Huntington’s Disease (HD) is a devastating genetic condition that encompasses the symptoms of Parkinson’s and Alzheimer’s Disease. HD is a cureless, progressive condition that continues to grow in effect throughout its stages. In the middle and late stages, patients’ quality of life decreases drastically - emotional outbursts and involuntary muscle movements are frequent, so patients require full-time caregivers to complete daily tasks and avoid injury.
This immense caregiving role often falls on family members of patients, and drastically decreases their freedom and causes consistent stress. The Guardian system is designed to increase patient safety, easily track patient symptoms, and lessen the burden on the caregiver so their own quality of life can improve during such a draining period.
Throughout this project, we conducted a total of six interviews with doctors who specialize in treating HD. Among the many difficulties that affect patients and caregivers, there were a few pain points that stood out as consequential and actionable.
First, patients are seen by doctors at most a few times a year, so doctors can miss out on a lot of valuable symptom data that can inform their course of action. This means that doctors are concerned mostly with the broad strokes, and less able to treat the day-to-day HD symptoms. Next, patients and caregivers alike struggle with decreasing levels of freedom - patients from the progression of the disease, and caregivers with the progression of their patient responsibilities. Once the patient is in the later stages of HD, caregivers often spend all of their waking hours looking after them, fearful that if they are away even a moment, a serious accident like a fall may occur.
From these interviews as well as a rigorous literature review, we created a stakeholder map for the HD patient, showing the specific ways in which they interact with and rely on their caregiver and medical professionals. We used this map to pinpoint exactly where our Guardian system could intervene and make improvements throughout the progression of the disease.
Patient safety is improved with the Fall Detection portion of the Guardian system. Patients fall frequently and some of these falls lead to serious injury. The Fall Detection feature automatically detects when a fall occurs (using Apple Watch) and is able to prompt the patient with a few simple but vital questions. Depending on how the patient answers these questions, text notifications are sent to the caregiver and Emergency Medical Services (EMS). Fall Detection is able to communicate with the central database using an MQTT protocol (which I took the lead on) and therefore the Guardian system has access to all fall data. The feature’s system architecture is mapped below.
Crafting a conversation user interface (CUI) for an HD patient is a difficult task, since verbal ability declines with the progression of the disease. Therefore, simplicity and flexibility are key to a usable CUI design. We prototyped a functional version of this CUI which was triggered by a real-time fall event, whose data was then sent to the Guardian database. While this was gratifying, and a promising prototype, there are some major limitations. Namely, we were not able to test the CUI with an HD patient - this is the most important part of developing the prototype further. Frankly, if an HD patient cannot use the CUI successfully, then then the system will not be able to improve life for the patient and their caregiver. A high level conversational flow is shown below for the Fall Response feature.
Along with fall data, there are other types of data sent to the Guardian database. These data are concerned with other symptoms, and track these over time for the caregiver and the doctor to make informed decisions about patient care. The purpose of tracking chorea (involuntary movement), sleep, and mental state data is to determine underlying patterns. Understanding these patterns allows caregivers and doctors to make data-driven care decisions to proactively improve the day-to-day experience of the patient, and consequently the caregiver.
The patient is currently able to wear an Apple Watch to detect falls and chorea as well as sleep on a pressure mat to track quality of sleep. However, patients and caregivers are responsible for manually inputting mental state data using the CUI or the graphical user interface (GUI) designed for the Guardian system. Moving forward, there needs to be more contextual computing implemented in the system to track events such as emotional outbursts and physical characteristics like gait and balance, all of which could potentially be built using camera sensors. The more that can be done automatically by the Guardian system allows more symptom data to be collected, ultimately improving the care for the patient and decreasing the burden on the caregiver.
