be careful, with numerous tests and staff working together to assess the best utilization of the technology. Machine learning will, of course, face obstacles, but if done correctly, it will change every stakeholder’s situation for the better.
References
Faggella, D. (2020). 7 Applications of Machine Learning in Pharma and Medicine. Emerj Artificial Intelligence Research. Web.
Futoma, J., Simons, M., Panch, T., Doshi-Velez, F., & Celi, L. A. (2020). The myth of generalisability in clinical research and machine learning in health care.The Lancet Digital Health, 2(9), e489–e492. Web.
IBM Research Editorial Staff. (2019). Featured patent: Machine learning models for drug discovery. IBM Research Blog. Web.
Mateen, B. A., Liley, J., Denniston, A. K., Holmes, C. C., & Vollmer, S. J. (2020). Improving the quality of machine learning in health applications and clinical research.Nature Machine Intelligence, 2(10), 554–556. Web.
Shah, P., Kendall, F., Khozin, S. et al. (2019). Artificial intelligence and machine learning in clinical development: a translational perspective. npj Digit. Med. 2(69).
Stevens, L. M., Mortazavi, B. J., Deo, R. C., Curtis, L., & Kao, D. P. (2020). Recommendations for Reporting Machine Learning Analyses in Clinical Research.Circulation: Cardiovascular Quality and Outcomes, 13(10). Web.
Weissler, E. H., Naumann, T., Andersson, T., Ranganath, R., Elemento, O., Luo, Y., Freitag, D. F., Benoit, J., Hughes, M. C., Khan, F., Slater, P., Shameer, K., Roe, M., Hutchison, E., Kollins, S. H., Broedl, U., Meng, Z., Wong, J. L., Curtis, L.,… Ghassemi, M. (2021). The role of machine learning in clinical research: transforming the future of evidence generation.Trials, 22(1). Web.