Modeling is the engineering act of producing a representation of a system for understanding, gaining an insight into the properties of that system, and predicting future outcomes. In medicine, modeling is useful in multiple tasks such as education, standardization, dissemination, innovation, and decision support [6]. Moreover, good systematic modeling of medical systems concludes with a formal model that, when applied, it may serve to improve quality, equity, optimization, and automation of processes within the modeled systems. In heath-care, medical practice (MP or the practice of medicine) is a varied and complex process that combines actions in a concrete health care setting that are performed by a set of professionals and caregivers in order to assist patients with their illnesses and ailments. This definition introduces MP at four different levels: functional (i.e., medical actions involved in MP), clinical (i.e., physical and technical requirements implementing MP), human resources (i.e., agents participating in MP), and medical case (i.e., sort of patient or disease addressed). Modeling these MP levels is seen not only as something beneficial, but also as a need in health care [14,31,10,6]. This is reflected in multiple models representing specific medical services (e.g., ER [20] or ICU [21]), tasks (e.g., diagnosis [7], treatment [8,11,16], and prognosis [34,25,1]), or diseases. However, only few studies exist that model MP at the functional level as a combination of diagnostic, therapeutic, and prognostic tasks. Some of them [15,27,14,31] use alternative names for MP, such as the art of medicine, the patient-centered model, the clinical procedure, the clinical methodology, or the medical function. The focus on the functional level is interesting because it provides a means for technological progress to improve medical practice. In this study, we present the results of an engineering process to construct a holistic functional model of MP for diagnosis, treatment and prognosis. The engineering process relies on a set of general hypotheses and it is composed of a sequence of steps. The hypotheses are: (1) there is a generic functional model underneath MP, (2) this model can be progressively approached and made explicit, and (3) the explicit approaches implementing the model can be supported by software technologies.
Read full abstract