Van Rooij et al.  describe a longitudinal study evaluating the presence and persistence of compulsive or problematic internet gaming in a large sample of Dutch school children. In many ways, this study goes beyond earlier research in attempting to identify and evaluate longitudinally ‘gaming addiction’. It examined a large sample, had a high response rate, employed a longitudinal design and included assessments of other presumably related constructs, such as psychosocial health.
‘Gaming addiction’ and the more broadly construed ‘internet addiction’ are currently not classified as psychiatric disorders in the Diagnostic and Statistical Manual for Mental Disorders (DSM), 4th revision, although discussions are under way regarding the existence of such a putative disorder . Many anecdotal reports exist regarding the excessive amounts of time spent in play, and adverse consequences of intensive play on individuals' personal, social and school or work functioning [e.g. 3]. Centers dedicated to the treatment of these conditions are also in operation .
However, there are not yet any clearly defined or well-established clustering of symptoms that characterizes gaming ‘addiction’. Expert consensus is not achieved in terms of what constitutes this putative disorder [5,6]. A variety of surveys exist to assess it, but none have well-established psychometric properties  and the research itself is inconsistent in its definition of problematic gaming. Some studies utilize diagnostic criteria for pathological gambling and reword them to assess gaming, other surveys integrate substance use disorder criteria, another set utilizes a combination of these two types of criteria, and still others propose new and unique ‘criteria’[e.g. 7–12]. Further complicating matters, many researchers and clinicians modify the existing instruments and employ different cut-points to identify problems [13,14].
Given the inconsistent methodology used to classify ‘gaming addiction’, it is not surprising that prevalence rates vary vastly depending upon the instruments and methods employed. Some reports suggest very low prevalence rates of ‘gaming addiction’[<1%; 15] and others very high rates [>35% 16]; rates may vary by culture, but even within countries high variability in prevalence rates of ‘gaming addiction’ are noted [c.f., 17,18]. Moreover, almost all the samples are derived from school- or university-based surveys, and representation of older and even young adult to middle-aged respondents in surveys assessing ‘gaming addiction’ are sorely lacking. When non-school-based surveys are conducted, most often it is a convenience sample of people recruited from gaming or internet websites [e.g. 19], thereby failing to be representative of the general population.
The van Rooij et al.  study was a school-based survey, and it had a high response rate. The longitudinal nature of the study design is critical to interpretation of the findings. Interestingly, only half the respondents classified initially as ‘addicted’ to gaming remained so 1 year later. Although the sample size of ‘addicted’ gamers was small, the lack of persistence of ‘addiction’ over this relatively short time-frame suggests that the behavior may be transitory. Further, classification as ‘gaming addicted’ was not associated with poor psychosocial functioning in this study, even when a variety of instruments was used to examine mental health along a number of dimensions. The modest level of impairment noted in these presumably ‘addicted’ gamers indicates that perhaps this may not be a true psychiatric disorder, at least not as defined in that study.
Much more scientific research is needed regarding gaming and internet ‘addictions’ more generally. Recommendations include establishment of a provisional expert consensus of the behavioral patterns underlying the condition(s). These criteria should be subjected to extensive psychometric testing to establish multiple forms of validity and reliability using samples drawn from the general population and clinical settings. Long-term prospective evaluations should be conducted using standardized criteria to determine temporal stability of the putative disorder(s). Familial and genetic patterns of the condition(s) should be evaluated, along with biomarkers. Finally, associations with other psychiatric disorders should be assessed to ascertain if they are unique disorders or manifestation of another (or other) psychiatric condition(s).
Until such data are collected, caution should be extended when attempting to classify excessive behavior patterns as psychiatric disorders. As we move forward in understanding more clearly how, why and who engages in problematic internet gaming, we should also consider the past. The internet gaming ‘addiction’ of the early 21st century may be reminiscent of video arcade and television ‘addiction’ from the previous generation. If we draw the threshold too low, all excessive behavior patterns, including eating too much chocolate and working too much, may become psychiatric disorders, minimizing the experiences of, and ultimately limiting treatment availability and coverage for, individuals with true psychiatric disorders.