Primary Care Physician Use of Cognitive Behavioral Techniques

Primary Care Physician Use of Cognitive Behavioral Techniques with Depressed Patients Patricia Robinson, PhD; Terry Bush, PhD; Michael Von Korff, ScD; Wayne Katon, MD; Elizabeth Lin, MD, MPH; Gregory E. Simon, MD, MPH; and Edward Walker, MD Seattle, Washington

Background. Although researchers are paying more at cian recommendations regarding planning pleasurable tention to the treatment o f depression in the primary activities, problem solving, challenging depressive care setting, little is known about the nature of psycho thoughts, and planning activities that boost confidence therapeutic interactions that occur between primary care were reported by 22% to 40% o f study patients. Older physicians and their patients in the context of a visit for patients reported fewer interactions about CB strategies. depression. In recent years, brief cognitive behavioral Primary care physicians’ suggestion o f CB strategies was therapy has been demonstrated to be efficacious, and associated with both patient use o f CB strategies in the the public has become more familiar with these tech months following the visit and better adherence to rec niques through media exposure and self-help books. ommended medication therapy during the first month of treatment. Methods. Depressed primary care patients were surveyed regarding the extent to which cognitive behavioral (CB) Conclusions. Many patients seem to recognize the oc techniques were suggested during the primary care visit currence of psychotherapeutic interactions during visits in which antidepressant medication was initially pre to their primary care physician in which an antidepres scribed. One hundred fifty-five patients completed re sant medication was prescribed, and patients’ recogni sponses to phone surveys 1 month and 4 months after tion of these interactions is associated with increased ad the visit. Patients were also surveyed regarding the rec herence to the recommended course of antidepressant ommendation o f counseling by the primary care physi prescriptions. cian. Key words. Depression; primary health care; physician- Results. The majority of patients (61%) reported that patient relations; patient compliance; patient adherence; their physician advised them to identify activities they cognitive therapy; cognitive-behavioral therapy. were already doing that helped them feel better. Physi (J Ram Pract 1995; 40:352-357)

Depression is one o f the most prevalent mental disorders ments of depression provided by primary care physi seen among primary care patients,1’2 and the majority of cians.4-10 depressed persons seeking care obtain mental health care Antidepressant medications are prescribed for over only from primary care physicians,3 yet little is known one half o f the patients seen in primary care clinics for depression,11 and 80% o f all the antidepressants used in about psychopharmacological and psychological treat- the United States are prescribed by primary care physi cians. While the duration and frequency o f primary care visits for depression preclude intensive cognitive behav Submitted, revised, December 20, 1994. ioral (CB) treatment of depression, patients may derive unique benefits when primary care physicians integrate From Mental Health Services (P.R.) and the Center for Health Studies (T.B., M.V.K., E.L., G.E.S.), Group Health Cooperative, and the Department o f Psychiatry brief CB techniques into primary care visits. For example- and Behavioral Sciences (W.K., E.W.), University o f Washington, Seattle. Requests psychological intervention during a supportive visit may for reprints should be addressed to Patricia Robinson, PhD, Center fo r Health Studies, Group Health Cooperative, 1730 Minor Ave, Suite 1600, Seattle, WA 98101-1448. enhance adherence to prescribed medication regimens.

Less depressed patients who are not being managed with are trained in family medicine; most of the remainder are medication also may benefit from brief CB interventions. trained in internal medicine. In addition to CB techniques, problem solving13 and interpersonal therapies14 are potentially useful techniques for the primary care physician. The model of interactions Study Design between patients and primary care physicians is shifting to Automated pharmacy data entered from September 1991 a more collaborative one, in which patients are given to April 1992, inclusive, were reviewed to identify pa greater information about, and responsibility for, choices tients receiving new prescriptions for an antidepressant in their treatment. This collaborative model is similar to from primary care physicians in two large primary care that used by mental health providers using CB treatment clinics. Prescriptions were considered new if the patient in specialty settings. These developments may be foster had not received an antidepressant prescription during ing transfer o f clinical techniques that have been used the 3 months preceding the patient identification date. predominantly by psychotherapists into the mainstream Medical charts of potential subjects were reviewed; 247 of primary care medicine.16-17 patients who were between the ages o f 18 and 75 years This paper provides information about the use of CB and had received a diagnosis of depression during an techniques during primary care visits in which an antide antidepressant prescription visit were invited to partici pressant medication is initially prescribed (the “initial pre pate in the study. scription visit” ). It also describes patient-physician inter Sixty-six percent (164) o f the eligible patients agreed actions about seeking counseling and factors that to participate, 21% refused, and 13% were unavailable. influence these interactions in the initial prescription visit, The patients who agreed to participate and those who did such as age and sex o f patients and their physicians; de not were similar in age, sex, treatment date, and type of pression severity; neuroticism; and whether the depres antidepressant medications prescribed. Nine o f those who sion is acute or chronic. We examined the relationship agreed to participate were excluded, seven o f whom never between patient recognition o f physician recommenda started taking the antidepressant that had been prescribed tion of CB techniques and (1) subsequent use o f these for them and two o f whom did not complete the 4-month techniques by patients, and (2) patient adherence to pre follow-up interview. The final sample consisted o f 155 scribed medication regimens. The major objective of subjects who started taking newly prescribed antidepres these analyses was to determine which CB techniques sants and completed both the original and the 4-month primary care physicians are currently using and for which interviews. patients they are being used. The efficacy o f brief CB Telephone interviews were conducted with 164 sub interventions for the treatment o f depression in primary jects between October 1991 and April 1992. The first care was not assessed. interview occurred approximately 1 month after the pre scription visit (mean, 25.3 days, standard deviation [SD] = 14.5). A second phone interview was completed Methods with 155 subjects 4 months after the original interview (mean, 119.9 days, SD = 11.7).

Study Setting Assessment Measures The study was conducted at Group Health Cooperative of Puget Sound (GH C), a large staff-model health main D e p r e s s io n S t a t u s tenance organization (HM O) serving approximately 390.000 residents in western Washington State. Group In the initial interview, patients were asked to rate the Health Cooperative provides comprehensive care on a severity of depressive symptoms during the 2 weeks be capitated basis, with enrollees typically receiving GHC fore the prescription visit. Depression was measured by coverage through employer-subsidized plans; GHC in means of a modified version o f the Inventory for Depres cludes approximately 45,000 Medicare enrollees and sive Symptomatology (IDS).18 The IDS was modified to 12.000 enrollees covered by Medicaid or by Washing ascertain severity o f patient self-reported symptoms of ton’s Basic Health Plan, a state program for low-income depression occurring during a 2-week period, as required residents. Primary care for adults is provided by approxi by the Diagnostic and Statistical Manual of Mental Dis mately 360 physicians, with each full-time physician being orders, Third Edition, Revised (D SM TII-R) criteria for a responsible for a defined panel of 1600 patients. More diagnosis o f major depression, and for administration than 95% of physicians providing primary care to adults over the telephone. The IDS was administered in the

The Journal of Family Practice, Vol. 40, No. 4(Apr), 1995 353 Cognitive Behavioral Techniques with Depressed Patients Robinson, Bush, Von Korff, etal

4-month follow-up interview with respect to depressive Table 1. Patient Depression Status Before the Prescription symptoms in the 2 weeks before the follow-up interview. Visit and at 4-M onth Follow-Up

The IDS is a 28-item measure o f depression based on Before At 4-Month a scale o f 0 (mild) to 3 (severe) and yielding a score Depression Status Prescription Visit Follow-up ranging from 0 to 84. Rush and colleagues18 evaluated IDS depression severity score, 31.226 (13.797) 15.252 (11,359) the IDS with a group of outpatients, including 211 with mean (SD)* unipolar depression or bipolar disorder, depressed phase, No. of DSM-III-R 4.723 (2.533) 1.736 (1.992) and 23 normal controls. Their results suggest that an IDS symptoms of major score of 36.5 (SD=9.7) is equivalent to a Beck Depres depression, mean (SD)f sion Inventory19 score of 26 (SD=9.3) and a Hamilton Patients with major depression 52.8 10.1 Depression Rating Scale20 score of 18 (SD=5.9). according to DSM-IITR criteria, %

P sychological T r e a t m e n t in t h e P rescription V i s i t Patients with previous episodes 70.8 t The questionnaire on CB techniques used in the present of depression, % study included items selected to parallel widely used CB Patients with dysthymia 2 6 .6 t interventions.21 Additionally, subjects were asked according to DSM-III-R whether their physician had discussed the possibility of criteria, % seeking counseling for depression. * Range o f IDS = 0 to 80, with higher scores indicating greater depression, f DSM-III-R diagnostic criteria fo r major depression include at least 5 of9depimn symptoms, present for at least 2 weeks. j These variables were measured only before the prescription visit. P sychological C o p in g T e c h n i q u e s IDS denotes Inventory o f Depressive Symptomatology18; SD, standard deviation: DSM-IITR, Diagnostic and Statistical Manual of Mental Disorders, Third Edi At the 4-month follow-up phone interview, subjects were tion, Revised. asked to indicate whether they had used any o f six psy chological coping techniques to overcome depression. These included planning regular participation in activities more likely to be married or living with a partner (77% of that (1) are pleasurable, (2) boost confidence, (3) help men, 55% of women, P=.012). Table 1 summarizes data with relaxation, and (4) are with other people; (5) using on level of depression before the prescription visit and at problem-solving techniques for problems in life (eg, the 4-month interview. Fifty-three percent of the subjects problems with work or personal relationships); and (6) were experiencing an episode o f major depression, as de recognizing negative thoughts and replacing them with fined by D SM -IIT R criteria, before the prescription visit. more positive thoughts. These techniques are simplified All other patients had substantial symptoms of depression descriptions o f strategies often used in brief psychological but not enough to qualify for a major depression diagno interventions with depressed patients.21’22 sis. By the 4-month follow-up, most patients remained symptomatic but there was substantial improvement: only 10% continued to meet criteria for major depression. Patient Adherence to Prescribed Medications Thirty-three different physicians were involved in the Patient adherence to prescribed antidepressant medica care of the 155 subjects. Fifty-one (33%) patients were tion regimens was assessed in both interviews. Early ad seen by female physicians and 104 (67%) by male physi herence was defined as continuing use o f antidepressant cians. Almost 60% o f the physicians were 45 years o f age medications for at least 31 days after the initial prescrip or older. Physician age was significantly related to patient tion visit.23 age (F= 11.66, P = .001): older patients had older physi cians.

Results Use o f Cognitive Behavioral Techniques Table 2 shows the percentage o f subjects reporting that Patient and Physician Characteristics their physician had recommended each o f the fiv e CB Most subjects were female, employed, married or living techniques. Overall, patients reported that the physician with a partner, and formally educated for more than 12 recommended a mean o f 2.48 strategies in the antidepres years. Male and female subjects differed significantly in sant prescription visit. The sum of responses to the five that men were significantly older (mean age, 52 years for items (l= n o , 2=yes) was used as a summary measured men, 45 years for women; P = .0 0 5 ), less likely to be em interactions about CB strategies to reduce depression ployed (54.8% of men, 77% of women; P = .006), and (Chronbach’s alpha=.77). We refer to this as the Cogni-

354 The Journal of Family Practice, Vol. 40, No. 4(Apr), ft(,: Cognitive Behavioral Techniques with Depressed Patients Robinson, Bush, Von Korff, et al

Table 2. Patient Recognition of Physician Use of Cognitive Older patients, less severely depressed patients, and Behavioral Techniques During the Prescription Visit patients with older physicians were significantly less likely

% o f Subjects to discuss seeking counseling during the primary care Recognizing visit. In a logistic regression, patient age was no longer Physician Use of Item-Total significant (P > .05), while depression severity (P = .02) Technique Correlation Cognitive Behavioral Technique and physician age (P = .05) continued to be significantly Identifying activities already doing 61 0.782 related to discussions about counseling. to help feel better

Planning pleasurable activities 40 0.728 Four-Month Follow-up Solving life problems (eg, those 34 0.632 related to jobs, relationships) U t il i z a t i o n o f C o g n it iv e B e h a v io r a l S t r a t e g ie s

Identifying and changing depressive 31 0.653 At the time o f the 4 -month follow-up, most patients were thoughts using a majority o f the strategies to reduce depression,

Planning activities that boost 27 0 .6 1 4 with 62% reporting use o f four to six CB strategies. Twen confidence ty-five percent were using all six strategies; only 6% denied using any. The mean number of CB strategies used by tive Behavioral Interaction Scale (CBIS) score (range, 5 subjects was significandy related to subject age (r = — 0.26, to 10 ). As shown in Table 2, item-total correlations were P=.001), with the young group reporting a mean of 4.24 high, ranging from 0.61 to 0.78. We separated the high strategies, the middle group reporting 4.15, and the older and low CBIS score groups, with high including subjects group reporting 3.08. whose scores were above the midpoint o f the CBIS score There were several significant relationships between range. CBIS scores were significantly associated with pa CBIS scores and patient self-reports on behavior at the tient age (F=5.8, P=.004). Younger patients (aged 18 to 4-month follow-up. Patients were more likely to be using 39 years) were more likely to report physician use of CB strategies if they had previously reported that their behavioral strategies (CBIS mean, 7.29) than were mid physicians had recommended CB techniques in the orig dle-aged (aged 40 to 55 years, mean, 7.07) and older inal prescription visit (t= — 3.149, d f= 127.3, P = .002). (aged 56 to 75 years, mean, 6.11) patients. This associa Subjects with low CBIS scores reported using a mean of tion remained significant after controlling for depression 3.48 (SD = 1.93) strategies at the 4-month follow-up, severity and physician age (F = 3 .6 , P = .03). CBIS scores while those with high scores reported a mean of 4.43 did not vary significantly by patient sex, level o f education, (SD = 1.57) strategies. Patient-physician interactions previous depression, or level of neuroticism. While there about a specific strategy correlated significandy with the was significant inter-physician variation in CBIS scores subsequent use o f two CBIS strategies: planning pleasur (F= 1.92, P = .007), physician sex and age were unrelated able activities (r= 0 .1 8 2 , P = .023) and discussions about to CBIS scores in multivariate analyses. planning activities to boost confidence (r= 0.238, P = .003). Discussions about solving problems (r= .141, P > 0.05) and discussions about identifying and challeng Interactions About Seeking Counseling ing negative thoughts (r= .0 5 6 5 , P > .05) were not signif Fifty-seven percent of the subjects reported that in the icantly associated with later use o f these strategies. prescription visit their physician discussed with them the option of seeking counseling. Depression severity at the Adherence to Prescribed Medications time of the prescription visit was strongly associated with discussions about seeking counseling (^2=12.314, de Twenty-seven percent of the subjects stopped using anti gree of freedom [df] = l, P<.001). Seventy percent of the depressants within 30 days o f the initial prescription visit. subjects with major depression reported discussing the Multivariate analyses indicated that after controlling for possibility o f seeking counseling, compared with 42.5% of covariates, higher CBIS scores were associated with the subjects with subthreshold depression. Chronicity of higher adherence rates in the first 31 days following the depression, on the other hand, was associated with a lower initial prescription visit. Seventy-eight percent of the 71 rate of interaction about seeking counseling (^ = 3.911, subjects with high CBIS scores reported continued use of df = 1, P = .048). Forty-four percent o f the dysthymic antidepressant medication for 31 or more days, while only group as compared with 63% o f the patients without a 60% of the 68 subjects with low CBIS scores continued history' of dysthymia reported discussions about seeking taking their medication (y2=4.79, df = 1, P = .03). When counseling. we evaluated the CB strategies individually, planning

The Journal of Family Practice, Vol. 40, No. 4(Apr), 1995 355 Cognitive Behavioral Techniques with Depressed Patients Robinson, Bush, Von Korff, et al

pleasurable activities, as a strategy suggested by the phy patient self-report, the validity o f which is unknown. Pa sician, was significantly related to adherence (^2 = 7.0, tient knowledge o f and previous use o f CB strategies may d f= l, P = .03) while the others were not significantly significantly influence not only the occurrence of interac related.21 Higher rates o f utilizing cognitive behavioral tions about these strategies with physicians but also the techniques in the months following the prescription visit subsequent use of these strategies. Patient perception of were also related to successful adherence during the first the physician recommendation o f counseling and CB 31 days of pharmacotherapy (t= —2.223, df= 66.9, techniques also may be influenced by the strength of the P= .03), with adherent patients reporting using a mean of physician-patient relationship, ie, frequency of visits and 4.12 strategies and nonadherent patients reporting using duration o f the relationship. Future studies in this area a mean o f 3.34 strategies. should formally assess and control the impact of the strength of the patient-physician alliance.24 Another limitation is that we measured a limited Discussion number of specific psychotherapeutic strategies rather than taking a more open-ended or comprehensive ap The present study is an attempt to examine the use of CB proach to measuring psychotherapeutic strategies. We techniques by the primary care physician from the per also did not control for potential important process vari spective o f the patient. Patients typically reported that ables, such as length o f the visit. Future studies should their physicians suggested two to five CB techniques in define and control process variables and wrestle with com the antidepressant prescription visit. Patient age appeared plex measurement and design issues. Such studies would to play a role in the occurrence o f interactions about CB enhance our understanding o f how primary care physi strategies, and in the actual use o f CB strategies subse cians use psychological techniques and would help define quent to the visit. More interactions occurred with the relationship between physician use o f CB techniques younger patients than with older patients, and after the and medication adherence by depressed patients. visit, younger patients utilized more strategies than did Primary care physicians see more depressed patients older patients. This age difference in rates o f discussing than any other professional caregiver group in the United CB strategies was not attributable to either depression States.25 The present study suggests that patients perceive severity or physician age. One possible explanation for physicians to be using CB techniques in conjunction with this trend is that younger patients may accept the diagno antidepressant therapy. It also shows that physician dis sis o f depression more readily and be more interested in cussion of CB strategies was associated with patient use of psychological treatments for depression. Increased pa tient acceptance of diagnosis and sophistication about these strategies and adherence to prescribed antidepres treatment options may provoke discussion o f CB and sant therapy subsequent to the initial prescription visit. other psychological treatments in the primary care visit. It This finding suggests that primary care physicians’ efforts is also possible that younger patients are engaging in more to educate depressed patients about CB strategies maybe CB techniques before making a primary care visit con worthwhile. Future researchers should examine the rela cerning depression. tion between the psychological aspects o f treatment of Slightly more than one half of the depressed primary depression (alone and in combination with pharmacolog care patients in the present study reported discussions ical treatment) and patient outcomes in primary care, about seeking counseling during the antidepressant pre Additional information about the efficacy o f these treat scription visit. Increased severity and lack o f chronicity ments in primary care may be obtained from randomized were associated with a greater likelihood of discussing controlled trials, in which standardized, brief cognitive counseling as a treatment option. This finding may sug behavioral treatments are delivered by primary care phy gest that physicians are discriminating between levels of sicians or mental health professionals, or both, working severity among depressed patients and responding with collaboratively in the primary care setting. more appropriate treatment planning. That older physi cians were less likely to discuss counseling with their pa tients may be related to the differences in training be Acknowledgments tween older and younger physicians, the tendency of older physicians to have older patients, the length and This research was supported by a National Institute of Mental Health grant No. N IM H M H 4 -1 7 3 9 , Rockville, Md, to study depression® strength of the patient-physician relationship, or a com primary care treatment centers. bination o f these factors. The present study has several limitations. First, to We would like to express our appreciation to Kirk Strosahl, PhD, for his measure cognitive behavioral interventions, we relied on help with earlier drafts o f this paper.

356 The Journal of Family Practice, Vol. 40, No. 4(Apr), 1$ Cognitive Behavioral Techniques with Depressed Patients Robinson, Bush, Von KorfF, et al

References 12. Delbanco T. The healing roles o f doctor and patient. In: Moyers B. Healing and the mind. New York, NY: Doubleday, 1993: 7 -2 4 . 1 Von KorfFM, Shapiro S, Burke JD , et al. Anxiety and depression in 13. Catalan J, Gath D H , Anastasiades P, Bond SAK, Day A, Hall L. a primary care clinic. Arch Gen Psychiatry 1 9 8 7 ; 4 4 :1 5 2 -6 . Evaluation o f a brief psychological treatment for emotional disorder 2 Barrett JE, Barrett JA, Oxman TE, Gerber PD. The prevalence of in primary care. Psychol Med 1991; 2 1 :1 0 1 3 -8 . psychiatric disorders in a primary care practice. Arch Gen Psychiatry 14. Schulberg H C , Coulehan JL, Block MR, Scott CP, Imber SD, Perel 1988;45:1100-6. JM. Strategies for evaluating treatments for major depression in 3. Schurman R, Kramer P, Mitchell J. The hidden mental health net primary care patients. Gen Hosp Psychiatry 1991; 1 3 :9 -1 8 . work. Arch Gen Psychiatry 1985; 42:89-94. 15. Emanual EJ, Emanual LL. Four models o f the physician-patient 4. Thompson J, Rankin H , Ashcroft G, et al. The treatment of depres relationship. JAMA 1992; 2 6 7 :2 2 2 1 -6 . sion in general practice: a comparison o f L-tryptophan, amitripty 16. Orleans C, George L, Houpt J, Brodie H. How primary care phy line and a combination o f L-tryptophan, amitriptyline, with a pla sicians treat psychiatric disorders: a national survey o f family prac cebo. Psychol Med 1982; 1 2 :7 4 1 -7 5 . titioners. Am J Psychiatry 1985; 142 :5 2 -7 . 5. Paykel E, Hollyman J, Freeling P, Sedgwick P. Predictors of thera 17. Katon W. Depression: the epidemiology of depression in medical peutic benefit from amitriptyline in mild depression: a general prac care. Int J Psychiatry Med 1987; 17:93-112. tice placebo-controlled trial. J Affect Disord 1988; 1 4 :8 3 -9 5 . 18. Rush J, Giles D E, Schlesser MA, Fulton C L, Weissenburger J, 6. Blackburn I, Bishop S, Glen A, et al. The efficacy of cognitive Burns C. The Inventory for Depressive Symptomatology (IDS): therapy in depression: a treatment trial using cognitive therapy and preliminary findings. Psychiatry Res 1985; 1 8 :6 5 -8 7 . pharmacotherapy, each alone and in combination. Br J Psychiatry 19. Beck AT, Ward CH, Mendelson M, Mock JE, Erbaugh JK. An 1981; 139:181-9. inventory for measuring depression. Arch Gen Psychiatry 1961; 7. Teasdale J, Fennel M, Hibbert G, Amies P. Cognitive therapy for 4 :5 6 1 -7 1 . major depressive disorder in primary care. Br J Psychiatry 1984; 20. Hamilton M. A rating scale for depression. J Neurol Neurosurg 1 4 4 :400-6. Psychiatry 1960; 12:56. 8. Schulberg H, McClelland M, Gooding W. Six-month outcome for 21. Lewinsohn P, Munoz RF, Youngren MA, Zeiss A. Control your medical patients with major depressive disorders. J Gen Intern Med depression. New York, NY: Prentice Hall, 1986. 1987;2:312-7. 22. de Shazer S, Kim Berg I, Lipchik E, Nunnally E, Molnar A, Gin- 9. Katon W, Schulberg C. Epidemiology of depression. Gen Hosp gerich W , Weiner-Davis M. Brief therapy: focused solution devel Psychiatry 1992; 1 4 :2 3 7 -4 7 opment. Fam Process 1986; 25:207-21. 10. Reiger DA, Hirschfield MA, Goodwin FK, Burke JD, Lazar JB, 23. Lin F2HB, Von KorfFM, Katon W, Bush T , Simon GE, Walker E, Judd L. The NIM H depression awareness, recognition and treat Robinson P. The role of the primary care physician in patient ad ment program: structure, aims, and scientific basis. Am J Psychiatry herence to antidepressant therapy. Med Care 1995; 3 3 (l):6 7 -7 4 . 1988; 145:1351-7. 24. Mallinckrodt B. Session impact, working alliance and treatment out 11. Olfson M, Klerman GL. The treatment of depression: prescribing come in brief counseling. J Counseling Psychol 1993; 40 :2 5 -3 2 . practices of primary care physicians and psychiatrists. J Fam Pract 25. Howard KI, Lueger R, Schank D. The psychotherapeutic service 1992; 35:627-35/ delivery system. Psychother Res 1992; 2 :1 6 4 -8 0 .

The Journal of Family Practice, Vol. 40, No. 4(Apr), 1995 357