A systems analysis of the Army Substance Abuse Program

Abstract

The Army Substance Abuse Program is a program within the United States Army that has the mission to deter, detect, and treat substance abuse by US Army Soldiers, Civilians, and Family Members. This thesis examines the program from a systems point of view, using the generic US Army installation for the system boundary, and how the system creates value for the installation enterprise. This thesis first explores the motivation for this research, drawing on contemporary reports from the US Army, published news articles, and my own personal experience. Secondly, I examine the system from a systems architecture perspective, employing design structure matrices or adjacency matrices, based on the normative state of the system codified in US Army Regulations. In doing so, I highlight the important architectural changes within the program since 2001 and determine what aspects of the architecture inhibit the program's performance. Thirdly, I examine the system's dynamic behavior over time and establish a causal loop diagram to explain that behavior, drawing on the US Army's reports, the literature surrounding management response to substance abuse in the workplace, and field interviews. I then examine whether commanders are actually adhering to the required processes and if key commander-driven processes are effective in deterring substance abuse. Concluding, I recommend specific actions that can drive more benefit from the program, particularly from the point of view of leader supervision. The research here suggests a degree of architectural dissonance within the program that may limit performance across the US Army. The choices of data capture, access, and authority across organizational boundaries inhibit real-time supervision through command channels and coordination of medical care. Although actions seem to be mostly in compliance with regulations, the rate of non-referrals after positive drug tests is a notable exception. The rate of testing soldiers seems to have the desired impact on behavior while the impact of other commander-driven actions is mixed.