In our latest blog series, AI in primary care we discussed what healthcare professionals deem important when it comes to AI-based tools. However, the patient perspective should not be overlooked. If the people seeking care are sceptical or even unwilling to use the technology, they will find other ways around it. So what do patients need in order to feel confident in using an automated solution? We gathered feedback from patients who tested our automated solution for triage and medical history at virtual clinics and physical primary care centres.
An automated triage and medical history solution can facilitate healthcare professionals’ work and create the conditions for secure diagnoses and effective appointments. Allowing patients to provide information about their condition and medical history can save healthcare a lot of time. However, in order for patients to want to use the tools, they need to be positive about the technology, trust the tools and see that the outcome meets their expectations.
But how is patient trust earned? We collected feedback from patients who tested our tool* and we identified 6 areas to consider:
1. Accessible and easy
An automated solution for history and triage should be easy to use, so that patients actually want to use it. It should also be available and accessible at all times, regardless of the patient’s location – which could be the sofa or the workplace.
‘Yesterday I felt so under the weather I would not have been able to go to the emergency room and wait for several hours. It was great to use something like this alongside an online consultation.’
Accessibility does not necessarily mean that you get an answer right away. In cases of a minor ailment, the care seeker is usually content with initiating contact with healthcare, by sending a message for example.
2. Clear communication and comprehensible language
Simplicity is key, not only with regard to technology but also when it comes to what kind of questions are asked and how. Clear communication and comprehensible language are of utmost importance and heavily affect the patients’ confidence in the solution. It’s not medical complexity or even accuracy that play the greatest role in creating trust, but a relevant, simple, and clear dialogue. In addition, delivering information in fairly few steps and with explanatory examples, if needed, is also a critical factor. Finally, patients should feel confident that they have understood the questions correctly and feel that they can answer without uncertainty.
3. Relevant follow-up questions and response options
The requirements for relevance apply to a great extent to follow-up issues. If a well-formulated, natural question is followed by an irrelevant question, the patient may lose confidence in the capability of the tool.
‘I understand that anxiety may be a necessary question, but maybe not in the case of a common cold.’
Finally, many people feel that the number of questions asked should be just the right amount – too many questions can test the patient’s patience.
‘It gets weird when it starts to ask about mental illness and asks too many questions about something that ultimately requires human contact.’
4. Credible sender
Another aspect that plays a big role in earning trust in an automated tool is the sender. Many believe that if the solution is provided by their own healthcare provider, whom they are confident in, it’s perceived as more reliable.
‘I wouldn’t have the same amount of trust if I had just googled it, but if I found it on my GP’s website, it would be fine.’
5. Acknowledging the patient’s knowledge
A lot of patients google their symptoms for guidance and information before they get in touch with their healthcare provider. Some also have recurring problems and are therefore pretty confident about their diagnosis. In this case, patients want their experience acknowledged, plus the option to use it as a starting point to confirm what they believe – as opposed to answering a bunch of unnecessary questions. The patient could then enter the diagnosis he/she suspects directly and move on in this track, instead of taking the same path as everyone else, i.e. describing symptoms and answering questions.
‘I work in healthcare myself and I would just like to confirm whether my suspicion of having pneumonia is correct.’
6. Guidance and information
What expectations do patients have of an automated tool? Most people agreed that such a solution should be able to guide them properly and inform them on what to do next: Go to the surgery, meet a doctor online, or contact someone else in healthcare.
In addition, some other patients want suggestions for possible diagnoses, but highlight the importance of the right tone:
‘It would be great if it can tell me what it could be potentially, but in a humble way – presented as a recommendation or suggestion rather than a fact.’
Others noted that the recommended steps should be clearly formulated – for example, if a video consultation is recommended, it should be specified that it’s not the end-station but a potential step down the road to a physical meeting.
‘It would have felt more secure if it was explained that the online consultation can be a first step that may lead to a physical meeting.’
At the same time, it is clear that an automated tool is a complement rather than a replacement of healthcare professionals.
‘It wouldn’t replace human contact, but it could replace googling.’
Finally, there’s no doubt that healthcare is facing a number of challenges, but patients are often overlooked in these discussions or solutions. There is a risk that we focus on the wrong things and develop tools that are way below the standard, and that can result in patient aversion. Today’s patients are not spineless puppets, but proactive individuals with both knowledge and the power to choose healthcare providers. If they don’t feel confident in a solution, they will soon look for other ways and into other options.
This post is the 5th part of our article series about AI in primary care. Click here to access all the articles.
*This compilation is based on evaluations of our pilot projects at virtual clinics and physical primary care centres in the Jönköping Region, during Q4 2019.