Abstract
Informed by data on the dose-response effect, the authors assessed use of psychotherapy in the Veterans Health Administration (VA). The authors identified 410,923 patients with newly diagnosed depression, anxiety, or posttraumatic stress disorder using VA databases (October 2003 through September 2004). Psychotherapy encounters were identified by Current Procedural Terminology codes for the 12 months following patients’ initial diagnosis. Psychotherapy was examined for session exposure received within the 12-month follow-up period and time (in days) between diagnosis and treatment. Of the cohort, 22% received at least one session of psychotherapy; 7.9% received four or more sessions; 4.2% received eight or more sessions; and 2.4% received 13 or more sessions. Delays between initial mental health diagnosis and initiation of care averaged 57 days. Patient variables including age, marital status, income, travel distance, psychiatric diagnosis, and medical-illness burden were significantly related to receipt of psychotherapy. Treatment delays and general underuse of psychotherapy services are potential missed opportunities for higher-quality psychotherapeutic care in integrated health care settings.
Keywords: psychotherapy, mental health, depression, anxiety, posttraumatic stress disorder
A significant number of patients in need of mental health treatment are not recognized and are either not treated or provided with inadequate care (Unutzer, Schoenbaum, Druss, & Katon, 2006; Young, Klap, Sherbourne, & Wells, 2001; Wei, Sambamoorthi, Olfson, Walkup, & Crystal, 2005). Although the efficacies and modalities of mental health treatments vary by condition, much evidence supports the provision of psychotherapy as a primary or secondary (adjunctive) treatment, especially for depression, anxiety, and posttraumatic stress disorder (PTSD; Westen & Morrison, 2001; for review see Lambert & Ogles, 2004). Despite the importance of psychotherapy as a treatment modality and evidence indicating that patients often prefer psychotherapy (57%) to medication (43%; Gum et al., 2006), research on the quality of mental health care has largely focused on pharmacotherapy (Busch, Leslie, & Rosenheck, 2004; Charbonneau et al., 2003, 2004), with little information available on the use and quality of psychotherapy practices, especially within integrated medical care settings.
Psychotherapeutic services provided within the United States have been described as “shallow” and of limited benefit, based on initial studies examining patterns of mental health care utilization (Olfson, Marcus, Druss, & Pincus, 2002). Using data from nationwide medical expenditure surveys in 1987 and 1997, Olfson et al. (2002) reported that 3.2% to 3.6% of community-dwelling U.S. civilians used outpatient psychotherapy services, of whom approximately 35% received only one or two sessions, and 40% received three to 10 sessions. Young et al. (2001), using a cross-sectional telephone survey and brief semistructured interview, reported that persons in the United States with probable depression or anxiety used counseling at rates of 29.9% and 20.3%, respectively. Notably, counseling was somewhat loosely defined by patient self-report and included not only specialty mental health encounters but also discussions with primary care providers about mental health issues. This study also assessed the frequency of psychotherapy sessions and found that 14.8% of patients with depression, 10.8% of patients with anxiety, and 28.9% of patients with depression and anxiety were provided with four or more sessions. In a separate naturalistic study examining adequacy, Hansen, Lambert, and Forman (2002) found that the average number of psychotherapy sessions was less than five in a cohort consisting primarily of patients from employee assistance programs and university counseling centers.
To better understand psychotherapy exposure, the dose-response effect literature provides an evidence-based measurement criteria (based on session frequency) that can be applied within larger systems of care to assess practice patterns and utilization. The provision of a psychotherapy session, although a crude unit of measurement, represents a quantifiable criterion that is roughly comparable across types of treatment (Howard, Kopta, Krause, & Orlinsky, 1986). Research has shown that, in general, the amount of exposure to psychotherapy (e.g., number of sessions) is positively related to beneficial outcomes (Orlinsky, Roonestad, & Willutzki, 2004). This dose-response effect as originally described by Howard et al. (1986) reflects the percentage of patients showing clinically significant improvements in relation to the number of sessions attended. Using a compilation of research studies, Howard et al. (1986) reported that approximately 50% of patients with depression and/or anxiety measurably improve after eight sessions of psychotherapy; while 75% improve after 26 sessions. More recent evidence indicates response rates are affected by patient-, provider- and system-level factors (Shapiro et al., 2003), and that, although eight sessions may suggest a lower limit for attaining positive changes in 50% of symptomatic patients, clinically significant change within naturalistic treatment settings may require 15 or more sessions (Hansen & Lambert, 2003; Shapiro et al., 2003).
We are not aware of any studies of patterns of psychotherapy use within a large, integrated medical care setting. Our aims were to examine utilization patterns of psychotherapy provided within the Veterans Health Administration (VA). The VA is the largest integrated health care system in the United States and provides a wide range of mental health services to a diverse population of veterans (Hankin, Spiro, Miller, & Kazis, 1999). In addition to obtaining general information related to psychotherapy in the VA, we sought to examine psychotherapy exposure (the number of sessions received) and time from diagnosis to initiation of treatment and to explore patient and institutional factors associated with receipt of psychotherapy.
Method
This retrospective administrative database cohort study utilized patient data from the VA out-patient treatment files (OPF), a national database maintained for all individuals receiving care within the VA system. The VA OPF contains encrypted patient identifiers which are attached to broad patient and service characteristics (e.g., demographic variables, clinical diagnoses, and clinical procedures). The accuracy and validity of these databases are monitored by the VA and the VA Information Resource Center (VIReC) serves as a resource for utilizing VA databases in health services research.
Patient Population
This study focused on patients receiving a new diagnosis of depression, anxiety, or PTSD in VA outpatient facilities during the 2004 fiscal year (October 1, 2003, to September 30, 2004). Using International Classification of Diseases–Ninth Edition–Clinician Modification (ICD-9-CM) codes, we extracted and categorized patients into three diagnostic groups, as follows: depression (293.83, 296.20–296.36, 300.4, 311), anxiety (293.84, 293.89, 300.00 –300.02, 300.09, 300.20 –300.23, 300.29, 300.3), and PTSD (308, 309.81). PTSD was separated from other anxiety disorders because of its potential for unique contributions to our analyses, as well as its high prevalence in this patient population. The construction of the database around patients with diagnosed depression, anxiety, and PTSD was chosen for two reasons: 1) these three conditions are highly prevalent and disabling within the VA patient population (Hankin et al., 1999) and 2) all three conditions have substantial evidence to support the use of psychotherapy as a primary or secondary treatment (Lambert & Ogles, 2004).
The cohort was restricted to new-onset diagnoses of depression, anxiety, and/or PTSD to evaluate time between initial diagnosis and receipt of psychotherapy (if any) and to standardize the extracted patient cohort. Restricting our cohort to newly diagnosed mental health conditions allowed for increased assurance that patients in the cohort may have benefited from a trial of psychotherapy. For example, patients with a longstanding mental health diagnosis may not require active treatment while newly diagnosed patients were more likely to need acute care.
A new onset condition was defined as 6 months without a related diagnosis before the index date (the date of first diagnosis during the study period). To avoid overlapping ICD-9 coding, conditions occurring within diagnostic categories (e.g., major depression, dysthymia, depression not otherwise specified) were evaluated collectively and considered as one condition. Notably, patients with multiple new onset conditions were classified into more than one diagnostic category (e.g., a patient with new onset depression and new onset PTSD would be classified into both categories for analytic purposes).
Our sample was restricted to patients with adequate opportunities to use outpatient mental health services by excluding those with 60 or more inpatient hospital days in the 180 days following the index date. Patients with a large number of inpatient bed days were deemed to have fewer opportunities for receipt of outpatient care and therefore may provide skewed data reflecting underuse of psychotherapy. Lastly, patients who died during the 12-month follow-up were excluded. The final cohort included 410,923 patients.
Psychotherapy, Session Frequency, and Time to Treatment
Patient use of mental health services (namely psychotherapy) was assessed during the 12 months following each patient’s index date. Total mental health encounters were assessed using the full spectrum of mental health Current Procedural Terminology (CPT) codes (90801–90911, 96100–96155). Use of administrative codes for assessing medical health services is considered valid and appropriate (Quan, Parsons, & Ghali, 2004).
Psychotherapy use was determined on the basis of CPT codes consistent with psychotherapy without medication management (90804, 90806, 90808, 90810, 90812, 90814, 90845, 90846, 90847, 90849, 90853, 90857, 90875, 90876, 96152, 96153, 96154, 96155). Psychotherapy was further classified by treatment modality into individual, group, and family (individual psychotherapy: 90804, 90806, 90808, 90810, 90812, 90814, 90845, 90875, 90876, 96152; group therapy: 90849, 90853, 90857, 96153; family therapy: 90846, 90847, 96154, 96155). Psychotherapy with medication management codes (90805, 90807, 90809, 90811, 90813, 90815) were not examined in this study, as these procedures were unlikely to utilize weekly psychotherapy sessions, would potentially inflate the number of patients getting low levels of psychotherapy exposure, and represented a form of mental health treatment that was outside the focus of the current study.
The total number of psychotherapy sessions was calculated for the 12 months following each patient’s index date. Psychotherapy exposure was examined based on various session cutoffs. Howard et al. (1986) reported that exposure to psychotherapy related to increasing sessions, with patient improvement occurring as follows: no sessions = 14% of patients improved; one session = 24%; two sessions = 30%; four sessions = 41%; eight sessions = 53%; 13 sessions = 62%; and 26 sessions = 74% improved. For the purposes of assessing exposure at a dichotomous level, a somewhat arbitrary yet liberal cutoff of eight or more sessions was used to define extended (vs. limited) exposure. The cutoff of eight sessions reflects the point at which 50% of patients improved in the original Howard work but may be an underestimate of need for naturalistic settings (Hansen & Lambert, 2003).
As a second measure, we assessed the time-liness of psychotherapy. To assess this construct, we calculated the number of days between the first psychotherapy encounter and the patient’s index date.
Covariates
Sociodemographic characteristics of patients included age, gender, marital status, income (estimated using the average adjusted gross income for each patient zip code—based on 2002 Internal Revenue Services data; Grant, O’Leary, Weilburg, Singer, & Meigs, 2004), and distance (in miles) to the nearest VA facility (calculated using patient zip code). Unfortunately, patient race was not highly populated in these administrative data files and therefore was unable to be utilized for descriptive or predictive analyses. Because of the increased access to care provided to veterans with service-related disabilities, patients were categorized into disability groups, including 0%, 1% to 49%, and 50%+. Illness burden was assessed using a diagnosis-based risk-adjustment methodology (DxCG Company, Boston) (Ellis et al., 1996), validated in the VA population (Petersen et al., 2005). An individual’s Relative Risk Score is his or her total predicted cost divided by the average predicted cost of the population, with a score of 1.0 reflecting an average risk (Ellis et al., 1996; Petersen et al., 2005).
Analyses
Descriptive analyses were used to report the proportion of patients receiving any form of mental health treatment and the mean number of visits for each patient over the 12 months following their index date. Subsequent analyses focused on the receipt and provision of psychotherapy, percentage of patients receiving such services (e.g., various levels of session exposure), and delays between initial mental health diagnosis and attainment of psychotherapeutic care.
Sociodemographic factors were examined as predictors in logistic regression analyses to ascertain variables significantly associated with receipt of psychotherapy. Logistic regression procedures are primarily used for the prediction of dichotomous as opposed to continuous dependent variables. Interpretation of logistic regression is focused around the concept of odds (the probability of membership in the target group) and the odds ratio (an estimate of change in the odds of membership in a target group for a one-unit increase in the predictor; Wright, 2003). Therefore, odds that fall below 1.0 reflect less odds of group membership while values above 1.0 reflect increased odds of membership. Logistic regression analyses in this study focused on predicting receipt of at least one session of psychotherapy (Model 1) and receipt of extended psychotherapy, defined as eight or more sessions within the 12-month follow-up period (Model 2). Predictor variables were entered using a block method where all variables entered the model at the same time, controlling for other potential predictor variables. Analyses were conducted using SAS version 9.2 (SAS Institute, Inc., Cary, NC).
Results
Mental Health and Psychotherapy Utilization
Table 1 lists the characteristics of the final cohort. The cohort was largely male (91%) with an average age of 59.2 years (SD = 14.4). Table 2 shows the total number of mental health encounters, as well as psychotherapy encounters for patients during the 12-month follow-up period. Of the 410,923 outpatients with a new-onset depression, anxiety, or PTSD diagnosis, 49% (n = 199,595) had at least one mental health encounter (CPT code) within 12 months of diagnosis. Twenty-two percent of the cohort received at least one session of psychotherapy. More patients received individual therapy than other psychotherapy modalities (e.g., group and family), though patients received a greater average number of sessions of group therapy (M = 8.9, SD = 12.0) than either individual (M = 3.4, SD = 4.2) or family therapy (M = 2.0, SD = 2.4).
Table 1.
Cohort Characteristics
| Full cohort (n = 410,923) | |
|---|---|
| Age in years (mean, SD) | 59.2 (14.4) |
| Gender (n, %) | |
| Male | 374,325 (91.1%) |
| Female | 36,598 (8.9%) |
| Marital status (n, %) | |
| Married | 235,851 (57.4%) |
| Not married | 175,072 (42.6%) |
| Income (n, %) | |
| <30,485 | 116,183 (28.3%) |
| 30,486–35,727 | 43,861 (10.7%) |
| 35,728–44,002 | 33,616 (8.2%) |
| >44,003 | 217,263 (52.9%) |
| Relative risk score (n, %) | |
| <0.26 | 92,434 (22.5%) |
| 0.26–1.00 | 179,428 (43.7%) |
| 1.01–2.00 | 73,406 (17.9%) |
| >2.00 | 65,655 (16.0%) |
| Distance to VA in miles (mean, SD) |
13.9 (15.2) |
| VA service connection (n, %) | |
| None | 257,360 (62.6%) |
| 1–49% | 75,785 (18.4%) |
| 50%+ | 77,778 (18.9%) |
| Comorbid mental health diagnosis (n, %) |
|
| Substance use | 42,101 (10.3%) |
| Schizophrenia | 3,847 (0.9%) |
| Cognitive disorder | 3,594 (0.9%) |
| Bipolar disorder | 2,506 (0.6%) |
Table 2.
Mental Health and Psychotherapy Visits, 12-Month Follow-up (N = 410,923)
| No. (%) patients receiving visits |
Mean (SD) no. of visitsa |
Median | Mode | Range | |
|---|---|---|---|---|---|
| Any MH CPT code | 199,595 (49) | 5.0 (7.5) | 3.0 | 1.0 | 1–269 |
| Psychotherapy | 88,763 (22) | 5.4 (8.7) | 2.0 | 1.0 | 1–180 |
| Individualb | 76,564 (19) | 3.4 (4.2) | 2.0 | 1.0 | 1–95 |
| Groupb | 25,696 (6) | 8.9 (12.0) | 4.0 | 1.0 | 1–179 |
| Familyb | 4,436 (1) | 2.0 (2.4) | 1.0 | 1.0 | 1–29 |
Note. CPT = Current procedural terminology; MH = mental health; SD = standard deviation.
Among those who had treatment.
Categories are not mutually exclusive; patients may have received multiple forms of psychotherapy.
Psychotherapy Timeliness
Table 3 lists the mean and median number of days between a mental health diagnosis and initiation of psychotherapy. For veterans who received psychotherapy, the time from initial mental health diagnosis to receipt of therapy ranged (by diagnosis) from a median of 50 to 65 days, with the total cohort averaging 57 days. Group comparisons examining treatment delays between diagnostic categories indicated that depression was associated with a longer median delay (relative to patients without depression; p < .001); anxiety was associated with a longer median delay (relative to patients without anxiety; p < .001); and PTSD was associated with a shorter median delay (relative to patients without anxiety; p < .001).
Table 3.
Time (Days) From Mental Health Diagnosis Until Initiation of Psychotherapy
| Mental health diagnosis | Mean (SD) days | Median days |
|---|---|---|
| Entire cohort (all diagnoses), n = 410,923 |
99.2 (106.3) | 57.0 |
| Depressive disorders, n = 268,039a |
101.6 (106.5) | 60.0 |
| Anxiety disorders, n = 97,831a |
105.6 (108.3) | 65.0 |
| PTSD, n = 77,743a | 92.6 (104.1) | 50.0 |
Note. PTSD = post-traumatic stress disorder; SD = standard deviation.
Patients may have received more than one new onset category diagnosis during the study period and therefore the total number of diagnoses exceed the full cohort number.
Psychotherapy Exposure
Table 4 presents percentages of patients for the full cohort and within each diagnostic category that had various levels of exposure to psychotherapy (calculated in number of sessions). For the full cohort, 322,160 (78%) did not receive any exposure to psychotherapy while 88,763 (22%) patients received at least one session. Examining the patients who received one or more sessions of psychotherapy (n = 88,763), 54% (n = 47,907) attended one or two sessions, 27% (n = 23,680) attended between three and seven sessions, and 19% (n = 17,176) attended eight or more sessions.
Table 4.
Psychotherapy Exposurea
| Full cohort n (%) |
Depression n (%) |
Anxiety n (%) |
PTSD n (%) |
|
|---|---|---|---|---|
| No exposure–0 Sessions | 322,160 (78) | 212,695 (79) | 82,411 (84) | 50,121 (64) |
| 1 Session | 33,729 (8) | 21,670 (8) | 6,483 (7) | 8,833 (11) |
| 2 Sessions | 14,178 (3) | 8,980 (3) | 2,668 (3) | 4,058 (5) |
| 3 Sessions | 8,457 (2) | 5,370 (2) | 1,511 (2) | 2,509 (3) |
| 4 Sessions | 5,760 (1) | 3,619 (1) | 1,033 (1) | 1,793 (2) |
| 5 Sessions | 3,984 (1) | 2,505 (1) | 669 (<1) | 1,290(2) |
| 6 Sessions | 3,030 (1) | 1,939 (1) | 445 (<1) | 1,041 (1) |
| 7 Sessions | 2,449 (1) | 1,499 (1) | 377 (<1) | 838 (1) |
| 8–12 Sessions | 7,268 (2) | 4,327 (2) | 1,090 (1) | 2,727 (4) |
| 13–25 Sessions | 6,523 (2) | 3,682 (1) | 801 (<1) | 2,830 (4) |
| 26 or more Sessions | 3,385 (1) | 1,753 (1) | 343 (<1) | 1,713 (2) |
Note. PTSD = post-traumatic stress disorder.
Patients may have received more than one new onset category diagnosis during the study period and therefore the total number of diagnoses exceed the full cohort number.
Examination of exposure to psychotherapy according to diagnostic categories showed that patients with PTSD were more likely to obtain psychotherapy and that their exposure was generally greater than that of patients with depression or anxiety. Patients with a non-PTSD anxiety condition were the least likely to obtain psychotherapy, and their exposure was less than that of patients with PTSD or depression. A chi square analysis examining no exposure versus any exposure across the three diagnostic groups was highly significant (χ2 = 80.50; p < .0001).
Exposure to individual and group psychotherapy was markedly different (see Table 5). Although a larger percentage of patients had at least one individual psychotherapy encounter, group therapy exposures tended to be of longer duration. For example, of patients who received individual psychotherapy (n = 76,564), 10.6% (n = 8,090) received eight or more sessions; of patients who received group psychotherapy (n = 25,696), 35.6% (n = 9,136) had eight or more sessions. The mean number of individual and group psychotherapy sessions was 3.4 and 8.9, respectively (see Table 2).
Table 5.
Psychotherapy Exposure by Individual and Group Modalities
| Individual psychotherapy n (%) |
Group psychotherapy n (%) |
|
|---|---|---|
| No exposure (0 Sessions) | 334,359 (81) | 385,227 (94) |
| 1 Session | 33,006 (8) | 7,678 (2) |
| 2 Sessions | 13,920 (3) | 2,796 (<1) |
| 3 Sessions | 8,168 (2) | 1,714 (<1) |
| 4 Sessions | 5,174 (1) | 1,484 (<1) |
| 5 Sessions | 3,582 (1) | 1,169 (<1) |
| 6 Sessions | 2,616 (1) | 926 (<1) |
| 7 Sessions | 2,008 (<1) | 793 (<1) |
| 8 or more Sessions | 5,084 (1) | 3,015 (<1) |
| 13 or more Sessions | 2,562 (1) | 3,845 (1) |
| 26 or more Sessions | 444 (<1) | 2,276 (<1) |
Prediction of Any Psychotherapy (One or More Sessions) and Extended Exposure (Eight or More Sessions)
We conducted separate logistic regression analyses to determine which patient factors predicted receipt of any psychotherapy (see Table 6) or extended exposure to psychotherapy (defined as eight or more encounters; Table 7). Predictor variables included: age, sex, marital status, income, VA service connection, new onset mental health diagnosis (anxiety, depression, and/or PTSD), comorbid mental health diagnosis (substance use, schizophrenia, cognitive disorder, and/or bipolar disorder), relative risk score, and distance (in miles) to the nearest VA health care facility. The prediction of receipt of one or more sessions of psychotherapy revealed significant contributions for age, gender, marital status, income, VA service connection, type of mental health diagnosis, comorbid mental health conditions, medical illness, and distance. Namely, young veterans with high medical illness burden who lived close to a VA facility had increased odds of receiving psychotherapy. In addition, patients with PTSD were almost three times more likely to obtain psychotherapy as other patients (odds ratio [OR] = 2.92), with the presence of a diagnosis of depression or anxiety reflecting relatively modest increases (OR = 1.27 and 1.04, respectively). Mental health comorbidities were also related to receipt of psychotherapy, with cognitive and bipolar disorders showing a slight increase (OR = 1.12 and 1.15, respectively), and substance use disorders relating to decreased use (OR = 0.87). The VA service connection variable indicated that veterans who had 1% to 49% service-connected disability were more likely to obtain psychotherapy (OR = 1.12), whereas veterans with 50% or more service connection used fewer psychotherapy services (OR = 0.76).
Table 6.
Logistic Regression Model: Receipt of Psychotherapy (One or More Sessions)a
| B | SE | Odds ratio | 95% CI | |
|---|---|---|---|---|
| Age group: | ||||
| <45 | referent | |||
| 45–54 | −0.37*** | 0.01 | 0.69 | 0.67–0.70 |
| 55–64 | −0.63*** | 0.01 | 0.53 | 0.52–0.54 |
| 65–74 | −1.35*** | 0.01 | 0.26 | 0.25–0.27 |
| 75 + | −1.53*** | 0.010 | 0.22 | 0.21–0.22 |
| Sex: | ||||
| Female | −0.06*** | 0.01 | 0.94 | 0.92–0.97 |
| Male | referent | |||
| Marital | ||||
| Married | referent | |||
| Unmarried | 0.09*** | 0.01 | 1.09 | 1.08–1.11 |
| Income | ||||
| <30,485 | −0.02* | 0.010 | 0.98 | 0.96–1.00 |
| 30,486–35,727 | −0.11*** | 0.01 | 0.89 | 0.87–0.91 |
| 35,728–44,002 | −0.10*** | 0.01 | 0.90 | 0.89–0.92 |
| 44,003+ | referent | |||
| VA service connection | ||||
| None | referent | |||
| 1–49% | 0.12*** | 0.01 | 1.12 | 1.104–1.14 |
| 50% + | −0.28*** | 0.01 | 0.76 | 0.74–0.77 |
| Mental health Dxb | ||||
| Anxiety | 0.04** | 0.01 | 1.04 | 1.01–1.07 |
| Depression | 0.24*** | 0.01 | 1.27 | 1.24–1.31 |
| PTSD | 1.07*** | 0.01 | 2.92 | 2.84–3.00 |
| Comorbid MH Dxb | ||||
| Substance use | −0.14*** | 0.01 | 0.87 | 0.85–0.89 |
| Schizophrenia | −0.001 | 0.04 | 1.00 | 0.93–1.07 |
| Cognitive disorder | 0.11** | 0.04 | 1.12 | 1.04–1.20 |
| Bipolar disorder | 0.14** | 0.04 | 1.15 | 1.05–1.25 |
| Relative risk score | ||||
| <0.26 | referent | |||
| 0.26–1.00 | 0.35*** | 0.01 | 1.41 | 1.39–1.44 |
| 1.01–2.00 | 0.83*** | 0.01 | 2.29 | 2.23–2.34 |
| >2.00 | 1.07*** | 0.01 | 2.91 | 2.84–2.98 |
| Distance to VA in miles | ||||
| 0–3.9 | referent | |||
| 4–9.9 | 0.01 | 0.01 | 1.01 | 1.00–1.03 |
| 10–19.9 | −0.11*** | 0.01 | 0.90 | 0.88–0.92 |
| 20+ | −0.28*** | 0.01 | 0.76 | 0.74–0.78 |
Note. CI = confidence interval; Dx = diagnosis; MH = mental health; PTSD = post-traumatic stress disorder; SE = standard error; VA = Veterans Affairs.
Predictor variables were entered as a single block, adjusting for the relationship of other variables.
The referent category for this variable represents the absence of the condition. For example, anxiety was examined as a dichotomous factor and represents a comparison of patients with and without (referent) anxiety.
p = 0.05.
p = 0.01.
p = 0.001.
Table 7.
Logistic Regression Model: Receipt of Eight or More Sessions of Psychotherapya
| B | SE | Odds Ratio |
95% CI | |
|---|---|---|---|---|
| Age group: | ||||
| <45 | referent | |||
| 45–54 | −0.14*** | 0.02 | 0.87 | 0.83–0.91 |
| 55–64 | −0.33*** | 0.02 | 0.72 | 0.69–0.75 |
| 65–74 | −1.42*** | 0.04 | 0.24 | 0.22–0.26 |
| 75 + | −1.89*** | 0.04 | 0.15 | 0.14–0.16 |
| Sex: | ||||
| Female | 0.08** | 0.03 | 1.09 | 1.03–1.15 |
| Male | referent | |||
| Marital | ||||
| Married | referent | |||
| Unmarried | 0.03 | 0.02 | 1.03 | 1.00–1.06 |
| Income | ||||
| <30,485 | −0.27*** | 0.02 | 0.77 | 0.74–0.80 |
| 30,486–35,727 | −0.22*** | 0.02 | 0.80 | 0.77–0.84 |
| 35,728–44,002 | −0.17*** | 0.02 | 0.84 | 0.81–0.88 |
| 44,003 + | referent | |||
| VA service | ||||
| None | referent | |||
| 1–49% | 0.11*** | 0.01 | 1.12 | 1.10–1.14 |
| 50% + | −0.29*** | 0.01 | 0.75 | 0.74–0.77 |
| Mental health Dxb | ||||
| Anxiety | −0.26*** | 0.03 | 0.77 | 0.73–0.81 |
| Depression | 0.03 | 0.02 | 1.04 | 0.98–1.09 |
| PTSD | 1.17*** | 0.02 | 3.22 | 3.01–3.39 |
| Comorbid MH Dxb | ||||
| Substance use | 0.05* | 0.02 | 1.06 | 1.01–1.10 |
| Schizophrenia | −0.55*** | 0.09 | 0.58 | 0.49–0.68 |
| Cognitive disorder | −1.04*** | 0.15 | 0.35 | 0.26–0.47 |
| Bipolar disorder | −0.08 | 0.08 | 0.92 | 0.79–1.07 |
| Relative risk score | ||||
| < 0.26 | referent | |||
| 0.26–1.00 | 0.35*** | 0.02 | 1.42 | 1.36–1.50 |
| 1.01–2.00 | 0.88*** | 0.03 | 2.41 | 2.28–2.54 |
| >2.00 | 1.27*** | 0.03 | 3.57 | 3.39–3.76 |
| Distance | ||||
| 0–3.9 | referent | |||
| 4–9.9 | −0.06** | 0.02 | 0.94 | 0.91–0.98 |
| 10–19.9 | −0.29*** | 0.02 | 0.75 | 0.72–0.78 |
| 20+ | −0.61*** | 0.02 | 0.54 | 0.52–0.57 |
Note. CI = confidence interval; Dx = diagnosis; MH = mental health; PTSD = post–traumatic stress disorder; SE = standard error; VA = Veterans Affairs.
Predictor variables were entered as a single block, adjusting for the relationship of other variables.
The referent category for this variable represents the absence of the condition. For example, anxiety was examined as a dichotomous factor and represents a comparison of patients with and without (referent) anxiety.
p = 0.05.
p = 0.01.
p = 0.001.
The prediction of extended exposure, defined as eight or more sessions, produced similar findings to those for the prediction of any psychotherapy with the following exceptions: women (OR = 1.09) and substance-use (OR = 1.06) patients were more likely to receive an extended dose, but persons with anxiety (OR = 0.77), or comorbid schizophrenia (OR = 0.58) or cognitive disorder comorbidities (OR = 0.35) were less likely to receive an extended dose of psychotherapy. Remaining strong were connections between extended dose and age, VA service connection, PTSD, medical burden, and distance. Notably, the odds ratios for PTSD, medical burden, and distance became stronger for extended exposure to psychotherapy (as compared with the one-session criterion, Table 6).
Discussion
Despite growing evidence of the efficacy, effectiveness, and potential for dissemination of psychotherapy in integrated medical care settings (Gilbody, Bower, Fletcher, Richards, & Sutton, 2006; Unutzer et al., 2006), we found that a large percentage of newly diagnosed patients (with depression, anxiety, or PTSD) in the VA did not receive any exposure to psychotherapy (78%) in the year following their diagnosis. Indeed, 95% of patients in our cohort received fewer than eight sessions. These results are generally consistent with those of other studies that found community-sample utilization rates ranging from 20% to 30% for anxious and depressed individuals (Young et al., 2001). Patients receiving psychotherapy experienced significant delays between initial mental health diagnosis and first psychotherapeutic encounter, with a median delay of 57 days. Collectively, these data suggest that important gaps exist in the timeliness and exposure to psychotherapy for veterans newly diagnosed with depression, anxiety, and PTSD.
Several patient factors associated with psychotherapy utilization (one or more sessions) were identified, including young age, male gender, unmarried status, high income level, PTSD diagnosis, increased medical burden/risk, and close proximity to a VA. Age, income, distance, and patient attitudes and beliefs about mental health service use are known challenges, requiring targeted efforts to improve engagement and adherence to care (Charney et al., 2003; Berk et al., 1995; Fortney, Lancaster, Owen, & Zhang, 1998; Klinkman, 1997). Research programs related to geriatric depression (Unutzer et al., 2002) and treatment of depression in rural patients (Fortney et al., 2006) provide evidence of care options available for wider implementation to improve outreach to these patient populations. Tailoring psychotherapeutic interventions by decreasing cost and increasing access (e.g., using telemedicine procedures, embedding psychotherapy in primary care sites) might alleviate some disparities seen in our data. However, distance- and age-related barriers to psychotherapy access are likely driven not only by patient-level factors, but also by provider- and system-level or supply-side factors that should be more fully explored to improve the mental health care provided to older persons with mental health needs (Klinkman, 1997).
Our marital status findings indicating that unmarried individuals have greater odds of obtaining psychotherapy are consistent with other recent research. Wang et al. (2005), using data from the National Comorbidity Survey Replication study, found that being married was related to longer delays and failure to make initial contact regarding mental health treatment and that this might be related to unmarried persons’ difficulties in social relationships, which add distress and increase patients’ incentive to seek care.
Contrary to study expectations, veterans who were 50% or more service connected (who receive increased access and benefits for VA care) were at increased risk for under-treatment. These findings require further investigation but post hoc analyses (data not presented) suggest that service connected veterans are less likely to received a mental health diagnosis (with the exception of PTSD) and therefore are likely service connected for medical rather than mental health issues. These patients, as evidenced by prior studies on barriers to mental health treatment, may posses attitudes and beliefs more averse to mental health care than those found in traditional mental health clinics (Van Voorhees et al., 2003) and may require focused clinical efforts to identify and treat their mental health symptoms within a physical health focus (Keeley et al., 2004). Further, post hoc analyses revealed significant differences in service and nonservice connected veterans on factors such as age, marital status, relative risk scores, and frequency of diagnosed mental health conditions. Although these group differences are likely related to psychotherapeutic treatment, they require further exploration to determine the overall impact of service connection factors on psychotherapy utilization.
Almost all diagnoses examined in our predictive models suggested increased odds of receiving at least one session of psychotherapy, relative to patients without each respective diagnosis. PTSD was associated with an almost threefold increase; substance use was the only diagnosis associated with decreased odds.
Factors predicting extended exposure to psychotherapy (defined as eight or more sessions) were similar to those identified for receipt of any psychotherapy except for female gender, which switched from a negative predictor in the first model to a positive predictor in the extended exposure model. Factors such as medical burden, distance, and the presence of a PTSD diagnosis were strongly associated with extended exposure. Comorbid mental health conditions such as schizophrenia and cognitive disorders placed patients at risk for limited exposure to psychotherapy (fewer than 8 sessions), which may be reflective of practice standards which downplay the importance of psychotherapy in these patient populations, rather than reflect the relatively recent evidence base that suggests that specific uses of psychotherapy are effective for persons with schizophrenia and cognitive disorders (Penn, Waldheter, Perkins, Mueser, & Lieberman, 2005; Terri, Logsdon, Uomoto, & McCurry, 1997; Terri et al., 2003).
Despite the potential “missed opportunities” for wide use and timeliness of psychotherapy, our data reflect several facets of mental health care that appear to have been positively affected by VA initiatives to improve the treatment of various mental health disorders. For example, it is the opinion and experiences of the authors that the VA has prioritized group treatment as a first-line option for psychotherapy, given the potentially increased clinical impact of patient-to-patient interaction and cost effectiveness associated with decreased clinician-to-patient ratios. Our data indicate that VA group psychotherapy recipients receive a greater average number of treatment sessions, and that a higher proportion of patients receive extended exposure to psychotherapy, as compared with patients who use individual psychotherapy. Other VA clinical care initiatives have attempted to increase screening and timely mental health care for veterans with PTSD and depression (Veterans Affairs Office of Research & Development, 2008). Findings from this investigation suggest that patients with depression, and especially those with PTSD, have higher odds of receiving extended exposure to psychotherapy than patients with anxiety conditions. Lastly, upcoming VA efforts to increase the number of mental health care professionals within primary care settings will likely increase psychotherapy engagement and reduce treatment delays (Bartels et al., 2004) and may also provide opportunities to address patient-level barriers, such as mental health treatment preferences and expectations, which have a lasting impact on patients’ initial decision to seek care and on longer-term adherence to treatment recommendations (Roness, Mykletun, & Dahl, 2005; Lin et al., 2005).
Study Limitations and Need for Future Research
The current investigation is limited in that we are unable to draw definitive conclusions about the apparent under use of psychotherapeutic services for several reasons, including: 1) the use of data from administrative databases which include limitations related to diagnostic and procedural coding reliability and validity; 2) the utilization of crude cutoffs for determining psychotherapy exposure; 3) an inability to account for non-VA mental health service use, which could lead to an underestimation of psychotherapy use; and 4) the study’s inability to control for pharmacological interventions, which, if offered, might decrease patients’ and providers’ perceptions of additional treatment needed.
Future research is needed to develop improved assessments of psychotherapy quality in health care settings, including more precise and up-to-date dose-response effect data as well as measures for establishing acceptable time intervals for the initiation of psychotherapy (delays) and between-session intervals. Further investigations which focus upon psychotherapy care processes at the system, practitioner, and patient levels may provide important information to improve the timely receipt and increased exposure to psychotherapeutic care for patients seeking care in integrated health care settings. Although data from other research suggest the presence of supply side barriers (Wei et al., 2005), patient and practitioner barriers such as poor mental health knowledge and stigma may also contribute to the limited numbers of patients engaging in and adhering to treatment (Klinkman, 1997).
Acknowledgments
Dr. Cully is a VA Health Services Research and Development Career Development Awardee (CDA-2: 05-288). Dr. Petersen is a Robert Wood Johnson Foundation Generalist Physician Faculty Scholar (grant number 045444), and an American Heart Association Established Investigator Awardee (grant number 0540043N). In addition, this work was supported in part by the Houston Center for Quality of Care & Utilization Studies (Houston VA HSR&D Center of Excellence [HFP90-020]). We thank Meghan Zimmer, Tracy Urech, and Myrna Khan for their contributions to this work.
Contributor Information
Jeffrey A. Cully, Houston Center for Quality of Care & Utilization Studies, Veterans Affairs Health Services Research and Development Center of Excellence, Baylor College of Medicine, Michael E. DeBakey Veterans Affairs Medical Center, and Veterans Affairs South Central Mental Illness Research, Education, & Clinical Center
Laura Tolpin, Michael E. DeBakey Veterans Affairs Medical Center.
Louise Henderson, Houston Center for Quality of Care & Utilization Studies, Veterans Affairs Health Services Research and Development Center of Excellence.
Daniel Jimenez, Menninger Department of Psychiatry and Behavioral Sciences, Baylor College of Medicine.
Mark E. Kunik, Houston Center for Quality of Care & Utilization Studies, Veterans Affairs Health Services Research and Development Center of Excellence, Baylor College of Medicine, Michael E. DeBakey Veterans Affairs Medical Center, and Veterans Affairs South Central Mental Illness Research, Education, & Clinical Center
Laura A. Petersen, Houston Center for Quality of Care & Utilization Studies, Veterans Affairs Health Services Research and Development Center of Excellence, Baylor College of Medicine, and Michael E. DeBakey Veterans Affairs Medical Center
References
- Bartels SJ, Coakley EH, Zubritsky C, Ware JH, Miles KM, Areán PA, et al. Improving access to geriatric mental health services: A randomized trial comparing treatment engagement with integrated versus enhanced referral care for depression, and at-risk alcohol use. American Journal of Psychiatry. 2004;161:1455–1462. doi: 10.1176/appi.ajp.161.8.1455. [DOI] [PubMed] [Google Scholar]
- Berk ML, Schur CL, Cantor JC. Ability to obtain health care: Recent estimates from the Robert Wood Johnson Foundation National Access to Care Survey. Health Affairs. 1995;14:139–146. doi: 10.1377/hlthaff.14.3.139. [DOI] [PubMed] [Google Scholar]
- Busch SH, Leslie D, Rosenheck R. Measuring quality of pharmacotherapy for depression in a national health care system. Medical Care. 2004;42:532–542. doi: 10.1097/01.mlr.0000128000.96869.1e. [DOI] [PubMed] [Google Scholar]
- Charbonneau A, Rosen AK, Ash AS, Owen RR, Kader B, Spiro A, III, et al. Measuring the quality of depression care in a large integrated health system. Medical Care. 2003;41:669–680. doi: 10.1097/01.MLR.0000062920.51692.B4. [DOI] [PubMed] [Google Scholar]
- Charbonneau A, Rosen AK, Owen RR, Spiro A, III, Ash AS, Miller DR, et al. Monitoring depression care: In search of an accurate quality indicator. Medical Care. 2004;42:522–531. doi: 10.1097/01.mlr.0000127999.89246.a6. [DOI] [PubMed] [Google Scholar]
- Charney DS, Reynolds CF, III, Lewis L, Lebowitz BD, Sunderland T, Alexopoulos GS, et al. Depression and Bipolar Support Alliance consensus statement on the unmet needs in diagnosis and treatment of mood disorders in late life. Archives of General Psychiatry. 2003;60:664–672. doi: 10.1001/archpsyc.60.7.664. [DOI] [PubMed] [Google Scholar]
- Ellis RP, Ash AS, Ayanian JZ, Bates DW, Burstin H, Iezzoni LI, et al. Diagnosis-based risk adjustment for Medicare capitation payments. Health Care Financing Review. 1996;17:101–128. [PMC free article] [PubMed] [Google Scholar]
- Fortney JC, Lancaster AE, Owen RR, Zhang M. Geographic market areas for psychiatric and medical outpatient treatment. Journal of Behavioral Health Services & Research. 1998;25:108–116. doi: 10.1007/BF02287506. [DOI] [PubMed] [Google Scholar]
- Fortney JC, Pyne JM, Edlund MJ, Robinson DE, Mittal D, Henderson KL. Design and implementation of the telemedicine-enhanced antidepressant management study. General Hospital Psychiatry. 2006;28:18–26. doi: 10.1016/j.genhosppsych.2005.07.001. [DOI] [PubMed] [Google Scholar]
- Gilbody S, Bower P, Fletcher J, Richards D, Sutton AJ. Collaborative care for depression: A cumulative meta-analysis and review of longer-term outcomes. Archives of Internal Medicine. 2006;166:2314–2321. doi: 10.1001/archinte.166.21.2314. [DOI] [PubMed] [Google Scholar]
- Grant RW, O’Leary KM, Weilburg JB, Singer DE, Meigs JB. Impact of concurrent medication use on statin adherence and refill persistence. Archives of Internal Medicine. 2004;164:2343–2348. doi: 10.1001/archinte.164.21.2343. [DOI] [PubMed] [Google Scholar]
- Gum AM, Arean PA, Hunkeler E, Tang L, Katon W, Hitchcock P, et al. Depression treatment preferences in older primary care patients. Gerontologist. 2006;46:14–22. doi: 10.1093/geront/46.1.14. [DOI] [PubMed] [Google Scholar]
- Hankin CS, Spiro A, Miller DR, Kazis L. Mental disorders and mental health treatment among U.S. Department of Veterans Affairs outpatients: The veterans health study. American Journal of Psychiatry. 1999;156:1924–1930. doi: 10.1176/ajp.156.12.1924. [DOI] [PubMed] [Google Scholar]
- Hansen NB, Lambert MJ. An evaluation of the dose-response relationship in naturalistic treatment settings using survival analysis. Mental Health Services Research. 2003;5:1–12. doi: 10.1023/a:1021751307358. [DOI] [PubMed] [Google Scholar]
- Hansen NB, Lambert MJ, Forman EM. The psychotherapy dose-response effect and its implications for treatment delivery services. Clinical Psychology: Science and Practice. 2002;9:329–343. [Google Scholar]
- Howard KI, Kopta SM, Krause MS, Orlinsky DE. The dose-effect relationship in psychotherapy. The American Psychologist. 1986;41:159–164. [PubMed] [Google Scholar]
- Keeley RD, Smith JL, Nutting PA, Miriam DL, Perry DW, Rost KM. Does a depression intervention result in improved outcomes for patients presenting with physical symptoms? Journal of General Internal Medicine. 2004;19:615–623. doi: 10.1111/j.1525-1497.2004.30611.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Klinkman MS. Competing demands in psychosocial care: A model for the identification and treatment of depressive disorders in primary care. General Hospital Psychiatry. 1997;19:98–111. doi: 10.1016/s0163-8343(96)00145-4. [DOI] [PubMed] [Google Scholar]
- Lambert MJ, Ogles BM. The efficacy and effectiveness of psychotherapy. In: Lambert MJ, editor. Bergin and Garfield’s handbook of psychotherapy and behavior change. 5th Ed Wiley; New York: 2004. pp. 139–193. [Google Scholar]
- Lin P, Campbell DG, Chaney EF, Liu CF, Heagerty P, Felker BL, et al. The influence of patient preference on depression treatment in primary care. Annals of Behavioral Medicine. 2005;30:164–173. doi: 10.1207/s15324796abm3002_9. [DOI] [PubMed] [Google Scholar]
- Olfson M, Marcus SC, Druss B, Pincus HA. National trends in the use of outpatient psychotherapy. American Journal of Psychiatry. 2002;159:1914–1920. doi: 10.1176/appi.ajp.159.11.1914. [DOI] [PubMed] [Google Scholar]
- Orlinsky DE, Roonestad MH, Willutzki U. Fifty years of psychotherapy process-outcome research: Continuity and change. In: Lambert MJ, editor. Bergin and Garfield’s handbook of psychotherapy and behavior change. 5th Ed Wiley; New York: 2004. pp. 307–389. [Google Scholar]
- Penn DL, Waldheter EJ, Perkins DO, Mueser KT, Lieberman JA. Psychosocial treatment for first-episode psychosis: A research update. American Journal of Psychiatry. 2005;162:2220–2232. doi: 10.1176/appi.ajp.162.12.2220. [DOI] [PubMed] [Google Scholar]
- Petersen LA, Pietz K, Woodard L, Byrne M. Comparison of the predictive validity of diagnosis-based risk adjusters for clinical outcomes. Medical Care. 2005;43:61–67. [PubMed] [Google Scholar]
- Quan H, Parsons GA, Ghali WA. Validity of procedure codes in International Classification of Diseases, 9th revision, clinical modification administrative data. Medical Care. 2004;42:801–809. doi: 10.1097/01.mlr.0000132391.59713.0d. [DOI] [PubMed] [Google Scholar]
- Roness A, Mykletun A, Dahl AA. Help-seeking behaviour in patients with anxiety disorder and depression. Acta Psychiatrica Scan-dinavica. 2005;111:51–58. doi: 10.1111/j.1600-0447.2004.00433.x. [DOI] [PubMed] [Google Scholar]
- Shapiro DA, Barkham M, Stiles WB, Hardy GE, Rees A, Reynolds S, et al. Time is of the essence: A selective review of the fall and rise of brief therapy research. Psychology and Psychotherapy. 2003;76:211–235. doi: 10.1348/147608303322362460. [DOI] [PubMed] [Google Scholar]
- Teri L, Gibbons LE, McCurry SM, Logsdon RG, Buchner DM, Barlow WE, et al. Exercise plus behavioral management in patients with Alzheimer disease: A randomized controlled trial. Journal of the American Medical Association. 2003;290:2015–2022. doi: 10.1001/jama.290.15.2015. [DOI] [PubMed] [Google Scholar]
- Teri L, Logsdon RG, Uomoto J, McCurry SM. Behavioral treatment of depression in dementia patients: A controlled clinical trial. Journal of Gerontology. Series B. Psychological Sciences and Social Sciences. 1997;52:159–166. doi: 10.1093/geronb/52b.4.p159. [DOI] [PubMed] [Google Scholar]
- Unutzer J, Katon W, Callahan CM, Williams JW, Jr., Hunkeler E, Harpole L, et al. Collaborative care management of late-life depression in the primary care setting: A randomized controlled trial. Journal of the American Medical Association. 2002;288:2836–2845. doi: 10.1001/jama.288.22.2836. [DOI] [PubMed] [Google Scholar]
- Unutzer J, Schoenbaum M, Druss BG, Katon WJ. Transforming mental health care at the interface with general medicine: Report for the presidents commission. Psychiatric Services. 2006;57:37–47. doi: 10.1176/appi.ps.57.1.37. [DOI] [PubMed] [Google Scholar]
- Van Voorhees BW, Cooper LA, Rost KM, Nutting P, Rubenstein LV, Meredith L, et al. Primary care patients with depression are less accepting of treatment than those seen by mental health specialists. Journal of General Internal Medicine. 2003;18:991–1000. doi: 10.1111/j.1525-1497.2003.21060.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Veterans Affairs Office of Research and Development [Accessed January 25, 2008];VA QUERI (Quality Enhancement Research Initiative), QUERI Fact Sheet, February 2008. 2008 http://www.hsrd.research.va.gov/publications/internal/mh_factsheet.pdf.
- Wang PS, Berglund P, Olfson M, Pincus HA, Wells KB, Kessler RC. Failure and delay in initial treatment contact after first onset of mental disorders in the National Comorbidity Survey Replication. Archives of General Psychiatry. 2005;62:603–613. doi: 10.1001/archpsyc.62.6.603. [DOI] [PubMed] [Google Scholar]
- Wei W, Sambamoorthi U, Olfson M, Walkup JT, Crystal S. Use of psychotherapy for depression in older adults. American Journal of Psychiatry. 2005;162:711–717. doi: 10.1176/appi.ajp.162.4.711. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Westen D, Morrison K. A multidimensional meta-analysis of treatments for depression, panic, and generalized anxiety disorder: An empirical examination of the status of empirically supported therapies. Journal of Consulting and Clinical Psychology. 2001;69:875–899. [PubMed] [Google Scholar]
- Wright RE. Logistic regression. In: Grimm LG, Yarnold PR, editors. Reading and understanding multivariate statistics. American Psychological Association; Washington: 2003. pp. 217–244. [Google Scholar]
- Young AS, Klap R, Sherbourne CD, Wells KB. The quality of care for depressive and anxiety disorders in the United States. Archives of General Psychiatry. 2001;58:55–61. doi: 10.1001/archpsyc.58.1.55. [DOI] [PubMed] [Google Scholar]