Article Preview
TopIntroduction
The modern medical sector is utilizing new innovative technologies extensively. The inclusion of digital technologies has added a new dimension to the medical sector. Online consultations, health blogs, YouTube, Facebook pages, and many more online platforms and tools are helping the doctors and patients to communicate with each other. The doctors can now easily disseminate knowledge through online platforms. However, the scenario is not ideal in developing and underdeveloped countries. The usage of digital technologies is limited in these countries. Lack of budgets from government, the remoteness of regions, and the poor economic condition of a country deteriorate the possibility of utilizing the scope of providing health care with the help of online technologies (Zishan, et al., 2019). Specially, in countries of low/lower middle income despite the rapid advancement in Information and Communication Technology the E-Health system is not properly upgraded. Place of residence has emerged as a crucial factor in the use of E-Health services, where rural people represented a remarkably low use of E-Health services (Sayed, et al., 2021). The major objective of this research is to design and develop an E-health system suitable for developing countries. This E-health system aims to facilitate all types of stakeholders in the medical sectors in providing and getting better medical services.
The E-health system will have several components such as disease identification model, specialist recommendation model, and medicine suggestion model. This paper describes all these 3 tasks listed but mainly focuses on disease identification model as the first step of the development of the proposed E-health system. For making the identification model more robust and effective, machine learning (ML) techniques have been implemented.
Machine learning models are becoming increasingly popular in medical sectors as disease identification, disease categorization, X-ray image analysis, recommendation models all of these can be designed and implemented using various machine learning and deep learning models reliably. Machine learning model has great potential for applications in epidemiology and disease outbreak studies (Salim, et al., 2021). With the help of ML predictive analysis, and pattern recognition, Google has been able to detect breast cancer with an accuracy of 89% whereas the accuracy of the human pathologists for the same task is 73% (Krieger, 2017). In another research, Stanford Artificial Intelligence Laboratory successfully designed a deep learning algorithm which is capable of detecting skin cancer from analyzing images. Their algorithm also matches with the accuracy of the human dermatologists in this task (Esteva, et al., 2017). The goal of ML is not to replace human doctors rather these technologies can be used for assisting in decision making, diagnosis, and identification of diseases.
In the medical sector, the number of skilled physicians, pathologists, lab technicians, and other workers are not equal to the required amount in most of the countries. Due to this scarcity, it is oftentimes not possible to provide medical services to a large portion of service seekers. This also makes medical services very expensive. As a result, a large number of people are exposed to different health issues. This lack of treatment can be lethal in many cases. Different innovative technologies are improving the situation gradually. However, the developing countries are still far away from fulfilling the ever-increasing demands in the medical sectors.
In such circumstances, the focus should be given on using technologies that are accurate and are capable of mitigating the absence of professionals. ML tools can achieve these objectives with efficacy and productivity. ML models can make the job easier for everyone. ML has been used successfully in industrial productions, automation, speech recognition, designing smart appliances, providing security, and in many other sectors where it was intricate to come up with solutions. ML systems are based on optimizing a single quality metric such as future predictions from previous experiences (Pawar, et al., 2020). It is undoubtedly proven that ML models can provide pragmatic solutions to real-world problems.
ML models in medical sectors can help the doctors, nurses, pathologists, and technicians in the following ways (Flatworld, 2020):