Discriminative Information Added by Wearable Sensors for Early Screening - a Case Study on Diabetic Peripheral Neuropathy

Wearable inertial sensors have demonstrated their potential to screen for various neuropathies and neurological disorders. Most such research has been based on classification algorithms that differentiate the control group from the pathological group, using biomarkers extracted from wearable data as predictors. However, such methods often lack quantitative evaluation of how much information provided by the wearable biomarkers contributes to the overall prediction. Despite promising results from internal cross validation, their utility in clinical practice remains unclear. In this paper, we highlight in a case study - early screening for diabetic peripheral neuropathy (DPN) - evaluation methods for quantifying the contribution of wearable inertial sensors. Using a quick-to-deploy wearable sensor system, we collected 106 in-hospital diabetic patients' gait data and developed logistic regression models to predict the risk of a diabetic patient having DPN. Adopting various metrics, we evaluated the discriminative information added by gait biomarkers and how much it improved screening. The results show that the proposed wearable system added useful information significantly $(\mathbf{p} < 3\mathbf{e-4})$ to the existing clinical standards, and boosted the C-index significantly $(\mathbf{p < 0.02)}$ from 0.75 to 0.84, surpassing the current survey-based screening methods used in clinics.

[1]  Shin J. Oh.  Clinical Electromyography: Nerve Conduction Studies , 1984 .

[2]  M. Pencina,et al.  Evaluating the added predictive ability of a new marker: From area under the ROC curve to reclassification and beyond , 2008, Statistics in medicine.

[3]  F. Harrell,et al.  Evaluating the yield of medical tests. , 1982, JAMA.

[4]  P. Zimmet,et al.  Diagnosis and classification of diabetes mellitus , 2002 .

[5]  G. Rossi,et al.  Diagnosis and Classification of Diabetes Mellitus The information that follows is based largely on the reports of the Expert Committee on the Diagnosis and Classification of Diabetes (Diabetes Care 20:1183–1197, 1997, and Diabetes Care 26:3160–3167, 2003). , 2008, Diabetes Care.

[6]  Jesus A. Gonzalez,et al.  Machine Learning for Imbalanced Datasets: Application in Medical Diagnostic , 2006, FLAIRS.

[7]  F. Piarulli,et al.  Reproducibility of Michigan Neuropathy Screening Instrument (MNSI): A comparison with tests using the vibratory and thermal perception thresholds , 1996, Diabetes Care.

[8]  Frank E. Harrell,et al.  Regression Modeling Strategies: With Applications to Linear Models, Logistic Regression, and Survival Analysis , 2001 .

[9]  Der-Chiang Li,et al.  A learning method for the class imbalance problem with medical data sets , 2010, Comput. Biol. Medicine.

[10]  E. Steyerberg,et al.  [Regression modeling strategies]. , 2011, Revista espanola de cardiologia.

[11]  Guang-Zhong Yang,et al.  Detecting Walking Gait Impairment with an Ear-worn Sensor , 2009, 2009 Sixth International Workshop on Wearable and Implantable Body Sensor Networks.

[12]  Guang-Zhong Yang,et al.  Body sensor networks , 2006 .

[13]  W C Stolov,et al.  Height is an independent risk factor for neuropathy in diabetic men. , 1992, Diabetes research and clinical practice.

[14]  Gareth Ambler,et al.  How to develop a more accurate risk prediction model when there are few events , 2015, BMJ : British Medical Journal.

[15]  Nancy R Cook,et al.  Quantifying the added value of new biomarkers: how and how not , 2018, Diagnostic and Prognostic Research.