Background and motivation for the research:
About me:
Personally, I’ve always been interested in our healthcare system and how it functions. Over the past decade I’ve had some personal events that have left me and my immediate family at the mercy of our Canadian health care system more times than I can count. From seeing various medical specialists ranging across provinces or having to spend countless days and weeks going in and out of emergency rooms, I have managed to gain a massive amount of experience from the perspective of a patient in our medical system. These many experiences (both good and bad), have allowed me to observe, not only the dynamics that typically occur between patients and health care providers, but also observe how other patients interact with one another as well.
What makes people do what they do?
My undergraduate degree was in sociology. Knowing this, it is quite understandable why I would be so interested in the study of people.
The dynamics of how people interact, why they do what they do and what motivates them, have always been concepts that personally fascinate me.
These interests lead me to conduct extensive research over the last academic year on the social acceptability of artificial intelligence from the point of view of clinicians. This was, along with my current research, funded by Montfort Hospital. However, even while I conducted my own research, I often supported many of my colleagues, whose post doctorate research ranged a myriad of medical related topics. These experiences helped to shape my current research focus.
As I said earlier in my About me section, something to which I have been constantly forced to do over recent years, was educate myself on various medically related issues, thereby raising my personal health literacy levels. This was important to do, as it became quickly apparent, that some medical professionals do not always know how to communicate their messages to patients in simple understandable terminology.
However, something odd I’ve noticed post further research, was that the idea of health literacy was rarely if ever, even extensively studied. For that reason, I chose to fill this information gap within the field. I combined my personal experience with health literacy, my experience studying various technological advancements in the field of medicine, finally combing it with my sociology background to create my final research question.
Research questions and objectives
The beginning of any real research project starts with a good question. Below, you’ll find mine:
Research Question:
The social acceptability of chatbots as a form of eHealth technology and how they ameliorate medical patients’ health literacy
What do you think?
Now that I’ve already stated why this topic interests me and why I was able to come up with this question, I need to outline some questions I’m entering this research problem with. I personally wonder whether or not varying levels of health literacy are at the core of why patients don’t follow the advice of their medical professionals when they are not around.
Simply put, I wonder if a reason some patients forgo the advice from their health care professional, is simply because they do not comprehend the medical importance behind the sort of advice they’re being provided.
Furthermore, I wonder whether there are differences between the levels of severity of my question, depending on the patient’s unique medical condition. Is a patient with Parkinson’s more likely to comply than someone with Celiac Disease? If so, how do these differences play a role at all?
Objectives:
Now that I know what questions I’m entering with, it becomes much easier to know what my research objectives are. To put it simply, I am aiming to discover whether the level of inclusion/usage of eHealth chatbots (when used as an intermediary by health care practitioners to speak to and support their patients), affect patient willingness to comply with doctor instructions. In addition, I aim to figure out if the amount of time chatbots are used correlate with patient health literacy levels. If they do, I hope to learn if this is more common among certain medically diagnosed patients or not (and why). Lastly, I hope to discover if the success or failure of increased literacy levels, affects the amount a eHealth chatbot is viewed as socially acceptable (in the eyes of patients). This will be understood by analyzing interviews, questionnaires, and quantifying patient satisfaction levels with regards to the chatbot.
Short Literature Review:
After defining key terms related to my study, such as social acceptability and health literacy, the next stage of my research lead me to analyze the usefulness of chatbots in the medical field currently. There I found that many benefits can arise from the use of chatbots within the healthcare system. Time-efficiency stands out as a benefit within a multitude of circumstances. As chatbot answers are ‘’at least as satisfying as answers given by a multidisciplinary committee” (Bibault et al., 2019), these artificial agents could prevent patient’s from having to physically go see their physician (Bibault et al., 2019). A direct consequence of this is that physicians can then spend more time with the patients that need their presence the most (Bott, 2019). Chatbots are also proposed to ameliorate physical symptoms in some cases, such as fewer falls and lower frequency of delirium (Bott, 2019).
With regards to the current state of knowledge surrounding my research topic, much of the literature indicated that there was not a lot of academic literature available that focused on chatbot use in the medical field (Palanica, 2019). Furthermore, not much research has been conducted evaluating medical physicians’ perspectives of chatbots within the healthcare system (Palanica, 2019). Due to this lack of literature, it has been proposed that at this stage, that artificial intelligence (AI) needs to ‘’supplement rather than replace medical professionals” (Powell, 2019).
Many limitations are highlighted within this newly emerging field of research. As a result of their lack of ability to replicate the limbic system as well as their inability to account for all of the personal factors associated with a patient for any particular case, ‘’many physicians [believe] that chatbots cannot effectively care for all… patients’ needs (Palanica, 2019). Furthermore, Palanica (2019) outlines that in certain cases, ‘’expert medical knowledge and intelligence is required, thus implying that chatbots’ had an inability to mirror these human skills. Lastly, self-diagnosing chatbots could pose a significant risk to patients according to many physicians, especially if they do not accurately understand the diagnosis presented, due to a low health literacy level (Palanica, 2019).
Given the limited number of peer-reviewed research studies on this specific topic, it is evident that larger experimental trials are needed (Bott, 2019). For future research, Montero (2010) stated how it would be important to remember the challenges arising from content development, technological limitations, and expectations management. In short, these challenges could be mitigated through “contingency planning, open communication, […] and training” (Montero, 2010).
Methodology:
My method consists of multiple phases:
1. Questionnaires
2. Interviews
3. Survey regarding chatbots
Each phase will be based within medical practitioner offices. All participants will be voluntary patients, during each phase. In order to adhere to ethical concerns, all questions will be screened by the physician beforehand to ensure there are no breaches of patient to doctor confidentiality. Each patient will need to undergo each phase of the study to be included in the final data set.
Phase 1:
Will be mainly focused on health literacy competency questions
It will also include general questions regarding past experiences using eHealth technologies and whether they feel it could be a viable option to improving their levels of care
Views on the current usefulness and social acceptability levels of chatbots in healthcare (from the perspective of patients) will be documented using a scale system
Phase 2:
These in person interviews will be conducted in the presence of a physician to respect potential medical ethical violations - Example: this could occur if a patient begins to share private information, they aren’t particularly supposed to share. In that event the patient would be informed by the physician and they would be able to make a choice as to whether to continue or not
25% of questions will be subjective to anecdotal responses from participants about their experiences with eHealth technologies and their relationship with their medical condition
50% of questions will address general questions inquiring patient views on the social acceptability of chatbots and how they feel about eHealth technology
25% of the interview will involve patient participating in a case study where they will interact with a provided chatbot to solve a problem
Participant experience/feedback will be recorded througout
Phase 3:
Health literacy competency will once again be tested
Satisfaction levels (days removed from the in person interview where they interacted and solved a problem with a chatbot in the case study), will be requested through use of a Likert scale
Other satisfaction questions regarding chatbot usefulness will be asked along with requesting their perceived impact on social acceptability and sense of reliability vs. actual doctors
Reflection questions addressing whether patients believed their health literacy was improved upon during this process will be requested
Furthermore they will be asked if they felt it affected their willingness to comply with doctor instructions more or less, in addition to whether the chatbots support played a role in that decision or not (there will be an open section at the bottom for general commentaries as to why if needed)
Open general feedback period for researchers and clinicians will be conducted.