1. Salcido, Richard "Sal MD"

Article Content

This editorial is a prologue to our continuing education activity, "Shear-Reducing Insoles to Prevent Foot Ulceration in High-Risk Diabetics Patients" on page 519. The investigators evaluate the effectiveness of shear-reducing insoles compared with a standard insole designed to prevent foot ulceration in high-risk patients with diabetes. Results suggest the shear-reducing insoles are more effective. This important article allows us to examine analytic methods used in patient-oriented research (POR).1

Figure. No caption a... - Click to enlarge in new windowFigure. No caption available.

Several methodological approaches can objectify the findings of POR, including quantitative and qualitative methods2 or a combination of both, aptly named "mixed methods."3 In this study, the investigators used quantitative analysis using a sophisticated linear regression model: the "proportional hazards model," named after Sir David Cox who developed the technique. The Cox regression model is a powerful modeling technique for "lifetime data analysis."4,5 Historically, the Cox regression model was used in the actuarial science field,4,5 which is commonly referred to as survival analysis. In the case of POR and this article, it is also used to measure "time to event" for patients in medical studies. For example, in this investigation, the time to the event is ulceration, with or without insole. Survival models can consist of 2 parts: the underlying hazard function, describing how the hazard (risk for ulceration) changes over time at baseline levels of covariates (a covariate is a variable that is possibly predictive of the outcome under study), and the effect parameters, describing how the hazard varies in response to explanatory covariates.4,5 A typical medical example would include covariates such as treatment assignment (shoe and/or insert), as well as patient characteristics such as age, gender, and the presence of other diseases. In biologic terms, the survival analysis has an event timeline-death, development of a disease, or exacerbation of the hazard.


The Cox regression model can examine the predictive value of survival, in terms of subjects (often patients in a medical setting) with covariates such as treatment, age, gender, height, weight, relative weight, smoking status, ethnicity categories, diastolic and systolic blood pressures, education, and income.6,7 The exponential of the coefficients from the Cox model gives the instantaneous relative risk for an increase of 1 unit for the covariate in question.


In this study, the Cox proportional hazard model was used to conduct a survival analysis based on the time to ulceration, comparing a shear-reducing insole with a standard insole. Several variables were included in the Cox regression analysis (risk group, treatment group, vibration perception threshold, ankle joint range, metatarsophalangeal joint range of motion, ankle-brachial index, foot deformity, race, type of diabetes, and age). The 2 significant factors elucidated from the Cox regression model were insole treatment and history of foot complications. The authors concluded that the standard therapy group was about 3.5 times more likely to develop an ulcer compared with the shear-reducing insole group (hazard ratio, 3.47; 95% confidence interval, 0.96-12.67).


The Human Machine Interface

The same methods used in industry to calculate failure-time analysis for machinery can be applied to potentially predict when the shoe insole will fail and require replacement before the risk for ulceration returns. The human machine interfaces in this particular study are the pathomechanical determinants or covariates, such as the shoe and insert, affecting shear during locomotion and other complex determinants of gait, and input from the afferent and efferent neurobiology of the foot and ankle during locomotion. In POR, we can use a combination of quantitative techniques to isolate covariates and determine their influence on increasing or decreasing the hazard of an event. Moreover, we can use these techniques to measure the performance or failure of devices used in patient care.


Richard "Sal" Salcido, MD

Figure. No caption a... - Click to enlarge in new windowFigure. No caption available.



1. Toto RD. Screening and evaluation of study subjects in patient-oriented research. J Investig Med 2010; 58: 608-11. [Context Link]


2. Moffatt CJ, Mapplebeck L, Murray S, Morgan PA. The experience of patients with complex wounds and the use of NPWT in a home-care setting. J Wound Care 2011; 20: 512, 514, 516. [Context Link]


3. Howe KR. Mixed methods, triangulation, and causal explanation. J Mixed Methods Res 2012; 6: 89-96. [Context Link]


4. Gill R. Understanding Cox's regression model. Experientia Suppl 1982; 41: 187-99. [Context Link]


5. Fox J, Laughton CD. Cox Proportional-Hazards Regression for Survival Data. An R and S-PLUS Companion to Applied Regression, Thousand Oaks, CA, Sage Publications Inc, 2002. [Context Link]


6. Telke SE, Eberly LE. Statistical hypothesis testing: associating patient characteristics with an incident condition: Kaplan-Meier curves, hazard ratios, and Cox proportional hazards regression. J Wound Ostomy Continence Nurs 2011; 38: 621-6. [Context Link]


7. Fisher LD, Lin DY. Time-dependent covariates in the Cox proportional-hazards regression model. Ann Rev Public Health 1999; 20 (6): 145-57. [Context Link]