Community Attitudes Toward People With Schizophrenia

Transcription

1 ORIGINAL RESEARCH Community Attitudes Toward People With Schizophrenia Heather Stu art, MA, PhD 1, Julio Arboleda- Flórez, MD, FRCPC, DABFP, PhD 2 Ob jec tive: We sur veyed public at ti tudes to ward people with schizo phre nia as part of a pi lot proj ect for the World Psy chi at ric As - so cia tion s Global Cam paign to Fight Stigma and Dis crimi na tion Be cause of Schizo phre nia. Meth ods: We con ducted random- digit tele phone sur veys with 1653 re spon dents (aged 15 years or over) re sid ing in 2 ad ja cent ru - ral and ur ban health re gions (71.9% re sponse rate). A brief in ter view col lected in for ma tion on ex pe ri ences with peo ple with a men tal ill ness or schizo phre nia, knowl edge of causes and treat ments for schizo phre nia, and lev els of so cial dis tance felt toward peo ple with schizo phre nia. Re sults: One- half of the sam ple had known some one treated for schizo phre nia or an other men tal ill ness. Of those able to iden tify a cause of schizo phre nia (two- thirds), most iden ti fied a bio logi cal cause, usu ally a brain dis ease. So cial dis tance in creased with the level of in ti macy re quired. One in 5 re spon dents thought they would be un able to main tain a friend ship with, one-half would be un able to room with, and three- quarters would be un able to marry, some one with schizo phre nia. Those over 60 were least knowl - edge able or en light ened and the most so cially dis tanc ing. Greater knowl edge was as so ci ated with less- distancing attitudes. When other fac tors were con trolled, ex po sure to the men tally ill was not cor re lated with knowl edge or at ti tudes, even among those who had worked in agen cies pro vid ing serv ices to the men tally ill. Con clu sions: Most re spon dents were rela tively well in formed and pro gres sive in their re ported understand ing of schizophrenia and its treatment. Clear sub group dif fer ences were ap par ent with re spect to age and knowl edge. Knowl edge of schizo phre nia, not ex po sure to the men tally ill, was a cen tral modifiable cor re late of stigma. (Can J Psy chia try 2001;46: ). Key Words: schizophrenia, stigma, public attitudes, mental health promotion Since the be gin ning of the com mu nity mental health move - ment, nega tive public at ti tudes have be come an es sen tial fac tor in the man age ment of mental ill ness. Suc cess ful com - mu nity re in te gra tion and ten ure de pend on the ex is tence of a tol er ant and sup por tive com mu nity en vi ron ment. Be cause nega tive and stig ma tiz ing pub lic at ti tudes to ward the men - tally ill have di rect im pli ca tions for pre ven tion, early de tec - tion, treat ment, re ha bili ta tion, and qual ity of life, it is im por tant to evalu ate, un der stand, and ul ti mately in flu ence the pub lic s at ti tudes to wards the men tally ill (2 4). Al though sev eral studies have ex am ined com mu nity atti - tudes to ward those with a mental ill ness (5), few have fo cused on par ticu lar dis or ders such as schizo phre nia (6). Yet, there is Manuscript received December 1999, revised, and accepted December Associate Professor, Department of Community Health and Epidemiology, Queen s University, Kingston, Ontario. 2 Professor and Head, Department of Psychiatry, Queen s University, King - ston, Ontario. Address for correspondence: Dr H Stuart, Department of Com mu nity Health and Epidemiology, Queen s University, Kingston, ON K7L 3N6 hh11@post.queensu.ca. evi dence that peo ple do dif fer en ti ate among groups (7) and may hold dif fer ent stereo typi cal un der stand ings of dif fer ent ill nesses (8). For ex am ple, para noid schizo phre nia is re ported to be the most rec og niz able and least ac cepted dis or der (9). Disorder- specific in for ma tion will be nec es sary to tar get pub - lic edu ca tion cam paigns more ef fec tively. This study ex am ines com mu nity at ti tudes to ward peo ple with schizo phre nia to in form a com mu nity an tis tigma pro gram un - der taken as the first (pi lot) wave of the World Psy chi at ric As - so cia tion s global cam paign to fight stigma and dis crimi na tion be cause of schizo phre nia (de scribed else - where (10,11). It de scribes public knowl edge, at ti tudes, and feel ings of so cial dis tance to ward peo ple with schizo phre nia. Al though re sults were origi nally in tended to as sist lo cal plan - ners to tar get an tis tigma in ter ven tions, they should also be of more gen eral in ter est to any one in volved in men tal health pro mo tion at the popu la tion level. Study Design Data were col lected by the Popu la tion Health Unit of the Cal - gary Re gional Health Author ity, a re search unit spe cial iz ing in popu la tion surveys for re gional health plan ning. Two Can J Psychiatry, Vol 46, April

2 246 The Canadian Journal of Psychiatry Vol 46, No 3 cross- sectional, random- digit tele phone sur veys using identi - cal sam pling and in ter view tech niques were con ducted, the first in August 1988 and the sec ond in March Be cause there were no note wor thy dif fer ences be tween the sur veys in knowl edge, at ti tudes, or so cial dis tance, re sults were com - bined for this pres en ta tion. This has in creased the num bers avail able for sub group com pari sons and im proved the sta tis - ti cal power of analyses. In di vidu als were eligible if they were aged 15 years or over, English- speaking, and liv ing in 1 of 2 ad ja cent health re gions in Al berta: a large met ro poli tan area of ap proxi mately resi dents, and a small ru ral re gion num ber ing about 5000 resi dents. Can di date tele phone num bers (excluding busi ness ex changes) were ran domly se lected from com put er - ized tele phone listings. In the ur ban area, where un listed tele - phone num bers oc cur fre quently, 1 digit was added to each 4- digit tele phone suffix to en sure that un listed num bers were in cluded. This pro ce dure proved un nec es sary in the ru ral sur - vey area be cause of the small number of un listed num bers used. New ex changes up dated within the year were not con - tained in the tele phone da ta base and were nec es sar ily ex - cluded from the sur vey. We used a stratified sam pling de sign, and ru ral resi dents were over sam pled to al low for sub se quent sub group analy ses. To maxi mize re sponse rates and en sure even so cio eco nomic cov er age, in ter view ers called during eve nings and week ends and al lowed up to 3 call backs for busy num bers and no- answers. Tele phone in ter view ers fol - lowed a computer- generated, struc tured ques tion naire. They read ques tions di rectly from the com puter, cod ing and en ter - ing live data as the in ter view pro gressed. In formed con sent was con sid ered to have been given when in di vidu als agreed to com plete the in ter view. In all, 1653 in ter views were com - pleted. Based on the number of eli gi ble re spon dents, the cal - cu lated re sponse rate was 71.9%. Ninety in di vidu als were con sid ered in eli gi ble be cause of a lan guage bar rier (71% of these were ur ban dwell ers). Re sults have been weighted (12) to the pro vin cial population by age group, sex, and re gion to ad just for the strati fied sam pling de sign and cer tain sampling dis crep an cies (de scribed be low). Interview Schedule The in ter view sched ule was ad min is tered in 10 to 15 min utes. We pre tested ques tions in tele phone in ter views con ducted as part of the in ter view ers train ing. In ad di tion to so cio de mo - graphic items, 10 ques tions as sessed knowl edge about the causes of schizo phre nia, its con se quences, and ap pro pri ate treat ment ap proaches. To elicit knowl edge of causes, re spon - dents were first asked to in di cate, to the best of their knowl - edge, what causes schizo phre nia. Un prompted first men tions were coded to re flect the fol low ing cate go ries: brain dis ease, an other bio logi cal fac tor, a social- psychological fac tor, ex act causes un known, and don t know. The next 9 ques tions re - quired struc tured re sponses fol low ing a 4- point or di nal scale rang ing from fre quently, of ten, rarely, to never. We also asked re spon dents to es ti mate the preva lence of schizo - phre nia (rounding to the near est per cent age point). To as sess the ex tent to which knowl edge and at ti tudes to ward those with schizo phre nia were as so ci ated with the level of re spon - dents ex po sure to those with a mental ill ness or to nega tive stereo types of the men tally ill, we asked about ex po sure to the me dia, to agencies pro vid ing serv ices to the men tally ill, and to peo ple re spon dents had known who had re ceived treatment for schizo phre nia or an other men tal ill ness. Six ques tions as - sessed so cial dis tance. Originally de vel oped by Bo gar dus to meas ure stig ma tiz ing at ti tudes to ward cul tural mi nori ties, social- distance ques tions have since been used to meas ure pub lic at ti tudes to ward the men tally ill (2,13 15). Ques tions evolve along a Gutt man scale of in creas ing per sonal in ti macy (16). In this case, the fol low ing is sues were in ves ti gated: be - ing em bar rassed if it were known that a fam ily mem ber had schizo phre nia, talk ing to some one with schizo phre nia, being co- workers with some one with schizo phre nia, rooming with some one with schizo phre nia, and mar ry ing some one with schizo phre nia. We ex pected that the high est levels of so cial dis tance would be ex pressed for those so cial situa tions in - volv ing the great est de gree of per sonal in ti macy. Fi nally, re - spon dents were asked how they would feel about hav ing a group home for 6 to 8 peo ple with schizo phre nia in their neigh bour hood. Results Sam ple Char ac ter is tics Ta ble 1 com pares the sam ple demo graph ics for ur ban and ru - ral sur vey re spon dents with their cor re spond ing re gional in - sur ance reg is try popu la tions, re flect ing all resi dents eli gi ble for pro vin cial health care cov er age (vir tu ally the en tire popu - la tion). Using chi square, ur ban and ru ral sam ples were sta tis - ti cally rep re sen ta tive with re spect to sex, al though ru ral women were slightly over sam pled, and both sam ples were sig nifi cantly skewed to ward younger ages. Eld erly re spon - dents liv ing in ru ral ar eas were un der sam pled by 4%, rep re - sent ing the larg est dis crep ancy in any sub group com pari son. To ad just for these dis crep an cies as well as for the strati fied sam pling de sign, re sults were population- weighted (using the reg is try popu la tion) based on age group, sex, and health re - gion of resi dence. Knowl edge and At ti tudes In for ma tion about how knowl edge and at ti tudes dif fer by popu la tion sub groups can help planners re fine the nature of the mes sage, tar get mes sages to spe cific audi ences, and iden - tify high-yield meth ods of trans mis sion (17). Ta ble 2 pres ents the population- weighted re sults (per cent ages and

3 April 2001 Community Attitudes Toward People With Schizophrenia 247 Characteristic Sex Male Fe male Age Group in years Miss ing Ur ban % (n) 48.3% (391) 51.7% (419) χ 2 = 1.86, df 1, P = % (231) 42.0% (332) 22.4% (177) 6.3% (50) 20 χ 2 = 8.79, df 3, P = 0.03 Table 1. Sample characteristics Cor re spond ing Re gional Population a % (n) 50.7% ( ) 49.3% ( ) 27.0% ( ) 46.4% ( ) 19.5% ( ) 7.1% (46 333) Ru ral % (n) 46.7% (394) 53.3% (449) χ 2 = 3.39, df 1, P = % (237) 41.0% (339) 23.6% (195) 6.8% (56) 16 χ 2 = 14.42, df 3, P = Cor re spond ing Re gional Popu la tion a % (n) 49.9% (19 177) 50.0% (19 219) 26.6% (10 218) 40.6% (15 594) 22.0% (8 443) 10.8% (4 141) a Based on the 1996 Al berta Health In sur ance Registry popu la tion of pro vin cial resi dents eli gi ble for pro vin cial health care cov er age (that is, vir tu ally all of the popu la tion). cor re spond ing stan dard er rors) for sur vey items by sex. (A su - per script b in di cates items of po ten tial sig nifi cance, re - flected by nonover lap ping 95%CIs). Given that mul ti ple com pari sons were made, in ter pre ta tions should be based on mean ing ful patterns, rather than on iso lated sta tis ti cal as so - cia tions. More de tailed re sults stratified by age group and place of resi dence are avail able from the authors, upon re quest. Ex po sure to Media Mes sages To re flect re spon dents over rid ing im pres sion, de scrip tions of people with schizo phre nia as por trayed in the news or on tele vi sion were elic ited through un prompted first mentions. These were sub se quently coded into fixed cate go ries. Nega - tive de scrip tions in cluded por tray als of peo ple with schizo - phre nia as vio lent or dan ger ous to oth ers, com mit ting crimes, home less, a public nui sance, and di shev eled or dirty in ap - pear ance. Posi tive de scrip tions in cluded human- interest por - tray als, de scrip tions of the symp toms or course of the dis ease, ac counts of peo ple with schizo phre nia as vic tims (either vic - tims of crime or of a se ri ous ill ness that re quires medi cal treat - ment), com men tar ies on the need for more or bet ter treat ment or com mu nity sup port sys tems, and re search ad vances or proj ects. Most re spon dents did not re mem ber seeing any thing re lat ing to schizo phre nia. Al though not shown, those over age 60 years watched sig nifi cantly more tele vi sion than their younger coun ter parts and were sig nifi cantly less likely to re - mem ber the tenor of the mes sages heard. Ex po sure to Peo ple With a Men tal Ill ness De spite gen er ally high lev els of ex po sure both to peo ple with an emo tional prob lem (50%) and to peo ple with schizo phre - nia (22%), most re spon dents in di cated that schizo phre nia had not touched their lives. Just un der 2% of the sam ple re ported that schizo phre nia touched their lives daily. Al though not shown, women were more likely to have had per sonal expo - sure to the men tally ill, ei ther be cause they worked at a serv - ice agency or be cause they, or some one they knew, had been treated for an emo tional prob lem or mental ill ness. Sig nifi - cant age dif fer ences also emerged, with those over age 60 years having the least ex po sure to the men tally ill. Knowl edge of Schizo phre nia Based on un prompted first men tions from an open- ended ques tion, two- thirds of re spon dents were able to iden tify a cause of schizo phre nia. Of these, most iden ti fied a bio logi cal cause, usu ally in di cat ing that schizo phre nia is a brain dis ease. Few iden ti fied a psy cho so cial cause such as poor upbringing by par ents, physi cal abuse, stress, or trau matic shock. In ad di - tion, most re spon dents re ported views that were con sis tent with a community- treatment phi loso phy. They rec og nized that peo ple with schizo phre nia are not of lower in tel li gence, can work at regu lar jobs, and need pre scrip tion medi ca tion to con trol symp toms. Al most one-half of the sam ple, how ever, still sub scribed to the con ven tional stereo type that peo ple with schizo phre nia suffer from split or mul ti ple per son ali ties. Bi vari ate analy ses of knowl edge items (not shown) showed that cor rect re sponses var ied by sex, age, and re gion; how - ever, the most pro nounced and con sis tent pat tern oc curred with age. Those over age 60 years re flected a less- enlightened view of schizo phre nia and its treat ment. They were less likely to be lieve that peo ple with schizo phre nia could be suc cess - fully treated in the com mu nity or main tain a regu lar job, and they were more likely to think that peo ple with schizo phre nia were a public nui sance be cause of pan han dling, poor hy - giene, or odd be hav iour. Women were more likely to un der - stand schizo phre nia as a brain dis ease and to rec og nize the im por tance of pre scrip tion medi ca tions in its treat ment. They also, how ever, tended to ex ag ger ate the oc cur rence of dis - turbed and dan ger ous be hav iours and the preva lence of schizo phre nia in the popu la tion, per haps in di cat ing greater fear. Ru ral resi dents were less in formed about the bio logi cal ba sis for schizo phre nia and tended to sub scribe to the con ven - tional stereo type that em pha sizes a split per son al ity and vio - lent be hav iour.

4 248 The Canadian Journal of Psychiatry Vol 46, No 3 Questionnaire Item: Table 2. Prevalence of exposure, knowledge, and social distance Exposure factors: Average hours of television per week (missing = 29) < > 20 In the past 6 months have seen, read, or heard something in the news about people with schizophrenia, where person was described (missing = 1) Nega tively (stereo typi cally) Posi tively (sym pa theti cally) Don t re mem ber Didn t see any thing in the news In the past 6 months have read or heard an advertisement or promotion about schizophrenia (missing = 1) On tele vi sion On the ra dio In a news pa per In a maga zine, pam phlet, or brochure Other me dia Don t re mem ber Didn t read or hear any ad ver tise ment or pro mo tion Have worked as an employee in an agency that provides services to people with emotional problems or mental illness (missing = 2) Know someone or have themselves been treated for (missing = 2) An emo tional problem or a mental illness Schizo phre nia To what extent does schizophrenia touch your life? (missing = 4) Not at all Some what Quite a bit All the time (deal with it al most daily) Knowledge factors: To the best of your knowledge what causes schizophrenia? (missing = 3) Brain dis ease Other bio logi cal factor Psychosocial factor Don t know or ex act causes are un known All things considered, people with schizophrenia frequently or often (missing = 1) Can be suc cess fully treated out side of hospital in the com mu nity Tend to be mentally re tarded or of lower in tel li gence Need pre scrip tion drugs to con trol their symp toms Can be suc cess fully treated with out drugs us ing psy cho ther apy or so cial in ter ven tions Suf fer from split or multiple per son ali ties Can work in regular jobs Are dan ger ous to the public be cause of violent be hav iour Are a public nui sance due to pan han dling, poor hy giene, or odd be hav iour Can be seen talk ing to them selves or shouting in city streets What percent of the population suffers from schizophrenia? (unprompted; missing = 2) < 1% 2% 10% > 10% Don t know Social distance factors: Portion of respondents who would definitely or probably (missing = 5) Feel ashamed if peo ple knew some one in your fam ily was di ag nosed with schizo phre nia Feel afraid to have a con ver sa tion with some one who has schizo phre nia Be up set or dis turbed about work ing on the same job with some one who has schizo phre nia Be un able to main tain a friend ship with some one who has schizo phre nia Feel up set or dis turbed about room ing with some one who has schizo phre nia Would not marry some one with schizo phre nia How would you feel about having a group home for 6 to 8 people with schizophrenia in your neighbourhood? (missing = 6) In favour Op posed In dif fer ent, that it does not mat ter Men (n = 785) % a (SE) 59.5 (2.6) 29.1 (2.4) 11.4 (1.6) 5.7 (1.2) 10.4 (1.6) 13.5 (1.8) 70.4 (2.4) 13.9 (1.8) 9.2 (1.6) 4.5 (1.1) 3.4 (1.0) 4.3 (1.1) 3.8 (1.1) 60.8 (2.6) 14.5 (1.9) 44.5 (2.6) 19.1 (2.1) 83.9 (1.9) 13.0 (1.7) 1.7 (0.7) 1.4 (0.6) 40.8 (2.6) 13.9 (1.8) 7.8 (1.3) 37.5 (2.5) 65.4 (2.5) 8.2 (1.3) 78.8 (2.1) 36.7 (2.5) 46.9 (2.6) 71.7 (2.3) 14.4 (1.8) 18.4 (2.0) 37.5 (2.5) 19.4 (2.1) 53.2 (2.6) 9.5 (1.4) 17.9 (2.0) 6.9 (1.3) 11.4 (1.6) 17.5 (2.0) 19.0 (2.0) 46.2 (2.6) 71.9 (2.3) 22.9 (2.2) 5.8 (1.1) 71.3 (2.3) Women (n = 868) % a (SE) 61.8 (2.4) 27.2 (2.3) 11.0 (1.5) 6.8 (1.3) 10.9 (1.6) 18.8 (2.0) 63.5 (2.4) 16.5 (1.8) b 4.3 (1.0) 8.9 (1.5) 7.7 (1.4) 4.3 (1.1) 2.6 (0.8) 54.9 (2.5) 23.2 (2.1) b 55.1 (2.5) b 24.6 (2.2) 81.2 (2.0) 14.7 (1.8) 2.4 (0.8) 1.7 (0.7) 47.7 (2.5) b 14.7 (1.8) 9.9 (1.5) 27.7 (2.1) 73.7 (2.1) 10.9 (1.6) 87.5 (1.6) b 29.9 (2.2) 47.5 (2.5) 72.3 (2.2) 20.5 (2.0) b 25.0 (2.2) 47.9 (2.5) b 10.8 (2.6) b 42.8 (2.5) b 17.7 (1.9) b 28.7 (2.2) b 6.6 (1.2) 11.9 (1.6) 14.9 (1.8) 17.1 (1.8) 47.7 (2.5) 78.3 (2.0) 27.5 (2.2) 9.2 (1.5) 63.3 (2.4) Total (n = 1653) % a (SE) 60.7 (1.8) 28.1 (1.6) 11.2 (1.1) 6.2 (0.9) 10.6 (1.1) 16.3 (1.3) 66.8 (1.7) 15.3 (1.3) 6.7 (0.9) 6.8 (0.9) 5.6 (0.9) 4.6 (0.8) 3.2 (0.7) 57.8 (1.8) 19.0 (1.4) 50.0 (1.8) c 21.9 (1.5) 82.5 (1.4) 13.9 (1.2) 2.1 (1.0) 1.5 (0.5) 44.4 (1.8) 14.3 (1.3) 8.9 (1.0) 32.5 (1.6) 69.6 (1.6) c 9.6 (1.0) 83.2 (1.3) 33.2 (1.7) 47.2 (1.8) 72.0 (1.6) 17.5 (1.6) 21.8 (1.5) 42.8 (1.8) 14.9 (1.3) 47.8 (1.8) 13.7 (1.2) 23.5 (1.5) 6.7 (0.9) c 11.6 (1.1) 16.1 (1.3) 18.1 (1.4) 47.0 (1.8) 75.2 (1.6) 25.3 (1.5) 7.6 (0.9) 67.1 (1.7) a Re flects weighted per centages; b Rep re sents po ten tially note wor thy dif fer ences with largely non overlapping 95% CI (cal cu l ated by adding or sub tract ing roughly 2 times the stan dard er ror from the es ti mate, or for more ac cu rate in ter vals, by using a mul ti ple of 1.96); c Each statement is independently calculated out of 100% so numbers will not total across statements.

5 April 2001 Community Attitudes Toward People With Schizophrenia 249 Items most di rectly as so ci ated with above- average knowl - edge were iden ti fied us ing lo gis tic re gres sion (18). A knowl - edge score was cre ated by sum ming the cor rect re sponses to the 10 knowl edge ques tions. Above- average knowl edge was de fined as any score ex ceed ing the mean score by at least 1 stan dard de via tion (SD), re flect ing those ob tain ing 8 to 10 cor rect an swers. Odds ra tios (ORs) and 95%CIs were de rived by si mul ta ne ously ad just ing for all fac tors in the model. Key in ter ac tions were as sessed and found to be non sig nifi cant. Three fac tors were note wor thy: age group (15 to 29, 30 to 59, and 60 to 90), re gion of resi dence (ur ban or ru ral), and so cial dis tance. Age and so cial distance were in versely as so ci ated with knowl edge. For ex am ple, us ing the old est group as base - line, those aged 30 to 59 years were 3.8 times more likely to have above- average knowl edge (95%CI, 1.6 to 4.5), and those un der the age of 30 were 1.8 times more likely to have above- average knowl edge (95%CI, 2.3 to 6.5). Simi larly, those who were least likely to dis tance them selves from peo - ple with schizo phre nia were the most knowl edge able. In fact, they were about 5 times more likely to have above- average knowl edge (95%CI, in the range of 1.5 to 10). Fi nally, those liv ing in ru ral re gions tended to be less knowl edge able over - all (OR = 0.6; 95%CI, 0.5 to 0.8). Sex, tele vi sion ex po sure, ex po sure to peo ple treated for schizo phre nia or an other men - tal ill ness, and having worked in an agency pro vid ing serv ices to the men tally ill were not as so ci ated with above- average knowl edge. This model ac counted for 11% of the vari ance. So cial Dis tance As ex pected, so cial dis tance in creased with the level of inti - macy re quired in the re la tion ship, fol low ing a Gutt man dis tri - bu tion (Ta ble 2). Few re spon dents in di cated they would feel ashamed if people knew that some one in their fam ily had been di ag nosed with schizo phre nia. More would feel un com - fort able having a con ver sa tion with a person with schizo phre - nia, and still more would be upset or dis turbed about working on the same job with some one with schizo phre nia. One in 5 would be un able to main tain a friend ship, al most 1 in 2 would be un able to room with, and three- quarters would be un able to marry, some one with schizo phre nia. Less than 10%, how - ever, ex pressed clear op po si tion to hav ing a group home for peo ple with schizo phre nia in their neigh bour hood, al though the domi nant sen ti ment was one of in dif fer ence rather than sup port. So cial distance did not vary by sex, but those over age 60 years and those liv ing in ru ral re gions re ported sig nifi - cantly higher lev els of so cial dis tance and less sup port for the es tab lish ment of neigh bour hood group homes. We ex plored the bi vari ate cor re lates of so cial dis tance us ing a cu mu la tive social- distance score. This was cal cu lated by tabu lat ing the number of de sir able re sponses for each of the 6 social- distance items, then cate go riz ing them to re flect 3 groups: low so cial dis tance (meaning all items were an swered de sira bly), mod er ate so cial dis tance (meaning 1 item was an - swered unde sira bly), and high so cial distance (mean ing 2 or more items were an swered unde sira bly). Using chi square, sev eral so cio de mo graphic and ex po sure fac tors were sta tis ti - cally as so ci ated with so cial dis tance: age group, re gion, expo - sure to ad ver tise ments or pro mo tions about schizo phre nia in the pre vi ous 6 months, tele vi sion ex po sure, and knowing some one who had been treated for schizo phre nia or an other emo tional prob lem or men tal ill ness. Every sin gle knowl edge ques tion reached sta tis ti cal sig nifi cance. The fac tors most as so ci ated with high so cial dis tance (as de - fined above to re flect 2 or more un de sir able an swers out of 6) were then iden ti fied using lo gis tic re gres sion. For this analy - sis, the cu mu la tive knowl edge score was cate go rized into 3 groups: below- average knowl edge (re flect ing a score of 0 to 3), av er age knowl edge (re flect ing a score of 4 to 7), and above- average knowl edge (re flect ing a score of 8 to 10). Be - low- and above- average groups re flected those who ex ceeded 1 SD from the mean. Only 2 fac tors were note wor thy in the con trolled analy sis. First, so cial dis tance in creased with age group. Those in the oldest age group were al most 3 times more likely to ex press high so cial dis tance, com pared with their young est coun ter parts. The 95%CIs did not in clude 0, sug gest ing that a dif fer ence this large is likely not due to chance. The re la tion of so cial dis tance to knowl edge was much stronger. The de clin ing ORs show that those with the high est knowl edge were the least so cially dis tanc ing. Stated an other way, those with the highest knowl edge of schizo phre nia were 10 times more likely to ex press highly tol er ant at ti tudes, com - pared with those with the least amount of knowl edge. When age and knowl edge were con trolled, ex po sure fac tors, place of resi dence, and so cio de mo graphic fac tors were not pre dic - tive. This model ac counted for 11% of the vari ance ex plained. Discussion Population- based lev els of knowl edge and at ti tudes were higher than ex pected, and most re spon dents were relatively well in formed and pro gres sive in their re ported un der stand - ing of schizo phre nia and its treat ment. For ex am ple, of the two- thirds who could iden tify a cause, most could cor rectly iden tify schizo phre nia as bio logi cally de ter mined, and onehalf said it was a brain dis ease. A com mu nity mental health phi loso phy was re flected in the majority- held be liefs that peo ple with schizo phre nia could be suc cess fully treated out - side of hos pi tal, needed pre scrip tion drugs to control their symp toms, could work at regu lar jobs, were not a threat to pub lic safety, and were not a public nui sance. Most thought

6 250 The Canadian Journal of Psychiatry Vol 46, No 3 they would not oppose (and more likely feel in dif fer ent to) a group home for 6 to 8 peo ple with schizo phre nia in their neigh bor hood. One ex pla na tion for these re sults is that social- desirability pres sures may have mo ti vated re spon dents to re flect more tol er ant at ti tudes (19), al though social- desirability bias would not ex plain the high level of fac tual knowl edge identi - fied though open- ended ques tions. Be cause this study was con ducted through a tele phone in ter view rather than a selfadministered ques tion naire or face- to- face in ter view, it was not pos si ble to in tro duce ques tions that could di rectly as sess the pres ence of a social- desirability re sponse ten dency. Thus, we can only specu late on the ex tent to which socialdesirability bias might ex plain the high levels of re ported tol - er ance and the lack of op po si tion (ac tu ally, in dif fer ence) to the lo ca tion of group homes when ex pe ri ence shows that com mu nity plan ners of ten face strong op po si tion to the de - vel op ment of community- based treat ment fa cili ties. Al ter na - tively, it may be that those ini tially ex press ing in dif fer ence may be swayed to op pose com mu nity mental health de vel op - ments by a vo cal, nega tive mi nor ity. If this is true, then com - mu nity plan ners could also sway public opinion through an tis tigma cam paigns con ducted in neigh bor hoods tar geted for community- based treat ment pro grams. In Eng land, a so - cio edu ca tional pro gram con ducted among 150 im me di ate neigh bours of sup ported houses for the men tally ill showed sig nifi cant in creases in so cial con tact with both staff and pa - tients. In the area re ceiv ing the ex peri men tal pro gram, 13% of neigh bours made friends with pa tients and in vited them to their homes, and 28% vis ited the pa tients in their group- home set ting. In the con trol com mu nity, none of the neigh bor hood resi dents made friends with the pa tients or staff, and only 8% had vis ited the pa tients. It was the more in ti mate so cial con - tact with the pa tients, rather than the edu ca tion per se, that ac - counted for the im proved at ti tudes (20). The more en light ened public at ti tudes and knowl edge re - ported in this re search may also be a re sult of secu lar trends; that is, im prove ments in public opinion to ward peo ple with schizo phre nia. The only Ca na dian evi dence bear ing on this ques tion evaluated changes in so cial dis tance to ward the men tally ill be tween 1957 and 1974 (15). Sta tis ti cally sig nifi - cant im prove ments were noted in 6 of the 12 social- distance items, but these in di cated only small shifts to ward moreaccepting at ti tudes to ward the men tally ill and tell us little about public at ti tudes to those with schizo phre nia. Nev er the - less, pre vi ous re search has iden ti fied sev eral fac tors that in - flu ence so cial dis tance to ward the men tally ill. Younger peo ple, those who are bet ter edu cated, and those who have had ex po sure to per sons with a men tal ill ness tend to be more en light ened, hu mani tar ian, and sci en tific in their un der stand - ing (15). These pat terns were only partly rep li cated, which per haps re flects the fact that the public may hold dif fer ent views, de pend ing upon the spe cific mental ill ness tar geted. Con sis tent with find ings reported else where (21 23), knowl - edge about schizo phre nia and its treat abil ity was sig nifi cantly re lated to so cial dis tance and age. The more knowl edge able peo ple were, the less stig ma tiz ing or dis tanc ing they tended to be. Older re spon dents were sig nifi cantly less knowl edge able and more dis tanc ing than were their younger coun ter parts. In con trolled analyses, how ever, in ter per sonal ex po sure was not as so ci ated with bet ter knowl edge or less- distancing at ti tudes. Al though un usual, such re sults have been re ported in other con trolled analy ses (20). Con sid er ing that both the preva lence and life time risk of de - vel op ing a mental dis or der is high, it may not be sur pris ing that al most one-half the re spon dents had known some one who had been treated for a men tal dis or der, and al most onefifth had known some one who had been treated for schizo - phre nia (ei ther a fam ily mem ber, friend, or co- worker). None the less, most re spon dents still did not view schizo phre - nia as having touched their lives in any im por tant way, al - though the pro por tion of those re port ing that they deal with schizo phre nia al most daily (1.5%) was gen er ally con sis tent with what we con sider the preva lence of schizo phre nia to be in the popu la tion. Simi larly, we had ex pected those who have worked in agencies pro vid ing serv ices to the men tally ill to be sig nifi cantly more tol er ant and less so cially dis tanc ing be - cause of their greater ex po sure. These find ings sup port the per cep tions of those with schizo phre nia that their most stig - ma tiz ing ex pe ri ences oc cur within the men tal health pro vider com mu nity, in which they have their most fre quent con tacts, and high light the im por tance of spe cifi cally tar get ing health and mental health pro vid ers in an tis tigma in ter ven tions. The im pli ca tions of our find ings for public edu ca tion cam - paigns are com plex. If audi ences are as well in formed as this sur vey sug gests, it will be dif fi cult to pro duce ad di tional im - prove ments (due to a ceil ing ef fect) and costly to meas ure them (be cause large sam ple sizes will be re quired to de tect small dif fer ences). Sec ondly, while knowl edge ap peared to be the most cen tral and modi fi able cor re late of stigma, it is worth while not ing that rela tively knowl edge able and en light - ened views co- existed in the popu la tion with so cially dis tanc - ing at ti tudes, high light ing their mul ti ple and com plex eti ol ogy. The small vari ance ex plained by our sta tis ti cal mod - els re in forces this view. Fur ther, the pat terns noted in the data sug gest that separate pro cesses may cre ate and maintain stigma within dif fer ent popu la tion sub groups (based on such fac tors as age, ur ban ic ity, or sex) or with re spect to dif fer ent men tal ill nesses. If this is true, our re sults ar gue against the ef - fec tive ness of large- scale, ge neric, fact- based popu la tion edu ca tional cam paigns de signed to re duce stigma by con vey - ing knowl edge about schizo phre nia. In ad di tion, be cause the gen eral popu la tion may hold dif fer ent stereo typi cal views of the men tally ill, de pend ing upon the ill ness in ques tion, cam - paigns that at tempt to im prove aware ness or knowl edge of

7 April 2001 Community Attitudes Toward People With Schizophrenia 251 men tal ill ness in gen eral one step re moved from those that at tempt to con vey in for ma tion tar geted to a spe cific ill - ness may be par ticu larly pow er less to show posi tive re sults. In an era char ac ter ized by evidence- based prac tice and im me - di acy of re sults, a cam paign s in abil ity to show posi tive gains may have longer- term con se quences for the per ceived vi abil - ity of mental health pro mo tion. Thus, edu ca tional cam paigns ought to be ap proached with some cau tion and con sid er able fore thought and plan ning. For ex am ple, prior to em bark ing on such a pro gram, the base line knowl edge and at ti tudes that are the ob ject of change must be es ti mated, the amount of change that could rea sona bly be ex pected with the type of mes sage pro posed must be con sid ered, and an evaluation strat egy that is pow er ful enough to de tect po ten tially small im prove ments must be de vel oped and im ple mented. Fi nally, it is im por tant to rec og nize that, de spite the relatively high levels of knowl edge and tol er ance self- reported by sur - vey re spon dents, schizo phre nia suf fer ers and fam ily mem - bers con tinue to ex pe ri ence so cial stigma as both the sin gle most im por tant fac tor un der min ing their qual ity of life and a key bar rier to care and treat ment com pli ance (24). The lack of an epi de mi ol ogy of felt stigma stigma as it is per ceived from the per spec tive of the suf ferer makes it im pos si ble ei - ther to meas ure the ex tent of the gap be tween what the public re ports and what per sons with schizo phre nia ex pe ri ence, or to lo cate ar eas of ob vi ous in con sis tency that could be the ob ject of more tar geted in ter ven tion. It also means that there is no di - rect meas ure of the im pact of public edu ca tion cam paigns where it counts the most. There fore, the lack of psy cho met ri - cally tested, well- validated in stru ments to cap ture so cial stigma from the per spec tive of schizo phre nia suf fer ers and their families is a ma jor limi ta tion (25). The de vel op ment of a range of such in stru ments would be an im por tant con tri bu tion to the field. Acknowledgements This work was undertaken as part of the activities of the World Psy - chi at ric As so cia tion s Global Program to Fight Stigma and Dis - crimi na tion Be cause of Schizo phre nia. Additional in for ma tion on the global ini tia tive may be obtained from the program s Web site, Funding for the Al berta pilot study was pro vided by Eli Lilly Inter - na tional and the Al berta Pro vin cial Mental Health Ad vi sory Board. The authors also acknowledge the important in tel lec tual and crea - tive support of the Lo cal Ac tion Committee for the Al berta pilot project and Closer Look Crea tive of Chi cago, US. References 1. Cro cetti G, Spiro HR, Saissi I. Are the ranks closed? At ti tu di nal so cial dis tance and men tal ill ness. Am J Psy chia try 1971;129: Malla A, Shaw T. At ti tudes to wards mental illness: the in flu ence of edu ca tion and ex pe ri ence. Int J Soc Psy chia try 1987;33: Ro sen field S. La bel ing mental illness: the ef fects of re ceived services and per - ceived stigma on life sat is fac tion. Ameri can So cio logi cal Re view 1997;62: Clinical Im pli ca tions Population-based campaigns that use generic messages are likely to be ineffective, given age-, sex-, and residence-based differences. Campaigns that provide fact-based knowledge may have difficulty registering gains if the public is as well informed as this sur vey sug - gests. Population-based campaigns should be approached cautiously be - cause they may reinforce the popular, but mistaken, belief that so cial stigma is not malleable to intervention. Limi ta tions Social-desirability bias may explain high levels of tolerance but not self-reported knowledge. The extent to which knowledge and positive attitudes reflect positive behaviours is unknown. Direct measures of stigma experienced by schizophrenia suf fer ers are unavailable, but they are ultimately key to understanding the di - rect effects of anti-stigma campaigns. 4. Me chanic D, McAlpine D, Ro sen field S, Davis D. Ef fects of ill ness at tri bu tion and de pres sion on the quality of life among per sons with serious mental ill ness. Soc Sci Med 1994;39: Brand RC, Clai born WL. Two stud ies of com para tive stigma: em ployer at ti tudes and prac tices to ward re ha bili tated con victs, men tal and tu ber cu lo sis pa tients. Com - mu nity Ment Health J 1976; 12: An ger meyer MC, Matschin ger H. Violent at tacks on public fig ures by persons suf - fer ing from psy chi at ric dis or ders. Eur Arch Psy chia try Clin Neu ro sci : Al brecht GL, Walker VG, Levy JA. So cial dis tance from the stig ma tized: a test of two theo ries. Soc Sci Med 1982;16: An ger meyer MC, Matschin ger H. Sozi ale dis tanz der bevölke rung ge genüber psy - chisch kranken. Gesund heits we sen 1996;58: Arkar H, Eker D. In flu ence of having a hos pi tal ized men tally ill member in the fam - ily on at ti tudes to ward men tal pa tients in Tur key. Soc Psy chia try Psy chi atr Epi de - miol 1992;27: Sar to rius N. Fight ing schizo phre nia and stigma: a new WPA edu ca tional pro gram [edi to rial]. Br J Psy chia try 1997;170: Du bey A. Dis pel ling the stigma of schizo phre nia: a global cam paign launches in Can ada. The Jour nal of Ad dic tion and Mental Health 1999;2(2): Dean JA, Cou lom bier D, Smith DC, Bren del KA, Ar ner TG, Dean AG. Epi Info. Ver sion 6. At lanta (GA): Cen tres for Dis ease Con trol and Pre ven tion; Cum ming E, Cum ming J. Closed Ranks. Cam bridge (MA): Har vard Uni ver sity Press; Leck wart JF. So cial dis tance: an im por tant vari able in psy chi at ric set tings. Psy - chia try 1968;31: D Arcy C. Opened ranks? Black foot re vis ited. In: Co burn D, D Arcy C, Tor rance GM, New P, edi tors. Health and Ca na dian so ci ety. 2nd ed. Rich mond Hill (ON): Fitz henry and White side; p Bo gar dus EM. Meas ur ing so cial dis tance. J Appl Sociol 1925;9: Flay BR, Bur ton D. Ef fec tive mass com mu ni ca tion strate gies for health cam paigns. In: At kin C, Wal lack L, edi tors. Mass com mu ni ca tion and pub lic health. New bury Park: Sage Pub li ca tions; p Rabe- Hesketh S, Ever itt B. A hand book of sta tis ti cal analy ses using Stata. New York: Chap man & Hall/CRC; Vogt PW. Dic tion ary of sta tis tics and meth od ol ogy. New bury Park: Sage Pub li ca - tions; Wolff G. At ti tudes of the me dia and the public. In: Leff J, edi tor. Care in the com - mu nity: il lu sion or re al ity? To ronto: John Wiley and Sons; p Hux ley P. Lo ca tion and stigma: a sur vey of com mu nity at ti tudes to mental illness. Part 1: en light en ment and stigma. Jour nal of Men tal Health 1993;2: Link B, Cul len F. Con tact with the mentally ill and per cep tions of how dan ger ous they are. J Health Soc Be hav 1986;27: Penn D, Guy nan K, Daily T, Spaulding W, Garbin C, Sul li van M. Dis pel ling the stigma of schizo phre nia: what sort of in for ma tion is best? Schizophr Bull 1994;20: Hol ley HL, edi tor. Qual ity of life meas ure ment in mental health. Ca na dian Jour nal of Com mu nity Mental Health 1998;(Spe cial Suppl 3). 25. Roman- Smith H. The de vel op ment of a self- report scale for the as sess ment of stigma and dis crimi na tion ex pe ri enced by in di vidu als with schizo phre nia [Msc The sis]. Uni ver sity of Cal gary: 1999.

8 252 The Canadian Journal of Psychiatry Vol 46, No 3 Rés umé Attitudes de la collectivité à l égard des personnes souffrant de schizophrénie Ob jec tif : Nous avons étudié les at ti tudes du pub lic à l é gard des per son nes souf frant de schizo phrénie dans le cadre d un projet pi lote de la cam pagne mondiale de lutte contre les stig mates et la dis crimi na tion causés par la schizo phré nie, menée par l As so - cia tion mon di ale de psy chia trie. Méth odes : Nous avons mené des sond ages té léphoniques à com po si tion aléa toire au près de répondants (de 15 ans ou plus) dans deux ré gions sani taires ad ja cen tes, l une ur baine et l autre ru rale (taux de réponse de 71,9 % ). Une courte en tre vue a per - mis de re cueil lir de l in for ma tion à pro pos de l expé ri ence des gens avec des per son nes souf frant de mala die men tale ou de schizo phré nie, des con nais sances des causes et des traite ments de la schizo phré nie et des ni veaux de dis tance so ci ale res sen tie en vers les schizo phrènes. Résul tats : La moitié de l é chan til lon con nais sait quelqu un qui avait été traité pour la schizo phré nie ou une autre mala die men - tale. Parmi ceux qui pou vaient iden ti fier une cause de la schizo phré nie (les deux tiers), la plupart men tion naient une cause bio lo - gique, habi tu el le ment une mala die du cer veau. La dis tance so ci ale s ac crois sait avec le de gré d in ti mité req uis. Un répon dant sur 5 croyait ne pas pouvoir en tretenir une ami tié avec une per sonne souf frant de schizo phré nie, la moitié se di sait in ca pa ble de part - ager un lo ge ment avec un ou une schizo phrène et les trois quarts n en épouse raient pas un ou une. Les per son nes de plus de 60 ans étaient les moins con nais seuses ou in struites, et celles qui mani fes taient le plus de dis tance sociale. Les con nais sances ac crues étaient as so ciées avec des at ti tudes moins dis tan tes. Lor squ on tenait compte d autres fac teurs, l ex po si tion à la mala die men tale n était pas corré lée avec les con nais sances ou les at ti tudes, même chez ceux qui avaient trav aillé dans des or gan ismes fournis sant des serv ices aux per son nes souf frant de mala dies men ta les. Con clu sions : La plupart des répon dants étaient rela tive ment bien in formés et évo lués dans ce qu ils di saient com pren dre de la schizo phré nie et de son traite ment. Des diffé rences de sous- groupes nettes étaient ap par en t es selon l âge et la con nais sance. La con nais sance de la schizo phré nie, et non l ex po si tion à la mala die men tale, était la prin ci pale corré la tion modi fi able des stig mates.