Yesterday (23 October), AIMed hosted a virtual launch event to provide an overview of AIMedConnect, the new AIMed community group designed to connect the healthcare industry to facilitate collaboration and drive forward the deployment of artificial intelligence (AI), with Jvion, a key healthcare clinical AI company coming on board as a founding member. The virtual event was marked by roundtable discussions on how to get started in AI, best practice to deploy AI, and ways to mitigate the impact of delayed care in the aftermath of COVID-19 pandemic.

Patients are avoiding the care that they should be receiving

In the session led by Dr. John Frownfelter, Chief Medical Informatics Officer at Jvion, he talked about how the pandemic has stopped patients, especially those with chronic diseases to seek help. The role of telemedicine in mitigating it and the possibility of using analytics to assist physicians in prioritizing the sickest patients shall receive care first.

Specifically, the public were being told to isolate as the COVID-19 pandemic caused large scale lockdowns and social restrictions. At the same time, patients are also afraid to come into the healthcare setting where they regard as a potential source of infection. When these two factors collide, people are avoiding care which in some cases, they truly need.

From a healthcare provider standpoint, healthcare processes are disrupted. There is a shift of focus on reacting and responding to the pandemic and loss of some usual resources. To a certain extent, there is also a decrease in emphasis on standards, particularly around documentations and the routine calling on patients with test results or following up on patients to schedule their next appointments and so on.

Nevertheless, some patches were in place as processes were broken up. For example, telehealth exploded and Dr. Frownfelter saw telemedicine as one of the “patches” to mend the broken workflow. However, they are incomplete. In addition, not everyone is comfortable engaging with technology. Some people in the older population may see it as a burden on top of their chronic diseases. All these factors will converge and lead to a steady loss of continuative care; quality of care being rendered; satisfaction of care received and eventually putting a part of the population at risks.

Apart from chronic conditions, routine health maintenance and disease screening were also put on temporary hold and is still done at a slower pace than before the pandemic. Dr. Frownfelter cited figures from a study done in England, whereby there is an estimated of 5-year mortality increase for colon cancer by 15% and breast cancer by 10%. “This is very significant and impactful because delays in diagnosis results in delays in treatment and more advanced diseases. Costs, morbidity and mortality are all worsen as a result… So, the further delays in care is a looming challenge in front of us and this is what we should focus on,” Dr. Frownfelter says.

Both Dr. David Marks, Interventional Cardiologist, Professor of Medicine and Radiology and Vice Chair for Clinical Affairs at the Medical College of Wisconsin and patient Rene Schanberger present at the roundtable discussions had expressed shared similar views.

A terrific opportunity for using analytics

Although the knowledge of delayed care has been established, Dr. Frownfelter added there is no good definition or summary around which population is falling through the cracks and how will these vulnerable individuals get picked up again as the COVID-19 pandemic comes to an end. At the moment, there are a lot of manual processes being done to ensure no patient is left behind as they missed their care in view of the pandemic. This pressing need and the volume of catchup work to be done turn into a terrific opportunity to experiment data-driven analytics.

Since telemedicine was deployed to flush out the volume of patients waiting for care, analytics should facilitate the understanding of who will benefit from interventions being done remotely rather than in-person visits. Being able to discriminate between the two groups is becoming important as telehealth becomes necessary from an operational standpoint.

The other thing analytics can do is to better understand the challenges around individual patient and population. Dr. Frownfelter believes this is crucial in a time like this, data-driven mechanisms will underline the population groups more at risks so that healthcare providers can reach out to these people first. There are many tools out there but they are meant for routine checking or typical risks stratification that are effective in usual times to account for those who are at higher risks of hospital admissions but will be destroyed when a global health crisis hit. Risk insights need to be far more precise and we need for more sophisticated analytic than simple predictive modelling.

In the long run, clinicians are able to learn from AI and machine learning the kind of outcomes that patients are going into and reasons behind them so that special interventions can be taken before any negative consequences such as deterioration or readmission emerged. This kind of data-driven, individually targeted approach forms the backbone of personalized medicine, which Dr. Frownfelter thought, is still a concept now. Yet, it should be the next step in the present AI and analytics journey, to create efficiency within the organization and enable healthcare to be better value, at a higher quality with lower cost.


Author Bio

Hazel Tang A science writer with data background and an interest in the current affair, culture, and arts; a no-med from an (almost) all-med family. Follow on Twitter.