Defining Process Capability Indices for Time Series Data from Patients with Chronic Diseases

Abstract

Adherence to long term therapies has been recognized as a major problem. Several quantitative measures have been developed to track the wellbeing of patients with chronic diseases. Most measures are however subject to high random variations. Many statistical techniques have been developed in signal processing for detecting abrupt changes. These techniques are being applied to a number of areas such as quality control, navigation and seismic data processing. Statistical process control is widely used to track time series data with the aim of achieving stable and capable processes as well detecting early signs of deterioration. An overall index of performance calculated using these charts is named process capability index and serves to quantitatively assess the overall performance of a process in terms of both variability and adherence to upper and lower specification limits. It is widely used in industry to determine whether a process is capable of generating results that meet requirements. To our knowledge, although statistical process control has been applied to a number of medical problems, the process capability concept has not yet been used. The aim of this paper is to introduce this concept for chronic diseases and argue that it can be a useful index for monitoring since this is similar to a quality control problem, In this case, the doctor and the patient cooperate to adjust several parameters (e.g. diet, medication, exercise) in order to achieve a process that delivers satisfactory physiological results. The process capability is used for continuous improvement in industrial processes. In this way, it is possible to document and standardize the improvements made to a process.В  In the chronic disease context, it can be a valuable metric for disease management by both the patient and the doctor in addition to instantaneously measured quantities. Normality and independence should be checked and evaluated for process capability measurements. В In this study, process capability was illustrated with a sample of daily home measured systolic blood pressure data over a month. В An individuals chart was first drawn and its stability tested. В Assuming an upper specification of 120 or 140 mmHg the one side process capability was calculated as Cpu=-0.69 and 0.34 respectively. According to quality control principles, these values are considered low and indicate nonconformance. In conclusion, this preliminary study presents a new index for monitoring and self-management of chronic conditions. It summarizes performance in a single index assuming process stability and representing the degree of capability to conform to specifications. It can be adapted to cases when multiple variables are monitored. These indices could also be further studied for their potential benefit in assessing therapeutic strategies, in prognosis, in predicting comorbidities and estimating physiological variables.В