L Biddle, D Gunnell, D Shrp nd J L Donovn Fctors influencing help seeking in mentlly distressed young dults: cross-sectionl survey Lucy Biddle, Dvid Gunnell, Debbie Shrp nd Jenny L Donovn SUMMARY Bckground: Young dults, especilly men, re mong those lest likely to consult helthcre professionls when mentlly distressed or suicidl. Aims: To investigte the help-seeking behviours of mentlly distressed young dults. Design of study: Cross-sectionl survey. Setting: Bristol nd surrounding res, including inner-city, suburbn nd urbn loctions. Method: A questionnire ws sent to smple of 3004 young dults ged 16 24 yers. This ssessed probble mentl disorder (using the 12-item generl helth questionnire [GHQ-12]), suicidl thoughts (GHQ-28 suicide subscle), nd help-seeking behviours. Results: Most responders who were ssessed s hving probble mentl disorders (GHQ cses ) hd not sought help. Help seeking ws more common in femle GHQ cses thn mle cses (34.8% nd 21.8%, respectively; P = 0.003) nd women with suicidl thoughts more commonly sought help thn men with suicidl thoughts (41.6% nd 30.9%, respectively; P = 0.15). Smll proportions of mle nd femle GHQ cses (7.5% nd 8.9%, respectively; P = 0.6), nd less thn one in five responders with suicidl thoughts, hd consulted generl prctitioner. In more femle thn mle cses, help ws sought from fmily nd friends (30.7% nd 18.4%, respectively; P = 0.004). GHQ score ws the strongest predictor of help seeking. Men hd higher threshold of severity t which they would seek help thn women. Recent experience of suicidl thoughts ppered to be stronger predictor of forml help seeking in mentlly distressed women thn mentlly distressed men. Conclusion: Distressed young dults re reluctnt to seek help. Men re prticulrly unlikely to do so unless severely distressed nd tend not to seek ly support. Sex differences in help seeking my be importnt in understnding the high suicide rte for men. Keywords: helth behviour; mentl disorder; ptient cceptnce of helth cre; suicide. Lucy Biddle, BA, MA, reserch ssocite in medicl sociology; D Gunnell, PhD, professor of epidemiology; J Donovn, PhD, professor of socil medicine nd helth services reserch, Deprtment of Socil Medicine; D Shrp, BA, MA, PhD, FRCGP, professor of primry helth cre, Division of Primry Helth Cre, University of Bristol, Bristol. Address for correspondence Lucy Biddle, Deprtment of Socil Medicine, University of Bristol, Cnynge Hll, Whiteldies Rod, Bristol, BS8 2PR. E-mil: Lucy.Biddle@bristol.c.uk Submitted: 11 June 2003; Editor s response: 17 October 2003; finl cceptnce: 13 Jnury 2004. British Journl of Generl Prctice, 2004, 54, 248-253. Introduction YOUNG people re mong those lest likely to consult helthcre professionls during times of emotionl crisis. 1 As few s one in six young dults with mentl distress seek help from helthcre professionl. 2 In the extreme, such ptterns of low consulttion re observed prior to suicide. Less thn third of young people who commit suicide consult their generl prctitioner (GP) in the month before deth 3,4 compred with 40% of ll dults. 5 Ptterns of low consulting for mentl disorder re most evident in young men. 6 These ptterns indicte need for greter understnding of the fctors contributing to help seeking in young people. Few studies hve ctully quntified young dults use of different help sources or focused on exploring fctors ssocited with the help-seeking behviours of mentlly distressed young dults in generl popultion smples. Existing dt tend to derive from psychitric morbidity surveys with more generl ims, nd which consider dults of ll ges. This study imed to investigte nd compre the help-seeking behviour of mentlly distressed young men nd women ged 16 24 yers. Method A rndom smple of 3004 young dults ged 16 24 yers, drwn from the popultion register of Avon Helth Authority, ws sent n eight-pge questionnire between December 2000 nd August 2001. The survey covered inner-city, suburbn nd urbn loctions. Two reminders were sent to non-responders. The questionnire contined the 12-item generl helth questionnire (GHQ-12) to identify possible cses of minor mentl disorder. Prticipnts scoring 4 or more on 12-point scle were clssified s GHQ cses. 7 A further three questions from the GHQ-28 were included to ssess the existence of suicidl thoughts: Hve you recently: 1. felt tht life isn t worth living? 2. found yourself wishing you were ded nd wy from it ll? 3. found the ide of tking your own life kept coming into your hed? A positive response to ny of these three questions ws tken s indicting the presence of suicidl thoughts. In ddition, the questionnire mesured pst nd current (in the previous 4 weeks) help seeking for psychologicl or emotionl problems, sources of help used, sociodemogrphic chrcteristics, nd perceived levels of socil support. The Duke-UNC functionl socil support questionnire 248 British Journl of Generl Prctice, April 2004
Originl ppers HOW THIS FITS IN Wht do we know? Mentlly distressed young dults re prticulrly unlikely to seek professionl help for their problems nd most young people who commit suicide do not seek help from helthcre professionl in the 4 weeks prior to suicide. Wht does this pper dd? Young men re less likely thn young women to seek some form of help when mentlly distressed. They tend not to seek ly support nd hve higher threshold of severity t which they will consult GP thn women. Less thn one in five young dults with suicidl thoughts seek help from their GP. Sex differences in help seeking my be possible contributor to sex differences in suicide. ws used to mesure socil support. 8 This is n eight-item scle ssessing two qulittive spects of socil support: 1. confidnt support, which refers to confiding reltionships where importnt issues nd life events re discussed; nd 2. ffective support, which refers to emotionl support nd cring. Both spects were exmined, but only the results for confidnt scores re tbulted, since this mesures most directly the issues surrounding the perceived vilbility of help sources. The socioeconomic position of study prticipnts ws mesured ccording to their prentl socil clss (coded s non-mnul [socil clsses I IIINM] or mnul (socil clsses IIIM V]). 9 Dt were nlysed using Stt (version 7). Student t-tests nd χ 2 tests were used to contrst the chrcteristics of responders nd non-responders. Logistic regression models were used to investigte fctors ssocited with help seeking. Initilly, ssocitions with ech individul fctor were ssessed in models controlling for cse severity (GHQ- 12 score) lone. GHQ score ws controlled for by fitting it s continuous term in the models. The effect of ge ws not controlled for since no ssocition ws found between this nd help seeking. Fctors ssocited (P<0.10) with help seeking in this preliminry model were entered together in finl multivrible model. Ethicl committee pprovl ws obtined from the following locl reserch ethics committees: United Bristol Helthcre Trust, North Bristol, nd Weston Locl. Results Response rte nd chrcteristics of responders Of the 3004 questionnires sent out, 340 (11.3%) were returned unopened becuse the ddress ws incorrect. The overll response rte mong the reminder of the smple ws 48% (1276/2664, 42% of men nd 54% of women) fter Tble 1. Summry tble of min chrcteristics of responders by sex. Vrible Mle (n = 560) Femle (n = 716) P-vlue (difference) Sociodemogrphic chrcteristics Age in yers b (men [SD]) 19.6 (2.3) 19.7 (2.3) 0.27 Socil support (confidnt support) c (men [SD]) 19.2 (4.5) 20.0 (4.7) 0.004 Living rrngements d,e (n [%]) With prents 435 (77.7) 475 (66.6) <0.001 With friends 91 (16.2) 107 (15.0) 0.54 With spouse or prtner 47 (8.4) 110 (15.4) <0.001 Occuption d,f (n [%]) Employed full-time 226 (40.4) 283 (39.7) 0.79 Studying full-time 268 (47.9) 316 (44.3) 0.20 Prentl socil clss g (n [%]) Non-mnul 328 (63.7) 428 (64.4) 0.81 Mnul 187 (36.3) 237 (35.6) Generl helth (self-rted) h (n [%]) Excellent/very good 297 (53.1) 367 (51.5) Good 194 (34.7) 267 (37.5) 0.54 Fir/poor 68 (12.2) 78 (11.0) Psychitric morbidity (n [%]) GHQ cseness (score of 4+ on GHQ-12) i 175 (31.5) 274 (38.5) 0.01 Suicidl thoughts (score 1+ on suicide score) j 81 (14.7) 91 (12.8) 0.34 Current psychologicl problems (self-rted) k 132 (24.1) 220 (31.2) 0.006 Usul psychologicl helth (self-rted) l Good 290 (52.2) 360 (50.7) Oky 212 (38.2) 283 (39.9) 0.83 Poor/very poor 53 (9.5) 67 (9.4) Help seeking for mentl disorder (n [%]) Current (previous 4 weeks) m 60 (10.8) 143 (20.2) <0.001 Pst n 169 (30.4) 335 (46.8) <0.001 The totl smple consisted of 1276 useble responses. However, missing dt on some items mens tht the totl n differs for ech vrible s indicted. b (n =1264). c (n = 1253). d Percentges do not dd up to 100% since not ll response ctegories re shown nd in ech cse the options given were not ll mutully exclusive. e (n = 1273). f (n = 1272)..g (n = 1180). h (n = 1271). i (n = 1268). j (n = 1261). k (n = 1254). l (n = 1265). m (n = 1266). n (n = 1270). SD = stndrd devition. British Journl of Generl Prctice, April 2004 249
L Biddle, D Gunnell, D Shrp nd J L Donovn Tble 2. Help sources used by GHQ cses nd those with suicidl thoughts in the 4 weeks before questionnire completion. Mle Femle P-vlue Help source b n (%) n (%) (difference) GHQ cses c Any help 38 (21.8) 94 (34.8) 0.003 Generl prctitioner 13 (7.5) 24 (8.9) 0.60 Counsellor 4 (2.3) 8 (3.0) 0.67 Fmily 20 (11.5) 51 (18.9) 0.04 Friend 25 (14.4) 71 (26.3) 0.003 Voluntry sector 1 (0.6) 2 (0.7) 1.00 Techer or employer 2 (1.1) 14 (5.2) 0.03 Other 6 (3.4) 12 (4.4) 0.60 Prticipnts with suicidl thoughts d Any help 25 (30.9) 37 (41.6) 0.15 Generl prctitioner 10 (12.3) 15 (16.8) 0.41 Counsellor 3 (3.7) 6 (6.7) 0.50 Fmily 11 (13.6) 15 (16.8) 0.55 Friend 14 (17.3) 25 (28.1) 0.09 Voluntry sector 1 (1.2) 2 (2.2) 1.00 Techer or employer 2 (2.5) 3 (3.4) 1.00 Other 4 (4.9) 5 (5.6) 1.00 The number of GHQ cses nd suicidl idetion cses given here is slightly lower thn tht given in Tble 1 since not ll cses indicted whether or not they hd sought help. b Ctegories in this tble re not mutully exclusive so numbers of specific sources dd up to higher totl thn the ny help ctegory. c Mle = 174, femle = 270. d Mle = 81, femle = 89. two reminders. Men, older subjects nd those living in more socioeconomiclly deprived res (s mesured by the Townsend deprivtion index) 10 were more likely to be nonresponders. Compred with mle responders, femle responders were more likely to be GHQ cses, hd higher levels of socil support, were more likely to be living with spouse or prtner, nd were more likely to consider they hd current mentl helth problem (Tble 1). Levels of suicidl thoughts were similr in men (14.7%) nd women (12.8%). There ws strong ssocition between suicidl thoughts nd GHQ score. Compred with those scoring less thn 4 on the GHQ scle, the odds of hving suicidl thoughts were 7.2 (95% confidence intervl [CI] = 4.5 to 11.4), 11.9 (95% CI = 7.1 to 20.1), nd 60.1 (95% CI = 31.6 to 114.5) in those with GHQ score of 4 6, 7 9, nd 10 12, respectively (P<0.0001, djusted for sex). Help seeking Altogether, 20.2% of women compred with 10.8% of men (difference = 9.4%, 95% CI = 5.5 to 13.3, P<0.001) hd sought help for psychologicl problem from friends, fmily, helth services, or other sources in the 4 weeks before completing the questionnire (Tble 1). Among those scoring 4 or more on the GHQ scle (Tble 2), femle GHQ-cses were lso more likely to hve sought some form of help in the previous 4 weeks thn mle cses (34.8% femle cses versus 21.8% mle cses; difference = 13%, 95% CI = 4.6% to 21.3%, P = 0.003). Similrly low proportions of mle nd femle cses hd consulted GP (7.5% mle cses, 8.9% femle cses; P = 0.6). The difference in help-seeking behviour between the sexes is minly ccounted for by women s greter use of fmily nd friends. Individuls experiencing suicidl thoughts were more likely to hve sought help thn GHQ cses, but help seeking ws still low (41.6% in women nd 30.9% in men) nd less thn one in five responders with suicidl thoughts hd consulted GP. Sex differences in help seeking were seen, though these differences were not significnt (Tble 2). Sex differences in the use of help sources persisted mong those who recognised tht they hd psychologicl problem. Of those scoring 4 or more on the GHQ-12, 52.9% of men nd 59.0% of women felt tht they were currently suffering from psychologicl or emotionl problem. A greter proportion of femle cses who perceived tht they hd problem sought some form of help (49.4%) compred with mle cses (37.1%) (difference = 12.3%, 95% CI = 0.5 to 25.0). This nlysis ws bsed on only smll subset of responders (n = 110), nd there ws only wek evidence of sex difference (P = 0.06). Help seeking from GP GHQ score ws the strongest predictor of seeking help from GP. Compred with those with GHQ score of 4 6, the odds rtio (OR) of consulting with score of 7 9 ws 3.4 (95% CI = 1.3 to 8.9) nd with score of 10 12 the OR ws 13.4 (95% CI = 5.3 to 33.9). The ssocition of GHQ score with help seeking differed in men nd women (P[interction] = 0.03). Men ppered to hve somewht higher threshold of severity (s indicted by GHQ score) before seeking help. The men GHQ score of mle help-seekers ws 10.0, compred with 8.5 in women, providing some (wek) evidence of sex difference (difference = 1.5, 95% CI = -0.2 to 3.1, P = 0.07). Tble 3 presents the findings of logistic regression nlysis (controlling for GHQ score) of fctors ssocited with seeking help from GP mong mle nd femle GHQ cses. Both mle nd femle cses ppered pproximtely four times more likely to consult GP if they hd previously sought some form of help (from ny source) for psychologicl or emotionl problems. Men from mnul clss origins were more likely to seek help from doctor thn those of non-mnul clss bckground, wheres the reverse ws true in women. Recognition tht they were suffering from mentl helth problem, suicidl thoughts, nd self-reported poor helth ll ppered to be stronger predictors of GP help seeking in women compred with men. No ssocition ws found between help seeking nd socil support for confidnt or ffective score. In multivrible model including both fctors ssocited with help seeking (P<0.10) in men (prentl socil clss, pst help seeking), there ws no evidence tht the ssocition with either of these fctors ws ttenuted by controlling for the other. In women, five fctors prentl socil clss, self-rted helth, recognition of current problems, suicidl thoughts, nd pst help seeking were ssocited with help seeking (P<0.10). In the multivrible model, ssocitions with prentl socil clss, self-rted helth nd pst help seeking were little chnged, but the ssocition with recognition of current problems nd suicidl thoughts ws much ttenuted (OR = 3.3, 95% CI = 0.7 to 16.2, P = 0.10 nd OR = 1.7, 95% CI = 0.6 to 5.0, P = 0.31, respectively). 250 British Journl of Generl Prctice, April 2004
Originl ppers Tble 3. Fctors ssocited with help seeking from generl prctitioner (GP) (previous 4 weeks) mong GHQ-12 cses (djusted for GHQ score). Mle (n = 168 174) Femle (n = 264 270) Vrible OR 95% CI P-vlue OR 95% CI P-vlue Living rrngements With prents 1.00 1.00 With spouse or prtner 1.00 0.08 to 11.70 0.27 0.03 to 2.16 With friends 1.46 0.28 to 7.61 1.84 0.59 to 5.70 Other b 2.21 0.32 to 15.02 0.86 1.12 0.29 to 4.34 0.28 Occuption Employed full-time or prt-time 1.00 1.00 Studying full-time or prt-time 0.72 0.13 to 3.87 0.74 0.28 to 2.00 Other c 1.62 0.25 to 10.61 0.97 1.33 0.42 to 4.22 0.61 Prentl socil clss d Mnul 1.00 1.00 Non-mnul 0.18 0.02 to 1.82 0.09 2.25 0.91 to 5.54 0.08 Socil support (confidnt score) 5 9 0.23 0.01 to 4.25 0.42 0.04 to 4.05 10 14 0.18 0.02 to 2.05 1.83 0.58 to 5.80 15 19 1.29 0.18 to 9.46 1.98 0.63 to 6.19 20 25 1.00 0.13 1.00 0.31 Generl helth (self-rted) Excellent/very good 0.44 0.08 to 2.39 0.11 0.02 to 0.54 Good 0.56 0.11 to 2.79 0.64 0.24 to 1.70 Fir/poor 1.00 0.61 1.00 0.004 Usul psychologicl helth (self-rted) Good 1.00 1.00 Oky 2.57 0.20 to 32.84 1.32 0.35 to 5.06 Poor/very poor 4.72 0.47 to 47.43 0.32 2.95 0.71 to 12.28 0.2 Current problems (self-rted) Yes 2.89 0.29 to 28.39 0.33 6.51 1.47 to 28.85 0.002 Suicidl thoughts Yes 1.22 0.30 to 4.86 0.78 3.02 1.16 to 7.82 0.02 Pst help seeking Yes 4.24 0.95 to 18.86 0.04 3.79 1.22 to 11.74 0.01 The smple size vries slightly for ech vrible due to missing dt on some items. b Includes those living lone, single prents nd smll number of miscellneous responses. c Includes unemployed, looking fter home/fmily nd not working due to sickness/disbility. d Missing dt on this vrible gve totl of 155 in men nd 244 in women. OR = odds rtio. Help seeking from friends nd fmily Altogether, 18.4% of mle GHQ cses nd 30.7% of femle GHQ cses hd sought help from friends or fmily (difference = 12.3%, 95% CI = 4.4 to 20.3, P = 0.004). As with seeking help from GP, men nd women who hd sought help from friends nd fmily were more likely to hve higher GHQ score thn those who hd not. The men GHQ score of those who hd sought help from friends nd fmily ws 7.6 in men nd 7.4 in women (difference = 0.2, 95% CI = -0.8 to 1.2, P = 0.69). Similr fctors were ssocited with help seeking from friends nd fmily in men nd women (Tble 4). These were the responders own recognition of their problems nd hving sought help in the pst, wheres suicidl thoughts nd socil support did not pper to influence the likelihood of seeking help from friends nd fmily in either sex. In multivrible model including terms for recognition of current problems nd pst help seeking, there ws no evidence in men or women tht ssocitions with either vrible were ttenuted when controlling for the other. Discussion Summry of min findings Young dults experiencing minor mentl disorder, s identified by score of 4 or more on the GHQ-12, hd notbly low rtes of help seeking nd were prticulrly unlikely to consult GP. Even when they perceived themselves s hving mentl helth problem, most did not seek help. The strongest predictors of help seeking were cse severity nd previous help seeking. There were two min sex differences in help-seeking behviour mong those with probble disorders (GHQ cses). First, men were significntly less likely to hve sought some form of help nd were much less likely to confide in the ly group thn women. Second, mle cses ppered to hve higher threshold of severity for help seeking thn femle cses, prticulrly for help from GP. British Journl of Generl Prctice, April 2004 251
L Biddle, D Gunnell, D Shrp nd J L Donovn Tble 4. Fctors ssocited with help seeking from friends nd fmily (previous 4 weeks) mong GHQ-12 cses (djusted for GHQ score). Mle (n = 168 174) Femle (n = 264 270) Vrible OR 95% CI P-vlue OR 95% CI P-vlue Living rrngements With prents 1.00 1.00 With spouse or prtner 0.29 0.03 to 2.41 0.94 0.43 to 2.04 With friends 1.89 0.74 to 4.81 1.06 0.48 to 2.34 Other b 1.41 0.36 to 5.50 0.22 1.29 0.54 to 3.05 0.94 Occuption Employed full-time or prt-time 1.00 1.00 Studying full-time or prt-time 0.69 0.29 to 1.69 1.01 0.56 to 1.81 Other c 0.37 0.09 to 1.46 0.32 1.45 0.66 to 3.19 0.6 Prentl socil clss d Mnul 1.00 1.00 Non-mnul 0.59 0.22 to 1.61 0.29 0.98 0.55 to 1.74 0.94 Socil support (confidnt score) 5 9 0.48 0.07 to 3.44 0.57 0.17 to 1.89 10 14 0.34 0.09 to 1.32 0.81 0.40 to 1.61 15 19 1.09 0.39 to 3.03 0.66 0.34 to 1.25 20 25 1.00 0.21 1.00 0.55 Generl helth (self-rted) Excellent/very good 0.81 0.28 to 2.32 0.6 0.28 to 1.27 Good 1.07 0.39 to 2.97 1 0.49 to 2.04 Fir/poor 1.00 0.83 1.00 0.19 Usul psychologicl helth (self-rted) Good 1.00 1.00 Oky 0.53 0.19 to 1.51 1.48 0.76 to 2.89 Poor/very poor 1.23 0.40 to 3.73 0.23 1.00 0.41to 2.39 0.35 Current problems (self-rted) Yes 7.75 2.13 to 28.22 <0.001 3.96 2.11 to 7.45 <0.0001 Suicidl thoughts Yes 1.25 0.53 to 2.93 0.6 0.93 0.50 to 1.72 0.81 Pst help seeking Yes 7.61 2.99 to 19.39 <0.0001 5.8 2.99 to 11.26 <0.0001 The smple size vries slightly for ech vrible due to missing dt on some items. b Includes those living lone, single prents nd smll number of miscellneous responses. c Includes unemployed, looking fter home/fmily nd not working due to sickness/disbility. d Missing dt on this vrible gve totl of 155 in mles nd 244 in femles. OR = odds rtio. Where this fits with other literture The prevlence of GP help seeking mong cses with mentl disorder in surveys of generl dult popultions hs been estimted s between 22.1% nd 39%. 11,12 The results of our study confirm the suggestion from Austrlin-bsed studies tht young dults re prticulrly low consulters. 2,13 Other reserch studies hve noted sex differences in help seeking for mentl disorder mong the generl popultion 14 nd dolescents nd young dults, 15 but hve focused on the use of trditionl medicl services. Sex differences in the use of rnge of help sources do not pper to hve been quntified elsewhere. The lck of ssocition between help seeking nd ge, residency, nd occuption in this study contrsts with findings from surveys of dults of ll ges tht hve found ge nd mritl nd employment sttus to be importnt. 1 This my be becuse such sociodemogrphic fctors re more influentil in middle or lter dulthood. The ssocition of severity of mentl disorder with help seeking hs been noted elsewhere, 1,16 lthough not in reltion to prticulr sources of help, nd sex effect in thresholds of severity for help seeking ppers to hve only been previously identified in minority disdvntged popultions. 17 Strengths nd limittions To the best of our knowledge, this is the first British study to focus specificlly on the help-seeking behviour of mentlly distressed young dults. It hs lso exmined help seeking from vriety of sources in ddition to trditionl medicl services, quntifying intuitive understndings bout young men s tendencies not to shre problems with fmily nd peers. Additionlly, wheres other studies of help seeking hve been criticised for considering sex solely s min effect vrible, this study hs lso explored the effect of sex on predictors of help seeking. 17 The study ws explortory nd hs identified issues for further reserch, though the reltively smll smple size imposed some limittions upon the power to detect cliniclly significnt ssocitions, prticulrly in multivrible nlysis. The response rte ws low (48%). This is probbly explined by the nture of the reserch subjects mobile 252 British Journl of Generl Prctice, April 2004
Originl ppers young popultion. Studies of non-response hve noted the difficulties of recruiting younger dults, prticulrly men, in popultion surveys. 18,19 In recent popultion survey, the response rte ws only 39% in those ged 16 24 yers, compred with 58% in 55 to 64-yer-olds. 20 Non-responders in the present study were more likely to be mle, older nd living in poorer res. In the bsence of further dt, crude ttempt ws mde to ssess further the likely chrcteristics of non-responders by compring the chrcteristics of those who responded fter the second reminder (n = 246) with those of other responders (n = 1030), on the ssumption tht these lte responders would more closely resemble non-responders. We found little difference in overll rtes of help seeking or GHQ cseness in lte responders compred with other responders. Limittions were imposed by the use of the GHQ to ssess psychitric morbidity, since it screens for probble disorder nd is sensitive to trnsient disturbnces. The prevlence of mentl disorder mong 16 to 24-yer-olds ws estimted s 14.2% in the United Kingdom Ntionl Survey of Psychitric Morbidity (2000), 12 which used the revised version of the clinicl interview schedule (CIS-R), 21 dignostic cse-finding tool, to screen responders. The sensitivity nd specificity of the GHQ-12 hve been estimted s 89% nd 80%, respectively. 7 Not ll of those identified s GHQ cses re likely to be suffering from mentl disorder, however, they do represent vulnerble individuls. Seeking help is complex process, nd survey pproch is limited in the extent to which it cn disentngle such complexity. Those identified s GHQ cses in survey of this nture will be t differing phses within both their illness nd help-seeking trjectories. Limittions of defining the smple ccording to GHQ cseness nd of the crosssectionl pproch my contribute towrds interpreting the strong ssocition between help seeking nd pst help seeking. Pst help seeking referred to ny episode of help seeking not in the previous 4 weeks, nd my in fct be seprting chronic or severe cses from trnsient disturbnces where current nd pst help seeking relte to continul process in response to one significnt episode. Implictions A high proportion of young people with mentl distress nd psychitric morbidity do not seek help. Observed sex differences in help seeking suggest importnt differences in the wys tht men nd women respond to mentl distress. Young men pper to be less likely to seek help until they rech more extreme levels of morbidity. This my increse the risk of their reching points of crisis nd suicide. Young men were lso less likely to use informl sources of help nd support, suggesting tht their help-seeking resources re more limited thn those of young women. Women s greter tendency to confide in friends nd fmily my lso hve n effect in incresing the likelihood tht those with mentl disorders rech erly medicl ttention, owing to the function of ly referrl provided by friends nd fmily. 22 The study results suggest tht we should focus reserch nd preventive ttention on the identifiction nd relxtion of brriers to help seeking. This is prticulrly the cse for young men nd should extend to help seeking from nonmedicl nd informl help sources, including friends nd fmily. Services designed to meet young people s needs nd preferences my increse their willingness to seek help. References 1. Bebbington PE, Meltzer H, Brugh TS, et l. Unequl ccess nd unmet need: neurotic disorders nd the use of primry cre services. Psychol Med 2000; 30: 1359-1367. 2. Rickwood D, Brithwite V. Socil-psychologicl fctors ffecting help-seeking for emotionl problems. Soc Sci Med 1994; 39: 563-572. 3. Vssils C, Morgn H. Generl prctitioners contct with victims of suicide. BMJ 1993; 307: 300-301. 4. Houston K, Hwton K, Shepperd R. Suicide in young people ged 15 24: psychologicl utopsy study. J Affect Disord 2001; 63: 159-170. 5. Gunnell D, Frnkel S. Prevention of suicide: spirtions nd evidence. BMJ 1994; 308: 1227-1233. 6. Foster T, Gillespie K, McClellnd R. Mentl disorders nd suicide in Northern Irelnd. Br J Psychitry 1997; 170: 447-452. 7. Goldberg D, Willims, P. A user s guide to the generl helth questionnire. Berkshire: NFER-Nelson Publishing Compny Ltd, 1988. 8. Brodhed WE, Gehlbch SH, De Gruy FV, Kpln BH. The Duke- UNC functionl socil support questionnire. Mesurement of socil support in fmily medicine ptients. Med Cre 1988; 26: 709-721. 9. Office for Ntionl Sttistics. Stndrd occuptionl clssifiction (volumes 1 nd 2). London: The Sttionery Office, 2000. 10. Townsend P, Phillimore P, Bettie A. Helth nd deprivtion: inequlity nd the North. London: Croom Helm, 1986. 11. Blnd RC, Newmn SC, Orn H. Help-seeking for psychitric disorders. Cn J Psychitry 1997; 42: 935-941. 12. Singleton N, Bumpsted R, O Brien R, et l. Psychitric morbidity mongst dults living in privte households, 2000. Norwich: HMSO, 2001. 13. Rickwood D, D Espignet E. Psychologicl distress mong older dolescents nd young dults in Austrli. Aust N Z J Public Helth 1996; 20: 83-86. 14. Kessler RC, Brown RL, Bromn CL. Sex differences in psychitric help-seeking: evidence from four lrge-scle surveys. J Helth Soc Behv 1981; 22: 49-64. 15. Gsquet I, Chvnce M, Ledoux S, Choquet M. Psychosocil fctors ssocited with help-seeking behviour mong depressive dolescents. Eur Child Adolesc Psychitry 1997; 6: 151-159. 16. O Neil MK, Lncee WJ, Freemn SJ. Help-seeking behviour of depressed students. Soc Sci Med 1984; 18: 511-514. 17. Albizu-Grci C, Alegri M, Freemn D, Ver M. Gender nd helth services use for mentl helth problem. Soc Sci Med 2001; 53: 865-878. 18. Pope D, Croft P. Surveys using generl prctice registers: who re the non-responders? J Public Helth Med 1996; 18: 6-12. 19. Cottler L, Zipp J, Robins L, Spitzngel E. Difficult-to-recruit respondents nd their effect on prevlence estimtes in n epidemiologicl survey. Am J Epidemiol 1987; 125: 329-339. 20. Turnbull J, Kersten P, Hbib M, et l. Vlidtion of the Frenchy ctivities index in generl popultion ged 16 yers nd older. Arch Phys Med Rehbil 2000; 81: 1034-1038. 21. Lewis G, Pelosi A, Ary R, Dunn G. Mesuring psychitric disorder in the community: stndrdised ssessment for use by ly interviewers. Psychol Med 1992; 22: 465-486. 22. Cornford CS, Cornford HM. I m only here becuse of my fmily. A study of ly referrl networks. Br J Gen Prct 1999; 49: 617-620. Acknowledgements We would like to thnk Aln Snders, Julie Cooper nd Helen Bond for fcilitting the use of the helth uthority s popultion register s smpling frme; Srh Polck nd Clre Morkne who provided clericl support; Professor Glyn Lewis, Dr Isbel Oliver nd Somerset Helth Authority survey stff for the dvice they hve given t vrious stges; Dr Dvid Kessler, Dr Michel Rossdle nd Dr Tom Smyth for ssisting with the pilot study; nd ll prticipnts who returned questionnires. British Journl of Generl Prctice, April 2004 253