IU-Bloomington Researchers:
IUPUI Researchers:
Currently, over 70 million Americans experience outbreaks of chronic pain [1], over 28 million have migraine headaches [2] and over 2 million suffer from Epilepsy [3]. When studying episodes of chronic conditions (i.e. headache, pain outbreak and seizure), researchers in health care are typically limited to obtaining information from patient recall, patient diaries and data collected in short periods in laboratory settings. Patient recall and diaries are often incomplete and incorrect, while lab experiments have a difficult time in recording sporadic and non-predictable events, such as migraine headaches and seizures. In addition, lab experiments cannot examine the rich details of the many factors that lead up to a particular event of interest. For example, the patient's diet, heart rate and stress levels in the day leading up to a migraine are inaccessible to researchers. As the numbers of people suffering from a variety of chronic conditions grows, health researchers' interest in this type of in-situ data will only increase.
Further, many people spend enormous amounts of time and energy trying to identify triggers for their chronic condition. Again, the typical method for identifying triggers is to keep track of suspected causes in a diary. Success is limited by the accuracy and completeness of the diary, and nearly impossible if the condition is caused by a combination of factors.
We are developing a toolkit that automatically gathers a variety of data as patients go about their normal daily activities. This data can be used in a several ways. The primary and most immediate purpose is to give health researchers access to rich datasets while studying chronic conditions. We also plan to use this toolkit as a medical intervention, both to assist patients in identifying triggers of their conditions as well as to analyze the data in real-time in order to predict and notify the patient of an impending episode so that they may take proactive measures. The toolkit is easily extended and tailored to study factors of interest for a variety of chronic conditions.
The primary difficult with diaries is that it is hard for the patient to remember to write everything down. This project is developing a mobile toolkit that assists in the real-time recording and analysis of several factors. The toolkit combines a PDA with the following recording methods:
A patient may still want to monitor only a few variables at a time. The toolkit allows the patient and their health providers to specify which factors to monitor. Our initial toolkit simply logs all data for later analysis. Future versions will experiment with applying filters to data to minimize the amount collected while maximizing utility, and will examine real-time analysis of the data for predicting episodes. This type of data manipulation, however, is dependent on the particular chronic condition and cannot be anticipated in advance. The following subsections describe the methodology for each component of our toolkit.
|
On-body sensors such as EEG, EKG, pulse, temperature and muscle tension have been used in laboratory settings for years. Researchers have used them to study a variety of conditions. Indeed, it has been shown that EEGs can predict epileptic seizures [7] and that patients can use the sensors to successfully treat some conditions using bio-feedback. To date, most sensors are wired, and as such, can only be used in a laboratory or clinical setting. Technology is just now becoming available to allow sensors to be wireless, allowing patients to move around as the sensors are no longer tethered to a machine. Most of the wireless sensors, however, transmit the signals to a nearby computer, still keeping the patient in range of the computer, typically in the same room. We are utilizing our expertise in mobile computing to take these wireless sensors and have them transmit to a portable computer: a PDA. This allows patients to wear the sensors anywhere by simply carrying a device that can fit in a purse or pocket. An example wireless device is an EEG cap equipped with Bluetooth technology (see Figure 1) [4]. In order to integrate a bluetooth sensor with the PDA, we are developing software that collects the data via Bluetooth. As the data becomes large and fills the memory on the device, our application securely uploads the data to a web service. Balancing the communication, memory and battery life is key for the success of this part of the toolkit. |
|
|
Nutrition is often a factor in many chronic conditions. However, even patients who must monitor their diet closely or risk a variety of complications, including death, have difficulty in keeping food diaries [8]. In conjunction with nurse researchers at IUPUI, the we have another project underway that uses a bar code scanner and other easy input mechanisms to help dialysis patients monitor their nutrition ( As shown in Figure 2, the nutrition monitoring tool uses a bar code reader to scan a UPC on a food label. It uses a database to link the UPC to the food item, then accesses the USDA nutritional database to determine the nutrients within the food. For the existing application, only certain nutrients are monitored and the patient is notified about their current consumption levels. For this project, we will alter the nutrition monitoring tool to record all of the nutrients and ingredients a patient consumes for later analysis. |
The experience sampling method (ESM) is a technique where subjects are prompted to respond to questions whenever certain events occur [6]. Time and location are the typical events that initiate questions, but questions could be asked at random times as well. In our toolkit, we use ESM to ask patients about activities, feelings and other items of interest that cannot be gathered automatically.
In this toolkit, ESM is used by having the PDA ask patients questions. The questions that are used will depend on the variables that are currently being monitored. For example, if stress is of interest, the patient will be asked to rate their stress level throughout the day. The key to successfully using ESM is to make it as unobtrusive as possible. As such, the interface is designed so that patients may quickly answer questions by tapping on a checkbox or choosing from a pull-down window.Essential to the success of this project is to have a graphical user interface (GUI) that makes it easy to specify which factors to monitor and how frequently to collect data. We are developing a configuration GUI which runs on a desktop machine. The GUI will guide researchers and patients through the process of selecting the data to record. The same GUI is used to configure sensors, nutrition monitoring and the ESM component. Our configuration GUI is being designed with input from researchers in healthcare and we utilize a participatory design process where we alter the GUI based on the specific needs and feedback of the target users.