1. Perceivedconnections betweeninformation and communication technology use and mental symptoms among youngadults - a qualitative
study
Abstract
Background
Prospective associations have been foundbetweenhigh use of information andcommunication technology (ICT) andreportedmental symptoms
amongyoungadult university students, but the causal mechanisms are unclear.Our aim was to explore possible explanations for associations
between high ICT use andsymptoms of depression, sleep disorders, and stress among young adults in order to propose a model o f possible
pathways to mental health effects that can be tested epidemiologically.
Methods
We conducteda qualitativeinterviewstudy with 16womenand16 men (21-28years),recruitedfroma cohort ofuniversity students on the basis
of reportinghigh computer(n = 28)or mobile phone (n = 20)use at baseline andreportingmental symptoms at the one-year follow-up. Semi-
structuredinterviews were performed, with open-endedquestions about possible connections between the use of computers and mobile phones,
andstress, depression,andsleep disturbances. The interviewdata were analyzedwith qualitative content analysis and summarized in a model.
Results
Central factors appearing to explain high quantitative ICT use were personal dependency, and demands for achievement and availability
originatingfromthe domains ofwork, study, social life,andindividual aspirations. Consequences included mental overload, neglect of other
activities andpersonal needs, time pressure, role conflicts, guilt feelings, social isolation, physical symptoms, worry about electromagnetic
radiation, andeconomic problems. Qualitative aspects (destructive communication and information) were also reported, with consequences
includingvulnerability,misunderstandings, alteredvalues, andfeelings of inadequacy. User problems were a source of frustration. Altered ICT
use as an effect of mental symptoms was reported, as well as possible positive effects of ICT on mental health.
Conclusions
The concepts andideas of the youngadults with highICT use andmental symptoms generateda model of possible paths for associations between
ICT exposureand mental symptoms.Demands forachievement andavailability as well as personal dependency were major causes of high ICT
exposure but also direct sources of stress andmental symptoms. The proposedmodel shows that factors in different domains may havean impact
and should be considered in epidemiological and intervention studies.
Background
Mental health problems including anxiety and sleep disturbances have increased among the general Swedish population over the past few
decades, with the highest increase seen in adolescents andyoung adults [1,2]. Mental disorders account for a large proportion of the disease
burden in youngpeople in all societies,and are a global public health challenge [3]. The causes of mental disorders are multi-factorial, and
cultural factors seem tohave aninfluence.One ofthe major changes in the exposure profiles of young people is the use of information and
communicationtechnology(ICT). In 2007,75% of16- to24-year-oldSwedes used the internet everyday, andmore than 90% hada computer at
home [4].Almost all hada mobile phone. In a qualitative study,Gustafsson et al [5] explored young adults' experience, attitudes, and health
beliefs in relation to information technology use and found that they perceived both opportunities and risks, including risks to physical and
psychosocial health. Positive healthaspects were also described. In a prospective cohort study of young adult college and university students,
participants displayinghigh ICT use at baseline hada higher prevalence ofreportedmental symptoms at the one-year follow-up, in comparison to
those with lowuse [6]. For women, a high combined use of computer and mobile phone at baseline was associated with an increased risk of
reportingsymptoms of depression and prolonged stress; online chatting was associated with prolonged stress; emailing and chatting were
associatedwith symptoms of depression; andinternet surfingwas associatedwith theprevalence of sleepdisturbances. For men, the number of
mobile phonecalls Short Message Service (SMS) text messages sent orreceivedperday were associatedwith sleep disturbances, and high SMS
use was also associatedwith symptoms of depressionat follow-up. The study concludedthat ICT use may have animpact on mental health but
that the causal mechanisms are unclear[6]. Earlier researchhas addressedthequestionof mechanisms. For example, it has been suggested that
amongsome people, computer work canleadtopsychophysiological stress reactions ("techno-stress") due to occupational strain, and that these
reactions can become conditionedtothe computer workenvironment leadingto symptoms associatedwith the computer [7]. Other researchers
have discussedpsychological factors, such as the role ofthe internet in relation to mental health[8,9], andpsychiatricfeatures, such as addiction
to the internet [10].Consideringthe wide use of ICT, thereis an urgent needto develop a model for possible explanat ions for the association
between ICT use amongyoungadults andsymptoms of depression, sleepdisturbances, andstress. A first step is to ask the young adults which
connections they themselves perceive between ICT use and mental symptoms.

2. Our overall aim was to explorepossible explanations for associations betweenhigh ICT use and symptoms of depression, sleep disorders, and
stress amongyoungadults in order topropose a model of possible pathways tomental healtheffects that can be tested epidemiologically. The
specific purpose ofthis study was to explore high ICT users' own ideas andperceptions of associations between computer and mobile phone use,
respectively, and stress, depression, and sleep disorders.
Methods
This explorative qualitativestudy was basedon semi-structuredinterviews with 32 youngadults who in a cohort survey had reported high ICT
exposure prior toreportingmental symptoms. The study was approvedby theregional ethics committee for the University of Gothenburg. All
participants gave written consent.
Study group
Sixteen women and16 men were recruitedfroma cohort of youngadults (20-28years)who hadrespondedto an annual questionnaire concerning
ICT exposure, demographics, andpsychosocial, physical, andmentalhealthfactors, in 2004 and2005 (n= 1843). The cohort participants were
recruitedfromstudent records at colleges anduniversities in the southwestern andsouthernregions of Sweden (IT -related, medical, or nursing
studies).
To be includedin the present study, participants must live in the southwestern region ofSweden, have reportedhigh computer or mobile phone
exposure in 2004,andhave confirmedat least two ofthe followingmental symptoms in 2005:prolongedstress,depressive symptoms, and sleep
disturbances. High ICT exposure was definedas the highest ranking reports (for men and women separately) of estimated total duration of
computeruse duringthe past week, numberof mobile phone calls andtext messages perday in the past week, or both. Particip ants must also
have reportedthat this exposurewas representative of their typical use. Mental symptomitems are shown in Table 1. To enhance the potential for
identifyingfactors or conditions connectingICT use with mental symptoms, ten subjects were strategicallyincludedbecause theyhadreported a
perceivedconnectionbetweenmental symptoms and IT use in response to a direct question in the cohort questionnaire. Participants were
recruitedconsecutivelyuntil32 individuals hadbeen enlisted. Potential participants receiveda postal letterof invitation with information about
the study. A week later they were telephonedandaskedto participate. Monetarycompensation was offered to make up for the loss of time or
salary. A total of 44individuals were contacted. Of these, elevendeclined to participate: six due to lack of time, three due to current travels
abroad, and two for no specified reason. One person agreed to participate but defaulted without further contact.
Table 1. Study demographics
The high ICT users formedtwo subgroups: high computer users (n = 28) andhigh mobile phoneusers (n = 20). Exposure in the upper half for the
technologyin questionwas consideredsufficient toqualify foreach subgroup. Sixteen participants qualified for both groups. See Table 1 for
study group demographics.
The high computeruse group comprised28 participants (15men and13 women, aged21-28).Of these,21 hadreportedregular weeklycomputer
use in the upper quartile of thetotal cohort. Twenty-four hadreportedall three mental symptomitems,andthe remainingfour hadreportedtwo of
the mental symptoms. All 15 men and10 of thewomenhada backgroundin computer-related study; two of the other women were nurses and
one was a medical doctor. Tenparticipants (six men andfour women) hadreporteda perceivedconnectionbetweenIT use andsubjective mental
symptoms in the cohort questionnaire.
The high mobile phoneuse group comprised20participants (8menand12 women,aged22-28). Of these,14 had reported regular daily mobile
phone use in the upper quartile ofthe total cohort.Seventeenhadreportedall three mental symptom items, andthe remainingthree had reported
two of the mental symptoms. Seven menandsix womenhada backgroundin computer-related study, one man and four women were medical
doctors or students, and two women were nurses.
Data collection
Individual semi-structuredinterviews were performedby the main author. The interviews took place from October 2005 to April 2006 at a
university hospital clinic.Theparticipants were askedopen-endedquestions about possible connections between theuse of computers andmobile
phones, andstress, depression, andsleep disturbances,forexample: "Doyouthink there is a connectionbetweenthe use of computers andstress?
If so, how? Have youexperiencedthat yourself? What about people in general?"In addition, direct questions were askedabout the participants'
own worries about personal ICT use, their experiences ofproblematic or destructive ICT use, and the impact of ICT use on their sleep. The
interviews lasted between 40 and 90 minutes and were tape-recorded.

3. In parallel with andblindedto the explorativeinterviews, a physicianperformedpsychiatric assessments ofthe participants. Psychiatricdisorders
were diagnosed in six men and six women [Edlund, M, unpublished data]. Mild depression was diagnosed in four participants, moderate
depression in two, anxietydisorder in four, mixedanxietyin one, andunspecifiedmoodsyndrome in one participant. None of these diagnoses
had previously been identified.
Analysis
All interviews were transcribedverbatimto enable qualitative content analysis [11-13].The interview texts were sorted at topic level to form
thematic domains andenteredintotables, mainly followingtheinterviewstructure. Onlydatafromhigh users of theICT technology in question
(computer ormobile phone) were usedin the analysis of eachdomain.A stepwise conventional qualitative content analysis [11,12] of the data
was performedby the main author (alsothe interviewer).The text was first condensedandcoded; codes were then sorted and categories formed
based on the factors andconditions describedby the respondents in each domain. The process focused on a low level of interp retation of the
manifest content,closelyfollowingthedictums ofthe subjects [13]. In thenext step, the categories were compared and sorted across thematic
domains to generate overall categories. Patterns and relations among the categories were sought in an iterative process, using the original
interviewtexts as additional verification. Finally, a model was developedto illustrate anddescribe the categories andtheir relations, in order to
present central features connectingICT andmental symptoms while accountingfor just about all factors mentioned by the subjects [13]. The
model was developed in conjunction with the co-authors.
Results
We first present results fromthe content analysis ofeach domain, followedby a comparisonofthe results between domains.Finally, we suggest a
model of proposedpathways toilluminate the connections betweencomputerandmobile phone use andmental symptoms, as perceived by the
young adults.
Computer useand stress
Most of the high computer users expressedideas about links between computer use andstress, though somestatedthat these only appliedto other
people, and not themselves. High quantitative use, that is, long periods spent sitting at the computer, was a central link between use and
symptoms. It was commonforparticipants to spendmore timethanplannedat the computer, either because tasks took longer th an expected
(leadingto time pressure,tight deadlines,andovertime work), orbecause theybecame involvedin internet surfing or game playing. Perceived
demands andexpectations, andtheir own desires, tobe available via chat or email were also a source of highquantitativeuse and spending more
time thanintendedat thecomputer. It was consideredeasy tolose perception oftime andtoneglect bodily signals of needs such as breaks, food,
drink, or physical activity,while sittingat the computer. Furthermore,time spent at thecomputerwas seen as time taken from other activities:
Particularly,in front ofthe computeryoucan lose perception oftime. Youdon't feel hunger, youdon't feel tired. And then if you sit in front of
the computer andlose trackoftime andlookat the clock - gosh! Well, then you end up getting stressed. Also, you don't get any energy or
nutrition.Youjust sit thereandyour fingers are ice coldandyouhaveno warmthin your feet.So in that way it couldvery well be harmful... and
stressful... and above all, you are tired when you eventually quit. (Man A, 24 years)
Overloadin communication was another problem. Chat or email messages interrupted other computer tasks, and it could be difficult to filter
important messages fromunimportant ones. It was difficult for participants to find time to answer all their messages as quickly as they felt
expected to, and not doing so often resulted in feelings of guilt, resentment, and stress:
If someone sends an email andyouhaven't answeredwithinan hour, you'll get "Didyoureadmy email or what?"[...] Youcan't concentrate on
anythingbecause youget interruptedall the time. Andalso, youare expectedtobe available somehowandthat can be stressful. (Woman A, 27
years)
Another stressor was managing several communications simultaneously:
Youhave your place in reality, where you're physicallyat,but then youoftenalso have maybe four other places in the virtual world [...] where
youare representedandwhere youneedto keep up with your friends there [...] Andyou have to keep up with all those worlds and keep them
alive and in your memory. (Woman B, 26 years)
Stress was also impliedby dependency andaddictionissues, such as compulsivelycheckingmessages or information, excessive game playing,
online poker addiction, andfeelingstressedwhen not connected: "It's like I'mso incredibly usedto the computer. I get kindof edgy when I'm not
at a computer. I feel almost naked." (Man B, 24 years)

4. High demands for work speed, perfection, efficiency, andbeingup to date, were experiencedin connection withcomputeruse. Computeruse was
also consideredto be part of increaseddemands in general. Physical aspects of computer use, such as lighting and monitor displays, could also
have an impact onstress (electromagnetic radiationwas mentionedas a possible stressor),as couldsome cognitive aspects, such as timelimits in
computergames or the pressure to process a great deal of data in a short time. User problems concerning software or hardware, including
frustrations due to slowperformance or harddisk crashes,as well as competence issues such as havingto learnnewsoftware, were considerable
sources of stress, and also increased the time spent at the computer.
Several participants made the point that the sourceof stress was not computer use in itself, but rather task-, work-, or study-related demands
connectedwith computer use, or computer use as part ofmeetingincreaseddemands in general. Stress concerningthefuture in connection with
computer use, in particular future job opportunities in IT, was also mentioned.
Alongwith all of the ways in which it was seen to add to stress levels, computer use was also considered to decrease stress, for example by
allowing more work to be done in a shorter time.
Computer useand depression
A central concern in discussions of computer use related to depression was high quantitative use, especially the idea of gett ing stuck in
unproductive activities, such as game playing, which led to the feeling of having wasted time.
Well, it's kindof sadthat youspendso many hours at the computer, as such. That's howI can feel afterI'veturnedthe computer off: What did I
do tonight?"Yeah, I sat threehours - afterspendingeight hours at work!" I couldhave goneout andhadcoffee with a friendinstead. That's what
I feel. (Woman C, 27 years)
Demands for andexpectations of availability were consideredtoresult in communicationoverload, andfeelings of guilt due t o not being able to
manage all the communication. Social isolation was another concernin relation to high computer use. A negative loop was suggested, in that
already-lonely people may have a preference for usingcomputers, which in turn could increase their tendency to lack real-life contacts and
relationships, and lead to even heavier computer use. Some participants identified themselves as being in this category.
It's probably more that youignore the otherlife. Andthat yousit at home toomuch anddon't meet any people [...] It feels safer to sit in front of
the computer forme sinceI findit easier to write thantocommunicatein real life andso I'll sit in front of the computer rather a lot. (Man C, 27
years)
High computeruse was also consideredto have a negative impact onphysical health, leadingtophysical symptoms such as musculoskeletal pain,
headaches, andtiredness, which in turn could increase depressive feelings. Dependency or addiction issues (e.g. online gambling) were also
considered risk factors for depression, as was stress on relationships caused by one partner's heavy computer use.
As well as high quantitative use, badquality or content ofuse, for example destructiveinformationandcommunications, was perceived as a link
between computer use anddepression.This includednot onlybador harmful information ormisunderstandings in chat communications, but also
feelings of inadequacy in the face of the unlimited possibilities and high achievements of others portrayed on the internet:
Well, if yousurf the internet a lot, then youcan findlots of things that canmake youfeel like a failure because yousee howsuccessful everybody
else is, or how good looking everybody is, or how much is happening there and you can't go there. (Woman D, 27 years)
User problems andthepossibilityof failingto meet excessiveexpectations of user competence were mentionedas factors for depression. On the
other hand, computer use was also considered to decrease depression through allowing easy access to social support even durin g times of
depression andby supplyingfun andentertainment. Furthermore,depressive feelings couldleadto alteredcomputeruse, such as more chatting,
because it couldseem easier to communicate via computer than in person,or decreasedchattingbecause of the wish not to communicate. More
time could also be spent on amusements or procrastination when one was feeling down.
Stress or sleep disorders, in relationtocomputer use and connected to study- or work-related demands, were considered possible links from
computeruse to depression. There were also participants who perceivedno associationat all between computer use and depressive symptoms.
Some pointed out that depression must depend on individual factors.

5. Computer useand sleep disorders
Many participants couldrelate tohavinginsufficient or dislocatedsleep aftersittingup late in front of the computer because of getting stuck in
tasks, meetingdeadlines, chatting, or game playing. It was sometimes compellingto keep up with other time zones for chatting or gambling.
Several experienceddifficulties relaxingafter intense computer use, which in turn gave rise to difficulties falling asleep because of being too
arousedor bringingtask-relatedproblems tobed. Some described seeing pictures of their recent computer activities (e.g. poker cards) when
trying to go to sleep.
Because if yousit - andI've made that mistake manynights that I sit up late andchat.Andthenyouget stuckin it - youtalktopeopleandit's fun,
andthen yousit andsurf arounda little bit andit takes extratime, andthen it's two o'clock - and you have to get up at seven. (Woman B, 26
years)
Andwhen workingon this project, for example,therewere intense deadlines all the time and I ended up staying awake two days in a row and
then maybe I wouldsleep for 24 hours after that. Or at other times I couldn't go tosleepat all, still so woundup that it wouldtake forever to fall
asleep. (Man D, 26 years)
The participants also reportedthat sleepcouldbe disturbedby the compulsiontoget up at night tocheckfor messages or information (addictive
or dependency factors) and by upsetting messages (destructive quality issues).
The physical impacts ofcomputeruse on sleep includednoise fromthe computer,although most subjects turnedoff the computer at night orslept
in another room. Less avoidable examples of physical impact were symptoms such as muscular pain or headaches relatedtocomputer use, as well
as a general lack of physical activity, all of which couldhave negative effects on sleep. There were also comments about radiation from the
computeras a possible cause of sleep disorders. Some pointedout that it was not thecomputeruse in itself that affectedsleep but task-, work-, or
study-relatedstress, which co-variedwith high quantitative use. Sleep disorders couldalso leadto altered computer use, for example using the
computer at night when unable to sleep. It was pointed out that sleep disorders are probably dependant on individual characteristics.
Subjectiveworry,possibly destructivepersonal use,and other effects of computer use
When askedabout their own worries about personal computer use andexperiences ofproblematic or destructive use, theparticipants mentioned
high quantitative use resultingin wastedtime, increasedpassivity, lost sleep, ergonomicproblems,physical symptoms and health effects, high
costs, addiction(e.g. compulsive chatting, gameplaying, information seekingor gambling), or illegal activities. However, most of those who
gambled considered themselves to be in control of the problem.
Other possible mental effects of computer use mentioned were altered values because of aggressive or role-playing computer games, and
vulnerability to bullyingor toexploitation of their accessible personal information. Positive effects of computer use included stimulation,
usefulness, entertainment, and facilitated communication.
Mobile phone useand stress
The dominant concept in interviews regardingmobile phone use was the perception by participants of demands and expectations that they be
available everywhere andat all times. This couldleadto highquantitativeuse, includinginterruptions of work, sleep, and other activities, the
annoyance of disturbing ring signals, the feeling of never being free, and difficulties separating work from leisure:
Sometimes youwant tobe alone, but youcan't turnit off.I never leave the apartment without the phone. It's stressful. You can't even take a
normal walk to clear your thoughts and relax without the phone and somebody calling. (Woman E, 26 years)
Some experiencedhavingso many phone calls andSMS messages that they couldnot make time to answer themall, which led to stress or even
feelings of guilt. Participants felt that theywere generally expected to explain whenever they were unavailable to answer a call or message.
You're expectedtoanswer. AndI findthat prettystressful. It rings andyoucan't answer, because you don't have time or you are busy. And so
many people expect youtocall back right away. But either youdon't feelup to it or youdon't have time to do it.Youare so mucheasier to reach
when youhave a mobile phone. Peoplealways expect youtoanswer,andthen when you do answer it's just, "What are you up to? Where are
you? Why haven't youansweredthe phone? I'vetriedcallingyoulots oftimes!" And then you're supposed to feel guilty for n ot always being
accessible. I find that pretty stressful. (Woman A, 27 years)

6. Furthermore, the flexibility of mobile phones could imply fragmentation - being overbooked and not being able to postpone. Feelings of
dependency or compulsiveness in relationtothe mobile phone were described, includingthe compulsion tocheckthe displayandfeelings of high
stress when not reachable (forgettingthephone at home, for example, or runningout ofbattery power or callingrange). Physical reactions, such
as headaches or heat sensations afterprolongeduse, were also stressful andledto worries about possible hazards of electromagnetic radiation.
Bad qualitative use (destructive communicationorinformation receivedvia the mobile phone) could also be a stressor. Relationship stress was
also mentioned, in terms ofthe possibilityof one partner keepingsecrets through communications on their private phone, leading to jealousy in
the other.
User competence problems mentionedas stressful includedhandlingall the functions of the phone ortrying to keep up with the latest models.
Other issues were costs, worry about losing the phone, and keeping the battery charged.
All but one of therespondents hadideas about associations between mobile phone use and stress, although several could only identify those
associations concerningotherpersons. Themobile phonewas considerednot only to increase, but also todecrease stress because of theflexibility
it provides. Some insisted that not having access was the main stressor.
Mobile phone useand depression
Respondents were less inclinedto linkmobile phone use to depressionthantostress. A possible pathway todepression was considered to be via
stress andsleep disorders, andit was suggestedthat the demandforconstant availabilitycouldinterfere withrecovery. Unreturned calls or SMS
messages couldlead to feelings of guilt. There was also a notionthat not beingavailable couldleadto being left out. Social isolation was more
evident when the phone didn't ring. It was also suggested that mobile phone use might decrease personal contact "irl" (in real life) and thus
increase social isolation. The qualityof information orcommunicationreceivedvia mobile phone could be destructive; for example, negative
informationreceivedat thewrongtime and place, or harassing phone calls or messages. It was considered easier to send negative messages
(breakingup a relationship,forexample) via SMSthan to deliver themfacetoface,andso there was a perception of a higher risk of receiving
bad informationvia mobile phone thanreceivingit through other channels.Jealousy in relationtoothers' phoneuse was also mentioned, as were
feelings of inadequacy in response to receivingmessages about others' goodfortune. Electromagnetic radiationwas suggested as another link to
depression.
Depression or feelingdown was seen as possibly leadingto alteredor higheruse through reachingout to talktoothers forsupport. In this sense,
the mobile phone was consideredtodecrease depressionbecause of the ease ofreachingsomeone totalkto; also,it was mentionedthat receiving
phone calls "makes you happy".
Mobile phone useand sleep disorders
The feelingthat one was requiredto be constantlyavailable was a central linkbetweenmobile phoneuse andsleep disturbances. A high quantity
of phone calls andmessages in general was perceivedtoleadto difficulties relaxing, andtherefore totoolittle sleepor decreasedquality ofsleep.
A high number of messages that hadnot been repliedtocouldalso leadto overloadandfeelings of guilt. Longphone calls before bedtime could
influence sleep negatively, because of headaches or tension. However, the most obvious aspect connecting mobile phone use with sleep
disturbances was being awakenedat night by phonecalls or messages. This meant wakingup, checkingthe phone, andmost likely answering or
replyingto a message.Evenwith thesoundswitchedoff, one couldbe woken by thesoundof a vibratingphoneor even by a blinkinglight. Some
felt obligedto checkthe phone if theywoke up at night, even if theyhadhadnoindication that this wouldbe necessary. Because a mobile phone
usually has only one user, some felt more obliged to answer:
If it is blinkingandI see it, if I don't havemy backtoit, then I wake up. And, yes,I want tosee who it is [...] But sometimes at night, if I wake up
(andI wake up many times)thenI get up andwalk over to the phonetosee if anybody has calledor sent anSMS[laughs]. (Woman E, 26 years)
The mobile phone is personal. Youknowyou'rereachinga specific person[...] Thenyouare morepressuredtoanswer,because youknowit's for
you, not just your house. (Woman F, 22 years)
The participants thought that people wouldbe more likely to make late-night calls andsendlate-night messages toa mobile phone,in comparison
to a regular phone. Oneopinion expressedwas that leavingthe mobile phoneon is voluntaryandimplies availability: "People think that if you
call a landline there is a certain time, like after ten,youcan't call. Withthe mobile phone theythinkthat if someone answers, then thephone is on
and the person is apparently available." (Woman G, 27 years)
Other factors that were seen as possible contributors tosleep disturbance were communications with disturbingcontent, the t houghts provokedby
a call or SMS, electromagnetic radiation, and worry about such radiation.

7. Subjectiveworry,possibly destructivepersonal use,and other effects of mobilephoneuse
There didnot seem tobe a general worryabout personal mobile phone use amongthe respondents. Some reflecteduponpossible health effects of
electromagnetic radiation.Personal dependencyon the mobile phone was a worry, andsomeparticipants sawtheirown compulsion (forexample
to constantly checkthe display)as alien. High cost was another area of concern. The most commonly mentionedpotentially destructive aspect of
personal use was dependency orcompulsiveness, but economic effects were also a factor. The increasedrisk ofbeingmeanto others with short
messages was mentioned. Some participants considered not being available or reachable to be the biggest problem.
Other effects of mobile phone use that couldhave a bearingonmental health were qualitative changes in communication,for example more, but
less personal,contacts; SMSlanguage style; less respect for others' privacy; spontaneity leading to impulsiveness; fragment ation; and risks of
misunderstandings and negative messages. Positive aspects included utility, flexibility, and easy access.
Summary of mobilephoneuseand mental symptoms
High quantity of use due to demands and expectations for availability at all times was a central area of concern. Demands for availability
originatednot onlyfrom workandthe social network, but also fromthe individual's own ambitions or desires. This resultedin disturbances when
busy or resting, the feelingof never beingfree,anddifficulties separatingwork andprivate life.Unreturnedcalls or messages ledto overloadand
feelings of guilt. Personal dependencywas an area of concern,as was worry about possible hazards associatedwith exposure t oelectromagnetic
fields. User competence issues were mentioned, as well as costs. Badquality of communicationwas anotheraspect mentioned, with the mobile
phone perceivedtoincrease the riskof receivingor sendingnegativemessages. Themajor stressorformany, however, was not being available.
Comparing the results: high computer useversus highmobile phoneuse
There were several similarities in the factors describedas linkingcomputer and mobile phone use, respectively, to mental sy mptoms. Issues
concerninghigh quantity ofuse, demands of availability,dependency, disturbedrecovery, andmental andcommunication overloadseemedto be
commontobothtechnologies. Concerns about thequality ofcommunications were also prevalent in relationtoboth.Work-related demands for
tasks andachievement were more highlyrelatedto computer use, while demands for availabilityregardless of time andspace were more related
to mobile phoneuse. Thoughts about increasedsocial isolationwere more evident in relation to computer use. Worry about electromagnetic
radiation was present in relationtoboth technologies, but more obvious in relationtomobile phones. User problems or competence issues were
prevalent in both areas, although computers seemed to generate more general frustrations.
In our opinion, there were sufficient similarities betweenthe perceivedeffects of the two technologies toallowthe proposal of a combinedmodel
of ICT, encompassing computer use, mobile phone use, and their perceived connections to mental symptoms.
Discussion
The results andthe proposedmodel are basedon the dictums andperceptions of theyoungadult high ICT users interviewed. The significance of
the results andtheplausibilityof themodel shouldbe discussed andverifiedin relationtostate-of-theart knowledge of factors influencingmental
health. The participants as a group perceivedconnections between ICT use andmental symptoms. Some were more prone to identify themselves
within the context of these connections, while others maintainedthat the connections didnot apply to themselves.Some respondents thought that
factors co-varyingwith ICT use, such as work- or study-relateddemands for achievement, orindividual characteristics, were the true causes of
stress or mental symptoms, regardless of ICT use.
The concept of (extrinsic) demands in our model is representedin well-known psychosocial stress models, for example the models of demand-
control [14] andeffort-rewardimbalance (ERI) [15]. Chronic psychosocial stress, as definedin these models, has an establishedassociation with
depression [15]. This also applies to young adults; in a prospective study, previously healthy young adults who were exposed to high
psychological demands at work hada twofoldrisk ofdevelopingmajordepressive disorderandgeneralizedanxiety disorder [16]. In our study,
demands relatedto workor study could result in periodic work or study overload with long hours at the computer. Computer tasks such as
programmingwere consideredtobe especiallydemanding. TheIT work field, withits limitless work hours and tight deadlines, is known to be
stressful [17]. Longworkhours in general have a broadimpact,leadingto less time for sleep andrecovery,longerexposuretoworkplace hazards
anddemands, andless time to attend to private life and family [18]. Other causes of high ICT use in the present study included perceived
demands for availability andsocial interactions.Constant availability viaICT couldbe a direct link tostress, implyingdifficulties separating the
domains of work andpersonal life andpossiblyleadingtorole conflicts andwork/non-workinterference. The boundless work situation offered
by the use of ICT puts high demands on theindividual's own capacity to set limits for work. This could be especially difficult for people with
high intrinsic demands in terms of achievement orperformance. The concept ofdemands originatingfrom theindividual's own aspirations in our
model is comparable toindividual characteristics such as overcommitment in the ERIstress model [15] andperformance-basedself-esteem [19],
the latter of which has been suggested as a risk factor for burnout.

8. The effects ofICT oncommunication andsocial relations are vast. ICT not only provides the means to develop, expand, andmaintainlarge social
networks via email, SMS, web forums, chat rooms, and so on, it also makes it possible to engage in several lines of communications
simultaneously.Chattingor instant messagingseemedto be a central andtimeconsumingactivityfor manyof the computer users in our study.
Social support is a factorthat promotes healthandbuffers the negativeeffects of high psychosocialdemands [20,21], and could be considered a
positive aspect of ICT use. However, ICT also provides themeans of communicationoverload. Distractions and dual-tasking are demanding on
workingmemory [22,23]; similarly, the participants in our study saidthat expectedimmediacyof communications, as evidenced by reminders
anddemands for explanations ofanyunavailability,addedto mental overloadandfeelings of guilt. There were also misgivings about the quality
of communications andinformationvia ICT,as it was thought to lead to misunderstandings, increased risk of sending or receiving negative
messages, andfeelings of vulnerability. Studies have shown prospective associations between chattingor instant messagingandperceived stress,
depression, andcompulsiveinternet use [6,24]. Furthermore, mobile phone use has been found to distract attention and affect driving and
pedestrian safety[25]. Frequently-ringingmobile phones have been shown to enhance allergic responses in patients with atopic eczema or
dermatitis syndrome[26], implyingthat the exposurerelatedtoavailability is stressful. Another view of this sees social isolation as a possible
consequence of high computer use. Some of theparticipants in our study who spent long hours at the computer said that they had difficulties
developingandmaintainingsocial relations outside of the internet. Theylongedformore real-lifefriends, but felt more secure at the computer.
Whether loneliness is an effect ofhigh computer use or computer use an effect of loneliness is an interesting topic for discussion. Morahan-
Martin et al [27]foundsupport for thepositionthat loneliness leads to increasedinternet use; lonelyindividuals were morelikelythantheir non-
lonely counterparts touse the internet for social interaction and emotional support. Lonely individuals also reported that t heir internet use
interferedwith real-life social activities, which couldimply a negative loop.Caplan [28] concluded that social anxiety rather than loneliness
explains thepreference foronlinesocial interaction. For socially anxious people, online communication could have a number of benefits in
comparison withface-to-face communication, includinghavingcontrol of self-presentation, phrasing, and the speed of the interaction, and
thereby feelingsafer and moreconfident thanduringinteractions in person. In a recent prospective study among adolescents, loneliness was
negativelyrelatedtoinstant messagingsix months later [24], indicatingthat those with a high level of loneliness were not moredrawn to this type
of communication. It seems that the connectionbetween ICT use andthe social network is complex; however, the impacts, both positive and
negative, should be of interest in our model.
Constructs such as internet addiction, problematicinternet use, andcompulsive internet use have been previously discussed[29],as has the term
"problem mobile phone use" [30].Internet addiction has evenbeen proposedas a specific psychiatric illness [31],though anotherviewholds that
problematic internet use shares elements with impulse control disorders and is related to specific activities like gambling or accessing
pornography,not tothe internet per se [10,29]. Inour study, some respondents claimed a compulsion towards the mobile phone or computer,
feelingan urgent needtocheck for messages or other information. Game playingwas a central activity, andsomeparticipants who spent a lot of
time playing computer games (online or offline) admitted to neglecting other activities such as social life, physical activit y, or sleep.
Another question raisedin our study was that ofpossible negative effects,such as increasedaggressionor altered perceptions of reality, due to
violent content in games. Grusser et al [32]foundonly weak evidence for the assumption that aggression is related to excessive gaming, but
concludedthat there is addictive potential in gaming. An evaluationof the short-term physiological andpsychological effects of playing violent
videogames, comparedtonon-violent games, revealedeffects onarterial pressure andstateof anxiety, but not on measures of hostility [33]. A
study amongmental healthprofessionals in the USA revealedthat when clients presented problematic internet use as a primary problem, this
most likely includeduse of pornography [34]. Another study [35] founderotica tobe thebest prospective predictorof compulsive internet use. In
contrast,use of pornographywas barely mentionedin our study. We didnot ask any direct questions aimedat the subject, but the interviewer did
ask questions about possibly destructivepersonal uses. Hence,eitherthe use of pornography was not perceivedby the participants as a problem,
or it was a subject they did not wish to discuss.
Online gamblingis anotherpossible hazard, though most of those in our study who gambled considered themselves to be in control of the
problem. Increasedonline gamblingaddictionhas been notedin Sweden. It is likely that easy access contributes to the increase; gambling can
take place in the privacyof the home, just "one click away". Anothergrowingproblemin Sweden is youngpeople indebt because of applyingfor
andacceptingbankloans or credit via SMS, or because of high phonebills. Dependency, compulsion, or addiction, whether it concerns chatting,
game playing, gambling, or other aspects of ICT use should be of relevance in our model. Addiction could be a link to mental ill-health,
regardless of ICT use. It couldalso be arguedthat ICT in itself is potentially addictive. Some participants in the study claimed that they felt
compulsive regarding their mobile phone, but not in any other areas of life.
User problems, includingtechnical frustrations andcompetenceproblems,are associated with stress [36] and were a cause of irritation in our
study, addingto workloadandcausingmoretime tobe spent at the computer thanplanned. Users' great dependency on ICT for work, leisure,
information, andcommunications makes user problems especiallydisturbing. Badergonomics andphysical symptoms associated with ICT use
were considered risk factors for mental health issues in this study. Computer use and mobile phone use have been associated with
musculoskeletal complaints [37,38], whichin turncan have a negative effect onmental health.Furthermore, the sedentarynatureof computer use
andthe neglect ofphysical activitywere commoncomplaints in our study.A high level ofsedentary leisure activities negatively affects mental
health [39],while physical activity has positive effects on mental health and is acknowledged as a possible complementary treatment for
depression andstress-relateddisorders [39,40]. The impact ofICT use on physical symptoms or physical activity should be of interest in our
model. Worryingabout possible exposure to electromagnetic radiation as a consequence of high ICT use was another source of stress in our
study. Some participants evenproposedradiationas a possible cause of mental ill-health. Exposure to electromagnetic fields due to ICT use is

9. currently not known to haveanymajor health effects [41], though there is some inconsistency in the published results [42]. A study of skin
complaints associatedwith computer work suggestedthat there were factors other than electromagnetic radiationcontribut ing to the symptoms
[7]. It was hypothesizedthat occupational strain resultedin psychophysiological stress reactions (termed techno-stress) that were conditioned to
the computer work environment, and that this was why symptoms appeared or worsened in connection with the computer. Perceived
electrosensitivityis associatedwith reportingsymptoms of depressionandworse general healththancontrols [43].Worrying about exposure to
electromagnetic radiation is probably of greater significance for mental health than actual exposure, and could be of interest in our model.
Prevention of mental disorders such as depression is of course of great importance. Psychosocial function in young adults is affected by
depression [44], andcognitive impairments are common[45]. Gender, socio-demographicfactors, general health, and major life events, as well
as individual factors such as copingskills, are all relatedto theincidence ofdepression [46-48]amongyoungpeople. Whether computer use or
internet use can leadtodepression has been the subject of debate; studies have shown mixedresults [8,49-51]. Bell [9]points out that the internet
is just a medium, andcannot in itselfbe consideredgoodor badfor mental health. Mentalhealthinformationcan be foundon the internet, as well
as online support groups, online therapy, andmore.For a reviewof internet-basedmental healthinterventions, see Ybarra andEaton[52].On the
other hand, the internet also provides informationandadvice onsuicide methods, enables interaction with other people contemplating suicide,
andcan provide the means to suicide via online shopping[53]. The internet can also be a medium for bullying via email and web pages, as can
mobile phones via text messages andthe misuse of picture-takingabilities [54]. Internet or mobile phone bullyingcan be done anonymously and
can affect victims in the privacyof their own homes. Almost onefifthof Swedish adolescents (12-16 years) hadexperiencedbeingbulliedvia the
internet [55]. Hence,negative content is another majorfactor in ICT use, with potential effects on depression and other mental symptoms.
Too littlesleep, dislocatedsleep, ordifficulties fallingasleepbecause of sittingup late at the computer were prevalent in our study, along with
sleep disturbances caused by late night phone calls or SMS messages. Intensive ICT use which negatively affects sleeping habits has been
associatedwith poor perceivedhealth amongadolescents [56].In a study of Korean universitystudents,30% reported insufficient sleep and of
these about a thirdpointedtovisual media includingcomputers as theprimaryreason[57]. Otherfactors associatedwith insomnia anddepression
amongyoungadults are negative family lifestress andacademic stress [58].The importanceof sleepandrecovery for physical as well as mental
health is well known. Insomnia affects psychosocial functioning and increases subsequent risks for somatic health problems, interpersonal
problems, andpsychological problems in adolescents [59]. Insufficient sleep has been shown to be a risk factor for reporting poor self-rated
health amongyoungadults [60].Furthermore, sleep disorders are a predictor ofmajor depressiononset, with anodds ratioof nearly4, according
to a prospective study of young adults [61]. The negative impact of ICT use on sleep should be of major concern in our model.
Limitations of thestudy
This interviewstudy was performedwith a highlyselectedstudy groupso as to enhance the potential toidentifyfactors or conditions that could
connect ICT use with mental symptoms. There is a possibilitythat the current mental states ofthe subjects affectedtheirperceptions. We had an
even gender distribution in thegroup as a whole, but thehigh mobile phone users includedmore women than men.We have not pursuedtheissue
of gender differences,though they certainly could exist. There are gender differences in ICT usage [62], as well as in prevalence of mental
symptoms [1]. The study groupwas also highly selectedin that it consistedof students from highachievingacademic programs. Neither socio-
economicfactors noracademic backgroundwere takeninto account in the present study. Another limitation is that the participants sometimes
spoke in moregeneral terms, rather than through personal experience, which means that some of the factors that emerged could be mere
speculation.On the otherhand, we were primarilyinterestedin concepts and ideas. We have not challenged the participants' thoughts about
stress, depression, and sleep disturbances.
Implications for futurestudies
Most of the factors andconditions in the model can be argued to influence mental symptoms. Some factors are probably more central; for
example, work load, psychosocial demands, addiction, impacts on sleep, and social support, as well as individual characteristics or
vulnerabilities. Different exposure profiles might co-occur withdifferent individual characteristics and lifestyle factors, leading to a variety of
possible pathways. One personcould be active and socially outgoing, carry a high workload and other demands, use computers and mobile
phones heavilyin bothwork andsocial life, andleada lifestyle withinsufficient recovery.Another personmight spendmuch ofthe day andnight
at the computer, playinginteractivegames online,carryingon a social life onthe internet but having few friends "irl", experiencing dislocated
sleep, andleadinga sedentary lifestyle.Methods of testingthe model could include combiningdifferent ICT exposurefactors (high/low), adding
psychosocial demands, dependencyandlifestyle factors (sleep, physical activity,social support), andusingmental symptoms as outcomes in an
epidemiological,preferably longitudinal, study.It seems likelythat tailoringis necessary whendesigninginterventions andintervention studies.
Conclusions
The concepts andideas expressedby the youngadults reportinghigh ICT use and mental symptoms generated a model of possible paths for
associations between ICT exposure andmental symptoms. Demands for achievement andavailabilityas well as personal dependencywere major

10. causes of high ICT exposure, but were also direct sources of stress andmental symptoms. The proposed model shows that factors in different
domains may have an impact and should be considered in epidemiological and intervention studies.
http://www.biomedcentral.com/1471-2458/10/66
Mobile phone use and stress, sleep disturbances, and symptoms of depression among young adults - a prospective cohort study
abstract
Background
Because of the quick development andwidespreaduse of mobile phones, andtheir vast effect oncommunicationandinteractions,it is important
to study possible negative healtheffects of mobile phone exposure. The overall aim of this study was to investigate whether there areassociations
between psychosocial aspects of mobile phone use and mental health symptoms in a prospective coho rt of young adults.
Methods
The study group consistedof youngadults 20-24years old(n = 4156), who respondedtoa questionnaire at baseline and1-year follow-up.Mobile
phone exposure variables includedfrequency of use, but also more qualitative variables: demands on availability, perceived stressfulness of
accessibility, beingawakenedat night by the mobile phone, andpersonal overuse of the mobile phone. Mental healthoutcomes included current
stress, sleep disorders, andsymptoms of depression. Prevalence ratios (PRs) were calculated for cross-sectional and prospective associations
between exposure variables and mental health outcomes for men and women separately.
Results
There were cross-sectional associations between high compared to low mobile phone use and stress, sleep disturbances, and symptoms of
depression for themen andwomen.Whenexcludingrespondents reporting mental health symptoms at baseline, high mobile phone use was
associatedwith sleep disturbances andsymptoms ofdepression for themenandsymptoms of depressionforthe womenat 1-year follow-up. All
qualitative variables hadcross-sectional associations with mental health outcomes. In prospectiveanalysis, overuse was associatedwith stress and
sleep disturbances for women, and high accessibility stress was associatedwith stress, sleep disturbances, andsymptoms of depression for both
men and women.
Conclusions
High frequency of mobile phoneuse at baseline was a risk factorformental health outcomes at 1-year follow-up among the young adults. The
risk for reportingmental healthsymptoms at follow-up was greatest among those who had perceived accessibility via mobile phones to be
stressful. Public healthpreventionstrategies focusingon attitudes couldinclude information and advice, helping young adults to set limits for
their own and others' accessibility.
Background
Mental healthproblems have been increasingamongyoungpeople in Sweden andaroundthe world[1,2]. Cultural andsocial changes in terms of
increasedmaterialismandindividualism have been discussed in relation to this [3,4], including the possibility of a decreasing stigma about
mental illness, improved screening for mental illness, and increased help-seeking behaviors [5]. Because of the quick development and
widespreaduse of mobile phones, andtheirvast effect on communication and interactions in work and private life, it is impo rtant to study
possible negative health effects of the exposure. Extensive focus has been onexposure to electromagneticfields (EMF). Self-reportedsymptoms
associatedwith usingmobile phones most commonlyinclude headaches,earache,andwarmth sensations [6,7], and sometimes also perceived
concentrationdifficulties andfatigue [6]. However, EMF exposuredue to mobile phone use is not currently known to have any major health

11. effects [8].Another aspect of exposure is ergonomics. Musculoskeletal symptoms due to intensivetextingon a mobile phonehave been reported
[9], andtechniques usedfor text entering have been studied in connection with developing musculoskeletal symptoms [10]. However, our
perspective is predominantly psychosocial.
In a previous study we foundprospective associations between high information and communications technology (ICT) use, including high
frequency of mobile phoneuse, andreportedmental health symptoms amongyoungadult college anduniversitystudents [11], but concludedthat
the causal mechanisms areunclear.Thestudy was followedby a qualitative interviewstudy with 32subjects with high computer ormobile phone
use, who hadreportedmental health symptoms at 1-year follow-up. Based on the young adults' own perceptions and ideas of associations, a
model of possible paths forassociations between ICT use andmental healthsymptoms was proposed[12], with pathways to stress, depression,
andsleep disorders via the consequences of high quantitative ICT use, negative qualityof use, anduser problems. Central factors appearing to
explain high quantitative use were personal dependency, anddemands forachievement andavailability originatingfrom domains of work, study,
the social network, and the individual's own aspirations. These factors were also perceived as direct sources of stress and mental health
symptoms. Consequences of highquantitative mobile phone exposure includedmental overload, disturbedsleep,the feelingof never being free,
role conflicts, andfeelings of guilt due to inabilitytoreturnall calls andmessages. Furthermore,addiction ordependencywas an area of concern,
as was worry about possible hazards associatedwith exposure toelectromagnetic fields. Forseveral participants in the study, however, a major
stressor was to not be available. The study concludedthat there are manyfactors in different domains that shouldbe taken into consideration in
epidemiological studies concerning associations between ICT use and mental health symptoms [12].
Based on the previous studies, we wantedto focus on someaspects ofmobile phone exposure other than mere quantity of use. For example,
demands on beingavailable or reachable, regardless of time andspace, couldbe arguedto be a stressor irrespective of actual frequency of use.
Another key determinant may be the extent to which a person actually perceives his or her own accessibility as stressful. Fur thermore,
accessibility implies the possibility to be disturbedat all hours, even at nighttime. Having one's sleep interrupted repeatedly can have direct
effects on recovery andhealth.In a study amongFinnishadolescents, intensive mobile phone use was linked to poor perceived health among
girls, both directly andthrough poor sleepandwaking-time tiredness [13]. Another area of concern could be addiction to the mobile phone.
Intensive mobile phone use has been associatedwith dependency onthe mobile phone [14,15], and problematic mobile phone use has been a
focus in the literature concerningpsychological aspects of mobile phone use, where criteria for substance addiction diagnoses or behavioral
addictions [16,17] have been used to define problematic use [18-24] including compulsive short messaging service (SMS) use [20]. In this
context, heavyor problemmobile phone use (overuse) has been associated with somatic complaints, anxiety, and insomnia [21], depression
[21,24], psychological distress [22], andan unhealthylifestyle [25].However, possible positive effects ofmobile phone use on mentalhealth can
also be hypothesized, for instancethe ease of reachingsomeone to talktowhen in need, implyingaccess to social support.Social support buffers
negative effects of stress [26], while low social support is a risk factor associated with mental health symptoms [27].
We have previouslystudiedICT use in relation to mentalhealthsymptoms amonghighlyselectedstudy groups (college and university students
studyingmedicine andinformation technology)[11,12]. Most investigations we have foundon mobile phone use and mental health outcomes
have been cross-sectional studies performedamongmainlycollege students (e.g., [15,19-23]). It is important to examine possible associations
between mobile phone use andmental healthoutcomes also in a moregeneralor heterogeneous population ofyoung adults, using a longitudinal
design.
Aims
The overall aim ofthis study was to investigate whetherthere are associations between psychosocial aspects of mobile phone use and mental
health symptoms in a prospective cohort ofyoungadults. Specific aims were toexamine whether the frequency of mobile phone use, but also
more qualitative aspects of mobile phone use (demands on availability,perceivedstressfulness of accessibility, being awakened at night by the
mobile phone, andperceivedpersonal overuse of the mobile phone), are associated with reported stress, symptoms of depressio n, and sleep
disturbances. Furthermore, we wanted to examine whether frequency of mobile phone use is associated with perceived social support.
Methods
Study populationand data collection
The study population consistedof a cohort of youngadults (Figure 1),20-24years old(correspondingto theUnitedNations' definition of young
adults [28]). Tenthousandmen and10 000 women, born between 1983and1987, were randomly selectedfromthe general population (from a
registry heldby the Swedish Tax Agency), 50% livingin theCountyof Västra Götaland, Sweden, and50% in the rest of thecountry.In October
2007, a questionnaire containingquestions about health, work,and leisure-relatedexposurefactors, backgroundfactors, andpsychosocial factors
was sent by post tothe selectedpopulation[29].Besides returningthe postal questionnaire it was also possible torespondto the questionnaire via
the webif desired. A lottery ticket (value approx.1 Euro)was attachedto the cover letterandcouldbe used whether the recipient participated in
the study or not.Two reminders were sent by post. The response rate at baseline was 36% (n = 7125). One year later,those respondents who had
indicatedthat they wouldaccept tobe offeredtoparticipate in further studies (n = 5734) were invitedtorespondtoan identical questionnaire, this
time administeredvia the web. The data collection process was otherwise similar to that at baseline, but with the addition of a third reminder

12. offeringa paperversionof thequestionnaire andtwo cinema tickets torespondents. The response rate at follow-up was 73% (n = 4163). After
excludingthose who failedtorespondto bothquestions concerningfrequency ofmobile phone and SMS use at baseline, 4156 remained in the
study group. All in all, non-participation and dropout from the study was 79% (Figure 1).
Figure 1. Participation process. The participation process from study population to study group
Mobile phone exposurevariables
Informationabout mobile phoneexposure was collectedfromthe baseline questionnaire. This includedtheaverage numberof mobile phone calls
made andreceived, andof SMSmessages sent andreceived, per day, but also more qualitative aspects ofmobile phoneuse, includinghow often
the respondent was awakenedat night by the mobile phone, howhe or she perceiveddemands on availability, and whether he or she perceived
the accessibilityvia mobile phones to be stressful, as well as perceptions regarding personal overuse of the mobile phone. Responses were
divided into high, medium, andlow categories, based on the frequency distribution of responses, except for overuse which was categorized
accordingto numberof items confirmed. A combinedquantitative mobile phone use variable was constructedby mergingthevariables frequency
of calls andfrequency of SMSmessages (Spearmancorrelationr = 0.35, p < 0.0001). The mobile phone use variable correlated well with the
original calls and SMS variables (r = 0.73, p <.0001, and r = 0.84, p <. 0001, respectively).
Mobile phone variables, questionnaire items, response categories, and response classifications are presented in Table 1.
Table 1. Mobile phone variables at baseline
Mental health outcomevariables
Information about mental health symptoms was collected from the cohort study questionnaire at baseline and at f ollow-up.
The outcome variable Current stress was constitutedby a validatedsingle-itemstress-indicator[30]: Stress means a situationwhena person feels
tense, restless, nervous, or anxious or is unable to sleepat night because his/her mindis troubledall the time.Are youcurrently experiencingthis
kind of stress? Response categories were: a = not at all, b= just a little, c = to some extent, d= rather much, e = very much. The responses were
divided into Yes (responses d-e) andNo (responses a-c),basedon frequency distribution while yet taking content of response categories into
account.
The Sleepdisturbances variable was constructedby includingthe most common sleep disorders (insomnia, fragmented sleep and premature
awakening) intoa single-item, adaptedfromthe The Karolinska Sleep Questionnaire[31]: How often have you had problems with your sleep
these past 30 days (e.g.,difficulties falling asleep, repeatedawakenings, wakingup tooearly)? Response categories were: a = never, b = a few
times per month,c = several times per week, andd= every day. The responses were divided into Yes (responses c-d) and No (responses a-b),
based on clinical significance.
Symptoms of depression(one item)andsymptoms of depression (two items) were made up by the two depressive items from the Prime-MD
screeningform[32]: Duringthe past month,have youoftenbeenbothered by: (a) little interest or pleasure in doing things? (b) feeling down,
depressed, or hopeless? Response categories were Yes andNo. It is proposedthat it is sufficient if one of the two items is confirmedin screening
to go forwardwith clinical assessment ofmooddisorder. This procedure has high sensitivity for major depression diagnosis in primary care
populations [32,33].In our cohort study group, approximately50% of themen andalmost 65% ofthe womenconfirmed at least one of the two
depressive items, which indicates that the instrument is probably very sensitive but has low specificity in our study group. Therefore, we
constructedtwo outcomes: Symptoms of depression (one item), in which the Yes category contained those who confirmed only one of the
depressive items,and Symptoms of depression(two items),in which the Yes category containedthose who confirmedboth depressive items. The
No category in both outcomes contained those who disclaimed the two depressive items.
Analysis
All analyses were performedusingthestatistical software package SAS, version 9.2 (SAS Institute, Cary, NC, USA). Spearman correlation
analysis was used to examine associations betweenthe mobile phone exposure variables, andbetweenmobile phoneuse and social support. The
Cox proportional hazardmodel (PHREGprocwith time set to1) was usedto calculate prevalence ratios (PRs) with a 95% confidence interval
(CI) for multivariate analysis of cross-sectional andprospective associations betweenexposurevariables andmental healthoutcomes.The robust
variance option(COVS) was used in the cross-sectional analysis to produce adequate CIs [35,36]. The low category in each exposure variable
was used as reference level. The PRs were adjustedfor backgroundfactors including relationship status, educational level, and occupation at
baseline. Missingvalues (non-responses toitems)were excludedfromthe analyses,which means that the n varied in the analyses. Prevalence
ratios with a CI not including1.00 (before round-off)were consideredstatisticallysignificant.In the prospectiveanalysis, subjects who reported

13. symptoms at baseline were excludedfromthe analysis of the mental healthoutcome variable concerned. All analyses were done separat ely for
the men and women.
The study was approved by the Regional Ethics Review Board in Gothenburg, Sweden (Reg. no. 191-05).
Results
Study group characteristics
Fifty-two percent of the men and34% of thewomenwere single at baseline. A majority of the respondents had completed upper secondary
school, 13% ofthe menand16% of the women had finished college or university studies, while 5% of the men and 6% of the women only
elementaryschool. Fifty-one percent of themenand41% ofthe womenreportedwork as mainoccupation, while 40% of themenand48% ofthe
women studied, and8% of the menand12% womenwere categorizedas other. Forty-three percent ofthe menand56% of the women reported
high social support, 41% of themenand 32% of the women reported medium social support, and 16% of the men and 13% of the wo men
reported low social support.
A little more than half ofthe participants were categorizedas having low mobile phone use (fiveor fewer calls and five or fewer SMS messages
per day) and22% of the menand24% of thewomen as having highuse (elevenor more calls or SMS messages per day) (Table 1). A massive
majorityreportedthat they were expected to be available on a daily basis and one out of four around the clock. Only a few p ercent found
accessibility via mobile phones verystressful, while about halfof theparticipants didnot findit stressful at all. Most participants were never, or
only on rare occasions,woken up by the mobile phone, andonlya fewreportedbeingwoken by the mobile phone on a weekly basis. Thirteen
percent of the men and 22% of thewomenindicatedthat theythemselves, orsomeone close tothem, thought that theyusedthe mobile phone too
much, and 6 and 14%, respectively, had tried, but failed, to cut down on mobile phone use (Table 1).
The women reportedstress almost twice as oftenas the men (29%comparedto16%) at baseline. Twenty-three percent of the men and 34% of
the women indicatedsleepdisturbances. Of the men, 27% reportedone and24% two symptoms of depression, andof thewomen, 30% reported
one and34% two symptoms of depression. Among participants who were symptom-free at baseline (in outcome variable concerned), the
prevalence at 1-year follow-up was as follows for the men andwomen,respectively; current stress: 10%and19%,sleep disturbances: 15% and
20%, symptoms of depression (one item): 24% and 28%, and symptoms of depression (two items): 12% and 18%.
Drop-out analysis
The drop-out group at the initial cohort baseline consistedof moremen(a difference of 17percentage points), were somewhat younger (an age
difference of <0.1 years),more often married (a difference of 1.4 percentage points), and more often foreign -born (8 percentage points),
comparedtothe study populationinvited toparticipate [29]. The final study group (n= 4156) consistedof almost twice as many women as men
(65% vs. 35%).Comparedtothe initial cohort baseline(n = 7125), the study group participants were slightlyless often single (40% compared to
42%), hada slightly higher educational level (with15% comparedto14% havingcollege oruniversity level education, and5% compared to 7%
havingcompletedonly elementaryschool), andwere less oftenworking(44%comparedto 48%) and more often studying (45% compared to
41%) at baseline. The level of mobile phone use was slightly lower in the study group; 54% were categorized as low mobile pho ne users
compared to 51% in the initial cohort baseline, while 23% compared to 26% were categorized as frequent (high) mobile phone users.
Discussion
Frequent mobile phone use was associatedwith current stress, sleep disturbances, andsymptoms of depressionamong the young adult men and
women in cross-sectionalanalysis. Prospective analysis indicatedthat high frequencyof mobile phone use couldbe a risk factor (or marker) for
developingsleep disturbances in the men,andsymptoms of depressionin both the men and women, at 1-year follow-up. The pattern of PRs
larger than 1.0was rather consistent (though not all statisticallysignificant), suggestinga robustness of results, andthere was even an indication
towards a dose-response relationshipbetween exposure andmental healthoutcomes (if lookingonly at PRs). It should be noted that the "high"
category ofmobile phoneuse in our study does not reflect an extreme part of the population, since almost 25% of the study group belonged to
this category.The use of the Coxregressionprocedure forestimating PRs gives wider than adequate CIs [35], which was corrected for in the
cross-sectional analysis by addingthe robust varianceoption. However,in the prospective analysis the CIs are still conservative. The results are
further supportedby the findingof prospective associations betweenhigh frequency of mobile phone use and mental health out comes in our
previous study among young adult university students [11].
The majorityof the youngadults reportedthat theywere expectedtobe reachable via the mobile phone all day or around the clock. One could
expect that this wouldfeel compellingandperhaps evenstressful, but most respondents didnot consider theaccessibilitytobe stressful,andthere
was no associationbetweenthe two variables.Yet,expectedavailabilityaroundtheclockwas associatedwith most mental health outcomes in
cross-sectional analysis (noclear prospective associations). Therisk forreportingmental healthsymptoms at follow-up was greatest amongthose

14. respondents who hadindicatedthat they perceivedthe accessibility to be rather or very stressful, and in cross-sectional analysis, it was even
sufficient toconsider the accessibilitytobe just a little stressful for higher prevalenceof mental healthoutcomes. The over-all low associations
between the mobile phonevariables suggest that availability demands andaccessibility stress not necessarily coincide with actual frequency of
use.
Reports in the media claimnightlydisturbances by mobile phonecalls or messages tobe a menacefor today's adolescents. This may be the case
amongyounger persons, but was not as obvious in our group of young adults, with only few being woken up regularly. However, there were
cross-sectional associations between being awakened a few times or more during the past month and all mental health outcomes (no clear
prospective effect).
It has been suggestedthat mobile phone use enhances social support [12,37], but, in our study,high frequencyof use hadlittle or no association
with perceived access to social support in private life.
Quite a fewparticipants reportedsubjectiveoveruse which couldindicate possible addictionto the mobile phone orits functions. Addictions can
consist of excessive behaviors ofall types, and some factors can be argued to be present in all types of addictions (e.g., salience, tolerance,
withdrawal, conflict, andrelapse)[17].Themost commonsymptomof problemmobile phone use amongadolescents in a study by Yen et al [24]
was "withdrawal symptoms without cellular phone use". Furthermore, impulsivity, especially urgency, has been related to mobile phone
dependency, andfeelingcompelledtoprovide forneeds as soon as possible has been suggested to increase the likelihood of using the mobile
phone in a destructive way, forexample when prohibited[15].There is also the risk for addiction through gambling on mobile phones [23],
which could be detrimental since the mobile phone enables gambling without time or space restrictions.
Methodological considerations
We knowlittleabout what time spanmaybe relevant when assessingpossible effects ofthe exposure on mental health, andwhether concurrent,
short-term, or long-termexposure andeffects areof interest. Wehave data frombaseline andfollow-up after 1 year, making it possible only to
performeithercross-sectional analysis (so that causal inferences cannot be made) orprospectiveanalysis with a 1-yearlatency period that could
be consideredrather long. The exposure duringthe latencyperiodis not known,andthe same applies to themental health outcomes, concerning
symptoms that arecommonin the populationandthat couldappearanddisappearin the latencyperiod. Consequently, it is difficult to draw clear
inferences about the effect of the exposure on the outcomes within the study design.
Usinga questionnaire tocollect informationon exposure as well as health aspects poses several limitations.It is important toemphasize that the
study concerns subjective symptom-reports andnot actual mentaldisorders or diagnoses. The prevalenceof reported depressive symptoms was
alarmingly high in our study group. The suggestedprocedure that it is sufficient if oneof the two PRIME-MD depressive items is confirmed in
screeningfor depression[32,33]proposes that about 20% of the study group would be clinically depressed (positive predictive value of 33%
[33]). Theprevalence of depressionis most likely lower in our population than in primary care populations as, for example, the 1-month
prevalence of depressionamongFinnishyoungadults (20-24years of age) was 9.6% [38]. Hence, the instrument seems too sensitive for our
population,andwe chose toanalyze one-itemandtwo-item responses as separate outcomes, with theexpectationthat the two-item outcome has
higher specificity than the suggested procedure.
Recall bias andrecall difficulties are most certainly present in the study,with, for example, difficulties tocorrectly specifythe average number of
calls andmessages sent andreceivedper day overthe past month. Furthermore,when mergingcalls andSMSmessages into one variable (mobile
phone use) we lose information about specific exposure. Also, while the high and low categories are distinct from each other, the medium
category overlaps to some extent with the highandlow categories, which means that,in someinstances, individuals in the medium category may
in fact have hada higherexposure(numberof calls andSMSmessages) than some individuals in the highcategory,or lower thansome in the low
category. There is a risk that misclassifications obscure results.
We have limitedour study topsychosocial aspects of mobile phone use. Possible biophysical pathways due to exposure toelectromagnetic fields
have not been considered. Furthermore, there might be factors, e.g. individual factors orpersonality traits, not accounted for in our study, which
co-varies with exposure variables andare "true" pathways to mental healthproblems. This couldparticularlybe the case concerning accessibility
stress which hadno association with availabilitydemands andlowassociationwith actual frequencyof use, but yet seemedto be thegreatest risk
factor among the mobile phone variables for developing mental health symptoms.
The study sufferedfroma high drop-out rate,which is fairly common whenperformingstudies via questionnaires in the general population. The
youngadult populationis probablyespecially difficult to recruit because more often than in another age group, their life situation undergoes
drastic changes, includingmovingmore often andtherefore beingmore difficult toreach.The drop-out analysis shows that especiallywomenand
native-bornSwedes are overrepresentedin the data. Earlierstudies, e.g. [13,14,21], have indicated gender differences in mobile phone usage,
therefore gender-specific analyses were performed. However,the results of the analyses were strikingly similar formenandwomenin the present
study. Thereis probably a healthyparticipant selectionbias, andthere is also an indicationof bias towards lower mobile phone exposure, which

15. couldaffect results in cross-sectional analyses but shouldhaveless influencein the prospective analyses. Even though the study group is more
representative in comparisontostudies amongonlycollege anduniversity students, caution must be used when generalizing the results to the
general population of young adults.
Implications
The place ofmobile phones as a technology distinct fromlandline phones onthe one hand, and from computers on the other, is declining, as
mobile phones increasinglyare takingthe placeof stationary phones andat the sametime are approaching computers in function. Therefore,
definingthe exposurebecomes difficult as technology andpossible uses are developingandchangingrather swiftly. The use of mobile phones
puts high demands on the individual's own capacity to set limits for use and accessibility. Norms on how to use mobile phones are set in
interaction withothers. Ifa youngpersonthinks that "all others" areavailable at all times, he/she might feel stress if not available. Attitudes are
probably an important factortofocus in preventionstrategies. This couldinclude informationtochildren,adolescents, and youngadults about the
importance of sleepandrecovery, andthe advice toset limits for accessibility(i.e., turnoffthe phone)at certaintimes such as at nighttime,when
needingto focus or rest, orwhen others need to focus or rest. Furthermore, shifts in attitude could also include limiting your demands and
expectations on others' availability, i.e., not expectingothers tobe available at all times. In our study, a clear risk factor for reporting mental
health symptoms was to perceive theaccessibility offered by mobile phones as stressful. Thus, actually perceiving something as a "problem"
couldindicate a more general problem, andcouldserve as a warningsignal for takingmeasures to preclude constant accessibility and overuse.
Conclusions
There were cross-sectional andprospectiveassociations between mobile phone variables andmental health outcomes among the young adults.
High frequency of mobile phoneuse at baseline was a risk factorforreportingsleep disturbances andsymptoms of depression for the men and
symptoms ofdepression for the women at 1-year follow-up.The riskforreporting mental health symptoms at follow-up was greatest among
those who hadreportedthat they perceivedtheaccessibilityvia mobile phones to be stressful. P ublic health prevention strategies focusing on
attitudes couldinclude information and advice, helping young adults to set limits for their own and others' accessibility by mobile phone.