Detecting psychological distress among patients attending secondary health care clinics

FLSEVIER

Detecting Psychological Distress Anm Attending Secondary Health Care Clinics Self-Report and Physician Rating Dina Feldman, M.A., Jonathan Rabinowitz, D.S.W., and Yoram Ben Yehuda, M.A. Abstract: A study was conducted to determine the prevalence of psychologicaldistress,as reported by patients and their physicians, in orthopedic, neurology, dermatology, and ophthalmology clinics; to study their accuracyin detecting psychological distress; and to determine if there is any connectionamong psychologicaldistress,accuracyof detecting distress,and useof mental health and primary health carephysicians’prognosisfor the somatic complaints. Five hundred and fifty-six patients, ages 18-21, responded to the Psychiatric Epidemiology Researchinterview Demorahzation Scale(PER&D), a measureof psychological distress, and to questions about their mental health and use of mental health and primary health services. Physicians,who were blind to patients’ responses,were asked to what extent they thought the causeof patients’ complaints was physical and to what extent they thought it was psychological in nature, and to prognosticate. Basedon the PER&D, about 25% of patients were distressed,this was lessfor females than malesand varied between clinics. Basedon self-reporting, about 14% of patients (males and females) were distressed. Based on physician reporting, about 17% (males Zess)were distressed.Physiciansidentified 35% of the PER&D-distressed casesand 79% of nondistressedcases.About 66% of patients identified their distress and 83% their lack of distress. lncreaseduse of primary health care and mental health care was related to distress. The prognosis was negatively related to distress. Basedon this study, there is a needfor more attention to psychologicaldistressamong secondaryhealth care patients. Patients’ ability to identify their distress suggests the importanceof involving the patient in the diagnostic process.Correct detection of distressalone does not appear to decreasethe use of primary medical and mental health services.

Mental Health Service of Medical Corps of IDF (D.F., Y.B.Y.), and Bar Ilan University (J.R.), Ramat-Gan, Israel. Address reprint requests to: Dr. Jonathan Rabinowitz, School of Social Work, Bar Ilan University, Ramat-Gan, Israel.

General Hospital Psychiatry 0 1995 Elsevier Science Inc. 655 Avenue of the Americas,

17, 425-432, New

York,

Introduction Some studies report that about 50% of patients in primary care suffer from emotional problems [l-3]. Others report 20% prevalence of well-defined psychiatric disorders and 40% prevalence of minor subthreshold disorders [P71. Little research has been done in the area of secondary health care. Recently, van Hemert et al. [S] found that 29% of internal medicine outpatients had a psychiatric disorder. They found psychiatric disorders in 15% of patients with a medical explanation for their presenting symptoms, in 45% with ill-explained symptoms, and in 38% with no medical explanation for their symptoms. Other studies of secondary health care patients have found pychiatric disorders in 33%-47% of patients [9-121. Persons with psychiatric disorders use medical services more than those without such disorders [13-171. Physical and psychiatric symptoms tend to cluster in some individuals who dominate the use of medical and psychiatric services [13,18,19]. There are also physical illnesses that are psychogenic in nature or are believed to have strong psychological components. Physicians have shown various degrees of success in detecting psychological distress among their patients. In most studies, success has not exceeded 50% [20-261. Similar results have been reported about their ability to recognize specific disorders [1,27]. These findings have been attributed to physicians’ lack of skills [28], short examination time [29,30], and lack of interest [22]. In contrast to the many studies of mental health aspects of pri-

425

1995 NY

10010

1SSN O&3-8343/95/$9.50 0163 -8343(95)00058-Y

D. Feldman et al

The sample included 556 patients who attended a secondary health care clinic of the Israel Defence Forces (IDF). They were between the ages of 18 and 21; 74.7% were men. They were referred by IDF primary care physicians to either a dermatology (N = 94, 16.9%), neurology (N = 87, 15.6%), orthopedic (N = 247,44.4%), or ophthalmology (N = 94, 16.9%) clinic. Referral clinic was unknown in 34 cases.

elevated scores on these scales like elevated temperatures, tell you that something is wrong.” PERI-D was designed to tap nonspecific indicators of distress and includes items common to rating scales measuring anxiety, depressive and psychosomatic symptoms. It is composed of fixedformat items about the frequency of psychological complaints in the past year. Responses are given on a 5-point scale ranging from “never” (weighted as 0) to “very often” (weighted as 4). It is scored by adding the responses to the 27 items and dividing by the number of completed items [34]. The higher the PERI-D score the more demoralization. Reliability and validity tests of PERI-D in the United States and Israel have shown satisfactory results [31]. PERI-D has been used in several studies in Israel [35-391 and high sensitivity and specificity levels have been reported [40]. In addition to the PERI-D, patients were asked questions about their mental health and use of primary medical and mental health services. Physicians were asked to complete an encounter form that included rating scales of their estimation of patients’ psychological and medical status and prognosis.

Instruments

Procedure

The Psychiatric Epidemiology Research Interview Demoralization Scale (PERI-D) was used to measure patients’ psychological distress. PERI-D is a 27-item instrument developed by Dohrenwend et al. [31] that measures demoralization. Demoralization is a type of psychological distress that has been defined as a predicament for which the person sees no solution [32]. PERI-D was designed and has been used as a screening instrument to measure nonspecific psychological distress among a wide spectrum of populations [31,33]. The concept of psychological distress and demoralization are not equivalent to a specific psychopathological entity. Instead, as Dohrenwend et al. 1331 explain, these concepts are “analogous in important ways to measures of temperature in physical medicine.” They point out that elevated scores on measures like the PERI-D have been observed in a wide variety of contrasting circumstances. For example, combat troops had higher levels than other soldiers, persons with more episodes of physical illness had higher levels than those with less, psychiatric patients had higher levels than nonpatients. They concluded that

For 1 month the PERI-D, and additional questions about mental health and service use, were given to every fifth patient to complete in the waiting room before the physician consultation. Instructions to patients explained that they were being asked to participate in a survey on health and mood, being conducted by the Medical Corps, and were assured that their responses would remain confidential. Physicians completed their encounter form without seeing patients’ questionnaire. No patients declined to participate. Results were first analyzed using average PERI-D scores and then using accepted demoralization cutting points as in 135,361 (1.23 for males, 1.51 for females) based on Shrout et al. [40].

mary health care, less work has been done on mental health aspects of secondary health care. The current study explores five questions: 1) What is the prevalence of psychological distress among patients in secondary health clinics; 2) How well can specialist physicians detect psychological distress; 3) How well can patients identify their own levels of distress; 4) To what extent does psychological distress relate to use of primary health care; and 5) To what extent does psychological distress relate to use of mental health care.

Methods Sample

426

I,

.

,

.

Results Prevalenceof PsychologicalDistress The mean PERI-D score was 1.01 (SD 0.65). The minimum (little distress) was 0.14 and the maximum (high distress) was 3.65. There was no statistically significant difference between mean

Detecting Psychological &tress PERI-D of males and females. Reliability, as measured by Cronbach’s alpha for the PERI-D was 0.93. Using demoralization cutting points described above, 26.4% (N = 110) of males and 16.3% of the females were demoralized (N = 23). Patients in the four specialty clinics had differences in average levels of psychological distress as measured by the PERI-D. From most distressed to least distressed were neurology 1.25 (SD 0.72, N = 87), dermatology 1.02 (SD 0.57, N = 94), ophthalmology 0.96 (SD 0.61, N = 94), and orthopedics 0.90 (SD 0.59, N = 247). We found the same order using percentage of demoralized patients per clinic. About 40% (N = 35) of the neurology patients were demoralized, 27% (N = 25) of dermatology, 23% (N = 22) of ophthalmology, and 21% (N = 51) of orthopedic patients.

Physician and Patient Detection of PsychologicalDistress According to physicians, 426 patients presenting problems had physical health problems (72% females, N = 101; 78% males, N = 325). Physicians were doubtful if the problem was medical in 29 cases (6% females, N = 9; 5% males, N = 20). In 55 cases (9% males, N = 38; 12% females, N = 17) physicians indicated that the problem was a mental and physical health problem. In 40 cases (6% males, N = 27, 9% females, N = 13) physicians indicated that the problem was entirely a mental health problem. The physicians’ impressions were consistent with the PERI-D scores. Physicians were asked whether the problem was medical, medical and psychological, or entirely psychological. There was almost a monotonic distribution of PERI-D scores starting with patient has “no medical problem” (0.96 SD 0.77, N = 40), “medical problem” (0.92 SD 0.57, N = 426), “doubtful if medical problem, possibly psychological” (1.13 SD 0.78, N = 29), and “psychological problem” (1.28 SD 0.71, N = 56). Physicians were also asked to estimate (in percentage) the extent to which the problem was medical and not psychological. There was a negative correlation between PERI-D and physicians’ estimated extent to which the problem was medical (r = -0.16, p = 0.001, N = 495). The average extent to which the problem was medical was signifi-

cantly lower for demoralized patients (78.8 SD 28.8, N = 132) than for nondemoralized patients (87.8 SD 24, N = 363, F = 12, p = 0.000). Patients who said that they had undergone mental health treatment were rated by physicians as having significantly less medical and more psychological problems than those patients who had not undergone mental health treatment (68.6 SD 36, N = 22 vs 84.1 SD 25.5, N = 303; F = 6.19, p - 0.01). There were no significant differences between average PERI-D scores of patients by levels of prognosis, but there was a monotonic distribution between demoralization and prognosis. About 36.5% (N = 15) of patients with low prognosis for recovery were demoralized, 30.4% iN = 112) of those with medium prognosis, and 2&l % (N = 76) of those with high prognosis. There was a similar distribution nt’ demoralization and physicians’ estimation of whether patient’s problem was “not a medical problem” (25%, N = 10, demoralized), a “medical problem” (20.8%, N = 132, demoralized), “doubtful if medical problem, possibly psychological” (34.5%, N = 10, demoralized), and “psychological problem” (44.6%, N = 25, demoralized). We dichotomized the patient and physician reported levels of psychological distress into low and high levels to correspond to the PERI-D demoralized and nondemoralized categories. Patients who reported no mental health problems were put in the low group and those with some or many problems were put in the high group. The “some“ problems and “many” problems groups were joined due to the closeness of their PERI-D scores. Patients whose physicians reported that they had only a physical health problem were considered low distress and those whose physicians considered them to have either entirely or partially a psychological problem were considered high distress. Using the above dichotomy of patient and physician responses, we found that patient and physician “low” distress groups were significantly less distressed, according to the PERT-D, than patient and physician “high” distressed groups. The mean PERI-D for the 14% (N = 79) of patients who reported being highly distressed was 1.69 (SD 0.73) and for the rest of the patients, who were low distress, it was 0.85 (SD 0.51; F = 159, df = 1, p = 0.0000). The mean PERI-D for the l?‘/o (N = 95) of patients whose physicians reported them as being highly distressed was 1.15 (SD 0.75), and for the rest of the patients whose physicians reported

427

D. Feldman et al. them as having low distress the PERI-D was 0.93 (SD 0.59; F = 9.6, df = 1, p = 0.002). It should be noted that the patient-reported high distress group had a higher mean PERI-D score than did the physician-reported high distress group. We found that physicians from different specialties estimated differently the extent to which a patient’s problem was a medical one. Here again, physicians reported that neurology patients had the greatest psychological component with their problems with an average of 77.4% (SD 27.7, N = 82) medical. They were followed by physicians’ reports about orthopedic patients whose problems were 82.2% (SD 27.6 N = 210) medical, ophthalmology patients whose problems were 91.4% (SD 22.9 N = 77) medical, and dermatology patients whose problems were 98% (SD 6.8 N = 90) medical. Patients were asked if they have mental health problems. Almost 85% (N = 473) responded “no” and about 15% (N = 80) responded that they had at least “some.” Of the patients who reported no problem, 19% (N = 90) were demoralized and 81% (N = 382) were not demoralized. Of the patients who had at least “some problems,” 70% (N = 56) were demoralized, and 30% (N = 24) were not. These differences were significant (Chi-square 91.4, df = 1, p = 0.00000). Also using the “low” and “high” patientreported and physician-reported distress levels, we examined gender differences. We found that about 14% of both sexes had high patient-reported distress levels. There was also no significant difference in PERI-D averages between these two groups. The average of both groups together was 1.70 (SD 0.73). Similarly, there were no significant differences between males and females based on physician ratings.

Physician and Patient Accuracy in Detecting Psychological Distress To measure physician and patient accuracy in detecting psychological distress we calculated sensitivity and specificity of patient-report and physician-report measures as compared to the PERI-D (Table 1). As can be seen in the sensitivity row, slightly less than half of the males and 10% of the females reported by physicians to be distressed (i.e., patients whose problem was entirely or at least partially psychological) were distressed based on PERI-D. Specificity was slightly higher for females than for males and sensitivity was considerably higher for males. Overall, the patient-report measures were the most accurate reflections of PERI-D demoralization, as reflected by Kappa values in the last row. Table 2 presents sensitivity and specificity of physician reporting by clinic. As in the previous table, specificity is considerably higher than sensitivity. In the analysis by clinic, there were a few gender differences in sensitivity and specificity of physicians’ assessment. In the orthopedic clinic, the sensitivity was about the same for males and females. Yet the specificity differed-for males it was 17.1% and for females it was 0. The kappa for male patients was 0.17 and for female patients it was - 0.11. Another difference was in the neurology clinic where specificity was 75% for males and 37.5% for females and the sensitivity was about 50% in both. Kappa for males was 0.22 and for females it was -0.07. We attempted to determine if the poor levels of accuracy were a result of demoralization cutting points that were not high enough, as suggested by the high false-positives. The male patients identified by physicians as having a psychological com-

Table 1. Sensitivity

and specificity of self-reported psychological distress and physician-reported psychological distress as compared with the PERI-D demoralization cutting points Self-reporting

Physician reporting Males Sensitivity Specificity Kappa

428

47.7% N = 31165 77.7% N = 272l350 0.20

Females

Total

10% N = 3130

35.8% N=34/95 78.7%

N = 43160 81.2%

89.2%

83.3%

N = 3631461

N = 2851351

N = 107/120

N = 394473

0.39

0.38

0.39

82%

N = 91/111 -0.09

0.12

Males 71.7%

Females 52.6%

N = lo/19

Total 66.2%

N = 53180

Detecting Psvchdogical Distress

Table 2. Sensitivity and specificity of physician-reported psychological distress as compared with the PERI-D demoralization cutting points by specialty clinic Physician reporting by clinic Ophthalmology (N = 94) Sensitivity Specificity

Kappa

28.6% N = 6121 74.0% N = 54173 0.02

Orthopedic (N = 247)

Neurology (N = 87)

13.3% N = 6145 94.6% N = 1911202 0.10

47.1% N = 16134 62.3% N = 33153 0.09

ponent to their problem had PERI-D scores that ranged from 0.11 to 3.56 with a mean of 1.24 (SD 0.83). Interestingly, the mean PERI-D of this group was almost the same as the demoralization cutting point for males (1.23). Based on an examination of the distribution even if the cutting point were dropped to 0.93, which would be low compared with PERI-D scores of nonclinical populations in Israel [36], the sensitivity would increase only slightly to 55%. Similarly, dropping the cutting point would bring about only a small improvement in accuracy of patient reporting. For males, the cutting point is between the 25th and 30th percentiles. The 20th percentile value is 1.07 and the 15th is

0.83. Relationship BetweenPsychologicalDistress and Use of Prima y Health Care Patients who visited primary care physicians more often tended to be more distressed than patients who made less visits. There was a monotonic distribution, with patients who reported going most often having higher PER&D scores (1.47 SD 0.77, N = 44) than those who defined themselves as “medium attenders” (1.12 SD 0.65, N = 133) and “low attenders” (0.88 SD 0.56, N = 331). For further analysis we divided use of primary medical care into “medium-high’ and “low.” We found that demoralized patients used primary medical care significantly more than nondemoralized patients. Of the demoralized patients, 47.5% (N = 57) were “low” users and 52.5% (N = 63) were “mediumhigh” users. For nondemoralized patients, 70.6% (N = 274) were “low” and 29.4% (N = 114) were

Dermatology (N = 94’j -4.5% N = 1122 87.5% N = 63172 -0.10

“medium-high.” These differences were significant (Chi-square = 21.6, df = 1, p = O.OOOOO). We attempted to learn the extent to which patients’ awareness of psychological problems may have modified their use of medical services. Patients who reported having psychological problems attended primary medical clinics as often as those patients who did not report such problems. Similarly, those demoralized patients whose physicians had not detected their distress used medical services as often as demoralized patients whose physicians had detected distress. Patients who were, or had been, in mental health treatment were significantly heavier users of medical services than patients who had not been in treatment. Fourteen out of 22 (64%) patients who had been in treatment were medium-high users; this compares

with 38% (N = 124) who had

not been in treatment (Chi-square 8.16, df = 1, p = 0.004). Of the 22 patients who were, or had been, in mental health treatment, 15 of them were demoralized. This was significantly more than the patients who had not been in mental health treatment, 24.5% of whom were demoralized (Chisquare 20.0, df = 1, p = 0.00001). Using physician

ratings based on a percentage scale of extent to which problems were medical or psychological, physicians rated 75.4% (SD 31.8) of all patients’ problems as being medical, 86.1% (SD 23) of the medium users, and 87% (SD 24.8) of the low users. One third (N = 127) of the patients rated by physicians as having a medical problem were medium-high users of primary care medical services. In cases where physicians were doubtful about whether the problem was medical or psychological, 48% (N = 13) were medium-high users of pri-

429

D. Feldman et al. mary medical care. Of patients rated by physicians as having a combined health-mental health problem 51.9% (N = 27) were medium-high users as were 23.7% (N = 9) who were judged to have a psychological problem. There was no significant difference in extent to which the problem was estimated to be medical for medium-high and low users.

Relationship BetweenPsychologicalDistress and Use of Mental Health Care PERI-D scores could differentiate between patients who reported having mental health problems, and treatment for those problems, from those who did not. Subjects were asked if they have mental health problems. Their responses were consistent with the PERI-D. PERI-D for those who reported “no problems” (N = 471) was 0.88 (SD 0.54), for those who reported “some problems” (N = 67) it was 1.71 (SD 0.66), and for those who reported “many problems” (N = 12) it was 2.07 (SD 1.09). These differences were significant (F = 83, p = 0.0000) and form a monotonic distribution. Subjects were also asked if they were, or had been, in mental health treatment during the last 2 years. The average PERI-D score for those who had been in treatment (1.82 SD 75, N = 22) was significantly higher than for those who had not been in treatment (1.00 SD 0.62, N = 342; F = 34, p = 0.0000). The 35% (N = 195) who did not respond to this question had an average PERI-D of 0.86 (SD 0.56). This was not significantly different from the patients who reported having had no treatment and significantly different from those who reported having had treatment (Scheffe multiple range test 0.05 level). Seventy-three percent (16 of 22) of patients who reported having had mental health treatment were demoralized as compared with 25.9% (89 of 342) of the patients who reported having had no treatment.

Discussion We found that almost one-quarter of the patients in four secondary health care clinics were demoralized, considerably lower than the 50% reported in the literature. This may be because all patients were soldiers who had already passed mental and physical health screening by the draft board. This may also explain why we did not find differences between males and females-in contrast to the lit-

430

erature-since military screening of females is more rigorous than for males even though military service is more demanding for males. Despite this careful screening some patients suffered from distress. Similar to Gilboa et al. [36], who studied prevalence of demoralization on a kibbutz, we found that about one-quarter of the patients were demoralized. In contrast, the female patients in this study had lower demoralization prevalence rates (16%) than the kibbutz females (30%) and the male patients had higher prevalence rates (26%) than the kibbutz males (20%). Female patients also had lower mean PERI-D scores (1.03 SD 0.50) than the kibbutz females (1.3 SD 0.5) and the male patients had higher mean PERI-D scores (1.00 SD 0.68) than the kibbutz males (0.9 SD 0.5). In another study conducted in primary health care in the IDF [20] it was found that about 50% of patients were distressed, about double of what we found in secondary health care clinics in the IDF. We attribute this difference to the screening by the referring primary health care physicians in the military unit. The prevalence of psychological distress among the patients in our secondary health care clinics was consistently lower than in van Hemert et al. [8] clinics. One possibility may be that our patients had passed draft board screening. Another possibility may be due to the types of clinics studiedvan Hemert et al. studied gastroenterological and internal medicine. We studied neurology, orthopedic, ophthalmology, and dermatology and found considerable differences in prevalence of psychological distress among these groups. We also found that rates of correct detection of distress differed among the clinics. The accuracy of physicians’ detection in the current study (ranging from 4.5% to 47%) was a little less than those who report about 50% [21-231 and similar to van Hemert et al. [8] who also studied secondary health clinics. Though correct detection can only be determined by measuring sensitivity and specificity, many studies focused exclusively on sensitivity and ignored base rates. Using both sensitivity and specificity, the physicians’ success rate in identifying psychological distress was similar to that of Maoz et al. [20] who found a sensitivity of 16% and a specificity of 89%. This is particularly interesting because Maoz et al.‘s study was also done in the IDF, among primary care physicians. We also found that it was important to take into

Detecting Psychological Distress

account the relationship to the distribution as expressed by kappa. For example, our high levels of specificity are to a great extent due to the fact that three-quarters of patients were not demoralized. Therefore the kappas, which reflect chance-adjusted levels of agreement, were very low. However, in other studies, kappa levels were not always reported and in most studies, almost twice as many patients were distressed compared with the current study. Clearly, the more patients are distressed, the more chance there is for correct positive detection. The results suggest that there is a need for attention to psychological distress among secondary health care patients. Patients with psychological distress need special attention regardless of whether the distress is the suspected cause of somatic complaint or a suspected result of it, indeed, there is probably a complex relationship between the two. The results also suggest that physicians should not rely totally on themselves as barometers of psychological distress, and that they should take into consideration patients’ perceptions, even if patients come with physical complaints. Physicians should also consider routinely using selfreport measures of distress. Distress appears to be an important contributor to heavy use of primary medical and mental health services. Yet correct detection of distress alone was not a sufficient condition for decrease in the use of primary medical and mental health services.

8.

9.

10. 11. 12. 13. 14. 15. 16. 17.

18.

frequency and diagnosis in four developing coun tries. Psycho1 Med 10:231-241, 1980 van Hemert AM, Hengeveld MW, Bolk JH, Rooij-. mans HG, Vandenbroucke JP: Psychiatric disorders in relation to medical illness among patients of a general medical out-patient clinic. Psycho1 Med 23: 167-173, 1993 Ford MJ, Miller PM, Eastwood J, Eastwood MA: Life events, psychiatric illness and irritable bowel syndrome. Gut 28:160-165, 1987 Colgan S, Creed FH, Klass SH: Psychiatric disorders and abnormal illness behavior in patients with upper abdominal pain. Psycho1 Med 38~887-892, 1.988 Lobo A, Perez-Echeverria J, Artal J, et al: Psychiatric morbidity among medical out-patients in Spain: a case study. J Psychosom Res 32~35-5-364, 1988 MacDonald AJ, Bouchier IAD: Non-organic gastrointestinal illness: a medical and psychiatric study. Br J Psychiatry 136:276-283, 1980 Hinkle LE, Wolff HG: The nature of man’s adaptation to his total environment and the relation of this to illness. Arch Intern Med 99:44246&j, 1957 Eastwood MR, Trevelyan MH: Relationship between physical and psychiatric disorder. Psycho1 Med 2~363-372, 1972 Lipowski ZJ: Psychosomatic Medicine and Liaison Psychiatry. Selected Papers. New York, Plenum Press, 1985 Lipowski ZJ: The interface of psychiatry and medicine: towards integrated health care clan J Psychiatry 32~743-748, 1987 Hankin JR, Steinwachs DM, Regier DA, Burnes BJ, Goldberg ID, Hoeper EW: Use of general medical care services by persons with mental disorders. Arch Gen Psychiatry 39:225-231, 1982 Fink I’: Mental illness and admission to general hospitals: a register investigation. Acta P%vchiatr Stand

82:45&462, 1990 19. Fink I’: Physical disorders associated with mental

References 1. Ormel J, Van Den Brink W, Koeter MW, et al: Recognition, management and outcome of psychological disorders in primary care: a naturalistic follow-up study. Psycho1 Med 20:909-923, 1990 2. Mari JJ, Williams P: Minor psychiatric disorders in primary care in Brazil: a pilot study. Psycho1 Med 14:223-227, 1984 3. Shiber A, Maoz B, Antonovsky

A, Antonovsky H: Detection of emotional problems in the primary care clinics. Fam Pratt 7:195-199, 1990 4. Kessler LG, Cleary I’D, Burke JD Psychiatric disorders in primary care. Arch Gen Psychiatry 42:583587, 1985 5. Bellantuono C, Firio R, Williams I’, Cortina P: Psy-

chiatric morbidity in an Italian general practice. Psychol Med 17:243-247, 1987 6. Barrett JE, Barrett JA, Oxman TE, Gerber PD: The prevalence of psychiatric disorders in a primary care practice. Arch Gen Psychiatry 45:1100-1106, 1988 7. Harding TW, Arango MV, Baltazar J, et al: Mental disorders in primary health care: a study of their

illness: a register investigation. Psycho1 Med 2&829834, 1990 20. Maoz B, Rabinowitz

S, Mark M, Antonovsky H, Ribak J, Kotler M: Physicians’ detection of psychological distress in primary-care clinics. Psvchol Rep 69:999-1003, 1991 21. Sartorius TB, Ustun JA, Costa e Silva JA, et al: An International Study of Psychological Problems in Primary Care. Preliminary Report from the WHO Collaborative Project on “Psychological Problems in General Health Care” 22. Marks J, Golberg D, Hillier V: Determinants of the ability of general practitioners to detect psychiatric illness. Psycho1 Med 9337-353, 1979 23. Goldberg DP, Steele J, Johnson A, Smith C: Ability of primary care physicians to make accurate ratings of psychological disturbance. Arch Gen Psychiatry 39:829-833, 1982 24. Jones LR, Badger LW, Ficken RP, Leeper JD, Ander-

son RL: Inside the hidden mental health network: examining mental health care delivery of primary care physicians. Gen Hosp Psychiatry 9287-293, 1987 2.5. Jones LR, Badger LW, Ficken RP. Leeper TD, Ander-

431

D. Feldman et al.

26.

27.

28. 29.

30. 31.

32. 33.

432

son RL: Mental health training of primary care physicians: an outcome study. Int J Psychiatry Med 18: 107-121, 1988 Borus JF, Howes MJ, Devins NP, et al: Primary health care providers: recognition and diagnosis of mental disorders in their patients. Gen Hosp Psychiatry 10:317-321, 1988 Von Korff M, Shapiro S, Burke JD, et al: Anxiety and depression in a primary care clinic: comparison of DIS, GHQ and practitioner assessments. Arch Gen Psychiatry 44:152-156, 1987 Taylor FK: The medical model of the disease concept. Br J Psychiatry 128:588-594, 1976 Leeman Cl’: Diagnostic errors in emergency room medicine: physical illness in patients labeled “psychiatric” and vice versa. Int J Psychiatry Med 6:533540, 1975 Knutsen E, DuRand C: Previously unrecognized physical illnesses in psychiatric patients. Hosp Community Psychiatry 42:182-186, 1991 Dohrenwend BP, Levav I, Shrout PE: Screening scales from the Psychiatric Epidemiology Research Interview (PERI). In Weissman MM, Myers JK, Ross CE (eds), Community Surveys of Psychiatric Disorders. New Brunswick, NJ, Rutgers University Press, 1986, p 349 Frank J: Persuasion and healing. Baltimore, Johns Hopkins University Press, 1973 Dohrenwend BP, Shrout PE, Galdys E, Mendelsohn

34.

35. 36. 37. 38.

39.

40.

FS: Nonspecific psychological distress and other dimensions of psychopathology. Arch Gen Psychiatry 37:1229-1236, 1980 Shrout PE, Lyons M, Dohrenwend BP, Skodol AE, Solomon M, Kass F: Changing time frames on symptom inventories: effects of the Psychiatric Epidemiology Research Interview. J Consult Clin Psycho1 56: 267-272, 1988 Fennig S, Levav I: Demoralization and social supports among Holocaust survivors. J Nerv Ment Dis 179:167-172, 1991 Gilboa S, Levav I, Gilboa L, Ruiz F: The epidemiology of demoralization in a kibbutz. Acta Psychiatr Stand 82:6(X4, 1990 Flaherty JA, Kohn R, Levav I, Birz S: Demoralization in Soviet-Jewish immigrants to the United States and Israel. Compr Psychiatry 29:58&597, 1988 Zilber N, Lerner Y: Psychological distress among recent immigrants from the former Societ Union to Israel: I. Mediating factors. Psycho1 Med, in press, 1995 Lerner Y, Zilber N: Psychological distress among recent immigrants from the former Societ Union to Israel. The effect of the Gulf War. Psycho1 Med, in press, 1995 Shrout PE, Dohrenwend BP, Levav I: A discriminant rule for screening cases of diverse diagnostic types: preliminary results. J Consult Clin Psycho1 54:314319, 1986