The future is now—using personal cellphones to gather data on substance use and related factors

Most of what is known in substance use research is based on retrospective answers in paper-and-pencil questionnaires or given online or in telephone interviews. Current smartphone technology, however, opens virtually unlimited possibilities for collecting data in real-time and real-life situations, including sounds, pictures or locations, and with (almost) no recall bias. Prospective diaries capture ‘life as it is lived’, minimizing recall bias 5. However, using paper-and-pencil diaries leads to other problems, such as forward- and backward-filling 6. Over the years, researchers improved diary protocols with (at that time) revolutionary technologies including interactive voice response 7, cellphone text messages (SMS) 8 or hand-held computers (palmtops) 9. These technical solutions offered many advantages, such as the possibility of prompting participants, stamping data records with the date and time of collection and programming alternative questions in response to previous answers 10. However, several drawbacks also became obvious, such as the intrusiveness of telephone calls or the character restrictions in SMS. Most importantly, the high costs, e.g. in terms of purchasing palmtops, training participants how to use them and replacing broken devices, made these data collection methods ill-suited for studies with large samples or particular populations, e.g. homeless people 10. In the last decade, cellphones have emerged as powerful data collection devices. First, cellphone ownership and use has proliferated globally. In 2013, 6.8 billion cellphone subscriptions were active world-wide, 96.2 subscription per 100 inhabitants overall, with 128.2 per 100 in developed countries and 89.4 per 100 in developing countries 11. Cellphone subscriptions exceeded landline telephone connections by a factor of 5.8 (i.e. 1.2 billion landline telephone connections) and 2.1 billion cellphone subscriptions even include mobile–broadband internet access. Secondly, most people carry a cellphone with them permanently and are familiar with its functions. Using personal cellphones, repeated measurements can thus be accomplished independently of time and location in a convenient manner, up to the point that they become part of the participants' everyday activities. Thirdly, smartphones combine the functionalities of any portable telephone (calls, SMS, etc.) with those of both multiple connectivity systems (mobile internet access, Bluetooth, WiFi, etc.) and hand-held computers (high-capacity data storage, e-mailing, web-browsing, installation of third-party software, etc.). Thus, in addition to being reachable almost any time and anywhere, smartphones provide a state-of-the-art data collection interface. Particularly dedicated applications (apps) have been used increasingly in mobile health-care management (the so-called ‘mHealth’) 12-15. Although most apps were designed for simple self-monitoring purposes, e.g. to survey chronic pain, women's menstruation cycles and diabetes treatments, their potential is unlimited in terms of collecting data on substance use and related factors. A major drawback of apps consists in the different operating systems (OS) of smartphones (Android, iOS, Windows Phone) and different versions of these. In January 2014, Android and iOS were world market leaders with shares of 45 and 24%, respectively 16. However, unless researchers decide that data collection will be limited to a given smartphone OS, different versions of the app need to be developed for each OS specifically, which is a costly enterprise. An alternative solution is to collect data using the cellphones' internet browsers. An example is the internet-based cellphone optimized assessment technique (ICAT), developed in 2009 17, 18. Independent of smartphones' OS, it consists of a series of internet-stored questionnaires that participants complete on their personal cellphone browsers. Prompts sent by SMS at any pre-programmed time-point contain a hyperlink that automatically opens the cellphone browser and uploads the questionnaire. For example, using ICAT, participants reported, within minutes and at multiple times during a single evening, how many drinks they had had during the previous hour, where, and with whom. Compared with the multiple in-the-event measures during Saturday evenings, the retrospectively indicated number of drinks consumed during the last four Saturdays before the study resulted in an underestimation of 45% for men and 39% for women 19. Smartphones offer virtually unlimited possibilities for substance use research. A wide range of data (e.g. scales, texts, sounds, pictures, videos) can be recorded with simple to highly complex study designs. To us, it appears that besides technical challenges (cellphone network coverage and speed, participants' media literacy and OS requirements, etc.) the main limitation for a more intensive use of this technology in the area of substance use is researchers' hesitation. While waiting for Google Glass to reach the same world-wide distribution that smartphones have nowadays, many years remain to explore more sophisticated research protocols combining self-reports with other data that can be collected on smartphones. GPS coordinates, sounds, pictures and videos are today just one fingertip away from drawing a more complete picture of what is actually happening before, during and after people's substance use behaviours, captured in the heat of their everyday life. None.