**1. Introduction**

Researchers have found that up to 79% of long-distance runners are expected to sustain a running-related injury in the lower extremities [1]. Such injuries could potentially be avoided if the long-distance runner runs within the boundaries of the recommended cadence throughout the entire run. Findings have shown that optimal control over one's cadence can aid the runner in reducing the impact forces on joints [2], reducing muscle soreness and fatigue [3] and increasing efficiency of oxygen use [4]; all of which reduce the possibility of injuries to the runner. Given the numerous advantages of maintaining an optimal cadence throughout a run for long-distance runners, we are interested in examining a mobile health (mHealth) solution to monitoring and coaching cadence for long-distance runners, aimed at minimizing their risk of injury.

There are many commercial apps that are widely used by runners, including MapMyFitness, Runtastic, Adidas miCoach, Nike+, RunKeeper and Endomondo [5–10]. The primary function of most of these apps is to monitor the runner's performance and to provide an interface for the runner to view statistics related to her or his runs. Some of these apps allow the runner to import workout plans that are aimed at motivating the runner and improving her or his performance. These apps, however, do not address minimizing the risk of injury, which creates a gap in the technology that we attempt to fill in this work.

To close this gap, we have designed an Android smartphone app, RunningCoach, to coach long-distance runners. In this work, we define "long-distance runners" as individuals who every week: (i) run for at least five kilometers (or three miles) in distance; or (ii) run at least one session that is one

hour or longer in duration; which is the definition of the Association of Athletics Federations (IAAF). RunningCoach is designed to monitor the runner's cadence, among other parameters, and to coach her or him based on the collected data. In this paper, we discuss the design and findings of a pilot study that aimed to explore the feasibility and usability of this app. This is the first of a series of studies that aim to refine the system and validate its efficacy regarding the reduction of injuries.

This paper extends on previous published work in [11]. Our previous paper was focused on the design and implementation details of the system. In this paper, we describe a feasibility and usability study for RunningCoach and report the findings of this study. Distinctly from the previously published paper, the contributions of this paper are (i) presenting evidence for the feasibility and usability of a coaching system based on remote monitoring for long-distance running; and (ii) reporting on lessons learned that developers can rely on to build robust mobile coaching solutions for fitness and health applications.

Concretely, we aim to understand the following factors related to the use of RunningCoach. First, we study the battery consumption incurred by the app as a usability factor. Second, this work examines general usability-related scenarios that are related to the robustness of the system, such as its ability to recover from faults (e.g., server is down, lost internet connection, etc.). Third, we examine the accuracy of the system in estimating the runner's cadence, speed, and other variables, as perceived by the runner. Fourth, we examine the privacy-related aspects of using this app and study the acceptability of the users to this technology through a post-study questionnaire. Finally, we explore a possible analysis relating cadence, speed and the gradient of elevation (in the path of the run) as a potential way to assess injury-related performance, which is a gateway to future directions of this research effort.

The rest of the paper is organized as follows. In Section 2, we survey the literature on related research. Subsequently in Section 3, we describe the study objectives and the study protocol. In Section 4, we give a brief summary of the architecture and implementation of RunningCoach before presenting and discussing the results of the study in Section 5. We finally close the paper in Section 6 by reflecting on our conclusions and directions of future research.
