Examining the Social Context of Alcohol Drinking in Young Adults with Smartphone Sensing

Abstract

According to prior work, the type of relationship between a person consuming alcohol and others in the surrounding (friends, family, spouse, etc.), and the number of those people (alone, with one person, with a group) are related to many aspects of alcohol consumption, such as the drinking amount, location, motives, and mood. Even though the social context is recognized as an important aspect that influences the drinking behavior of young adults in alcohol research, relatively little work has been conducted in smartphone sensing research on this topic. In this study, we analyze the weekend nightlife drinking behavior of 241 young adults in a European country, using a dataset consisting of self-reports and passive smartphone sensing data over a period of three months. Using multiple statistical analyses, we show that features from modalities such as accelerometer, location, application usage, bluetooth, and proximity could be informative about different social contexts of drinking. We define and evaluate seven social context inference tasks using smartphone sensing data, obtaining accuracies of the range 75%-86% in four two-class and three three-class inferences. Further, we discuss the possibility of identifying the sex composition of a group of friends using smartphone sensor data with accuracies over 70%. The results are encouraging towards supporting future interventions on alcohol consumption that incorporate users' social context more meaningfully and reducing the need for user self-reports when creating drink logs for self-tracking tools and public health studies.

Opportunities for Smartphone Sensing in E-Health Research: A Narrative Review

“…A further six studies examined the use of smartphone sensing to detect intake and intoxication due to substances such as alcohol [ 60 , 61 , 62 ], tobacco [ 63 ], and cannabis [ 12 , 64 ]. They utilized location and activity data to understand the contexts of consumption and its impact on mobility.…”

“…The majority of studies (93%) utilized passively sensed data, by using smartphone sensors for data collection. While there were a few common in-built hardware sensors such as the accelerometer [ 3 , 11 , 53 ] and GPS/location sensor [ 9 , 60 , 69 ] used in most studies (62% and 53%, respectively), some also looked at software features such as application usage [ 61 , 91 ]. Studies used a range of sensors and software features to collect various types of contextual information.…”

“…Firstly, it is arguable that mental health conditions are a particularly important concern to measure, not only due to their prevalence over the globe [ 116 , 117 ], but also due to limitations that exist around actively collecting data with participants experiencing mental health concerns. For example, the subjectivity of self-reports [ 2 , 20 , 100 ], reliance on user input and compliance [ 61 , 99 ], and recall bias [ 70 , 81 ] might make passive sensing a more favorable approach to monitor mental health. Secondly, one of the limitations around smartphone sensing that was discussed was the impact of smartphone placement on physical data [ 2 , 3 , 11 , 103 ].…”

Sensing Eating Events in Context: A Smartphone-Only Approach

“…In essence, this means that smartphone sensing features have shown to be promising in inferring attributes that have shown to be part of the eight dimensions related to eating events. Such relationships have been used in prior ubiquitous computing (ubicomp) studies regarding the eating and drinking behavior [1], [9], [20], [33], [38]- [40]. Leveraging these relationships, this study seeks to examine whether smartphone sensing could be used to directly infer eating events by taking situational context and behavior sensed via smartphones as proxies for eating events.…”

“…According to Figure 2, 2X time window means that the inference can be done X minutes after the T anc . Using a larger time window like one hour (X = 30 minutes) is common for ubicomp studies regarding eating and drinking behavior that considers the context and behavior of participants in addition to the actual eating/drinking episode [9], [20], [33], [40]. This does not mean that the time of inference should be 30 minutes after the end of eating.…”