This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Public Health and Surveillance, is properly cited. The complete bibliographic information, a link to the original publication on https://publichealth.jmir.org, as well as this copyright and license information must be included.
Digital technology use is nearly ubiquitous among young adults; this use provides both benefits and risks. The risks of technology use include maladaptive technology use or technology addiction. Several conceptualizations of these addictions have emerged, each with its own assessment tools. These conditions include problematic internet use (PIU), internet gaming disorder (IGD), and social media addiction (SMA). These conditions have been associated with health outcomes such as problematic alcohol use, sleep disorders, and mental illness. These maladaptive technology conditions have been most commonly studied in isolation from each other.
The aim of this study is to examine PIU, IGD, and SMA together to better inform future research approaches and provider screening practices for young adults.
This cross-sectional survey study was conducted using Qualtrics panel-based recruitment and survey hosting. We recruited US young adults aged 18-25 years. The survey assessed PIU, IGD, and SMA. Survey measures also included assessments of problematic alcohol use, sleep, depression, and anxiety. We evaluated the frequency of and overlap in positive screening scores among PIU, IGD, and SMA and modeled each condition using multivariate logistic regression. Finally, we calculated sensitivity and specificity, as well as the positive predictive value and negative predictive value of the screening tools using the most prevalent maladaptive technology type.
Our 6000 participants had an average age of 21.7 (SD 2.5) years. Of these 6000 participants, 3062 (51.03%) were female, 3431 (57.18%) were Caucasian, 1686 (28.1%) were in a 4-year college program, and 2319 (38.65%) worked full time. The mean PIU score was 3.5 (SD 3.1), and 53.58% (3215/6000) of participants met the criteria for PIU. The mean IGD score was 2.7 (SD 2.6), and 24.33% (1460/6000) of participants met the criteria for IGD. The mean SMA score was 7.5 (SD 5.7), and 3.42% (205/6000) met the criteria for SMA. Across all 3 maladaptive technology use diagnoses, there were varied associations with demographic variables and similar overlap with health outcomes. The sensitivity of PIU screening to detect IGD was 82% and to detect SMA was 93%, whereas the specificity and positive predictive value were much lower (37%-54% specificity; 6%-37% positive predictive value).
This cross-sectional survey screened a large national sample of adolescents and young adults for PIU, IGD, and SMA to determine prevalence and overlap, demographic associations with each, and associations between these technology-related conditions and health outcomes. There was overlap across PIU, IGD, and SMA in some associated demographic variables and health outcomes. However, the patterns in the associated variables demonstrated unique qualities of each of these conditions.
Adolescents and young adults (AYAs) are often considered digital natives as they are growing up in a highly immersive technological society. Most US adolescents own their own personal smartphones, providing constant access to communication, information, and social networks [
Efforts toward understanding maladaptive technology use began in the 1990s, with a focus on the internet and its overuse. Two initial conceptualizations were based on the existing Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition disorders: substance abuse and dependency and pathological gambling [
Since these initial conceptualizations, the field has changed both through additional evidence and conceptualizations of maladaptive technology use and on changing technology. Conceptualizing maladaptive use has moved forward as inclusive of constructs beyond overuse and now includes constructs to represent risky use and individual impairment. Furthermore, studies have introduced conceptualizations of specific addictions related to new technologies, such as video games and social media. In most cases, these conceptualizations all center on these behaviors as “pathological forms of normal and necessary behaviors” [
A 2012 study developed a conceptual framework for PIU using empirical data and defined it as “internet use that is risky, excessive or impulsive in nature leading to adverse life consequences, specifically physical, emotional, social or functional impairment” [
Maladaptive video game use is most commonly referred to as IGD, given that most video gaming occurs on the web. In 2013, the American Psychological Association proposed IGD as a disorder in need of further study by [
SMA or addiction to social networking sites (SNSs) was defined in 2014 as “being overly concerned about SNSs, to be driven by a strong motivation to log on to or use SNSs, and to devote so much time and effort to SNSs that it impairs other social activities, studies or job, interpersonal relationships, and/or psychology health and well-being” [
A previous study examined the relationships between several behavioral addictions, including internet addiction, video game addiction, and Facebook addiction and the 5-factor model of personality [
Although tremendous strides have been made in the past several decades, several critical gaps remain in the literature. The divergent areas of focus for maladaptive technology use, including PIU, IGD, and SMA, mean that it remains unclear whether these diagnoses have similar associations with health outcomes. Although the literature has supported associations between PIU, IGD, and SMA and mental health conditions in particular, formal testing has not been conducted to evaluate the strength of these associations across these conditions. Furthermore, health care providers are increasingly called upon to screen for maladaptive technology use [
Building on these gaps in the literature, this study has 2 goals. The first goal is to understand the overlap in prevalence, demographic factors, and health outcomes across PIU, IGD, and SMA. If these 3 diagnoses are truly distinct, we would expect little overlap in a study population when screening for all 3. Furthermore, we would expect to see unique patterns in the associated demographic factors with PIU, IGD, and SMA. Finally, we would expect to see distinct associations with health outcomes. For this study, we will focus on health outcomes that have been studied in previous work with maladaptive technology use, including problematic alcohol use, sleep, depression, and anxiety [
The second goal of this study is to consider our findings and their impact on screening options for health care providers. If these 3 diagnoses are unique, then individual screening for PIU, IGD, and SMA is warranted. If there is overlap, it may be possible to identify a screening instrument with optimal sensitivity to facilitate universal screening. Then, when a positive screen emerges for a given patient, individual screening tools can be provided to improve specificity.
The aim of this study is to examine PIU, IGD, and SMA to better inform future research approaches and provider screening practices for young adults.
A web-based cross-sectional survey was conducted using Qualtrics panel-based recruitment and survey hosting. This study was approved by the University of Wisconsin institutional review board.
Our goal was to achieve a purposeful national sample of young adults to complete a closed web-based survey. Compared with traditional survey approaches, such as in-person, phone, or mail recruitment, web-based survey panels offer broader reach and lower costs in data collection [
The target population for this study was 18-25-year-olds who were US residents and English-speaking. We established parameters for Qualtrics to recruit a sample consistent with race or ethnicity representative of the US census population for 18-25-year-olds who could speak or read in English [
The survey was hosted on Qualtrics. Participants who provided consent were allowed to begin the survey. After informed consent was obtained, each measurement tool was provided on a single webpage. Demographic information was collected on a single webpage. Participants were allowed to skip questions or scales if they were uncomfortable or did not want to complete certain items. Participants could move backward and forward in the survey before submitting the results.
Participants were provided Qualtrics points as an incentive for survey completion. Survey results were delivered to the investigative team without identifying information.
Our 3 technology-related scales included assessments for PIU, IGD, and SMA.
The Problematic and Risky Internet Use Screening Scale (PRIUSS) [
IGD was measured using the IGD Scale [
SMA was measured using the Bergan SMA Scale [
Our health behavior and condition measures included problematic alcohol use, sleep, depression, and anxiety.
Problematic alcohol use has been associated with maladaptive use in previous work [
Robust literature, including a systematic review and a meta-analysis [
Depression has consistently been associated with maladaptive technology use across multiple studies [
Anxiety has been linked to maladaptive technology use in previous work [
Demographic measures included age, gender, and race and ethnicity. We asked the participants about the highest grade they had completed, with answer options including some high school, high school graduate, some college, technical school or associate arts degree, college degree, some graduate school, completed a graduate degree, and others. We asked about current education or employment status as well (eg, part-time or full-time employment).
Statistical analyses were conducted using SAS software (SAS Institute Inc), version 9.4. All
Age was categorized as younger (18-20-year-olds) versus older (21-25-year-olds) AYAs. Race or ethnicity was categorized for analyses based on 2 goals. One goal was to leverage statistical power by collapsing some groups with smaller proportions of participants. This categorization led to several larger groups. Our second goal was to ensure that the groups identified in previous studies as at risk for problematic technology use were included in the analyses. Thus, we then reviewed previous studies of problematic technology use [
To address our first study aim, we evaluated the frequency of and overlap in positive screening scores among PIU, IGD, and SMA. We calculated the proportions of participants who met the clinical criteria for PIU, IGD, and SMA using those scales’ validated score cutoffs.
To determine associations between demographic variables and maladaptive technology use diagnoses, as well as maladaptive technology use and health behavior and conditions, we used multivariate logistic regression.
Finally, we calculated the typical measures to assess the value of screening tests in clinical settings [
Our 6000 participants had an average age of 21.7 (SD 2.4) years. Of the 6000 participants, 3062 (51.03%) were female, 3431 (57.18%) were Caucasian, 1686 (28.1%) were in a 4-year college program, and 2319 (38.65%) worked full-time.
Demographic information of young adult participants recruited using Qualtrics panels (N=6000).
Variables | Values | |
Age (years), mean (SD) | 21.7 (2.4) | |
|
||
|
Female | 3062 (51.03) |
|
Male | 2841 (47.35) |
|
Other | 91 (1.5) |
|
||
|
Caucasian or White | 3431 (57.18) |
|
Asian | 304 (5.07) |
|
Hispanic or Latino | 502 (8.37) |
|
Black or African American | 793 (13.22) |
|
Other | 957 (15.95) |
|
||
|
Some high school | 101 (1.68) |
|
High school graduate | 2149 (35.82) |
|
Some college | 1682 (28.03) |
|
Technical school or associate arts degree | 311 (5.18) |
|
College degree | 1123 (18.72) |
|
Some graduate school | 232 (3.87) |
|
Completed graduate degree | 314 (5.23) |
|
Other | 68 (1.13) |
|
||
|
Part-time work | 1853 (30.88) |
|
Full-time work | 2319 (38.65) |
|
Part-time school | 486 (8.1) |
|
Full-time school | 1263 (21.05) |
Descriptive data for all measures are provided in
Descriptive data from young adult participants.
Variables | Mean (SD; range) |
PIUa | 3.5 (3.1; 0-12) |
IGDb | 2.7 (2.6; 0-9) |
SMAc | 7.5 (5.7; 0-24) |
Problematic alcohol use | 2.6 (2.6; 0-12) |
Depression | 9.4 (7.2; 0-27) |
Anxiety | 8.4 (6.1; 0-21) |
Sleep | 9.1 (5.2; 0-24) |
aPIU: problematic internet use.
bIGD: internet gaming disorder.
cSMA: social media addiction.
For PIU, the mean PRIUSS score was 3.5 (SD 3.1), and 53.58% (3215/6000) of participants met the criteria for PIU. The mean IGD score was 2.7 (SD 2.6), and 24.33% (1460/6000) met the criteria for IGD. The mean SMA score was 7.5 (SD 5.7), and 3.42% (205/6000) met the criteria for SMA. Of the 6000 participants, 1959 (32.65%) met the criteria for only PIU, 266 (4.43%) met the criteria for only IGD, and 13 (0.22%) met the criteria for only SMA. Approximately 40.03% (2402/6000) did not meet the criteria for any of these diagnoses.
Overlap in screening rates for problematic internet use, internet gaming disorder and social media addiction among a young adult population.
There were varied associations with demographic variables across all 3 maladaptive technology use diagnoses. There were no noted differences across younger and older AYAs in our study population for PIU, IGD, or SMA. Males were more likely to meet the criteria for PIU (odds ratio [OR] 1.2, 95% CI 1.0-1.4) and IGD (OR 2.9, 95% CI 2.5-3.5) than females. There was no gender association noted for SMA.
In evaluating race and ethnicity, Asian participants were more likely to meet the criteria for PIU than Caucasian participants (OR 1.7, 95% CI 1.3-2.4). Similar findings were present for IGD; Asian participants were more likely to meet the criteria for IGD than Caucasian participants (OR 1.6, 95% CI 1.2-2.2). Hispanic participants were also more likely to meet the criteria for IGD than non-Hispanic Caucasian participants (OR 1.3, 95% CI 1-1.7).
Higher educational levels were positively associated with PIU (OR 1.4, 95% CI 1.2-1.6) and SMA (OR 1.4, 95% CI 1-2) compared with lower educational levels. Being employed was positively associated with IGD (OR 1.3, 95% CI 1.2-1.5) than not being employed.
For health behavior and condition variables, problematic alcohol use was positively associated with PIU, IGD, and SMA. Sleep issues were associated with PIU and IGD across all 3 sleep outcomes (mild, moderate, and severe sleepiness), whereas SMA was only associated with severe sleepiness (OR 4.6, 95% CI 2.9-7.2). Depression screening showed a similar pattern: there were associations with PIU and IGD across all categories of depression, although SMA was only associated with severe depression (OR 2.5, 95% CI 1.3-5.3). Anxiety also showed a similar pattern. Anxiety was associated with PIU and IGD across all categories of anxiety, and SMA was only associated with severe anxiety (OR 5.2, 95% CI 2.5-10.9).
Given that PIU was the most prevalent maladaptive technology use diagnosis, we tested PIU as our defined gold standard in this study and its capacity to predict IGD and SMA. The sensitivity of PIU for detecting IGD was 82% (95% CI 80%-84%), and the specificity was 54% (95% CI 53%-56%). The PPV was 37% (95% CI 35%-39%), and the NPV was 90% (95% CI 89%-91%). Thus, the overall accuracy was 61% (95% CI 60%-62%).
The diagnostic accuracy of PIU for predicting SMA included sensitivity of 93% (95% CI 88%-96%), with a specificity of 47% (95% CI 46%-48%). The PPV was 6% (95% CI 5%-7%), and the NPV was 99% (95% CI 99%-100%). The overall accuracy was 48% (95% CI 47%-50%).
This cross-sectional survey screened a large national sample of AYAs for PIU, IGD, and SMA to determine prevalence and overlap, demographic associations with each, and associations between these technology-related conditions and health outcomes. There was overlap across PIU, IGD, and SMA in some associated demographic variables as well as health outcomes. However, the patterns in the associated variables demonstrated unique qualities of each of these conditions.
Our first finding was that PIU was the most prevalent condition among our study population, and screening for PIU captured many of the participants who screened positive for IGD and SMA. Given that PIU is more nonspecific than IGD’s focus on video games and SMA’s focus on social media, this finding may seem logical. There is overlap in the emotions and behaviors asked in each of the scales for PIU, IGD, and SMA; however, the PRIUSS asked participants to consider their internet use more broadly. Placing these findings in the context of previous work, we can consider the early conceptualizations of problematic technology use. Our findings align with an early cognitive behavioral model with focused attention on the impact of an individual’s thoughts on their development of problematic behaviors. This approach separated internet overuse into
It is notable that prevalence rates for PIU in this study, using a brief screening tool, indicated that just over half of participants screened positive for PIU using the short PRIUSS-3 screen. Previous studies of college students using the longer PRIUSS-18 have suggested prevalence rates among 4-year college students of around 4% [
A second consideration is whether these prevalence findings represent differences based on our broad study population of young adults. This study recruited a general population of young adults across the United States, a different approach than that of many previous studies that have focused on specific populations of college students [
A third possible consideration is whether our findings suggest an increasing prevalence of PIU within society over time. It can be argued that an increasing number of daily activities now occur on the web, such as shopping, viewing recipes, or learning new skills. Thus, the continued infringement of the internet into lives may lead to increasing reliance or dependence on web-based connections. However, given the stark differences in the prevalence for this study compared with others using the PRIUSS-18 and studying 4-year college student populations, this final consideration is not likely able to fully address or explain our findings.
We found strong correlations between PIU and the mental illnesses of anxiety and depression in the 0.4 range, and similar findings for SMA were found. Correlations for gaming and anxiety and depression were lower, in the 0.27-0.34 range. For correlations with alcohol use, we found lower correlations with all our technology diagnoses; the highest correlation was with PIU at 0.21. For sleep, correlations were similar across PIU, IGD, and SMA, all in the 0.3 range. These findings support screening for depression and anxiety concomitantly with maladaptive technology use. These findings also illustrate subtle differences among PIU, IGD, and SMA, such as how participants who screened positive for IGD also demonstrated a lower correlation for anxiety compared with depression, which differed from the pattern for PIU and SMA. These findings should be explored in future studies.
This study’s results may not generalize beyond a study population of young adults recruited via Qualtrics. Recruiting from a web-based panel meant that we could designate the study population size and criteria; however, it limited our ability to assess external validity. However, the Qualtrics platform and panels have been used in previous technology-focused studies [
Given that our study is cross-sectional and not longitudinal, we cannot conclude the directionality of these correlations. It is possible that our study illustrates preferential web-based activities among young adults with depression or anxiety. Furthermore, our study did not test or posit mechanisms by which problematic technology use and health conditions may be related. Although previous literature has identified associations between problematic technology use and certain conditions, an exploration of the mechanisms to support these associations remains understudied. This area of inquiry will be important for future work.
Finally, our study represents a unique approach for measuring multiple maladaptive technology conditions in a single study; thus, placing the findings in the context of existing literature is more challenging because of this unusual approach. Future studies should consider longitudinal study designs incorporating more than one measure of maladaptive technology use to further understand the similarities and differences across these conditions.
Despite these limitations, our study has several intriguing implications. The first implication is that the findings suggest that PIU, IGD, and SMA are more alike than different. We found some small differences with demographic factors, which may support group or cultural differences rather than different underlying mechanisms for these types of maladaptive technology use. Across our health outcomes, we found similar positive correlations across PIU and SMA with anxiety and depression, with IGD showing some small differences.
One clinical implication centers on our finding that PIU was the most common condition and that screening for PIU captured many participants who screened positive for IGD and SMA. Given that PIU was identified as having sensitivity to detect IGD and SMA at the levels of 82% and 98%, respectively, it is likely that screening for PIU may be a valid approach to detect concerns related to technology. Then, if the screen is positive, follow-up screening could include specific evaluations for IGD or SMA depending on the patient’s history. As the PRIUSS-3 is designed to optimize sensitivity, this tool is likely a reliable candidate for this type of initial screening. Furthermore, as the PRIUSS-3 includes 3 questions, this brief screening tool could be incorporated into clinical flows without undue patient or clinic staff burden.
Our study found that PIU, IGD, and SMA were also commonly associated with positive screens for depression and anxiety. Thus, another clinical implication is that for clinics that do not conduct routine depression and anxiety screening for all patients, a positive screen for PIU should also prompt a screen for depression and anxiety. Early and ongoing screening for PIU may be a potential approach for identifying young adults at risk for other mental illnesses, and it may prompt further evaluation.
In conclusion, our study advances the understanding of PIU, IGD, and SMA by demonstrating their strong overlap in meeting diagnostic criteria and similarities in demographic risk factors. We also found similarities in health behavior and conditions across PIU and SMA in particular. If these 3 conditions were truly distinct, we would anticipate little overlap in a population when screening for all 3 maladaptive technology conditions. We can conclude that PIU, IGD, and SMA are more similar than different. Finally, we identified that an efficient screening approach might be to conduct initial screening for PIU, followed by technology-specific screening assessments and screening for depression and anxiety.
Associations between demographic factors and health outcomes with maladaptive technology conditions.
adolescent and young adult
internet gaming disorder
negative predictive value
odds ratio
problematic internet use
positive predictive value
Problematic and Risky Internet Use Screening Scale
social media addiction
social networking site
None declared.