Intelligent online monitoring and support for chronic patients

Treatment for chronic diseases often consists of a combination of lifestyle advices and medication. Many patients experience difficulties in following treatment recommendations [WHO, 2003]. Adherence to these recommendations is often far from optimal, especially in patients with chronic diseases. According to [Claxton et al, 2001], therapy adherence of patients with chronic diseases ranges between 70-80%. As a result of this widespread adherence problem, substantial numbers of patients do not get the maximum benefit of their medical treatment - with poor health outcomes, lower quality of life and increased health care costs as a result [Burke, 2001; Einarson, 1993]. Patients with chronic diseases require long-term adherence. Adherence can be improved by patient counseling as is common in disease management programs. However, personal counseling is expensive, especially given the fact that chronic patients will often require treatment for the rest of their life. Permanent personal counseling is therefore often not realistic. However, adherence is not a static property and as such, it is likely to decline when counseling is ended [Lee, 2006]. In this project, we will develop an intelligent online self-monitoring and self-support system, with the aim of improving adherence to therapy, both with respect to lifestyle changes and medication intake. This system can be used within disease management programs as a new element or as long-term follow-up to support retaining the effects of personal counseling. The core of the system consists of a mobile phone and web application that gives a patient personalized advice, focused information, appropriate reminders and motivational messages. Computer-tailored health education has shown to be an effective way to change health behavior [Kroeze et al, 2006]. The personal advice is generated by a computer, using a computational model of patient behavior (change). This model will be built using psychological literature and domain expert knowledge. The computer will use this model to reason about the current psychological state of the patient. The input for this reasoning process consists of medication intake data and reports about the personal goals of a patient and his achievements. For generating the information we will use 'Semantic Web' technology, which means that information will be described with meta-data and a computer readable representation of the meaning of information. This is then used to specify how information for patients should be composed from different pieces of text. The medication intake data will be collected using an electronic medicine box, which automatically registers the fact that medication has been taken and then stores this data in a central database. For the self-reports, we will develop an application that will run on modern mobile phones. This program will allow patients to answer a few questions about their goals and achievements in a simple way. The elements of the model and the system are meant to be generally applicable to chronic diseases. In order to evaluate the feasibility, we will test the system in this project in two different patient groups. One group consists of Diabetes type 2 patients, and the other of HIV patients. These groups are chosen because the importance in the therapy of medication and lifestyle changes differs: for diabetes patients, durable lifestyle changes are most important, whereas high compliance to medication prescription is especially important for HIV patients. The ultimate goal is that the system provides an improved adherence to therapy (both to medication intake and life style changes), eventually resulting in a reduction in the risk of developing late-stage complications. First, a pilot study with 40 patients will be done to validate the model and fine-tune the application. During this phase, and also during the development of the model, we plan to consult patient and expert groups such as the Dutch Diabetes Association (DVN) and the "Aids Fonds". The above mentioned hypothesis will be tested in two controlled studies, in which the adherence to therapy of 75 chronic patients that use the system is compared to the adherence of 75 patients that receive the usual treatment. The two studies differ in the type of disease of the patients (i.e. diabetes / HIV). Patients will be recruited from Dutch community pharmacies using the "Service Apotheek" formula. Some pharmacies specialized in HIV (plusapotheken.nl) will be used to recruit HIV patients. The system will be developed by a multidisciplinary team, primarily consisting of psychologists and artificial intelligence experts, but also of people with a medical, pharmaceutical and social scientific background (NIVEL, KNMP, the "Service Apotheek" formula, CZ Health Insurance, VUmc). Part of the technical implementation is done in collaboration with Evalan BV, a company specialized in innovative communication technology in health care.