German 
Compact report
To the Federal Ministry of Health, Germany
Project duration: 09/01/2011 to 03/31/2013
Funding number: IIA5-2511DSM230
Funding amount: € 252,960
Gallus Bischof, Anja Bischof, Christian Meyer, Ulrich John & Hans-Jürgen Rumpf
Lübeck, August 2013
Project management: PD Dr. Hans-Jürgen Rumpf
Contact:
PD Dr. Hans-Jürgen Rumpf, University of Lübeck, Clinic for Psychiatry and Psychotherapy, Ratzeburger Allee 160, 23538 Lübeck, Tel: 0451 / 500‑2871
hans-juergen.rumpf@ uksh.de
This compact report summarizes the main findings of the PINTA-DIARI study. Findings and data are not described in detail. These can be found in the detailed final report.
1. Background
Internet addiction is a relatively new disorder for which there is little scientifically proven data on how it can be diagnosed and which risk characteristics exist. In particular, there is a lack of data from the general population. Most studies have looked at groups at risk or people receiving treatment for the disorder. Frequently, only questionnaire procedures were used and there was usually no detailed diagnosis and assessment of the clinical significance of the disorder. Such data are necessary for understanding this disorder, estimating the frequency of its occurrence, and planning prevention and therapy.
Such data are necessary for understanding this disorder, estimating the frequency of its occurrence, and planning prevention and therapy. A statistical method was used for this, which determined classes based on the answers to a questionnaire on Internet behavior (latent class analysis). The results showed that two groups were suspicious of the response patterns, one that showed signs of risky internet use and a second that showed evidence of internet addiction. This last group comprised 1% of the population between 14 and 64 years of age (Rumpf et al., In press). While significantly more male respondents had internet addiction in older studies, the PINTA study showed rather small and insignificant differences between men (1.2%) and women (0.8%), with women and girls more social Reported networks as the main internet use and the male participants online role-playing games. Overall, higher rates were found in younger age groups (2.4% of the 14 to 24 year olds and 4.0% of the 14 to 16 year olds) and here female participants were affected even more frequently.
As a consequence, the findings showed that more precise recording methods were necessary to clarify the relevance of the PINTA results. It remained unclear whether some of the findings could possibly be attributed to the fact that, especially in young people, the Internet triggers a form of fascination, the appearance of which can be confused with signs of addiction. It was necessary to clarify whether these people actually experience significant impairments from excessive Internet use.
2. Objectives of the study
The main aim of the present study was to carry out a more precise diagnosis with the help of a standardized clinical interview in order to provide a better picture of the disorder and to assess the associated impairments. The goals included:
- The prevalence estimate made in PINTA should be reviewed.
- For the Compulsive Internet Use Scale (CIUS; Meerkerk, Van Den Eijnden, Vermulst, & Garretsen, 2009), which was used in PINTA, a cut-off value should be determined that differentiates between people with and without dependency.
- Risk factors should be investigated which could play a role in the development of the disorder.
- For the identified people with internet addiction, it was necessary to examine whether this disorder is also associated with impairments.
- In addition, the aim was to compare main activities on the Internet (computer games, social networks and a group with other activities such as compulsive research or the consumption of erotic and pornography) in terms of their significance for Internet addiction.
3. Methodical approach
Participants in the PINTA study received a follow-up questionnaire in PINTA-DIARI, in which a standardized, fully structured diagnosis of Internet addiction was carried out in detail. All those who showed increased values of 21 or more points in the CIUS were included. The follow-up survey period was between 12 and 30 months and averaged 21.5 months. The interviews were carried out nationwide by clinically experienced interviewers (mostly psychologists). A total of 196 people could be followed up.
The diagnosis consisted of an extensive interview, in which suggestions for the criteria of internet addiction were integrated. In particular, the proposal of the American Psychiatric Association announced during the project period could be mapped. This proposal was initially introduced as a research diagnosis “Internet Gaming Disorder” in the Diagnostic and Statistical Manual for Mental Disorders (DSM‑5; American Psychiatric Association, 2013) limited to the field of computer games (Petry & O’Brien, epub 2013) and within the framework used in the present study as the basis for diagnosing Internet addiction; this means that the criteria were also used here for other activities on the Internet and not just for computer or online games. According to DSM‑5, 5 out of 9 criteria must be present. In addition to this diagnosis, risk factors and impairment measures as well as other psychological disorders were recorded. The recording of the still existing (comorbid) mental disorders also included personality disorders and was carried out through proven, standardized, partially and fully structured interviews. The Munich Composite International Diagnostic Interview (M‑CIDI; Wittchen et al., 1995) was used to diagnose disorders in the area of axis I according to DSM-IV (Saß, Wittchen, & Zaudig, 1996). Personality disorders were diagnosed with the SKID II (structured clinical interview for DSM-IV, axis II; Fydrich, Renneberg, Schmitz, & Wittchen, 1997).
4. Results
4.1 Prevalence Estimation
The prevalence estimate from PINTA (see Table 1) was essentially confirmed. In the follow-up sample, the proportion of people who showed an internet dependency at the two measurement times using the different measurement methods (LCA in PINTA and diagnostic interview with the criteria according to DSM‑5 in PINTA-DIARI) was almost identical. When comparing gender- and age-specific subgroups, only a few deviations were found when comparing the two diagnoses, so that a correction of the data from PINTA does not seem justified. The largest deviation was found in the group of 14- to 24-year-old women with higher prevalence rates at the follow-up examination.
Table 1: Prevalence of internet addiction from PINTA
| Age group | Total (%) | Feminine (%) | Masculine (%) |
|---|---|---|---|
| 14–64 | 1,0 | 0,8 | 1,2 |
| 14–24 | 2,4 | 2,4 | 2,5 |
| 14–16 | 4,0 | 4,9 | 3,1 |
4.2 Cut-off for the CIUS
For the CIUS questionnaire, the results from the first survey time (PINTA) could be compared with the interview data from the second (PINTA-DIARI). These were used to determine a cut-off value. A distinction should be made between a recommendation for the purpose of case identification (e.g. in the context of prevention or intervention) and the prevalence estimate. The results suggest that different values should be recommended here.
To identify the case, the data of the people examined (who all had at least 21 points in PINTA, i.e. represent a selected group) could show that a cut-off of 28, as suggested by the authors of the test (Meerkerk et al., 2009 ), or 30, as the data from PINTA suggest, is unsatisfactory. This would, if you z. For example, using the values for those dependent on PINTA-DIARI Lifetime, only 35% (cut-off 28) or 27% (cut-off 30) can be identified. Overall, a value of 24 is useful for the purpose of case finding if you want to discover at least 70% of the target persons.
A higher cut-off should be used for epidemiological purposes (prevalence estimation), because otherwise there is a risk of overestimating the prevalence due to the low specificity. A cut-off of 30 is recommended. The PINTA-DIARI data suggest a better approximation of the actual prevalence. This cut-off is in agreement with the PINTA results. Methodological restrictions apply to the choice of cut-offs due to the time interval between the surveys. Overall, caution is advised when using short tests to estimate prevalence, as the estimate depends on the relationship between sensitivity, specificity and true prevalence and can lead to massive overestimates or underestimations.
4.3 Risk Factors
A number of risk factors have been found that are associated with the diagnosis of internet addiction. As expected, the addicts indicated longer usage periods on the Internet. Furthermore, the main use of online games is a risk factor. While no gender differences with regard to the presence of Internet addiction were found within the group with excessive use, the previous findings of clear gender differences with regard to the preferred Internet applications were replicated among Internet addicts. While dependent computer games are primarily found in (younger) men, characteristics of addiction dominate when using social networks in (younger) women. There was a connection to other mental illnesses. Of the DSM-IV axis 1 disorders, the presence of at least one disorder was increased, as was the presence of mood disorders, but not of anxiety disorders. In particular, PINTA-DIARI was able to show that high rates of personality disorders occur comorbidly. This applies to the overall presence of at least one disorder as well as to clusters A, B and C. The rates are increased by at least three times for the addicts. Another comorbid disorder was ADHD as a risk factor. Sub-areas of impulsivity (lack of persistence and cognitive instability) were also more pronounced in the addicts.
4.4 Impairments
The DSM‑5 addicts showed significant impairments in a number of characteristics. Of 15 variables that record the effects of the Internet, the values for addicts were significantly worse in 11 areas. The negative consequences included wide areas of life such as health, performance and social contacts.
Table 2: Impairments of Internet consumption among addicts and non-addicts
| Impairment last 12 months | Internet-dependent ( > 4 criteria according to DSM‑5) MW (SD) | Not internet dependent ( < 5 criteria according to DSM‑5) MW (SD) | significance |
|---|---|---|---|
| Household work | 4,0 (2,6) | 2,8 (2,1) | .002* |
| Ability to work | 2,9 (2,7) | 1,3 (2,0) | <.001** |
| Ability to develop close relationships | 2,3 (2,3) | 1,4 (2,3) | <.001** |
| Social life | 2,7 (2,5) | 1,1 (1,7) | <.001** |
| Days of complete internet-related incapacity for work | 5,6 (24,6) | 0,1 (0,1) | <.001** |
| Days of light restriction of normal activities | 30,7 (64,6) | 3,3 (13,5) | <.001** |
* significant (p < 0.01)
** significant (p < 0.001)
When asked specifically about impairments due to Internet use in the last 12 months, all characteristics were significantly more pronounced among the addicts: Limitations in the household, in the ability to work, in the ability to enter into close relationships and in social life were stated. Furthermore, the number of days with slight restrictions on normal activities or complete internet-related inability to work was significantly increased (Table 2). To have a comparison, data from a WHO study can be used (Alonso et al., 2011): People with depression reported 4.1 days on which they were completely unable to fulfill their obligations, Drug abusers or addicts 1,2 and the internet addicts in PINTA-DIARI 5,6. The findings therefore suggest that this is a clinically relevant disorder.
4.5 Differences in main activities on the Internet (computer games, social networks, other activities)
Of the group of addicts, 36.6% said their main activity was online gaming, 36.6% said social networks and 26.8% other internet applications. There were hardly any differences between these three groups, also with regard to impairments (see Table 3), coping with life and characteristics of dependency. The online gamers, however, had the longest period of use. Women were significantly more likely to be dependent on social networks.
Table 3: Impairments caused by internet consumption among addicts, broken down according to main activities
Impairment last 12 months
| Impairment last 12 months | Computer games MW (SD) | Social networks MW (SD) | Others MW (SD) | significance MW (SD) |
|---|---|---|---|---|
| Household work | 2,8 (2,7) | 3,9 (2,6) | 4,3 (2,5) | .735 |
| Ability to work | 2,3 (2,4) | 3,2 (2,9) | 3,1 (2,7) | .458 |
| Ability to develop close relationships | 1,9 (2,2) | 2,3 (2,4) | 2,7 (2,1) | .325 |
| Social life | 2,9 (2,6) | 2,5 (2,5) | 2,9 (2,4) | .675 |
| Days of complete internet-related incapacity for work | 7,5 (34,7) | 6,3 (20,9) | 2,0 (7,8) | .690 |
| Days of light restriction of normal activities | 21,6 (71,6) | 27,2 (55,8) | 47,8 (66,9) | .308 |
5. Conclusions
As far as we know, this is the first time that data is available on the criteria for Internet gaming disorder proposed in DSM‑5. The criteria were also transferred to other applications (social networks and others). It is also a follow-up survey that comes from a representative sample of the population and is therefore not — unlike the majority of other studies — limited to subpopulations such as students or selected online users.
The following conclusion can be drawn:
- The prevalence from PINTA can essentially be confirmed. There are no indications of systematic deviations that justify a correction.
- Two cut-off values could be proposed for the CIUS, depending on whether a case identification (24) or a prevalence estimate (30) should be carried out.
- Longer periods of use, computer games and the presence of comorbid mental illnesses were factors that were associated with the presence of addiction.
- The results also show that Internet addiction is associated with significant negative effects and that this applies to online gaming as well as to social networks and other Internet applications.
- Overall, there were hardly any deviations with regard to the applications on the Internet. The data suggest that in addition to computer games, social networks and other applications can lead to dependency.
- Women show a greater risk from social networks, male participants more from computer games. From this it can be deduced that the intervention offers, which have so far mostly been tailored to the clientele of computer gamers, need to be expanded to include the application of social networks.
The data provide a good basis for developing specific measures with regard to prevention and therapy. The PINTA-DIARI study is, even if a subpopulation from PINTA was questioned, not an actual longitudinal study that can depict the course of the disorders. No data was available from the first measurement time for such a purpose. For a better understanding of the disorder, such studies are desirable in the future.
Literature
Alonso, J., Petukhova, M., Vilagut, G., Chatterji, S., Heeringa, S., Ustun, T. B., et al. (2011). Days out of role due to common physical and mental conditions: results from the WHO World Mental Health surveys. Molecular Psychiatry, 16(12), 1234–1246.
American Psychiatric Association (Ed.). (2013). Diagnostic and Statistical Manual of Mental Disorders, fifth edition. Washington, D.C.: American Psychiatric Association.
Fydrich, T., Renneberg, B., Schmitz, B., & Wittchen, H.-U. (1997). SKID-II. Strukturiertes Klinisches Interview für DSM-IV Achse II: Persönlichkeitsstörungen. Göttingen: Hogrefe.
Meerkerk, G. J., Van Den Eijnden, R., Vermulst, A. A., & Garretsen, H. F. L. (2009). The Compulsive Internet Use Scale (CIUS): Some Psychometric Properties. Cyberpsychology & Behavior, 12(1), 1–6.
Petry, N. M., & O’Brien, C. P. (epub 2013). Internet gaming disorder and the DSM‑5. Addiction, epub 2013.
Rumpf, H. J., Vermulst, A. A., Bischof, A., Kastirke, N., Gürtler, N., Bischof, G., et al. (in press). Occurence of internet addiction in a general population sample: A latent class analysis. European Addiction Research.
Saß, H., Wittchen, H.-U., & Zaudig, M. (1996). Diagnostisches und Statistisches Manual psychischer Störungen DSM-IV. Göttingen: Hogrefe.
Wittchen, H.-U., Beloch, E., Garczynski, E., Holly, A., Lachner, G., Perkonigg, A., et al. (1995). Münchener Composite International Diagnostic Interview (M‑CIDI), Version 2.2. München: Max-Planck-Institut für Psychiatrie.