Canadian Psychiatric Inpatient Religious Commitment: An Association With Mental Health

Transcription

1 Orig i nal Re search Canadian Psychiatric Inpatient Religious Commitment: An Association With Mental Health Mar i lyn Baetz, MD 1, Da vid B Larson, MD 2, Gene Marcoux, MD 3, Rudy Bowen, MD 4, Ron Grif fin, PhD 5 Back ground: Re search in di cates that re lig ion may have a posi tive ef fect on cop ing and pos si bly en hance clini cal out comes. This study aims to de ter mine the level of re lig ious in - ter est of psy chi at ric in pa tients and to as sess whether re lig ious com mit ment has an im pact on se lected out come vari ables. Meth ods: There were 88 con secu tive adult pa tients (50% men) who were ad mit ted to a Ca na dian ter ti ary care psy chia try in pa tient unit and were in ter viewed about their re lig ious be liefs and prac tices. Pa tients with a Beck De pres sion score of 12 or more were in cluded for out come analy sis. Re sults: A to tal of 59% be lieved in a God who re wards and pun ishes, 27% had a high fre - quency of wor ship at ten dance, and 35% prayed once or more daily. More fre quent wor ship at tenders had less se vere de pres sive symp toms, shorter cur rent length of stay, higher sat is - fac tion with life, and lower rates of cur rent and life time al co hol abuse ( P < 0.05), when com pared with those with less fre quent or no wor ship at ten dance. In con trast, pri vate spiri - tu al ity was as so ci ated with lower de pres sive symp toms and cur rent al co hol use only ( P < 0.05), and prayer fre quency had no sig nifi cant as so cia tions. Dis cus sion: This study in di cates that cer tain re lig ious prac tices may pro tect against se ver - ity of symp toms, hos pi tal use, and en hance life sat is fac tion among psy chi at ric in pa tients. This is the first known Ca na dian study that ex am ines re lig ious com mit ment among psy chi - at ric in pa tients. (Can J Psy chia try 2002;47: ) Clini cal Im pli ca tions Ad dress the spiri tual as pect of pa tients lives whether posi tive, nega tive, or neu tral in or der to ad dress the per son as a whole. Sup port this po ten tially sig nifi cant cop ing re source. In volve spiri tual ad vi sors, such as the clergy or chap lains, when needed. Limi ta tions The small sam ple size was lim ited to 1 geo graphic lo ca tion. Cross- sectional as sess ment of re lig ious ness may not re flect past ac tiv ity. Only psy chi at ric in pa tients were sur veyed, so gen er aliz abil ity is lim ited. Key Words: spir i tu al ity, re li gion, men tal health, out comes, length of stay, life sat is fac tion W Can J Psy chia try, Vol 47, No 2, March

2 The Ca na dian Jour nal of Psy chia try Orig i nal Re search Over the last few de cades, Gal lup polls in the US and opin - ion polls in Can ada have con sis tently shown that a high pro por tion of the adult pop u la tion has mod er ate to high lev els of re li gious be liefs. A 1996 US Gal lup Poll (1) found that the 88% sur veyed felt re li gion was fairly to very im por tant in their lives. More over, a Ca na dian opin ion poll in 1993 (2) found that 78% affirmed a belief in God, with 67% as crib ing to the ba sic ten ets of Chris tian ity. Many stud ies have fo cused on the im pact of re li gious be liefs and practices on health out comes. These stud ies have been conducted with increasingly stringent accounting for con - found ing vari ables, namely, age, sex, so cial sup port, so cio - eco nomic sta tus, and health (3). Lit er a ture re views in di cate that, in general, higher lev els of re li gious in volve ment are moderately as so ci ated with better phys i cal and men tal health sta tus (3 8). Many of these stud ies have been epidemiologic in na ture, ini tially study ing well pop u la tions and de ter min ing in ci dence of phys i cal or men tal di ag no ses. In sev eral large stud ies, reg u lar wor ship at tend ers had lon ger life ex pec tancy (9,10), de creased ad verse con se quences of stressful life events on psy cho log i cal well-being (11,12), and en hanced psychosocial well-being (3). Cau sal ity can not be de ter mined, due to the cross-sectional na - ture of many of the stud ies. There is, how ever, an in creas ing body of ev i dence to in di cate that some forms of re li gious or spir i tual prac tices do en hance men tal health (13), in clud ing sev eral re cent pro spec tive anal y ses (14). Strawbridge and oth - ers (15), for ex am ple, in a lon gi tu di nal study with a 29-year fol low up in Alameda county, found weekly wor ship at ten - dance as so ci ated with im prov ing and main tain ing good men - tal health, increasing so cial re la tion ships, and enhancing mar i tal sta bil ity. Re cent re search has be gun to look at clin i cal pop u la tions, both in a cross-sectional and pro spec tive man ner. This has been pri mar ily among the med i cally ill el derly. Koenig and oth ers (16) have shown that, for the el derly, hav ing higher lev els of in trin sic re li gious ness not only en hanced their abil ity to cope with the stress of phys i cal ill ness but also showed faster re cov - ery times from major depressive illness in those who were hos pi tal ized with a comorbid med i cal ill ness. Among the el - derly who use hos pi tal ser vices and in those who were more fre quent church at tend ers, Koenig and Larson showed lower use of hos pi tal ser vices. For those who used them, lengths of stay were shorter, with un der stand able health-cost im pli ca - tions (17). Studies of re li gious be liefs and prac tices on hos pi tal ized psy - chi at ric pa tients found sim i lar rates of religious be liefs; in fact, in some cases, they were higher than those found in con - trol groups. These rates were based on small num bers and di - verse psy chi at ric di ag no ses (18,19). How these be liefs spe cif i cally im pact outcome has not yet been addressed. Finally, nearly all of these stud ies have been done on US sam - ples. It is im por tant to de ter mine whether these find ings gen - er al ize to psy chi at ric pop u la tions in other coun tries. This study of hos pi tal ized pa tients on a psy chi at ric ward fo - cuses on 2 ques tions. First, what is the level of re li gious com - mit ment? Sec ond, does re li gious com mit ment have an im pact on the level of men tal health, sat is fac tion with life, hos pi tal use, and al co hol use? Methods Pa tient Se lec tion A con sec u tive sam ple of pa tients 18 years or above, ad mit ted to ei ther of 2 ter tiary care psy chi at ric fa cil i ties be tween June and Au gust 1999, were in cluded in the study. All sub jects pro - vided writ ten in formed con sent. Ex clu sions were as fol lows: cog ni tive in ca pac ity, in abil ity to un der stand Eng lish, or too ill to par tic i pate based on the opin ion of the pa tient s nurse. Of 128 el i gi ble pa tients, 28 re fused to give con sent. To ex clude pa tients with ma nia or so cial ad mis sions such as those with shel ter needs, those with a Beck De pres sion score of less than 12 were re moved. Thus, there were 88 pa tients in cluded in the study. Pro ce dure The re search as sis tant was a med i cal stu dent trained in ad min - is ter ing the ques tion naires by a se nior psy chi a trist. We ob - tained con sent, and pa tients were in ter viewed within 72 hours fol low ing ad mis sion. Char ac ter is tics of the Sam ple De mo graphic variables were ex tracted from the patient s chart. So cial sup port was cat e go rized as 1) poor, 2) mod er ate, or 3) good, based on pa tient re port and the at tend ing phy si - cian s as sess ment. The study s psy chi a trists de ter mined the di ag no ses by chart re view, in de pend ent of knowl edge of the ques tion naire. The fol low ing mea sures were used. Mea sures of Men tal Health and Sat is fac tion We used the Beck De pres sion In ven tory (BDI) (20) as a self-report mea sure of de pres sive symp toms and to screen for those with scores of less than 12. Chart re view on dis charge de ter mined cur rent hos pi tal length of stay. Re view of prior hos pi tal re cords, as well as pa tient in ter view, de cided life time past psy chi at ric hos pi - tal stay. The Sat is fac tion with Life Scale (SWLS) (21), a 5-item scale that as sesses the sub jec tive-cognitive judg men tal as - pects of gen eral life sat is fac tion, scores each item from 1 to 7 in terms of strongly dis agree to strongly agree. Test re test re li abil ity of 0.82 and in ter nal con sis tency of al pha 0.87 has been found (21). 160 W Can J Psy chia try, Vol 47, No 2, March 2002

3 Canadian Psychiatric Inpatient Religious Commitment: An Association With Mental Health We de ter mined cur rent and past al co hol abuse by re view - ing hos pi tal re cords and con duct ing pa tient in ter views. Mea sures of Re li gious Com mit ment Re li gious com mit ment re fers to the par tic i pa tion in or en - dorse ment of prac tices, be liefs, at ti tudes, or sen ti ments that are as so ci ated with an or ga nized com mu nity of faith (6). The fol low ing items were cho sen to al low for com par i son with na - tional data and for time ex pe di ency, so that it would not bur - den the pa tients. Re li gious be liefs were ex am ined us ing 3 ques tions that were drawn from Gallup s Poll of Religious Be liefs (1). These items asked re spon dents to rate their be lief in a per sonal God who re wards and pun ishes, in the need to be born again, and in ev ery part of the Bi ble as the ac tual word of God. Re - sponses were on a 5-point scale from 1) not at all to 5) very much so. Re li gious prac tice was as sessed by the 5-item Duke Re li gion In dex (22). To as sess or ga ni za tional re li gious ness or so cial par tic i pa tion, we asked, How of ten do you at tend church/re - li gious meet ings? Fur ther, to as sess nonorganizational re li - gious ness, we asked the ques tion, How of ten do you spend time in pri vate re li gious ac tiv i ties that is prayer, med i ta tion? In trin sic re li gious ness was mea sured as a con tin u ous vari able by com bin ing 3 ques tions into a sin gle in dex that in cluded in my life I experience the di vine, re li gious be liefs are my whole ap proach to life, and I try hard to carry re li gion into all deal ings in life. The re sponses were rated on a scale of 1 to 5, from def i nitely not true to def i nitely true, for a max i mum score of 15. Cronbach s al pha for the Duke Re li gion In dex in the cur rent sam ple (n = 88) was The Re li gious Coping In dex (RCI) is a 3-item scale that mea - sures how much the pa tient re lies upon re li gion to help man - age the stress of ill ness (23). Item 1 is an open-ended ques tion that asks how in di vid u als cope with their cur rent stress. A re li - gious re sponse is scored a value of 10, with any other re - sponse, scor ing 0 to give equal weight with the other 2 items in the in dex. Item 2 asked pa tients to rate, on a scale of 0 to 10, their over all as sess ment of how much they use re li gion to cope. In Item 3, on a scale of 0 to 10, the in ter viewer rates to what ex tent pa tients use religion to cope with their life stresses. Data Anal y sis We used Pearson cor re la tions to de ter mine the bivariate re la - tion be tween the mea sures of re li gious com mit ment and the out come vari ables. Next, we used or di nary least squares re - gression tech nique to con trol for covariates. Causal mod els were de vel oped for each of the de pend ent vari ables, in clud ing BDI, SWLS, cur rent length of stay, past length of stay, cur rent alcohol abuse, and past alcohol abuse. The first model in - cluded both the de mo graphic (age, sex, ed u ca tion, and mar i tal sta tus) and so cial-support vari ables. The sec ond model added the BDI as a mea sure of se ver ity. The third model then added the re li gious com mit ment vari ables. Results Char ac ter is tics of the Sam ple There were 88 patients (50% men), with a mean age of 39 years (SD 13, range 18 to 86 years) who were in cluded in the study. Of the pa tients, 32% were mar ried or liv ing com mon law, while 68% were single, di vorced, or widowed. There were 50% who were de ter mined to have poor so cial sup port. Fur ther, 73% of the pa tients had com pleted high school or above. Of the pa tients, 67% were di ag nosed with a ma jor de pres sive ep i sode (26% bi po lar and 41% uni po lar). Of the 88, 33% had de pres sion comorbid with an other Axis I di ag no sis (that is, panic dis or der, eat ing dis or der, or sub stance abuse). The at - tend ing psy chi a trist di ag nosed 49% (n = 43) as hav ing a per - sonality dis or der or traits such as an ti so cial, borderline, or de pend ent. Mea sures of Men tal Health and Sat is fac tion There was a mean score of 32 (SD 11) on the BDI, which is in the se vere range. The range was 12 to 55, be cause only those in di vid u als with a BDI equal to or greater than 12 were in - cluded in the study. The cur rent length of stay was 12.1 days (SD 8.2, range 1 to 39). There was a mean of 3 prior life time hos pi tal iza tions for an av er age of 38 (SD 60, range 0 to 410) days; 33% ( n = 29) had never been pre vi ously hos pi tal ized. The SWLS had a mean value of 13 (SD 7, range 5 to 35). For com par i son, normed val ues are 23 to 25 in the el derly and col - lege pop u la tions, re spec tively (21). There were 16 (22%) who had a cur rent al co hol-abuse his tory, and 46 (53%) who had a past al co hol-abuse his tory. Re li gious Com mit ment Mea sures Re li gious be liefs, as mea sured by the ques tions adapted from the Gal lup Poll, re vealed that 59% ( n = 52) be lieve in a God who re wards and pun ishes, com pared with 78% of the Ca na - dian pop u la tion who be lieve in God (2). More fun da men tal be liefs, such as the need to be born again and be lief in the Bi - ble as lit eral, were 30% and 51%, re spec tively, com pared with 15% of the Ca na dian pop u la tion (2,24). Mea sures of re li gious prac tices in clude wor ship at ten dance, prayer fre quency, and 3 di men sions of in trin sic re li gious ness. There were 27% (n = 23) who at tended wor ship meet ings once or more weekly, and 35% (n = 31) who spent time in pri vate religious ac tiv ity (prayer, med i ta tion, or scrip ture study) more than once daily. In con trast, 25% of the Ca na dian pop u la tion W Can J Psy chia try, Vol 47, No 2, March

4 The Ca na dian Jour nal of Psy chia try Orig i nal Re search Table 1. Correlations of religious commitment with mental health and hospital-use outcomes Beck De pres sion In ven tory Length of stay in days Sat is fac tion with life scale Al co hol abuse Cur rent Past Cur rent Life time Worship attendance 0.22 a b 0.26 a 0.25 a Intrinsic religiousness 0.27 a b 0.31 b Prayer frequency 0.24 a a 0.26 a Religious coping 0.23 a a b P < 0.05; P < 0.01 Table 2. Correlations of religious commitment with mental health and hospital-use outcomes controlled for covariates (standardized betas) Beck De pres sion In ven tory Length of stay in days Sat is fac tion with life scale Al co hol abuse Cur rent Past Cur rent Lifetime Worship attendance 0.21 a 0.27 a b 0.23 a 0.18 a Intrinsic religiousness 0.23 a a 0.16 Prayer frequency Religious coping a 0.24 a a P < 0.05; b P < 0.01 en gages in these activities at about the same level (2,24). In - trin sic re li gious ness mea sured by the 3 items dis cussed above, had a mean of 9.42 (SD 3.7, me dian 9). The RCI, which is a com pos ite mea sure re flect ing a spon ta ne - ous re sponse and self- and ob server rat ings, had a mean of 9.7 (SD 8.2) on a scale of 0 to 30. For com par i son, a sur vey of US med i cally ill, el derly male pa tients had a mean RCI of 14.3 (SD 8.9) (25). There were 14% who spon ta ne ously re plied that re li gion was the pri mary fac tor that helped them cope when asked the open-ended ques tion, How do you cope with stress? On a vi sual an a logue scale of 0 to 10, 33% of pa tients rated them selves as 7.0 or higher in re sponse to the ques tion, To what ex tent does re li gion help you cope with your cur rent sit u a tion? Im pact of Re li gious Com mit ment Cor re la tions of re li gious com mit ment mea sures and the men - tal health and sat is fac tion as so ci a tions are re ported in Ta ble 1. Wor ship at ten dance, prayer fre quency, in trin sic re li gious - ness, and re li gious cop ing sig nif i cantly as so ci ate with sev eral out come vari ables, while be lief in God, be lief in the need to be born again, and be lief in the Bi ble as lit eral, do not. For this rea son, the lat ter 3 are not in cluded in Ta ble 1. Ta ble 2 in cludes the re sults of the hi er ar chi cal mul ti ple re gres - sion anal y sis. Each of the stan dard ized beta co ef fi cients re - ported rep re sents the degree of as so ci a tion be tween the par tic u lar di men sion of re li gious com mit ment and the men tal health or sat is fac tion out come when the other vari ables are held con stant. Wor ship at ten dance had a sig nif i cant im pact on more mea - sures of men tal health and ill ness than did any other di men - sion of re li gious com mit ment. Those who at tended wor ship ser vices more reg u larly re ported sig nif i cantly lower scores on the BDI, had shorter cur rent lengths of stay, and had lower rates of al co hol abuse, both cur rently and over their life time. Con versely, their pa tient coun ter parts who re ported low rates of wor ship at ten dance also re ported higher scores on the de - pres sion scale, higher lengths of hos pi tal stay, and more al co - hol abuse. In ad di tion, those who at tended wor ship ser vices more reg u larly also re ported sig nif i cantly greater sub jec tive sat is fac tion with life than did those who were in fre quent at - tend ers. In creases in the ex plained vari ance, with the in clu - sion of wor ship at ten dance into the base model, were sig nif i cant at the P < 0.01 level for the outcome vari - ables with the most sub stan tial be ing life sat is fac tion. 162 W Can J Psy chia try, Vol 47, No 2, March 2002

5 Canadian Psychiatric Inpatient Religious Commitment: An Association With Mental Health In trin sic re li gious ness also im pacts sig nif i cantly on both the rates of de pres sion and cur rent rate of al co hol abuse. Only fre - quency of prayer failed to reg is ter any sig nif i cant im pact on some as pect of psy cho logic dis tress. Re li gious cop ing was the only fac tor found to have a sig nif i - cant im pact on psy chi at ric length of stay, both cur rently and in the past. In ex am in ing the com po nents, we found the stron gest pre dic tor to be the ini tial re sponse to the ques tion, How do you cope with stress? Those re port ing a re li gious re sponse as the most im por tant coping fac tor also have a significantly lower past length of stay. In fact, add ing this 1 el e ment to the model in creased the ex plained vari ance in past length of stay by 9.5% (from 2.6% to 12.1%). Discussion Levels of Re li gious Com mit ment This group of psy chi at ric in pa tients ex pressed lev els of re li - gious com mit ment that were sim i lar to or higher than those in the Ca na dian pop u la tion. While there is no available Ca na - dian data for di rect com par i son to the ques tion about be lief in God who re wards and pun ishes, about 78% of the Ca na dian pop u la tion re ports a be lief in God (2). In our study, it is rea - son able to as sume that the ad di tion of re wards and pun ishes would lower the num ber, be cause fewer peo ple be lieve in Hell or the Devil (2,24). Sim i larly, Kroll and Sheehan re ported on a sur vey of 52 psy - chi at ric in pa tients in Min ne sota that found high rates of re li - gious be lief and ex pe ri ence and prac tice which cor re sponded to the lev els in the American population (18). In the UK, Neeleman and Lewis (19) also re ported that psy chi at ric pa - tients have higher lev els of re li gious be liefs and at ti tudes than do nonpsychiatric com par i son groups. The pa tients sur veyed in cluded 75 psy chotic in pa tients, de lib er ate self-harm in pa - tients, de pres sion out pa tients, and a com par i son group of 25 orthopedic out pa tients. The in ten sity of re li gious beliefs was most pro nounced among the more se verely ill sub jects. Our psy chi at ric in pa tient sam ple had a high in ci dence of psy chi at - ric comorbidity and prior hos pi tal iza tions, in di cat ing a high de gree of chronicity and se ver ity. These find ings, in clud ing our own, con trast with ear lier re ports that pa tients with men tal ill ness pro fess or prac tice less re li gious com mit ment (26). Out comes and Re li gious Com mit ment Our re sults re veal the types of re li gious com mit ment that sig - nif i cantly in flu enced the se lected out come vari ables (de pres - sive symp toms, length of hos pi tal stay, sat is fac tion with life, and al co hol use). These in cluded fre quency of wor ship at ten - dance, use of re li gion as an iden ti fied cop ing re source, and in - trinsic re li gious ness. Those re li gious measures that did not sig nif i cantly im pact out come vari ables were prayer fre quency and ex pressed lev els of re li gious be liefs. Re li gious Com mit ment and De pres sive Symp toms. This study dem on strates, af ter con trol ling for covariates, that the lev els of de pres sive symp toms were lower for those pa tients with more fre quent wor ship at ten dance and higher lev els of in trin - sic re li gious ness. In ac cor dance with our re sults, many stud ies have shown wor ship at ten dance to as so ci ate with lower lev els of de pres sive symp toms (3,6,7,27). In a strin gent re view of 29 pri mar ily nonclinical stud ies that ex am ine the cross-sectional re la tion of wor ship at ten dance and de pres sive symptoms, McCullough (5) found 24 were as so ci ated with lower lev els of de pres sive symp toms. Af ter con trol ling for 2 or 3 covariates, the association dropped, yield ing re gres sion co ef fi cients in the range of β = Con trolling for 6 covariates in our clin - i cal study, how ever, did not sig nif i cantly change the as so ci a - tion of wor ship attendance or in trin sic re li gious ness with de pres sive symp toms, and val ues re mained sig nif i cant at lev - els around β = In the clin i cal sit u a tion, wor ship ser vice at ten dance may have a more im por tant role, given the low mar i tal rates and poor so cial sup ports seen in our study. Re li gious Com mit ment and Length of Stay. A no ta ble find ing in our group of in pa tients was that the length of stay in the psy - chi at ric ward was sig nif i cantly shorter for those pa tients with more fre quent wor ship at ten dance and for those who used re li - gious thoughts or ac tiv i ties as the most im por tant strat egy to cope with their ill ness. Few stud ies have in cluded a re li gious vari able as a pre dic tor of men tal health ser vices out come (for ex am ple, length of stay). Re li gious af fil i a tion was found to be sig nif i cantly re - lated to length of stay in a gen eral hos pi tal psy chi at ric unit in Brazil (1992) (28), with out con trol for covariates. In a cost anal y sis of the im pact of re li gious faith and prac tice on pa - tients suf fer ing from de pres sion, church at ten dance showed an in verse re la tion to cost for health care (1995) (29). Koenig and Larson found that in a group of med i cally ill el derly, the more frequent church at tend ers had fewer days in hospital when they were ad mit ted and had fewer ad mis sions 1 year af - ter the in dex ad mis sion (17). Why worship at ten dance or religious cop ing should im pact length of stay or men tal-health re cov ery in gen eral is a com - plex question, and studies are vir tu ally nonexistent (30). There are at least 3 pos si ble mech a nisms through which wor - ship at ten dance may work to pro mote better men tal health and quicker recovery from mental illness (30,31). First, it may reg u late life style and be hav iours (for ex am ple, dis cour ag ing use of ad dic tive sub stances). Sec ond, it may en hance so cial and cop ing re sources, be cause fel low ship is of ten an ex plicit part of the or ga ni za tional man date of many faith groups. In fact, many of this sam ple were liv ing alone, and 50% had poor so cial sup ports. Third, re li gion may pro vide a sense of co her - ence and mean ing to life. Un der stand ing one s mean ing or W Can J Psy chia try, Vol 47, No 2, March

6 The Ca na dian Jour nal of Psy chia try Orig i nal Re search pur pose in life and find ing mean ing in suf fer ing were found to ac count for a sig nif i cant pro por tion of the re la tion be tween re - li gious in volve ment and health (30,31). Re li gious Com mit ment and Sat is fac tion with Life. Wor - ship-attendance fre quency was the only re li gious pre dic tor of in creased life sat is fac tion. Our re sults re veal that re li giously ac tive peo ple re port greater life sat is fac tion or hap pi ness a find ing that has been generally consistent in pre vi ous re - search. In a metaanalytic study of pa pers that are now over 20 years old, Witter and oth ers (32) found re li gious ac tiv ity pos i - tively re lated to well-being. More re cently, na tional polls in both the US (1995) (33) and Can ada (1996) (34) found highly re li gious peo ple nearly twice as likely as those low est in re li - gious com mit ment to de clare them selves very happy. Fur ther, re gard less of ex pe ri enc ing ma jor neg a tive life events, those who re ported a closer re la tion ship with God in di cated more hap pi ness and life sat is fac tion (26). Re li gious Com mit ment and Al co hol Abuse. This study adds to a wealth of lit er a ture that has con sis tently re vealed an in verse re la tion be tween re li gious ness and al co hol use in both pop u la - tion sur veys and clin i cal stud ies (11,27,35,36). Cur rent al co - hol abuse was sig nif i cantly and neg a tively re lated to wor ship at ten dance and in trin sic re li gious ness. More over, wor ship at - tendance was negatively associated with life time al co hol abuse. It is pos si ble that wor ship at ten dance es pe cially if a life long ac tiv ity may be im por tant when de cid ing to start us - ing sub stances such as al co hol. Ex pe ri encing higher lev els of in trin sic re li gious ness may be more im por tant than at ten dance alone in achiev ing the abil ity to quit drink ing or to main tain low lev els of use (11). Re li gious Be liefs and Prayer Fre quency. The re sults of this study re veal that in trin sic re li gious ness can sig nif i cantly im - pact the rates of de pres sion and cur rent al co hol use. We can only spec u late, how ever, upon why re li gious be liefs are not correlated with outcome but re li gious prac tices are. Mea - suring be liefs us ing a sim ple ques tion naire may be vague and thus not be mean ing ful as an out come mea sure (for ex am ple, 88% of the US population be lieves in God). In a study of nearly 2000 twins, Kendler (11) also found that re li gious be - liefs had no cor re la tion with out come but that re li gious prac - tices did pre dict lower lev els of de pres sive symp toms, cur rent and life time risk for al co hol ism, and buff er ing of depressogenic ef fects of stress ful life events. We found that prayer frequency did not sig nif i cantly in flu - ence de pres sion rates. Cor re sponding to our find ings, McCullough (5) noted that, in the study of de pres sion, pri vate re li gious ac tiv ity has a ten u ous re la tion and any as so ci a tions are small. It had been sug gested that prayer can be used in var i ous ways, in clud ing a pas sive man ner, which leaves the re spon si bil ity for re solv ing a cri sis en tirely up to di vine in ter - ven tion with out mo bi liz ing per sonal re sources (3). Limitations Lo ca tion at 1 site and the small sam ple size limit the study; how ever, it does have equal sex dis tri bu tion. Al though all pa - tients had de pres sion symp toms, there was a range of comorbid di ag no ses, which may con ceal pat terns as so ci ated with a more ho mo ge neous group, but cer tainly con firms the se ver ity of this study sam ple. It is lim ited to an in pa tient pop u - la tion, and we do not know if these are generalizable to psy - chi at ric pa tients in general. There were 28 pa tients who re fused; there fore, it is pos si ble that there was a se lec tion in - fluence and that those with religious in ter ests were more likely to agree to par tic i pate. The re sults pre sented also rely on the as sess ment of re li gious com mit ment at a sin gle time point and may not re flect past lev els of ac tiv ity (11). This is per ti nent when ex am in ing re li - gious com mit ment in re la tion to past length of stay and past lev els of al co hol abuse. These lat ter 2 mea sures are also af - fected by re li ance, in part, on pa tient re port, de spite our at - tempt to cor rob o rate find ings ob jec tively in as many cases as pos si ble. As we noted in the in tro duc tion, while the cross-sectional nature does not allow for de ter mi na tion of cau sal ity, it is con sis tent with the many studies that have shown a pos i tive men tal health ben e fit (13), in clud ing more re cent lon gi tu di nal stud ies (14,15). The so cial-support mea sure ment was based on a sub jec tive eval u a tion by pa tients and their psy chi a trists. Given the im - por tance and fre quent low level of this vari able, an as sess ment of pa tient psy chi a trist concordance might have been valu - able. Finally, worship-attendance fre quency had a con sis - tently pos i tive im pact on out come, but not know ing which func tion ac counts for the salutory mech a nism may un der es ti - mate the to tal in flu ence of the re li gious vari able (31). Conclusions Re li gious in ter ests of pa tients are not usu ally con sid ered in psy chi at ric care. This study is the first known Ca na dian study to ex am ine lev els of spir i tu al ity and re li gious ness among psy - chi at ric in pa tients. We have dem on strated a level of re li gious com mit ment in psy chi at ric in pa tients that is at least sim i lar to or higher than that found in the gen eral pop u la tion. Our re sults show that re li gious com mit ment has a sig nif i cant im pact on de pres sive symp toms, sat is fac tion with life, hos pi tal use, and al co hol use. These find ings add to the lit er a ture on pos i tive 164 W Can J Psy chia try, Vol 47, No 2, March 2002

7 Canadian Psychiatric Inpatient Religious Commitment: An Association With Mental Health men tal health ben e fits with re li gious ac tiv ity and ex pand it to a clin i cal psy chi at ric pop u la tion, al beit on a small level. Fur ther stud ies, lon gi tu di nal in na ture, based on larger num - bers of pa tients, and more ho mo ge neous pa tient groups, while ac count ing for covariates, will help to fur ther im prove our un - der stand ing of the im pact of re li gious be liefs and spir i tu al ity on men tal health. These find ings re veal the need to be gin ad - dress ing pa tient spir i tu al ity, whether helpful, neu tral, or harm ful and, par tic u larly when harm ful or com plex, to in - volve chap lains or clergy (37 39). Given the US find ings of the lack of re la tion be tween psy chi a try and spir i tual ad vi sors (37), we hope Can ada will not fol low a sim i lar tra jec tory, but will in stead work toward improving col lab o ra tion be tween these im por tant groups. Ac knowl edge ments Sup port for the writ ing of this paper was pro vided by the John Templeton Foun da tion, Radnor, Penn syl va nia and Mon arch Phar - ma ceu ti cals, a wholly-owned sub sid iary of King Phar ma ceu ti cals, Inc, Bris tol, Ten nes see. Thanks to Kerri Schellenberg Med IV for her as sis tance in in ter view ing and data col lec tion. References 1. Prince ton Re li gion Re search Cen ter. Re li gion in Amer ica. Gal lup Poll Prince ton (NJ). 2. Nemeth M, Under wood N, and Howse J. Spe cial re port: the re li gion poll God is alive. MacLean s 1993 April: Ellison CG, Levin JS. The re li gion-health con nec tion: ev i dence, the ory, and fu - ture di rec tions. Health Educ Behav 1998;25: Koenig HG. Hand book of re li gion and men tal health. San Diego (CA): Ac a demic Press; McCullough ME, Larson DB. Re li gion and de pres sion: a re view of the lit er a ture. Twin Res 1999;2: Matthews DA, McCullough ME, Larson DB, Koenig HG, Swyers JP, Greenwold Milano M. Re li gious com mit ment and health sta tus: a re view of the re search and im pli ca tions for fam ily med i cine. Arch Fam Med 1998;7: Koenig HG, Hays JC, George LK, Blazer DG, Larson DB, Landerman LR. Modeling the cross-sectional re la tion ships be tween re li gion, phys i cal health, so - cial sup port, and de pres sive symp toms. Am J Geriatr Psy chi a try 1997;5: Larson DB, Sherrill KA, Ly ons JS, Crai gie FC, Thielman SB, Greenwold MS. As so ci a tions be tween di men sions of re li gious com mit ment and men tal health re - ported in the Amer i can Jour nal of Psy chi a try and Ar chives of Gen eral Psy chi a - try: Am J Psy chi a try 1992;149: Hum mer RA, Rog ers RG, Nam CB, Ellison CG. Re li gious in volve ment and US adult mor tal ity. De mog ra phy 1999;36: McCullough ME, Hoyt WT, Larson DB, Koenig HG, Thoreson C. Re li gious in - volve ment and mor tal ity: a meta-analytic re view. Health Psychol 2000;19: Kendler KS, Gardner CO, Prescott CA. Re li gion, psychopathology, and sub - stance use and abuse: a multi-measure, ge netic-epidemiologic study. Am J Psy - chi a try 1997;154: Wil liams DR, Larson DB, Buck ler RE, Heckmann RC. Re li gion and psy cho log i - cal dis tress in a com mu nity sam ple. Soc Sci Med 1991;32: Koenig HG, McCullough ME, Larson DB. Hand book of re li gion and health. New York: Ox ford Uni ver sity Press; Koenig HG, Larson DB. Re li gion and men tal health: ev i dence for an as so ci a tion. Int Rev Psy chi a try 2001;13: Strawbridge WJ, Shema SJ, Co hen RD, Kaplan GA. Re li gious at ten dance in - creases sur vival by im prov ing and main tain ing good health be hav iours, men tal health, and so cial re la tion ships. Ann Behav Med 2001;23: Koenig HG, George LK, Pe ter son BL. Re li gi os ity and re mis sion of de pres sion in med i cally ill older pa tients. Am J Psy chi a try 1998;155: Koenig HG, Larson DB. Use of hos pi tal ser vices, re li gious at ten dance, and re li - gious af fil i a tion. South ern Med i cal Jour nal 1998;91: Kroll J, Sheehan W. Re li gious be liefs and prac tices among 52 psy chi at ric in pa - tients in Min ne sota. Am J Psy chi a try 1989;146: Neeleman J, Lewis G. Re li gious iden tity and com fort be liefs in three groups of psy chi at ric pa tients and a group of med i cal con trols. Int J Soc Psy chi a try 1994;40: Beck AT, Steer R, Garbin M. Psychometric prop er ties of the Beck de pres sion in - ven tory: twenty-five years of eval u a tion. Clin Psychol Rev 1988;8: Diener E, Emmons RA, Larsen RJ, Grif fin S. The sat is fac tion with life scale. J Pers As sess 1985;49: Koenig HG, Parkerson GR, Meador KG. Re li gion in dex for psy chi at ric re search. Am J Psy chi a try 1997;153: Koenig HG, Co hen HJ, Blazer DG, Pieper C, Meador KG, Shelp F. Re li gious cop ing and de pres sion among el derly, hos pi tal ized med i cally ill men. Am J Psy - chi a try 1992;149: Bibby R. Frag mented Gods: the pov erty and po ten tial of re li gion in Can ada. To - ronto: Irwin Pub lishing; Koenig HG, Co hen HJ, Blazer DG, Kudler HS, Krishnan KRR, Sibert TE. Re li - gious cop ing and cog ni tive symp toms of de pres sion in el derly med i cal pa tients. Psychosomatics 1995;36: Ken nedy GJ: Re li gion and de pres sion. In: Koenig HG, ed i tor. Hand book of re li - gion and men tal health. San Diego (CA): Ac a demic Press; p Gart ner J, Larson DB, Al len GD. Re li gious com mit ment and men tal health: a re - view of the em pir i cal lit er a ture. Jour nal of Psy chol ogy and The ol ogy 1991;19: Dalgalarrondo P, Gattaz WF. A psy chi at ric unit in a gen eral hos pi tal in Brazil: pre dic tors of length of stay. Soc Psy chi a try Psychiatr Epidemiol 1992;27: Berg GE, Fonss N, Reed AJ, VandeCreek L. The im pact of re li gious faith and prac tice on pa tients suf fer ing from a ma jor af fec tive dis or der: a cost anal y sis. J Pas to ral Care 1995;49: George LK, Larson DB, Koenig HG, McCullough ME. Spir i tu al ity and health: what we know, what we need to know. Jour nal of So cial and Clin i cal Psy chol ogy 2000;19: Levin JS, Chat ters LM. Re search on re li gion and men tal health: an over view of em pir i cal find ings and the o ret i cal is sues. In: Koenig HG, ed i tor. Hand book of re - li gion and men tal health. San Diego (CA): Ac a demic Press; p Witter RA, Stock WA, Okun MA, Haring MJ. Re li gion and sub jec tive well-being in adult hood: a quan ti ta tive syn the sis. Rev Relig Res 1985;26: Myers DG, Diener E. The pur suit of hap pi ness. Mys teries of the mind Sci en tific Amer i can 1996;70: Clark W. Re li gious ob ser vance: mar riage and fam ily. Sta tis tics Can ada 1998; Cat a logue 11008XPE: Booth J, Mar tin JE. Spir i tual and re li gious fac tors in sub stance use, de pend ence, and re cov ery. In: Koenig HG, ed i tor. Hand book of re li gion and men tal health. San Diego (CA): Ac a demic Press; p Koenig HG, George LK, Meador KG, Blazer DG, Ford SM. Re li gious prac tices and al co hol ism in a South ern adult pop u la tion. Hos pi tal and Com mu nity Psy chi - a try 1994;45: Larson DB, Hohmann AA, Kessler LG, Meador KG, Boyd JH, McSherry E. The couch and the cloth: the need for link age. Hosp Com mu nity Psy chi a try 1988;39: Puchalski CM, Larson DB, Lu FG: Spir i tu al ity courses in psy chi a try res i dency pro grams. Psychiatr Ann 2000;30: Josephson AM, Larson DB, Juthani N. What s hap pen ing in psy chi a try re gard ing spir i tu al ity? Psychiatr Ann 2000;30: Manuscript received February 2001, revised August 2001, and accepted November As sis tant Pro fes sor, De part ment of Psy chi a try, Uni ver sity of Sas katch e wan, Saskatoon, Sas katch e wan. 2 Di rec tor, Na tional In sti tute for Healthcare Re search, Rockville, Mary land. 3 Clin i cal As sis tant Pro fes sor, De part ment of Psy chi a try, Uni ver sity of Sas - katch e wan, Saskatoon, Sas katch e wan. 4 Pro fes sor and Head, De part ment of Psy chi a try, Uni ver sity of Sas katch e wan, Saskatoon, Sas katch e wan. 5 As sis tant Pro fes sor, De part ment of So ci ol ogy, St Thomas More Col lege, Uni ver sity of Sas katch e wan, Saskatoon, Sas katch e wan. Address for cor re spon dence: Dr Mar i lyn Baetz, De part ment of Psy chi a try, Room 111, Ellis Hall, University of Saskatchewan, 103 Hospital Drive, Saskatoon, SK S7N 0W8 W Can J Psy chia try, Vol 47, No 2, March

8 The Ca na dian Jour nal of Psy chia try Orig i nal Re search Rés umé : Engagement religieux des patients psychiatriques canadiens hospitalisés : une association avec la santé mentale Con texte : La re cher che in dique que la re lig ion peut avoir un ef fet fa vor able sur l a dap ta tion et qu elle pour rait éven tu el le ment amé liorer les résul tats clin iques. Cette étude vise à déter mi ner le ni veau d inté rêt re li gieux chez des pa tients psy chia tri ques hos pi tal isés et à éval uer si l en gage ment re li gieux a un ef fet sur des vari ables de résul tat choisies. Méth odes : Il y avait 88 pa tients adul tes (50 % des hommes) qui ont été hos pi tal isés de façon consé cu tive dans une unité ca na di enne de soins psy chia tri ques ter ti aires et qui ont été i n ter viewés sur leurs croy ances et leurs pra tiques re li gieuses. Les pa tients qui ont eu 12 ou plus au ques tion - naire de dépres sion de Beck ont été in clus dans l ana lyse de résul tat. Résul tats : Un to tal de 59 % croyaient en un Dieu qui ré com pense et qui pu nit, 27 % as sis taient fréquemment aux serv ices re li gieux, et 35 % pri aient une fois par jour ou plus. Ceux qui as sis taient sou vent aux serv ices re li gieux avaient des symptômes dépres sifs moins graves, une du rée moins longue du séjour cou rant, une plus grande sat is fac tion quant à la vie, et des taux moins élevés d abus d al cool cou rant et de du rée de vie ( P < 0,05), com para tive ment à ceux qui fréquen tent moins ou pas du tout les serv ices re li gieux. Par con tre, la spiri tu al ité privée était as so ciée seule - ment aux symptômes dépres sifs moin dres et à la con som ma tion d al cool cou rante (P < 0,05). La fréquence de la prière ne pré sen tait pas d as so cia tions sig ni fi ca tives. Dis cus sion : Cette étude in dique que cer taines pra tiques re li gieuses peu vent pro té ger con tre la gravité des symptômes et le re cours aux hôpi taux, et qu elles peu vent ac croître la sat is fac t ion face à la vie des pa tients psy chia tri ques hos pi tal isés. La pré sente est la pre mière étude ca na di enne con - nue qui ex am ine l en gage ment re li gieux des pa tients psy chia tri ques hos pi tal isés. 166 W Can J Psy chia try, Vol 47, No 2, March 2002