| | Treatment seeking for obsessive-compulsive disorder: Role of obsessive-compulsive disorder symptoms and comorbid psychiatric diagnoses☆☆☆★Abstract Previous research has indicated that individuals afflicted with obsessive-compulsive disorder (OCD) have a very low rate of seeking help from mental health professionals. From standardized psychiatric interviews of 7,214 residents of Edmonton, Canada, we identified 172 subjects with a lifetime diagnosis of OCD; 63 (36.6%) had consulted a doctor about their symptoms. Total number of OCD symptoms (odds ratio [OR] = 2.23) and severe obsessions of violence and other unpleasant thoughts (OR = 2.52) were significantly associated with treatment seeking in multivariate analysis. The absence of association between compulsions and treatment seeking was corroborated by a very low prevalence of treatment seeking (17.3%) in individuals suffering from compulsions only. Our findings suggest that there is a need to teach the public about compulsions, and such a strategy may enhance future public health education programs. Copyright 2003, Elsevier Science (USA). All rights reserved.
Once considered to be a relatively rare illness, obsessive-compulsive disorder (OCD) has been demonstrated to be moderately widespread with a lifetime prevalence of 2.6% according to the Epidemiologic Catchment Area (ECA) study,1 and cross-national epidemiology reporting a lifetime prevalence worldwide of approximately 2%.2 This figure is considerably higher than the previously recognized prevalence rate of 0.05%, which was the most quoted estimate until the early 1980s.2, 3 In addition, OCD shows high rates of comorbidity with several other disorders: the most prevalent is major depression, the two coexisting in 67% of OCD cases.2 OCD also tends to co-occur with simple phobia (22%), social phobia (18%), eating disorders (17%), alcohol abuse or dependence, panic disorder, Tourette's syndrome, and schizophrenia.2, 4
The emergence of OCD can greatly reduce an individual's quality of life5; it is frequently chronic and disabling, with significant impact on the lives of OCD sufferers, their families, friends, employers, and society in general. Left untreated, obsessions and compulsions cause a significant amount of distress, are time consuming, and have a substantial negative effect on the sufferer's interpersonal relationship and career, often culminating in isolation.
Although OCD was initially considered among the most intractable of the neurotic disorders, recent studies have ascertained that OCD can be successfully treated.5 However, despite the immense social and economic costs associated with the disorder and the availability of treatment options, it has been shown that individuals afflicted with OCD have a very low rate of seeking help from mental health professionals.5, 6, 7 The aforementioned ECA study found that only 34% of individuals with OCD had ever mentioned their symptoms to a doctor,8 and that only one fifth had sought help from a mental health specialist.9
Consequently, an important question pertaining to OCD is: if the rate of impairment is so high, why is the rate of service utilization low? Prior studies have examined prevalence rates of treatment seeking and user services, but our literature search revealed that no previous studies have investigated the potential predictors of help-seeking. In this study, we examine the association of OCD symptoms and comorbid psychiatric diagnoses with treatment seeking.
Method  The data for this study originated from a survey of adult household residents of Edmonton, Canada, using a two-stage random design, with sampling of households in the first stage, and subsequent selection of one respondent in each household. To be eligible for the survey, respondents had to have been 18 years of age or older at the time of interview and a usual occupant at the address. All subjects participated voluntarily after verbal and written consent was obtained. The design and field methods are described in further detail elsewhere.10 The interviews were completed in two stages: first, a sample consisting of 3,258 individuals were administered the Diagnostic Interview Schedule (DIS version III) between January 1983 and May 1986; second, a sample of 3,956 individuals were administered the DIS (version IIIa) between December 1984 and February 1989 using identical methods. Therefore, a total of 7,214 individuals were interviewed. Data were screened to exclude the presence of double interviews (i.e, the same individual being recruited in both stages). The response rate (the proportion of eligible addresses at which an interview was obtained) was equal to 71.6%. Nonresponse was categorized as due to refusal (17.3%), contact not being made at an address (8.8%), and language barriers (2.3%). A detailed examination of the prevalence of psychiatric diagnoses observed in the first stage showed that the great majority of the results were very close to ECA findings.11 Instrument The DIS instrument is a structured and standardized diagnostic interview, administered by a trained lay interviewer, which elicits information about psychiatric symptoms at the time of interview and during the lifetime of the respondent, including questions about onset and offset of psychiatric illnesses and counts of symptoms for a particular illness.12 The DIS data were analyzed using a computer program available from the authors of the instrument. This program produces diagnoses of psychiatric disorders that are the DIS version of hierarchy-free DSM-III diagnoses. The DIS/DSM-III diagnoses covered in version III of the DIS include OCD, mania, major depression, dysthymia, phobia, somatization, alcohol abuse or dependence, drug abuse or dependence, antisocial personality, schizophrenia, and schizophreniform disorder. Version IIIa also includes items pertaining to generalized anxiety disorder, post-traumatic stress disorder (PTSD), and bulimia, and incorporates slightly revised diagnoses for alcohol abuse/dependence, dysthymia, and schizophrenia. OCD section of the DIS The DIS is based on a hierarchical structure. If an interviewee responds positively or negatively to a question he/she may be exempt from other questions or conversely he/she may have to answer more questions on that topic. Consequently, not all respondents will be asked the same questions. In the OCD section of the DIS, a minimum of three and a maximum of five OCD symptom questions are asked of each respondent. The questions are organized such that the maximum number of OCD symptoms a respondent can be assigned is three. A positive diagnosis requires at least one obsession or one compulsion. The DIS includes two questions concerning obsessions.13 The first inquires whether the respondent has had any persistent and unpleasant thoughts. The example given is “…the persistent idea that you might harm or kill someone you loved, even though you really didn't want to.” If the respondent does not report having such thoughts, a second question is asked that gives two examples of obsessions: thoughts that one's hands are dirty or have germs no matter how much they are washed, and that relatives who are away may have been hurt or killed. If the respondent replies affirmatively to the first question (obsessions of violence), this second question is not asked. The same duration (3 weeks) and severity (continued thoughts despite attempts to get rid of them) criteria apply to both questions. For individuals who report either of these obsessions, the age of onset and recency of the obsession are determined. The DIS contains three questions concerning compulsions.13 The first question states: “Some people have problems with feeling that they have to do something over and over again even though they know it is really foolish—but they can't resist doing it—things like washing their hands again and again or going back several times to be sure they've locked a door or turned off the stove. Have you ever had to do something like that over and over?” If the respondent replies negatively, another question is asked: whether she/he has to do something in a particular order, such as getting dressed, and has to do it over if it is done out of order. All respondents are then asked a third question pertaining to compulsive counting, despite trying to prevent such behavior. The example of counting floor tiles is presented. The same duration criterion (3 weeks) applies to the three questions, and age of onset and recency of the compulsion(s) is determined. The reliability and validity of the OCD section of the DIS have been extensively reviewed by Karno et al.13 Reliability appears satisfactory while validity is more limited: when compared to physician diagnosis, the lay-administered DIS showed an excellent specificity but a low sensitivity. The relatively low level of clinician-DIS agreement has been attributed to the low base rates of the disorder in the community and to the tendency for positive cases to be less severe than in clinical settings, with community cases clustering near the threshold of the diagnostic criteria to a greater extent than cases seen in clinical settings.13 Definition of treatment seeking In our study, treatment seeking was defined as having ever consulted a doctor (including medical doctors, psychiatrists, osteopaths, and medical students) about one's obsessions and/or compulsions. A non-treatment seeker was an individual afflicted with the disorder who has never discussed his or her OCD symptoms with a doctor. It should be noted that the fact that a respondent brought his/her OCD symptoms to the attention of a doctor does not imply that this subject was diagnosed and received treatment for these symptoms. In other words, we do not know whether this treatment-seeking behavior led to subsequent use of medical services. Definition of comorbidity Previous research has demonstrated that comorbidity can increase the likelihood of treatment utilization.14 It was therefore hypothesized that the likelihood of treatment seeking for OCD could be influenced by the existence of a comorbid disorder. Furthermore, we postulated that the influence of the comorbid disorder would be greater if there was evidence of a time overlap between the two disorders. For the purpose of this study we defined disorders as comorbid with OCD if the age of onset of the later disorder was lower or equal to the age of offset of the earlier disorder, indicating a temporal overlap between the two disorders. Detailed examination of the data indicated that, among those with other lifetime diagnoses, the proportion of cases that did not show a time overlap with OCD and thus were excluded by our comorbidity definition was low: null for four diagnoses (phobia, somatization, schizophreniform, bulimia), less than 10% for six other diagnoses (depression, mania, panic disorder, alcohol abuse/dependence, antisocial personality, schizophrenia), and between 10 and 20% for the three remaining diagnoses (drug abuse/dependence, generalized anxiety, PTSD). The inclusion of these cases of comorbidity without temporal overlap in our analyses would have resulted in a slight reduction of the degree of significance for a few of the associations between comorbid diagnoses and treatment seeking. Statistical analysis Data analysis was conducted using SPSS version 9.0 (SPSS Inc, Chicago, IL). Comparisons of proportions and means in the treatment seeking and non-treatment seeking groups were made using the chi-square test and independent samples ttest, respectively. Multivariate logistic regression (both backward elimination and forward selection) was carried out with treatment seeking (yes or no) as the outcome variable and a number of explanatory variables, including all of the symptoms and comorbid diagnoses associated bivariately with treatment seeking with a P value of .10 or less. The only exception to this strategy concerned the question about “severe obsessions of contamination and/or doubt,” which was only asked to 66 participants because of the hierarchical nature of the DIS. We ran a multivariate analysis with these 66 persons and this variable did not reach significance; it was consequently excluded from further analysis. The P value for inclusion in the multivariate models was set at .05, and a P value less than .051 was required to remain in the models. The multivariate logistic regression was carried out on two separate groups of subjects: with the entire sample of 172 subjects with a DIS diagnosis of OCD, and then with the subset of 69 participants who were administered version IIIa of the DIS. This was done in order to include those diagnoses not contained in DIS version IIIa, i.e, generalized anxiety disorder, PTSD, and bulimia. The data used in this study were not weighted. Weighting was not necessary because we examined associations within the sample rather than forming general population estimates. In addition, weights were generated in the original survey with respect to age and gender, and both variables were used as adjustment variables in our multivariate analysis. Results Demographic characteristics and frequency of symptoms In order to meet the criteria for OCD, the study participant must have been afflicted with at least one obsessive-compulsive symptom. In total, 172 of the 7,214 respondents (2.4%) met this criterion and were therefore included in our sample. The frequency distribution revealed that 138 participants (80.2%) had one obsessive-compulsive symptom, 23 (13.4%) had two symptoms, and 11 (6.4%) had three OCD symptoms. The sample was composed of 60 males (34.9%) and 112 females (65.1%). The mean age at interview was 36.3 years, and the mean age of onset was 19.2 years. Frequency of comorbid diagnoses The frequency distribution for the total number of comorbid disorders in our sample was as follows: 19.2% had no comorbid disorders, 24.4% had one, 24.4% had two, 14.0% had three, and 18.0% had four or more comorbid disorders. The frequency at which OCD co-occurred with other DIS/DSM III psychiatric diagnoses was analyzed. As previously mentioned, comorbidity was operationalized as a temporal overlap between the two disorders. Lifetime rates in our OCD sample in rank order of prevalence were as follows: depression 45.9%, phobia 37.8%, alcohol abuse/dependence 34.3%, generalized anxiety 31.9%, drug abuse/dependence 20.9%, panic disorder 12.8%, PTSD 11.6%, schizophrenia 9.9%, antisocial personality 9.3%, mania 6.4%, bulimia 2.9%, schizophreniform 0.6%, and somatization 0.6% (the percentages for generalized anxiety disorder, PTSD, and bulimia were derived from those subjects who were administered the DIS version IIIa, n = 69). Multivariate analysis In a first step, a backward elimination procedure was performed using the entire sample of 172 participants interviewed with DIS version III, thereby excluding comorbid generalized anxiety and PTSD. This procedure resulted in a final model containing four variables: sex and age (both forced into the model), “total number of OCD symptoms,” and “severe obsessions of violence and other unpleasant thoughts” (Table 4).
| | |  | Variables in the Final Model* | Adjusted OR (95% CI) | P Value in Model | |
|---|
| Sex (female v male)† | 1.20 (0.60, 2.41) | .60 | |
| Age of subject† | 1.00 (0.98, 1.03) | .69 | |
| Severe obsessions of violence and other unpleasant thoughts | 2.52 (1.24, 5.14) | .011 | |
| Total no. of OCD symptoms | 2.23 (1.24, 4.02) | .008 | |
| *Variables not retained in the final model: comorbid depression, comorbid mania, comorbid panic disorder, total number of comorbid diagnoses. †Control variable (forced into the model). | | | | |
None of the comorbid diagnoses significant in bivariate analyses were retained in the final model, and the total number of comorbid diagnoses was also removed in the multivariate procedure. In a second step, a forward selection procedure was also performed in order to validate these findings. The results were identical to those of the backward elimination procedure using the full sample. In a third step, a backward elimination procedure was performed exclusively on the 69 participants who were interviewed with DIS version IIIa, thereby including the diagnoses of generalized anxiety and PTSD. The results were very similar to those obtained on the entire sample, with the two additional comorbid diagnoses not retained in the final model. Discussion To the best of our knowledge, this is the first study to examine which specific symptoms and comorbid psychiatric diagnoses influence treatment seeking for OCD. A further advantage of our study is the large sample of participants, which was derived from the general population. In addition, a structured and standardized diagnostic procedure was utilized, which enhances the reliability and validity of our findings. In this study, 2.4% (172 of 7,214) of respondents met the criteria for OCD according to the DIS. This finding is consistent with the recent literature indicating that lifetime prevalence of OCD is approximately 2% in the general population.2 We found that 37% (63/172) of the participants with OCD had brought their OCD symptoms to the attention of a doctor or other health professional over the course of their lifetime. This finding is congruent with the ECA study, in which it was reported that 34% of individuals with a lifetime diagnosis of OCD consulted a doctor or other health professional regarding their symptoms.8 OCD generally lies in the mid range among disorders with respect to likelihood of treatment seeking. According to the ECA study, somatization disorder (100%), panic disorder (73%), depression (61%), and schizophrenia (47%) have higher rates of treatment seeking in comparison to OCD, whereas mania (22%), phobia (22%), drug abuse (18%), alcohol abuse (15%), and antisocial personality (11%) have lower rates.8 At the bivariate level, several variables were significantly associated with treatment-seeking behavior. These included total number of OCD symptoms, total number of comorbid diagnoses, having had severe obsessions of violence and other unpleasant thoughts, having had severe obsessions of contamination and/or doubt, and OCD comorbidity with panic disorder, mania, generalized anxiety, PTSD, and depression. It is somewhat intuitive that the total number of symptoms and total number of comorbid diagnoses would lead to an increase in treatment-seeking behavior, as level of distress likely increases when the weight of morbidity and comorbidity is higher. Despite the fact that OCD sufferers tend to mask their discomfort and illness, it is probable that they would at some point attempt to alleviate their discomfort and seek medical help, especially if the severity of their illness is high. It is interesting to note that in our sample there was no significant gender difference regarding treatment-seeking behavior. Although contradicting the notion that women in general exhibit a greater propensity to seek professional help for psychiatric morbidity,15 this finding is in line with what has been observed for other anxiety disorders. Thompson et al.16 reported an absence of differences between male and female subjects in their use of services for phobia, and the same pattern has been observed for health care seeking for panic attacks17; regarding social anxiety disorder, Weinstock18 reported that men are more likely to seek treatment than women. Regarding the role played by specific OCD symptoms, it is not surprising that “severe obsessions of violence and other unpleasant thoughts” and “severe obsessions of contamination and/or doubt” were positively associated with treatment seeking. Given that the obsessions are persistent and cause distress or significant impairment, and the individual recognizes that the obsessions are excessive or unreasonable, it seems likely that the individual would eventually resort to treatment seeking. What was somewhat unexpected is the finding that the three primary compulsive symptoms (pertaining to hand washing and checking, symmetry and precision, and counting) were not significantly related to treatment-seeking behavior. The absence of association between compulsions and treatment seeking was also reflected by the low prevalence of treatment seeking in subjects suffering from “compulsions only.” It may be that the compulsive behavior itself actually reduces the individual's anxiety/distress and hence the need to consult a health professional. Alternatively, subjects suffering from “compulsions only” may not recognize their symptoms as pathological and requiring treatment. This finding also suggests that subjects suffering from “compulsions only” are under-represented in clinical samples; in fact, 30% of our sample suffered from “compulsions only,” but 14.3% of the treatment seekers were in this situation. It should be noted that, in our sample, the majority of the subjects diagnosed with OCD were suffering from either obsessions or compulsions, but not both. This finding is consistent with what has been observed in some other parts of the world and particularly with the ECA study, where the symptomatology distribution was 50% for obsessions only (v our 51.8%), 34% for compulsions only (v 30.2%), and 16% for both (v 18%).19 This has led some authors to suggest that the “obsessive-compulsive” label is a misnomer.20 At the multivariate level, two variables were significantly related to treatment seeking: total number of OCD symptoms and severe obsessions of violence and other unpleasant thoughts. Since both the backward elimination and forward selection procedures resulted in the same final model, this appears to be a fairly robust finding. It is interesting to note that treatment-seeking behavior in our study was explained primarily by OCD symptoms and not by comorbid disorders. This finding contrasts with what has been observed in a study of treatment seeking in depression,21 where few depressive symptoms were associated with treatment seeking but several comorbid diagnoses were strong predictors. The fact that, in our study, some comorbid diagnoses were significant at the bivariate level but not retained in the multivariate model may be attributable to confounding. It is conceivable that the severity of OCD symptoms is associated with both comorbidity and treatment seeking. For example, in the case of depression, it could be that the severity of OCD symptoms results in both sufferers feeling depressed and being more likely to bring their OCD symptoms to the attention of a doctor. As a result, depression may be associated with treatment seeking at the bivariate level, but this association is in fact explained by the severity of the OCD symptoms. This study had several limitations. Due to data limitations, we were not able to investigate several potential confounding variables. Employment status, race/ethnicity, housing (urban v rural), and marital status have all been hypothesized to play a role in the manifestation of OCD1 and could possibly influence treatment-seeking behavior as well. In addition, the OCD section of the DIS contains few questions concerning which specific symptoms might be present. Prevalent symptoms such as sexual imagery, pathologic doubt, and hoarding are not investigated and therefore individuals suffering from these symptoms may be incorrectly diagnosed as unaffected. A further problem is that the symptoms tend to be grouped together: subjects are asked if they have ever suffered from obsessions of contamination and/or doubt, and they are also asked about compulsions pertaining to hand washing and/or checking. By grouping symptoms together we cannot determine the specific influence of each symptom (contamination v doubt, or hand washing v checking) on treatment-seeking behavior. An additional drawback of the DIS is its duration criterion. A period of 3 weeks is used, but perhaps a more meaningful approach would be to examine the number of hours per day during which the subject experiences symptoms and the impact this has on an individual's global functioning. In conclusion, this study is the first, to our knowledge, to examine predictors of treatment seeking for OCD. The absence of association between compulsions and treatment seeking and the very low prevalence of treatment seeking in individuals suffering from “compulsions only” suggests that these subjects are seldom seen in health care settings. Our findings suggest that there is a need to teach the public about compulsions, and such a strategy may enhance future public health education programs. Acknowledgements We thank Jean-François Boivin and Lucy Boothroyd for helpful comments on the manuscript. References 1.
1
Karno M, Golding JM.
Obsessive compulsive disorder.
In:
Robins LN, Regier DA editor.
Psychiatric Disorders in America: The Epidemiologic Catchment Area Study. New York, NY: The Free Press; 1991;p. 204–219. 2.
2
Sasson Y, Zohar J, Chopra M, Lustig M, Iancu I, Hendler T.
Epidemiology of obsessive-compulsive disorder: a world view.
J Clin Psychiatry. 1997;58(Suppl 12):7–10. 3.
3
Hollander E.
Obsessive-compulsive disorder: the hidden epidemic.
J Clin Psychiatry. 1997;58(Suppl 12):3–6. MEDLINE 4.
4
Rasmussen SA, Eisen JL.
The epidemiology and differential diagnosis of obsessive compulsive disorder.
J Clin Psychiatry. 1994;55(Suppl 10):5–10. MEDLINE 5.
5
Hollander E, Kwon JH, Stein DJ, Broatch J, Rowland CT, Himelein CA.
Obsessive-compulsive and spectrum disorders: overview and quality of life issues.
J Clin Psychiatry. 1996;57(Suppl 8):3–6. 6.
6
Whitaker A, Johnson J, Shaffer D, Papoport JL, Kalikow K, Walsh TB, et al.
Uncommon troubles in young people: prevalence estimates of selected psychiatric disorders in a nonreferred adolescent population.
Arch Gen Psychiatry. 1990;47:487–496. 7.
7
Hatch ML, Paradis C, Friedman S, Popkin M, Shalita AR.
Obsessive-compulsive disorder in patients with chronic pruritic conditions: case studies and discussion.
J Am Acad Dermatol. 1992;26:549–551. Abstract |
CrossRef
8.
8
Robins LN, Locke BZ, Regier DA.
An overview of psychiatric disorders in America.
In:
Robins LN, Regier DA editor.
Psychiatric Disorders in America: The Epidemiologic Catchment Area Study. New York, NY: The Free Press; 1991;p. 328–366. 9.
9
Leon CL, Portera L, Weissman MM.
The social costs of anxiety disorders.
Br J Psychiatry. 1995;166:19–22. MEDLINE |
CrossRef
10.
10
Orn H, Newman SC, Bland RC.
Design and field methods of the Edmonton survey of psychiatric disorders.
Acta Psychiatr Scand. 1988;77(Suppl 338):17–23.
CrossRef
11.
11
Bland RC, Orn H, Newman SC.
Lifetime prevalence of psychiatric disorders in Edmonton.
Acta Psychiatr Scand. 1988;77(Suppl 338):24–32.
CrossRef
12.
12
Robins LN, Helzer JE, Orvaschel H, Anthony J, Blazer D, Burnam A, et al.
The Diagnostic Interview Schedule.
In:
Eaton WW, Kessler LG editor.
Epidemiologic Field Methods in Psychiatry: The NIMH Epidemiologic Catchment Area Program. New York, NY: Academic Press; 1985;p. 145–190. 13.
13
Karno M, Golding JM, Sorenson SB, Burnam A.
The epidemiology of obsessive-compulsive disorder in five US communities.
Arch Gen Psychiatry. 1988;45:1094–1099. 14.
14
Galbaud du Fort G, Newman SC, Bland RC.
Psychiatric comorbidity and treatment seeking.
J Nerv Ment Dis. 1993;181:467–474. MEDLINE |
CrossRef
15.
15
Kessler RC, Brown RL, Broman LL.
Sex differences in psychiatric help-seeking: evidence from four large-scale surveys.
J Health Soc Behav. 1981;22:49–64. MEDLINE 16.
16
Thompson JW, Burns BJ, Bartko J, Boyd JH, Taube CA, Bourdon KH.
The use of ambulatory services by persons with and without phobia.
Med Care. 1988;26:183–198. MEDLINE |
CrossRef
17.
17
Katerndahl DA.
Factors associated with persons with panic attacks seeking medical care.
Fam Med. 1990;22:462–466. MEDLINE 18.
18
Weinstock LS.
Gender differences in the presentation and management of social anxiety disorder.
J Clin Psychiatry. 1999;60(Suppl 9):9–13. MEDLINE 19.
19
Weissman MM, Bland RC, Canino GJ, Greenwald S, Hwu HG, Lee CK, et al.
The cross national epidemiology of obsessive compulsive disorder.
J Clin Psychiatrv. 1994;55(Suppl 3):5–10. 20.
20
Kolada JL, Bland RC, Newman SC.
Obsessive-compulsive disorder.
Acta Psychiatr Scand. 1994;89(Suppl 376):24–35.
CrossRef
21.
21
Galbaud du Fort G, Newman SC, Boothroyd LJ, Bland RC.
Treatment seeking for depression: role of depressive symptoms and comorbid psychiatric diagnoses.
J Affect Disord. 1999;52:31–40. Abstract | Full Text |
Full-Text PDF (84 KB)
|
CrossRef
Centre for Clinical Epidemiology and Community Studies, Sir Mortimer B. Davis-Jewish General Hospital, Montreal; Department of Psychiatry, McGill University, Montreal, Quebec, Canada; Department of Psychiatry, University of Alberta, Edmonton, Alberta, Canada ☆ Supported by the National Health Research and Development Program, Alberta Mental Health Research Fund, and Alberta Health. Preparation of this paper was also supported by an Investigator Award (Fonds de la Recherche en Santé du Québec) to G.G.F. ☆☆ Address reprint requests to Guillaume Galbaud du Fort, M.D, Ph.D, Centre for Clinical Epidemiology and Community Studies,, The Sir Mortimer B. Davis-Jewish General Hospital, 3755 Côte Ste-Catherine, Montreal, Quebec, Canada H3T 1E2. ★ 0010-440X/03/4402-0021$30.00/0 PII: S0010-440X(03)00025-7 doi:10.1053/comp.2003.50005 © 2003 Published by Elsevier Inc. | |
|
FULL TEXT