SMURF: Systematic Methodology for Unveiling Relevant Factors in Retrospective Data on Chronic Disease Treatments

Abstract

Deciding on the continuous treatment of chronic diseases is vital in terms of economy, quality of life, and time. We present a holistic data mining approach that addresses the prediction of the therapeutic response in a panoramic and feedback way while unveiling relevant medical factors. Panoramic prediction makes it possible to decide whether the treatment will be beneficial without using previous knowledge and without involving unnecessary treatments. Feedback prediction can be more accurate prediction since it considers the results of previous stages of the treatment. A novel label encoding called simulated annealing and rounding (SAR) encoding is also proposed to help improve the accuracy of prediction in both approaches. To unveil the medical factors that make the treatment effective for patients, various techniques are applied to the prediction models found through the proposed approaches. Finally, this methodology is applied in the realistic scenario of analyzing electronic medical records of migraineurs under BoNT-A treatment. The results show a significant improvement in accuracy due to the use of SAR encoding, from close to 60% (baseline) to 75% with panoramic prediction, and up to around 90% when using feedback prediction. Furthermore, the following factors have been found to be relevant when predicting the migraine treatment responses: migraine time evolution, unilateral pain, analgesic abuse, headache days, and the retroocular component. According to doctors, these factors are also medically relevant and in alignment with the medical literature.INDEX TERMS Multi-target prediction, classification algorithms, data mining, simulated annealing.

SOFIA: Selection of Medical Features by Induced Alterations in Numeric Labels

“…In addition, in [21], a multi-layer multi-target regression is proposed for the prediction of cognitive assessment from multiple neuroimaging biomarkers, allowing an early detection of Alzheimer's disease. Finally, in the SMURF methodology [9], a panoramic prediction is contemplated for predicting the responses of a multi-stage treatment for the migraine. That approach is based on the improvement of multi-target classification models to give a general idea of how the patient will evolve after receiving certain medication.…”

“…In the aforementioned SMURF methodology [9], a data transformation that improves the accuracy percentages in multi-target prediction models is presented. This transformation consist of the multiplication between a feature weighting vector and the numeric labels of the columns in a clinical dataset.…”

“…Due the promising accuracy results achieved in [9] when performing SA for a single-objective optimization, the NO method can be extended to the multi-objective scenario, allowing a parallel execution to reduce the computational cost [17]. In addition, this work can explore the use of Multi-Objective Evolutionary Algorithms (MOEAs) when finding the best data transformation for multi-target prediction models [23,24].…”

“…For example, Armañanzas et al [7] use the Feature Subset Selection (FSS) technique to reveal the most important factors when predicting the severity of symptoms in patients with Parkinson's disease [8] using non-motor symptoms. In other works [2,9], the authors propose a data transformation for carrying out the selection of relevant features while achieving high percentages of accuracy when predicting treatment responses in migraine. In these works, a single-objective Simulated Annealing (SA) [10] was applied to find that data transformation instead of employing a multi-objective optimization metaheuristic [9,11].…”

“…In other works [2,9], the authors propose a data transformation for carrying out the selection of relevant features while achieving high percentages of accuracy when predicting treatment responses in migraine. In these works, a single-objective Simulated Annealing (SA) [10] was applied to find that data transformation instead of employing a multi-objective optimization metaheuristic [9,11]. Moreover, SA has a high computational cost [12].…”