AI led Decision-making in the Healthcare Industry

Why we desperately need to up the pace of tech adoption?

In a world that’s increasingly more reliant on artificial intelligence, it is no longer a matter of whether we can implement it the healthcare industry or not; it is just a matter of when and how. While the usage of AI in healthcare is still considered controversial by certain groups (especially in the clinical assessment space), many AI algorithms are already working seamlessly in aid of a large number of auxiliary areas and a few in the critical areas too. The COVID-19 pandemic has brought to the fore the shortage of healthcare infrastructure including beds, equipment, PPE kits, and medicines. Moreover, the fact that doctors and nursing staff across the globe are overworked / stretched and stressed, indicates that we urgently need to focus on boosting the efficiency of the healthcare systems using AI.

The healthcare industry is one of the most complicated yet critical industry because it connects to every living person’s prized possession: their health. While AI may not solve every problem of such a complex structure in the short term, it can immensely increase the efficiency in many parts of a healthcare system, from clinical and administrative decisions to executive dashboarding and operational monitoring. We hypothesize that the combined effect of these improvements will be equal to thousands of lives saved without a substantial change in facility capacities.

In this article, we will explore three areas critical to decision-making in healthcare which can be improved with AI.

1.     Clinical Decision Support

2.     Operations and Administration

3.     Executive Dashboarding and Monitoring.

Three areas where AI can help

Clinical Decision Support

Diagnosis

The first step of the clinical decision process is diagnosis. Accurately diagnosing a patient and doing so before it is too late is the first crucial step towards appropriate treatment. AI driven models can anticipate the behaviors and circumstances of a patient preliminarily and diagnose the patient based on historical patterns of data from past customers (symptoms, family history, clinical diagnosis data and demographic data). It can ask follow-up questions to the patients after records of specific symptoms to further improve the chances of successful diagnosis and store the data in the patient’s dataset, which directly increases its success and efficiency over time. Even if it just remains as an aid to the physician, the benefit in eliminating biases may save many lives.

Of course, when it comes to healthcare, most people want to be in the hands of a professional instead of artificial intelligence. Even though studies show that deep learning algorithms can match the diagnostic effectiveness of healthcare professionals, when compared to AI, some people still prefer a regular doctor.

Earlier this year, deep neural networks were used to detect COVID19 from X-ray scans. The software was trained to differentiate chest congestion in COVID and in pneumonia. The accuracy of the models was an astonishing 98%.

Given low doctor-to-patient ratio prevalent due to the pandemic in many hospitals and the rapid progress of COVID19, AI can add the much-needed speed to diagnosis.

Treatment and care pathway

The same AI technologies can subsequently be used to create a more coordinated and efficient treatment and care pathways for patients. In an ideal scenario, a medical database would contain billions of data points belonging to millions of patients, and the AI would use that data pool to draw precise conclusions about a patient’s needs quickly. These data points can be gathered using multiple data sources – ranging from electronic health records (EHRs) to remote sensors and wearables – and can help clinicians make informed care decisions and improve patient outcomes.

For combating COVID19, the wearables maker Fitbit collects nearly 250,000 data points per day, including vitals like heart rate, sleep cycle, and body temperature. Findings have detected subtle changes in patterns before the COVID19 symptoms become evident. While the study is WIP, the efforts offer hope to reduce the spread and check progress of the disease by identifying signals in time.

Chronic disease management

Chronic disease management (CDM) is another problematic area that AI can be a big help. CDM is a continuous process which includes regular check-ups, monitoring, and patient education. Often requiring input and care from multiple specialists, CDM gets complex too. These steps can be automated by the advanced application of AI algorithms, which can help take the guesswork out of chronic disease management. Since chronic conditions such as diabetes, cancer, and cardiovascular diseases are the leading causes of death and disability in the world, increasing the efficiency of the CDM through complex AI systems would simultaneously free up much-needed healthcare capacity while saving possibly millions of lives from chronic diseases.

Intensive care monitoring

Intensive care monitoring can also harness the power of AI given the fact that an abundance of clinical, physiological, and laboratory data is gathered from the intensive care units (ICU) constantly. Of course, this would mean that the AI must be trained according to the extremely challenging environment of intensive care. It must adapt to the rapidly changing patient status in real-time and provide the appropriate treatment. While this poses challenges for the development of AI for intensive care, if successful, it can respond to patient anomalies faster and more accurately, an early warning even a couple of minutes earlier in intensive care could mean thousands of saved lives.

Remote OPD assistance

Restrictions on mobility has led to a sudden spurt in online consulting and remote treatment. AI can improve healthcare even outside hospitals with remote outpatient assistance. Telemedicine is an already popular service for remote patient care and monitoring, and the role of AI in telemedicine is only expanding, from tele-assessment and tele-diagnosis to tele-interactions and telemonitoring. To achieve wider adoption, the AI algorithms used in these remote care applications need to be developed further and the public should be educated about the fields of usage of these applications.

Administrative /Operations Support

Customer relationship management

On the administrative side of the healthcare industry, a significant amount of time and resources can be saved by utilizing AI on customer relationship management (CRM). Compared to clinical decisions that directly tie to a patient’s health, most steps of the CRM services are much more basic and can be safely automated through artificial intelligence. AI chatbots that utilize large knowledge libraries can easily respond to customer queries with relevant information such as booking appointments or informing customers about a hospital’s doctors, timings, report details, hospital location and much more.

Health insurance verification and assistance

Similarly, AI can also help patients select appropriate insurance plans and help hospitals verify patients’ health insurance. It is not very rare to see patients paying for a treatment that they didn’t know was covered by their insurance. Given the vast number of third-party insurance providers, this can also waste the hospital staff’s valuable time as they spend hours trying to match and validate patients’ insurance plans. An AI-led solution can verify patient insurance, saving both the patient’s and the hospital’s time and money, which can be later used for more productive and important purposes.

Boosting overall efficiency

Correct use of AI in hospital operations and administrative processes means a more efficient workflow, which directly translates to more patients treated with the same capacity. We can achieve this outcome through a variety of methods. AI could detect patients in critical conditions and patients with complex medical conditions more effectively, thus prioritizing them to further increase the recovery rate. This general increase in efficiency would equal to faster patient discharges, which would again equal to more patients treated in less time with the same facility capacity.

Executive Dashboarding and Operational Monitoring

Enhanced reporting and data analysis

Manual data reporting, storage, and analysis is an error-prone process. Even if hospitals utilize digital solutions for data management, as long as the data is reported and analyzed manually, it will be prone to human error. Simple AI algorithms can combine, store, and analyze data from operations, patient history, contracts, and other areas with fewer errors. This will minimize inefficiencies by completing and improving missing and redundant data much faster than human applications since AI networks can communicate, learn, and act significantly faster than humans.

Operational insights

Likewise, an AI solution could provide valuable insights into the hospital’s commercial efficacy and the trends in patient satisfaction scores, and do it more transparently and efficiently than a human reviewer. This would help the hospital’s executive and operational team to make more educated and beneficial decisions, which would directly translate into improved revenue and patient satisfaction scores for the hospital and better care and experience for the patients. A simple win-win for everyone involved, made possible with an investment in artificial intelligence.

Promoting transparency

Implementations of AI for the sake of transparent monitoring would also help to promote transparency within a healthcare facility and create a collaborative environment among peers, specialists, and other staff members. Aware of the existence of a neutral AI in the operational monitoring system of the hospital, every member of the hospital would be incentivized to follow the rules, including the executive staff. Internal conflicts rooted in the hierarchy of the hospitals can be prevented and healthcare facilities can be moved away from harmful bureaucratic tricks.

Blik.ai with its intuitive dashboard helps executives with a cohesive, easy and quick to understand report covering all vital stats for operational monitoring of hospitals. With a situation as dynamic and rapidly evolving, priorities and performance metrics that need monitoring by leadership may rapidly change. In-built flexibility of Blik.ai will help the leadership team customize the dashboard as per their needs, with little outside help. While an integration with existing ERPs within the healthcare space are not yet live, we do have rich experience in the healthcare space and also have pre-defined templates that can be deployed on Blik.ai within weeks (instead of going through a complete development cycle of a few months).

Conclusion

No matter what you think about the ethical and social questions that artificial intelligence raises, there is a simple truth that we need to accept sooner or later: AI is already here, and it is not going anywhere. If it is going somewhere, that place is deeper into our lives. If you haven’t noticed it already, AI has already improved our lives in many areas. From human resource management and banking to business and communication, it’s helping make many day-to-day processes seamless and efficient.

Understandably, people are more concerned when it comes to AI applications in the healthcare industry. Leaving the decisions about your health to artificial intelligence can seem like a dystopian concept at first glance, but when you escape from the grasp of suspicion that is caused by unfamiliar methods, you will see that appropriate utilization of artificial intelligence in healthcare facilities will only save lives and is the ultimate answer to healthcare shortages that become more prevalent every day.

Delay in treatment of COVID due to resource and infrastructure crunch is resulting in death of many people. An increase in efficiency powered by AI, may prove out to be a boon to many whose lives can be saved with timely intervention.

Author Profile:

Randhir Hebbar is one of the founders of Convergytics — Asia’s fastest growing and leading analytics brand. Randhir heads Digital, BI, AI and Products at Convergytics and Blik.ai is a solution that he has conceptualized, designed and built with his team of team of developers, BI specialists, data engineers and data scientists. You can read some of his other posts here.