| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
Correspondence: Address correspondence to Patricia A. Arean, PhD, University of California San Francisco, Department of Psychiatry, 401 Parnassus Avenue, San Francisco, CA 94143. E-mail: pata{at}lppi.ucsf.edu
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Depression • Geriatrics • Primary care • Psychotherapy
Because older adults are more likely to seek mental health services in primary care medicine (Arean & Miranda, 1996), a growing number of health care settings, such as the Veterans Affairs (VA) system and health maintenance organizations (HMOs), are beginning to provide brief mental health treatment within primary care. A dilemma faced by these settings is that even the briefest interventions (such as cognitive behavioral therapy) were not designed with the primary care setting or patient in mind. Primary care settings are not designed for visits longer than 30 minutes, and they do not typically employ staff with mental health expertise. Problem-solving therapy for primary care (PST-PC; Mynors-Wallis, Gath, Lloyd-Thomas, & Tomlinson, 1995) is a psychotherapeutic intervention created specifically to address the time and resource issues present in primary care medicine; it is a brief intervention, lasting between four and eight sessions, and is adapted so that non-mental-health providers can learn to administer the intervention. According to the protocol, monthly group meetings for patients who respond to treatment can also be provided. Research finds that PST-PC is as effective as antidepressant medication in treating major depression in younger primary care patients (Mynors-Wallis et al., 1995; Mynors-Wallis, Gath, Day, & Baker, 2000), and traditional forms of PST (Arean et al., 1993) have been found to be effective in treating depression in older adults (Alexopoulos, Raue, & Arean, 2003; Arean et al.; Charney et al., 2003; Kendrick et al., 2005; Mynors-Wallis et al., 2000). However, PST-PC has not been investigated in older primary care patients with major depression, and it has not been compared with psychotherapy that older adults would typically receive from community providers.
Our purpose in this study is to compare a primary-care-based psychotherapy, namely PST-PC, to community-based psychotherapy in treating depression in older, primary care patients. In this study, community-based psychotherapy includes counseling or psychotherapy that is available to older adults who receive primary care in the VA system, HMO system, private provider organizations, and the county health care system. Because PST-PC has been adapted for medical patients and is tailored to the preferences and needs of these patients, and because the treatment is integrated into the patient's primary source of health care, we hypothesize that people treated with PST-PC will have better depression and functional outcomes than people treated with community-based psychotherapy. To test this hypothesis, we ran two analyses; one included participants who were also treated with and without medication, and one excluded everyone who received antidepressant medications.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Participants
Recruitment and Consent
Researchers recruited primary care patients who were 60 years or older from 18 diverse primary care clinics in six cities (five states). Participants were recruited either through provider referral, clinic screening, or by mail. Participants who were interested in participating were explained the study procedures by a research assistant who was trained to provide quality informed consent by the study-coordinating center (University of California at Los Angeles). All participants signed written informed consent forms approved by the Institutional Review Boards at the study-coordinating center and all study sites.
Sample Size
Researchers randomized 1,801 primary care patients with major depression or dysthymia to collaborative care or usual care; all patients spoke English. There were 269 participants in the collaborative care arm who received PST-PC, and 85 received PST-PC only (no antidepressant medications). In the usual care arm, 164 participants received CBP, of whom 46 had no medication in addition to psychotherapy. Therefore, a total of 433 participants received psychotherapy, and of those 131 received psychotherapy only. For the purposes of this article, we analyze both subgroups, those who received both psychotherapy and antidepressant medication (n = 433) and those who received psychotherapy only (n = 131). For a detailed description of recruitment, screening, and sampling strategies, see the article by Unutzer and colleagues (2000). Subject flow is presented in Figure 1.
|
|
PST-PC was only available to participants in the first year following randomization. Thus, in the second year of the study, those who received PST-PC had access to community psychotherapy.
Interventions: Therapist Training in PST-PC
There were 11 DCSs (8 clinical nurse specialists, 2 clinical psychologists, and 1 social worker) who were trained in PST-PC by two licensed clinical psychologists who are expert in the use of this therapeutic model. Training consisted of a provision of a treatment manual, as well as a daylong workshop to introduce the principles of PST-PC and to role-play the intervention. Shortly thereafter, DCSs identified between three and five training cases. The first, third, and last sessions of PST-PC were recorded for each training case and were reviewed by the two trainers. The trainers assessed adherence to PST (if therapists administered PST as planned) and competence (therapist skill in delivering PST-PC) by using the Problem Solving Treatment Provider Adherence Checklist (Hegel et al., 2004). After DCSs demonstrated proficiency on the checklist, they were considered certified in the use of PST-PC. DCSs reviewed all their patients with a study psychiatrist on a weekly basis to monitor treatment outcome, and they participated in a monthly supervision call with the PST-PC trainers and the Principal Investigator for the IMPACT study. The Patient Health Questionnaire–9 (Kroenke, Spitzer, & Williams, 2001) was completed at each PST-PC session. We used the scores to track patient improvement over time and to inform treatment options for those patients who did not show improvement. Thus, ongoing consultation on PST-PC cases was an integral part of the intervention. Finally, to determine fidelity to treatment, we required DCSs to audiotape all their PST-PC sessions for review at random by the trainers. Average quality scores based on the problem solving treatment provider adherence and competency (PSTPAC) were maintained at 4 for the entire study period.
Interventions: CBP
Control subjects were encouraged to continue in care as usual. This could have been continuing care with their primary care provider and any mental health specialty provider of their choosing, or not receiving any mental health treatment at all. No services were withheld from this group and the study team provided no special services. Counseling or psychotherapy provided to the usual care patients who elected to be treated with these services consisted of treatment they would normally have access to through their health plans. The health plans represented in this study employ masters- to doctoral-level clinicians trained in psychotherapy. Of the eight study sites, two study sites were VA sites, four sites were group model HMOs, one site was a county hospital with a mental health outpatient service, and the final site was a preferred provider organization.
Further, the patients included in the analyses are those who indicated that they received psychotherapy or counseling from a social worker, psychologist, or any other mental health provider. Although the quality of the therapy and therapists are likely to vary, they are also likely to represent the typical psychotherapy older adults would encounter in the community. Patients in CBP reported having received psychotherapy, but no information as to the specific type of psychotherapy was collected by the study. We suspect that the type of therapy received was fairly representative of the psychotherapy available in the diverse communities from which our study participants were recruited and represent a mix of eclectic therapy, CBT, psychodynamic therapy, and supportive therapy (Halpern, Johnson, Miranda, & Wells, 2004; Kavanagh, Littlefield, Dooley, & O'Donovan, 2007; Olfson & Pincus, 1994a, 1994b, 1994c; Olfson, Pincus, & Dial, 1994). However, it is important to note that this study is not a comparison of PST-PC versus any specific therapy, but a comparison of PST-PC versus counseling-based services that older adults are likely to encounter in the communities in which they reside.
Data Collection
Data reported here are from baseline and from 12-month and 24-month follow-up interviews. Baseline data were collected before randomization by trained interviewers using a computer-assisted interview. A telephone survey research group conducted blind follow-up interviews at both time periods. Overall response rates were 83% at 12 months and 77% at 24 months.
Measures
In addition to information on demographics, we also collected data on clinical diagnosis (baseline only), depression severity (baseline and follow-up), functional impairment (baseline and follow-up), and service utilization (baseline and follow-up). To determine the participants' eligibility for inclusion, we had the Structured Clinical Interview for DSM-IV (SCID; J. B. Williams et al., 1992) administered at baseline. The SCID is a semistructured clinical interview that has been used widely in clinical trials. Intake clinicians administered the SCID and were trained by use of the SCID training tapes and interrater methods.
The Hopkins Symptom Checklist–20
We assessed depression outcomes in two ways. One way was the severity of depression, and the other was the number of depression-free days over a period of time. We determined each outcome by using the Hopkins Symptom Checklist–20 (HSCL-20; Sheehan, Harnett-Sheehan, & Raj, 1996). The HSCL-20 incorporates the 13 items from the Hopkins Symptom Checklist Depression Scale and 7 additional items intended to more completely assess criterion symptoms and improve responsiveness. To calculate depression severity, the list rates each item on a 5-point scale, from 0 (not at all) to 4 (extremely), and the total score is computed as the average item response. Williams and colleagues (J. W. Williams, Stellato, Cornell, & Barrett, 2004) evaluated the HSCL-20 in a primary care population of depressed adults and reported good reliability (Cronbach's alpha = 0.86). In the present study, the HSCL-20 demonstrated good reliability, with Cronbach's alpha = 0.83.
Depression-Free Days
We used the HSCL-20 scores measured at baseline and follow-up assessments to compute the main outcome for this study, which is the estimated number of depression-free days (DFDs) during the 12-month and 24-month follow-up periods. The purpose of estimating DFDs is to put statistical measures of outcomes into a clinically meaningful context. Estimating DFDs from depression severity scores has been used in several trials of depression treatment as a means of determining clinical significance (Araya, Flynn, Rojas, Fritsch, & Simon, 2006; Ciechanowski et al., 2006; Lave, Frank, Schulberg, & Kamlet, 1998; Liu et al., 2003; Lynch et al., 2005; Mallick, Chen, Entsuah, & Schatzberg, 2003; Montgomery & Andersen, 2006; Pyne, Tripathi, Williams, & Fortney, 2007; Revicki et al., 2005; Simon et al., 2001, 2002; Trivedi et al., 2004). Further, DFDs are directly related to quality-adjusted life years, and thus they can also serve as a measure of improvement in quality of life (Pyne et al.; Shanti & Tripathi, 2004). We calculate DFDs by using linear interpolation to estimate the daily depression severity across assessment points, baseline, 3 months, 6 months, 12 months, 18 months, and 24 months in this study (Lave et al.).
We considered days with HSCL-20 scores below 0.5 to be fully depression free; we assigned days with scores between 0.5 and 1.7 a score between 0 and 1; and we considered scores above 1.7 to make zero contribution to the DFD estimate (they are fully depressed days). We estimated DFDs between two consecutive assessments by multiplying the average of the two DFDs scores by the number of days during the interval. For example, a participant whose HSCL-20 score at baseline and 3 months was converted to 0.1 and 0.9 on the DFD scale, respectively, would yield 45 (the average of 0.1 and 0.9 times 90 days) DFDs in the 0- to 3-month interval. We then summed the DFDs of the 0- to 3-month, 4- to 6-month, and 7- to 12-month intervals to yield the total DFDs during the 12-month period. We derived the 0- to 24-month DFDs by adding the DFDs of the 13- to 18-month and 19- to 24-month intervals to the 0- to 12-month DFDs. Estimates of DFDs have been reported using the Hamilton Depression Rating Scale, the Beck Depression Inventory, and the HSCL-20. No standards exist for establishing scale cutoffs for the interpolation process. We adapted our HSCL-20 thresholds for computing DFDs from the research by Simon and colleagues (2002). In their article, they reported 0.5 and 2.0 as the thresholds for "one" DFD and zero DFD, respectively. We used an upper cutoff point of 1.7, the mean baseline HSCL-20 score of IMPACT participants, to better reflect this sample's reported depression severity, all of whom met SCID criteria for major depression or dysthymic disorder at the time of study entry.
Sheehan Disability Assessment Scale
We used the Sheehan Disability Assessment Scale (Sheehan et al., 1996) to determine functional impairment. This is a brief, analog disability scale, which uses visual-spatial, numeric, and verbal anchors. The scale has been validated in medical and psychiatric populations with a variety of psychiatric diagnoses (Leon, Olfson, Portera, Farber, & Sheehan, 1997; Sheehan et al.). The scale has been used in depressed medical patients and has been found sensitive to change following patients' antidepressant treatment (Franchini, Zanardi, Gasperini, & Smeraldi, 1999; Unutzer, Simon, Belin, et al., 2000).
Service Utilization
We determined the use of psychotherapy and medication in two ways. At all assessment times, patients in both conditions were asked, "in the past 3 months, have you seen a psychologist, social worker, or counselor for therapy for depression?" If the participants responded "yes," then they were then asked the number of times they saw this counselor in the past 3 months. We included participants who responded yes to the initial question in this analysis. Although self-reported use of services has historically been considered a biased method of determining service use (Marshall, Grayson, Jorm, & O'Toole, 2001), recent research has found that this method of determining service utilization is in fact reliable, when it is compared with claims data (Wallihan, Stump, & Callahan, 1999). However, it is limited in the degree to which one can determine the type and quality of service provided, and we were not able to determine what type of psychotherapy (CBT vs IPT; family vs individual) the participants in the usual care arm had access to, or the type of provider that the participants saw.
Statistical Methods
Descriptive statistics of the demographic variables as well as baseline, 12-month, and 24-month outcome measures are provided. We performed an analysis of variance and chi-square tests to compare the continuous and categorical variables, respectively, between the PST-PC and CBP groups. We used multiple linear regression models to compare the effect of PST-PC to the effect of psychotherapy for CBP participants on DFDs, HSCL-20 scores, and functioning, after we adjusted for use of medication, baseline HSCL-20 score, organization that delivered the treatment, the method of recruitment, age, gender, ethnicity, education, and marital status. We used similar models on the smaller sample in which participants did not use any antidepressants during the 12-month period. We also compared the 12- and 24-month outcomes between the PST-PC patients who elected to participate in maintenance PST-PC and the PST-PC participants who did not elect to participate. Only those PST-PC participants who showed greater than 50% improvement in the Patient Health Questionnaire measure at the fourth or later PST-PC sessions were eligible for this analysis. In order to explore the differences between intervention participants who selected PST-PC and those who did not, we also compared the baseline demographic variables and baseline, 12-month, and 24-month outcome variables between these two groups. We again performed an analysis of variance and chi-square tests for comparing continuous and categorical variables, respectively.
The Pearson correlation coefficients among different outcomes measured at 12 months and 24 months are also provided. In addition, we examined the association between the number of PST-PC sessions in the first 12 months and the 12- and 24-month outcomes by using linear regression models.
We performed all of our analyses with Stata 8.1 (Stata Corporation, College Station, TX) and SAS 9.1 (SAS Institute Inc., Cary, NC).
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Second, participants who received PST-PC used an average of 6.4 sessions in the first year, and participants who received CBP had an average of 6.0 sessions in the first year. The difference is not statistically significant. In the second year, however, participants who received PST-PC had an average of 0.8 psychotherapy sessions, whereas the participants who received CBP had an average of 2.7 sessions. The number of psychotherapy sessions over time was not related to DFDs (β = 0.77, ns), depression severity (β = 0.01, ns), or functioning (β = –0.04, ns).
Third, because of the differences in accessibility of treatment between the two study groups, we looked for differences in psychotherapy visits across the five assessment periods. Our analyses revealed no significant differences in the number of visits in each 3-month assessment interval between the two study groups in the first year. However, between Months 12 and 24 of the study (the follow-up period), the CBP participants received significantly more psychotherapy visits than did the PST-PC participants (see Table 2). Given the short assessment window (3 months), and the fact that there were no differences in the number of visits, the data suggest that there was no significant difference in how long participants waited before engaging in treatment.
|
Generalizability of the Sample: Differences in Participants Who Selected Psychotherapy Versus Those Who Did Not
To further describe our sample, we ran analyses to determine whether there were any baseline differences between IMPACT study participants who selected psychotherapy as their choice of depression treatment and those participants who did not. As stated in Table 1, the only differences we found between CBP and PST-PC participants was age and ethnicity, both of which we controlled for in the longitudinal analyses. For participants in usual care, the only baseline difference between those who chose psychotherapy and those who did not was that those who selected psychotherapy were more depressed than those who did not. In the collaborative care group, we found that those who selected PST-PC were more likely to be White, have a higher education, and to be less disabled than those who did not select PST-PC (see Table 3).
|
|
|
Effects of Maintenance PST-PC on 24-Month Outcomes
There were 184 participants who were eligible for maintenance PST-PC, of which 75 elected to participate. As we already stated, maintenance PST-PC was offered to participants who had made a significant recovery from depression after receiving a course of PST-PC. Those who elected to participate in maintenance PST-PC did not differ on baseline or posttreatment characteristics from those who did not elect to participate. The results indicate that those who participated in maintenance PST-PC had no additional effect on depression and functioning at 12 months or 24 months.
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
It is common for outcome comparisons to be conducted as a cross-sectional comparison (i.e., evaluating only the difference on the outcome measure at a particular point in time). Our analysis, using the estimated parameter of DFDs, illustrates that a treatment that produces a quicker response may have a profound effect on patient well-being. Although there was no significant difference between depression scores at 24 months, the PST-PC participants had an estimated 95 additional DFDs, representing more than 3 months' worth of symptom relief during the evaluation period. Furthermore, this effect was nearly twice as large for participants who elected to receive only counseling (vs with antidepressant medication); in that case, the PST-PC participants experienced an additional 182 DFDs compared with CBP participants. Hence the nonsignificant difference in cross-sectional comparison of depression scores at 24 months should not be interpreted as an equivalent outcome. Finally, the data indicate that participation in monthly maintenance groups for PST-PC has no discernible effect on maintenance of treatment gains; improvement from PST-PC tends to be persistent over a 2-year period, with or without the monthly booster sessions. Finally, number of sessions in PST-PC does not affect treatment outcomes.
Clinical Implications
First, the implication that a very brief intervention delivered in primary care medicine can treat the symptoms and functional consequences of depression in older adults better than treatments delivered in the community is of some importance. PST-PC can be feasibly delivered in health care settings, the setting where older adults are likely to seek treatment for depression. It is also a relatively cost-effective intervention in that it can be delivered by existing health care providers within the primary care setting, provided they are given ample training and guidance. Furthermore, given that the effects of treatment appear to occur sooner in the course of treatment than they do in community-based settings, PST-PC may be an intervention of choice for older adults, either inside or outside the primary care setting.
We do acknowledge that the level of expert support that DCSs received around the delivery of PST-PC is more than that which is generally available to most therapists, and the differential treatment effects may be due in part to these differences in access to expert consultation. Research on decision support in psychotherapy has found that, even in community-based treatment, providing ongoing assessment of clinical outcomes and consultation for individuals who are not responding can substantially improve clinical outcomes in psychotherapy patients (Lambert, 2005). Unfortunately, these processes are not readily available to community therapists (Aegistadottir et al., 2006). The data from this study, coupled with recent findings about improving CBP outcomes, suggest that current standards of mental health care may be improved simply by supporting the use of guideline treatment and outcome monitoring. Thus, we anticipate that the community care outcomes may have occurred more rapidly if providers had the same degree of access to ongoing treatment monitoring and consultation as was available to the PST-PC specialists. Future research should clarify this issue by comparing PST-PC as it was delivered in this study to CBP that is enhanced with expert consultation and outcome monitoring.
A second implication from these findings is that once participants improve from the acute-phase treatment of PST-PC, ongoing monthly maintenance sessions do not influence the maintenance of treatment gains. These findings should be interpreted with caution, however. First, participation in maintenance PST-PC was voluntary, at the discretion of the participant and DCS. Although we found no significant baseline or postacute treatment differences between those who elected to participate in maintenance PST-PC and those who did not elect to participate, differences such as confidence in executing PST-PC without further guidance from the DCS may have been a factor that contributed to both the decision to participate and in ultimate treatment outcome. That is, those who did not attend maintenance groups may have felt quite comfortable with PST-PC, and those who did attend felt less secure in implementing the new skills.
In addition, all participants who received PST-PC were followed by their DCS monthly after acute treatment. It is possible that participants who did not attend the maintenance groups still received maintenance support from the DCS individually, thus diluting group differences. Future research should further elucidate the role that maintenance PST-PC has on treatment outcome by randomizing those who respond to acute PST-PC to maintenance PST-PC or no ongoing support.
Another important finding is the eventual lack of group difference between PST-PC and CBP at 24 months. The long-term effects of PST-PC begin to diminish whereas the long-term effects of CBP begin to improve. It is important to note a few things from these findings. Although the PST-PC group shows some increase in depression symptoms and functioning by 24 months, these symptoms never return completely to baseline levels, suggesting some overall sustained improvement over time. Although there is improvement in CBP by 24 months, the mean depression symptoms and functional outcomes never quite reach the level that was evident in PST-PC at 12 months. Further, increased improvement in CBP may be due to participants having used significantly more therapy sessions between 12 and 24 months of the study. Implications from this data suggest that PST-PC has strong, positive effects on depression acutely and during the year after treatment, but once support from a depression care manager is no longer available to continue the use of PST-PC, we eventually see relapse that is most common after discontinuation of treatment. Continued use of psychotherapy, as was found in the CBP group, resulted in improvements in depression and functioning over time.
Limitations
Although this is one of the first studies to our knowledge to compare primary-care-based psychotherapy to CBP, there are limitations that require mention to put this data into perspective. First, although participants were randomized to receive collaborative care in primary care or usual care, participants were not randomized to psychotherapies, and therefore this design is quasi-experimental rather than a true randomized trial, introducing some threats to internal validity. We found that those who select psychotherapy for the treatment of their depression may be different from those who select other forms of treatment. This study at a minimum provides information about the effectiveness of psychotherapy among patients who self-select for psychotherapy. Second, we did not have information as to the type of therapy CBP participants received; thus, we can only make interpretations based on the relative effects of PST-PC compared with psychotherapy that is available to patients with health plans represented in this study. Therefore, this study does not provide information about the relative effects of PST-PC compared with other evidence-based treatments; it provides information only about the effects of PST-PC compared to what is offered in these communities. A final limitation, which we already discussed, is the fact that PST-PC was delivered in the context of a broader care management program that included ongoing patient monitoring and consultation. Future research should therefore investigate the added effect of this support on the effects of psychotherapy for treating depression in older primary care patients.
Conclusions
Despite the methodological limitations of this study, the data presented here suggest that depression in older primary care patients can be successfully treated in the primary care setting by using therapies developed for these settings. In addition, access to expert consultation may also improve the quality of depression care for these patients. The long-term impact of these therapies, however, indicates that either ongoing support from a DCS or modifications to PST-PC are warranted to secure continued remission of depression. These promising data should provide the field with sufficient impetus to investigate the issues that this study raises: Does the quality of community-based care have to be improved to better address the treatment of depression in older primary care patients? What aspects of primary-care-based treatment (brief structured therapy or ongoing consultation) influence the quality of care for depressed older patients? This study is a first step in the process of improving depression psychotherapies for a majority of older people in need of care.
|
Footnotes |
|---|
1 Department of Psychiatry, University of California, San Francisco.
2 School of Medicine, Dartmouth University, Hanover, NH.
3 Department of Pyschiatry, University of Washington, Seattle.
Decision Editor: William J. McAuley, PhD
Received for publication March 8, 2007. Accepted for publication August 27, 2007.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
