THE APPLICATION OF TECHNOLOGY IN DATA COLLECTION AND ANALYSIS
The use of technology in behavioral interventions is not limited to intervention delivery. Technologies can also provide advantages to the behavioral intervention researcher with respect to data collection and analysis. New technologies provide researchers with new tools to collect and analyze data and also afford researchers opportunities to collect new types of data that were previously thought to be unobtainable or difficult to obtain, for example, tracking the behavioral patterns of an individual or real-time continuous records of health indicators. The following section presents examples of possibilities of new technologies that are available with regards to data collection and assessment. The possibilities in this area are rapidly expanding as technology continues to evolve and improve. Our intent is to provide some insight into the possibilities offered by technology.
One of the more noticeable and positive developments in data collection afforded by technology is the move from using paper-and-pencil methods of collecting and recording data to performing these activities using a device such as a tablet. Use of devices such as tablet computers to record data (in comparison to a traditional paper-and-pencil method) can increase the efficiency and quality and decrease the costs associated with data collection and management. For example, collecting data via a tablet allows for the immediate transfer of data to the data management system for a study and eliminates the need for an additional data entry step in the process. These technologies (when designed properly) can be intuitive and easy-to-use for professionals of varying skill levels (Abernathy et al., 2008).
Another popular development with regard to technology and data collection is the proliferation of Web-based survey research. With traditional mail-in survey and telephone response rates decreasing dramatically over the past several decades (Dillman, Smyth, & Christian, 2009), Web-based surveys present researchers with a new method of engaging participants and distributing assessment instruments to a large number of people. Web-based surveys can be distributed through various avenues such as being e-mailed directly to potential research subjects, being shared on social media and in online forums, or being embedded into a website. Questionnaires can be built and programmed by the researcher to be tailored to the potential respondent and fit the needs of the study, and there are also a number of free online options (such as SurveyMonkey) that allow users to create surveys using premade templates. While survey response rates for Internet-only questionnaires may not be higher than those in traditional survey methods (Kaplowitz, Hadlock, & Levine,
2004), researchers may use a mixed-methods approach that incorporates Web- based instruments in the hopes of increasing overall response to a questionnaire. As noted in Chapter 13, issues related to informed consent can also be challenging.
In the case of Internet-based behavioral interventions, real-time data on the use of these interventions can be tracked remotely. As an example, if a researcher is evaluating an online intervention that is intended to relieve symptoms of depression, the researcher would be able to gather real-time information on the frequency with which the study participants accessed the intervention, when participants accessed the intervention, and the components of the intervention that they accessed on the site (such as opening a behavioral intervention module or posting a message on a support group message board). We tracked real-time participant usage data in the videophone study (Czaja et al., 2013). This type of data provides valuable information on how an intervention might need to be modified—for example, which components of the intervention engaged participants or were of low usage. In addition, it can help facilitate dosage-outcome analyses. There is software that exists (e.g., Morae) that allows researchers to record and visually track all activity completed on a device; a researcher conducting a study determining what a participant does on a behavioral intervention website can not only see what pages the participant visits or what links the participant clicks, but may also see the cursor movements in real time and gain additional insight as to the decision-making process used by participants when navigating the site. This opens up another exciting realm of analytics and level of understanding concerning how participants interact with treatment elements that has not hitherto been possible.
Mohr et al. (2013) outline another development in data collection that involves mobile technologies: passive data collection. In passive data collection, rather than having the technology user manually log in data regarding specific health practices or behaviors, the technology itself will log data using built-in or externally connected sensors. Examples of passive data collection can be GPS sensors that track the location of a user (which scientists can use to see where a user is and how much he or she is traveling, a potentially important piece of information in physical activity interventions) or sensors that record heart rate.
While mobile technologies can allow for passive data collection, other wearable technologies allow for similar data collection. A study done by Najafi, Armstrong, and Mohler (2013) tested accelerometers (in this case, sensors designed to monitor physical activity) to see whether sensors that could be worn as easily as inside a shirt could successfully track walking movements. They found that such technologies could actually help identify older adults who were at risk for falls. Many wearable technologies exist in the marketplace that can help track participant characteristics such as fitness level (e.g., Fitbit). Most of these devices also provide feedback to the user, which is in turn an intervention. New developments in sensing and wearable technologies are on the horizon. Examples include implantable devices such as implantable cardiac monitoring devices and stimulating devices and smart home applications that involve integrated networks of sensors—which may include a combination of safety, health and wellness, and social connectedness technologies— installed into homes or apartments to simultaneously and continuously monitor environmental conditions, daily activity patterns, vital signs, and sleep patterns.
Behavioral interventions conducted in an online environment also allow for the collection of qualitative data in addition to more quantitative survey responses and health measures. The most obvious example of online qualitative data is that found on social media sites or in online forums. In a study conducted by Frost and Massagli (2008), comments made by members of an online community called PatientsLikeMe were qualitatively analyzed to determine how members communicated their health status and opinions to other members and how members used the site to answer inquiries on their own health and behaviors. An advantage of using online discussions and forums is that online qualitative data produce “automatic transcripts”; there is no need for the researcher to use a recording device to capture an interview with a research subject nor is there a need for the researcher to be furiously scribbling down notes, because with analysis of online communication the “dialogues” between people are already written and ready for analysis (Im, 2006).