Study on the diagnosis and prevalence of internet addiction

Prevalence of Internet Addiction — Diagnostics and Risk Profiles (PINTA-DIARI)

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

PD Dr. Hans-Jür­gen Rumpf, Uni­ver­si­ty of Lübeck, Cli­nic for Psych­ia­try and Psy­cho­the­ra­py, Rat­ze­bur­ger Allee 160, 23538 Lübeck, Tel: 0451 / 500‑2871

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:

  1. The pre­va­lence esti­ma­te made in PINTA should be reviewed.
  2. 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.
  3. Risk fac­tors should be inves­ti­ga­ted which could play a role in the deve­lo­p­ment of the disorder.
  4. For the iden­ti­fied peop­le with inter­net addic­tion, it was necessa­ry to exami­ne whe­ther this dis­or­der is also asso­cia­ted with impairments.
  5. 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 groupTotal (%) Femi­ni­ne (%) Mas­cu­li­ne (%)
14–64 1,00,81,2
14–24 2,42,42,5
14–16 4,04,93,1

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 mon­thsInter­net-depen­dent ( > 4 cri­te­ria accord­ing to DSM‑5)
Not inter­net depen­dent ( < 5 cri­te­ria accord­ing to DSM‑5)
House­hold work4,0 (2,6)2,8 (2,1).002*
Abi­li­ty to work2,9 (2,7)1,3 (2,0)<.001**
Abi­li­ty to deve­lop clo­se relationships2,3 (2,3)1,4 (2,3)<.001**
Social life2,7 (2,5)1,1 (1,7)<.001**
Days of com­ple­te inter­net-rela­ted inca­pa­ci­ty for work5,6 (24,6)0,1 (0,1)<.001**
Days of light restric­tion of nor­mal activities30,7 (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 mon­thsCom­pu­ter games
Social net­works MW (SD)Others
House­hold work2,8 (2,7)3,9 (2,6)4,3 (2,5).735
Abi­li­ty to work2,3 (2,4)3,2 (2,9)3,1 (2,7).458
Abi­li­ty to deve­lop clo­se relationships1,9 (2,2)2,3 (2,4)2,7 (2,1).325
Social life2,9 (2,6)2,5 (2,5)2,9 (2,4).675
Days of com­ple­te inter­net-rela­ted inca­pa­ci­ty for work7,5 (34,7)6,3 (20,9)2,0 (7,8).690
Days of light restric­tion of nor­mal activities21,6 (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.


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.

Leave a Reply