How medtech AI can be used to improve patient diagnosis
Irving Loh, co-founder and chief medical officer at Infermedica, explores how artificial intelligence (AI) can be used to streamline the consultation process.
For many, consultations are the most common contact with health services; with doctors in the US seeing on average 20 patients a day. While the route to official diagnosis and treatment can include a range of moving parts and stakeholders, when assessing a patient’s health condition, physicians tend to follow a step-by-step process. Clinical experience allows them to analyse a multitude of signs and symptoms, employing probabilistic logic to determine a diagnosis. This is a challenging task considering the number of patients and potential diagnoses across generations and geographies, as well as the acceleration of telehealth during the pandemic which decreases in-person interaction.
With health providers stretched and populations increasing in age and in need of greater services, the challenge facing physicians will only grow. So, the question is: can medtech that simulates how doctors think assist them to reach quicker outcomes and improve patient diagnosis?
Streamlining the triage process and improving patient experience
Examinations can be both lengthy and exhausting. Doctors don’t always have accurate information prior, meaning they require in-depth questions as well as time to analyse what’s being said. They also follow stringent processes to collate information after the consultation. This costs the healthcare system money, with the average appointment costing the NHS £30, for example.
Such situations generate a few challenges, with the most pressing being that the health providers’ finite resources are spent on patients who don’t really need to be there. One in five visits are deemed as inappropriate which extends waiting times, increases costs and reduces the time and money available for those in need of urgent care.
This is where medtech, and particularly AI, can streamline the triage process. It seamlessly integrates into the medical workflow and uses the same logic physicians embrace at medical school, delivering enhanced insights by calculating the likelihood of particular conditions based on the patient’s symptoms. This helps to improve the consultation process and, with that, patient experience. AI can facilitate a faster, automated route to outcome by reducing the steps that doctors need to take, ultimately meaning quicker answers and patient recovery.
Further, using AI ensures that full medical histories are collected with the patient entering their details. This in turn boosts the collection and maintenance of electronic health records, which previously was largely reliant on manual upload by the healthcare providers. Up-to-date, complete records give physicians greater insight to make informed decisions, while the time saved on data collection and entry provides more time to improve patient experience through fostering stronger relationships.
What’s more, the pandemic has highlighted that triage from home is crucial for patient safety. Indeed, some health providers are seeing patient levels drop because people are unwilling to come in even if they do require help. This is because pathogens and other illnesses are concentrated in waiting rooms, so AI tools that can help patients to evaluate whether it’s worth the risk of attending in person – or if self-care is more appropriate – are extremely valuable. This has positive knock-on effects on the wider healthcare system by mitigating queues and demands on resources.
Enhancing doctors’ experience
Burnout is a very real risk for medical professionals. In the US alone, there is a burnout rate of 42%. It’s an alarming statistic – if that number of doctors suddenly stopped working, it would devastate patients and the entire healthcare system.
Adopting fast-learning AI into the diagnosis process means that physicians can feel supported and gives them greater ability to focus their time into effective services, improving their experience.
For instance, by shortening the signposting process, they can spend time making sure their practices are aligned with other professionals and working together to deliver more efficient outcomes for patients, improving the healthcare system for all. This goes hand-in-hand with enabling doctors to provide treatment to those who need urgent care faster, as they are not lost to time-intensive triage processes of patients who don’t need them.
Additionally, using AI and automated administrative processes can save time, with a quarter of doctors time currently spent on paperwork. Currently, many are acting more as data clerks, interfacing with computers rather than eye-to-eye with patients in-person.
As more patients return to in-person facilities and face-to-face consultations become incentivized, it’s important that doctors don’t fall into old habits. Human contact is important, but it takes a lot of work to take care of patients, from filling out registry forms, to retrieving medical history, analysing, drafting and sharing reports -- technology can do the leg work. AI can not only be a great source of intelligence, but it can manage the many administrative processes that really take away the experience of patient care from doctors.
Eliminating unlikely illnesses
Doctors think in probabilities. They are presented with symptoms, consider the next logical question, and look to eliminate or strengthen the likelihood of it being a certain condition, before directing patients to the correct level of care.
AI works similarly, analysing multiple conditions at once and eliminating the least likely illnesses. This saves the physician time by providing the correct range of diagnoses much faster. Not only this, but it can determine the acuity level, a difficult skill for even the most experienced medical professionals. Additionally, AI can analyse trends and predict issues. It can compare population data to anticipate wider issues at play, for example, a task that would be near impossible for a single doctor to manage within their day-to-day responsibilities.
Adopting AI also eliminates the need for additional expensive tests and procedures. For instance, it’s reported that around one in three routine consultations for outpatients could be done online, saving the NHS nearly half a billionpounds every year. It can also prevent unnecessary stress for patients who do not need to be presented with ‘worst case scenario’ information which is far too readily available on the internet as they wait for results. In the long run, this maximises experiences for patients, physicians, and healthcare services, creating a win-win situation for multiple stakeholders.
Any form of technology needs to integrate seamlessly into the clinical workflow to be effective. AI not only improves patient diagnosis by simulating how medical professionals think, but also enhances their experiences, ultimately improving patient care and health outcomes. As outlined by the UK Government, technology must be harnessed for the long-term sustainability of the healthcare system – understanding how medtech AI can be used to improve patient diagnosis is key to this.