ADEQUACY OF TREATMENT RECEIVED BY PRIMARY CARE PATIENTS WITH ANXIETY DISORDERS

Abstract

Background

We examined the adequacy of pharmacotherapy and psychotherapy received by primary care patients with anxiety disorders over up to 5 years of follow-up.

Method

Five hundred thirty-four primary care patients at 15 US sites, who screened positive for anxiety symptoms, were assessed for anxiety disorders. Those meeting anxiety disorder criteria were offered participation and interviewed again at six and 12 months postintake, and yearly thereafter for up to 5 years. We utilized existing definitions of appropriate pharmacotherapy and created definitions of potentially adequate psychotherapy/cognitive-behavioral therapy (CBT).

Results

At intake, of 534 primary care participants with anxiety disorders, 19% reported receiving appropriate pharmacotherapy and 14% potentially adequate CBT. Overall, 28% of participants reported receiving potentially adequate anxiety treatment, whether pharmacotherapy, psychotherapy, or both. Over up to five years of follow-up, appropriate pharmacotherapy was received by 60% and potentially adequate CBT by 36% of the sample. Examined together, 69% of participants received any potentially adequate treatment during the follow-up period. Over the course of follow-up, primary care patients with MDD, panic disorder with agoraphobia, and with medicaid/medicare were more likely to receive appropriate anxiety treatment. Ethnic minority members were less likely to receive potentially adequate care.

Conclusions

Potentially adequate anxiety treatment was rarely received by primary care patients with anxiety disorders at intake. Encouragingly, rates improved over the course of the study. However, potentially adequate CBT remained much less utilized than pharmacotherapy and racial-ethnic minority members were less likely to received care, suggesting much room for improved dissemination of quality treatment.

INTRODUCTION

Anxiety disorders are the most common mental health problems in the United States,^[1] affecting ~30% of the population. Individuals with anxiety disorders experience impaired quality of life^{[2, 3]} and educational/occupational functioning,^[4–6] and higher suicide rates.^[7] Fortunately, efficacious treatments for these disorders exist, including pharmacotherapy^{[8, 9]} and cognitive-behavioral therapy (CBT).^[10–12]

It is important to examine anxiety treatment adequacy in primary care patients. Anxiety disorders are among the most common mental health problems in primary care;^[13–15] where they are more common than in the general population.^[16] Primary care patients with anxiety disorders demonstrate considerable impairment in functioning.^[17–19] Among primary care patients, anxiety disorders are associated with significantly increased utilization of primary care resources, even after controlling for chronic medical conditions.^[20] Further, more than half of patients with a psychiatric problem receive treatment from a primary care provider (PCP).^{[21, 22]}

Few studies describe the treatment for anxiety disorders received by primary care patients. Previously reported cross-sectional data from our sample suggest that approximately half of primary care patients with anxiety disorders received any mental health treatment.^[23] Limited research to date indicates that most US primary care patients with anxiety disorders do not receive appropriate care. In an epidemiological sample, minimally adequate treatment was received by 34% of US adults with anxiety disorders overall; 52% of those receiving treatment in mental health settings, and 13% of those treated in primary care.^[24]

Other research has reported somewhat more optimistic rates. Two studies conducted primarily in the western United States examined the quality of care of primary care patients with anxiety disorders, specifying CBT as adequate psychotherapy. Between 31 and 41% of patients were receiving appropriate care for anxiety disorders.^{[25, 26]}

Differing definitions of adequate treatment have been used in primary care anxiety disorder research. Whereas potentially adequate pharmacotherapy has been fairly consistently defined (i.e. appropriate medication, at an appropriate dose, for an appropriate duration), potentially adequate psychotherapy has at times been determined by the number e.g.^[24] rather than the content, of sessions. Existing research has been cross-sectional, inquiring only about treatment in the 3–12 months prior to inquiry. This may be misleading as PCPs, having a long-term relationship with patients, may employ “watchful waiting”^{[27, 28]} a recommended strategy particularly for patients with less severe symptoms, in determining when to begin treatment or refer a patient.

In the present study, we sought to add to the scarce literature on anxiety treatment adequacy, including the utilization of CBT techniques, by primary care patients. To our knowledge, this is the first such study conducted in the northeastern United States and the first to examine the adequacy of treatment received longitudinally.

METHODS

STUDY DESIGN

Data are from the Primary Care Anxiety Project (PCAP), an observational, longitudinal study of primary care patients with anxiety disorders.^[27] Participants were assessed at baseline, 6 and 12 months postbaseline, and then yearly for up to 5 years.

PCAP was conducted at 15 primary care practices (see Table 1) in New Hampshire, Massachusetts, Rhode Island, and Vermont. No site had fully integrated behavioral health, though two (sites 7 and 14) had some colocated services. Thus, the vast majority of psychotherapy was received outside of primary care.

TABLE 1.

Recruitment site characteristics and intake N from each site

Site number	N	State	Setting	Provider type
1	10	VT	Suburban	Internal medicine
2	8	VT	Rural	Mixed: internal and family
3	34	VT	Rural	Mixed: internal and family
4	52	VT	Rural	Family medicine
5	31	MA	Urban	Family medicine
6	131	MA	Urban	Internal medicine
7	3	MA	Urban	Family medicine
8	15	RI	Urban	Internal medicine
9	63	RI	Urban	Family medicine
10	6	RI	Suburban	Family medicine
11	59	RI	Suburban	Internal medicine
12	17	NH	Rural	Family medicine
13	52	RI	Urban	Internal medicine
14	49	MA	Urban	Internal medicine
15	4	MA	Suburban	Internal medicine

PARTICIPANT RECRUITMENT AND ASSESSMENT

Participants were recruited in primary care waiting rooms. Inclusion criteria were: a general medical appointment the day of recruitment, age ≥18, English proficiency, and currently meeting DSM-IV criteria for one or more intake anxiety disorder; panic disorder with (PDA) or without (PD) agoraphobia, social phobia, generalized anxiety disorder (GAD), posttraumatic stress disorder (PTSD), agoraphobia without history of panic disorder, mixed anxiety-depressive disorder, or GAD features exclusively during a mood disorder. Exclusion criteria included active psychosis, no current address and phone number, and pregnancy.

A research assistant asked all eligible patients if they wished to participate in a study of stress or nervousness. Interested patients completed an anxiety screener.^[29] Patients screening positive were offered a diagnostic interview. After complete description of the study, written consent was obtained. Institutional Review Boards at each of the sites approved the protocol. A detailed description of recruitment has been published.^[29]

MEASURES

Anxiety Screener

A 32-item self-report measure assessed key features of DSM-IV anxiety disorder criteria. To avoid excluding potentially eligible participants, the screener was designed to be highly sensitive. A validation study see^[28] found a sensitivity of 1.0 and a specificity of .67.

Clinical Interview

Participants were diagnosed using the Structured Clinical Interview for DSM-IV (SCID).^[30]

Psychosocial Functioning

Global Social Adjustment (GSA) is part of the LIFE-base interview.^[31] The LIFE has good interrater reliability; intraclass correlation coefficients (ICC) range from .58 to .91.^{[32, 33]} After assessing functioning in specific domains (e.g., work relationships), the interviewer rates GSA; overall psychosocial adjustment during the past week from 1 (no impairment) to 5 (marked impairment).

Nonpsychiatric medical problems were assessed using a Medical History Form designed for the study^[34] that inquires about the presence self-reported medical problems. A dichotomous (yes/no) variable was constructed indicating the presence of a major medical illness including current asthma, cancer, diabetes, epilepsy, heart disease, kidney disease, liver disease, lung or respiratory illness, stroke, and/or thyroid disease.

Treatment Received

At intake, participants were questioned about treatment received over the past three months. At each follow-up, they reported all treatment since the previous interview. Use of psychotropic medications, including the average weekly dose, was captured on the psychotropic treatment section of the LIFE.^[31] Participants receiving pharmacotherapy at intake also reported whether the prescriber was their PCP or a psychiatrist. The Psychosocial Treatment Interview-Revised (PTI-R)^[35] inquired about the receipt of specific psychotherapy techniques (e.g., in-vivo exposure, homework assignments) during treatment. The behavioral and cognitive scales showed good internal consistency and good to very good interrater reliability (behavioral scale: Cronbach’s alpha = 0.80, kappa = 0.68, ICC = 0.88; cognitive scale: Cronbach’s alpha = 0.79, kappa = 0.75, ICC = 0.91).

Definitions of Potentially Adequate Pharmacotherapy

We utilized definitions provided by Stein et al.^{[25, 26]} Pharmacotherapy was considered potentially adequate, if: (1) the medication has known efficacy for anxiety disorders; and was taken (2) at an appropriate dose; (3) for an appropriate amount of time. Appropriate medication type was any selective serotonin reuptake inhibitor (SSRI), serotonin norepinephrine reuptake inhibitor (SNRI), or benzodiazepine. Additionally, participants with GAD were considered to be receiving an appropriate medication type if they received buspirone, gabapentin, or pregabalin. Appropriate dosages were those previously determined by Stein et al.^{[25, 26]} Appropriate duration was a minimum of eight consecutive weeks.

Definitions of Potentially Adequate Psychotherapy

We de-fined potentially adequate psychotherapy as the receipt of empirically supported cognitive and/or behavioral treatment (CBT) techniques, endorsed on the PTI-R. Receipt of cognitive therapy (CT) was coded if the participant’s therapist: (1) helped them to identify maladaptive or problematic patterns in their thinking, AND; (2) asked them to consider and substitute different, more rational/adaptive thoughts, or otherwise help them to restructure their thinking. Behavior therapy (BT) was defined as receiving (1) in vivo, interoceptive, and/or imaginal exposure, AND; (2) homework practice. Participants whose psychotherapy met criteria as CT and/or BT, were coded as receiving potentially adequate psychotherapy.

STATISTICAL METHODS

Descriptive statistics were used to characterize the receipt of medication and therapy techniques during the three months prior to intake and over follow-up. Chi-square examined treatment adequacy at intake by medication provider type. To examine predictors of receiving potentially adequate pharmacotherapy and/or psychotherapy, six dependent variables were created; receipt of: (1) potentially adequate pharmacotherapy at intake; (2) potentially adequate psychotherapy at intake; (3) any potentially adequate treatment (pharmacotherapy or psychotherapy) at intake; (4) potentially adequate pharmacotherapy during the follow-up period; (5) potentially adequate psychotherapy during follow-up; and (6) any potentially adequate treatment during follow-up. Possible predictors of each outcome were examined. The initial pool of candidate predictors included age, gender, race/ethnicity, college education (yes/no), insurance type (private, medicaid/medicare, or uninsured), income (< or ≥ $20,000 per year), marital status at intake (yes/no), functioning as measured by the GSA, presence of major depressive disorder (MDD), presence of alcohol/substance use disorder, presence of each type of index anxiety disorder, presence of a major nonpsychiatric medical illness, and anxiety disorder illness age of onset. Using stepwise logistic regression, the final logistic regression models were examined including all variables that entered the multivariate model at a 0.05 level of significance.

RESULTS

SAMPLE CHARACTERISTICS

A detailed description of sample selection, including refusal rates, has been published.^[34] Five hundred thirty-nine primary care patients were enrolled in PCAP between 1997 and 2001. We examine 534 participants with complete intake treatment data (see Table 2). Average age was 39. The majority was female and Non-Latino White. Panic disorder (combining with and without Agoraphobia) and PTSD were the most common anxiety disorders.

TABLE 2.

Demographic and diagnostic characteristics of the study sample at intake (N = 534)

Characteristic	N/Mean	%/SD
Age	39	11.58
Gender
Female	408	76.40
Male	126	23.60
Ethnicity
Non-Latino White	442	82.77
Latino	19	3.56
African American	41	7.68
Native American	9	1.69
Asian	7	1.31
Other	16	2.30
Currently Married	264	49.44
Education
Four year college degree or more	123	23.03
Less than 4 year college degree	411	76.97
Employed full-time outside the home	216	41.78
Posttraumatic stress disorder	196	36.70
Social phobia	180	33.71
Panic disorder with agoraphobia	148	27.72
Generalized anxiety disorder	133	24.91
Panic disorder without agoraphobia	85	15.92
GAD features exclusively during the course of a mood disorder	29	5.43
Agoraphobia without history of panic disorder	22	4.12
Mixed anxiety-depressive disorder	10	1.87
More than one anxiety disorder	252	47.19
Comorbid major depressive disorder	217	40.64
Comorbid alcohol/substance use disorder	55	10.30
Comorbid eating disorder	56	10.49

For analyses predicting adequate treatment over follow-up, we required participants to have a minimum of two years of follow-up data. Therefore the N for these analyses is 419. We ran dropout analyses, comparing these 419 participants with those who had ≤ 1 year of follow-up data. There were no significant differences on gender, baseline GAF, number of lifetime anxiety diagnoses, insurance, or receipt of potentially adequate treatment. Those lost to follow-up before 2 years were significantly younger (M = 35.7, SD = 11.3) than those with ≥2 years of data (M = 39.9, SD = 11.5) t = −3.48, P < 0.0006). Therefore age was included as a candidate covariate for the forward stepwise regression in follow-up predictor analyses. Many participants had less than 5 years of follow-up data, and some participants completed nonconsecutive follow-up interviews (see Table 3).

TABLE 3.

N and rates of potentially adequate treatment at each year of follow-up

N	Ad CBT	Ad pharm	Any Ad Tx
Year 1
395	84 (21.32%)	154 (38.99%)	184 (46.58%)
Year 2
366	75 (20.78%)	136 (37.16%)	177 (48.36%)
Year 3
309	48 (15.89%)	120 (38.83%)	140 (45.31%)
Year 4
276	39 (14.18%)	118 (42.75%)	135 (48.91%)
Year 5
252	39 (15.48%)	103 (40.87%)	121 (48.02%)

TREATMENT RECEIVED AT INTAKE (see Fig. 1)

Figure 1.

Percent of sample receiving potentially adequate psychotherapy, pharmacotherapy, or any potentially adequate anxiety treatment.

At intake, 37.45% (n = 200) reported taking an appropriate type of antianxiety medication. This included 26.40% (n = 141) receiving an SSRI/SNRI, 17.42% (n = 93) a benzodiazepine, and 1 participant (0.19%) taking buspirone. No participants took gabapentin or pregabalin at intake. Only 19.10% (n = 102) had taken an antianxiety medication at an adequate dose for at least eight consecutive weeks prior to baseline and were considered to have received potentially adequate pharmacotherapy. Data on medication provider were available for a subset of 193 of these individuals; 51% received pharmacotherapy from their PCP, 49% from a psychiatrist. There were no differences in rates of adequate pharmacotherapy by provider type (χ² = 0.36, df = 1, P = 0.5485).

Just under a third of participants (30.34%; n = 162) received psychotherapy within the 3 months prior to intake. Only 68 participants (12.76%) reported receiving CT and only 28 (5.26%) received BT. Seventy-seven participants (14.42%) received potentially adequate psychotherapy. Overall, 151 (28.28%) participants reported receiving potentially adequate anxiety treatment, whether pharmacotherapy, psychotherapy, or both, during the three months prior to study intake.

Predictors of Receiving Adequate Treatment at Intake

Forward, stepwise, logistic regression was used to examine predictors of receiving adequate treatment at study intake (see Table 4).

TABLE 4.

Significant predictors of receiving potentially adequate anxiety treatment at intake

	Odds ratio (95% CI)	Wald χ2	DF	P-value
Model 1. Potentially adequate pharmacotherapy
Medicaid or other public insurance	3.84 (1.86–7.93)	13.15	1	.0003
Private insurance	2.27 (1.09–4.72)	4.79	1	.0287
Panic w/agoraphobia at intake	4.00 (2.49–6.44)	32.63	1	<.0001
Model 2. Potentially adequate psychotherapy
Global social adjustment	1.91 (1.37–2.65)	14.70	1	.0001
Married at intake	0.36 (0.20–0.66)	10.81	1	.0010
College education	3.70 (1.99–6.85)	17.24	1	<.0001
Ethno-racial minority	0.44 (0.19–0.99)	3.95	1	.0469
Model 3. Potentially adequate anxiety treatment
Ethno-racial minority	0.45 (0.24–0.84)	6.21	1	.0127
College education	1.93 (1.13–3.30)	5.78	1	.0162
Uninsured	0.26 (0.09–0.70)	6.96	1	.0083
Married at intake	0.62 (0.39–0.98)	4.17	1	.0411
Global social adjustment	1.59 (1.23–2.05)	12.68	1	.0004
Panic w/agoraphobia at intake	2.99 (1.89–4.75)	21.70	1	<.0001
Income less than $20 K	1.68 (1.05–2.69)	4.63	1	.0315

Results from final forward stepwise model including all variables that entered at P ≤ .05.

Potentially Adequate Pharmacotherapy

Individuals with PDA and with insurance were more likely to receive potentially adequate medication.

Potentially Adequate Psychotherapy

GSA (being more functionally impaired), being unmarried, and completing college were associated with receipt of potentially adequate psychotherapy at intake. Minority members were less likely to receive potentially adequate psychotherapy.

Any Potentially Adequate Anxiety Treatment (Pharmacotherapy and/or Psychotherapy)

Having PDA, a college education, and worse functioning were associated with an increased likelihood of receiving potentially adequate anxiety treatment. Being married, uninsured, and an ethnic minority were associated with a decreased likelihood of adequate care. Interestingly, having a yearly income less than $20,000 was associated with an increased likelihood of potentially adequate treatment.

TREATMENT RECEIVED OVER THE FOLLOW-UP PERIOD (see Fig. 1)

Exclusive of treatment received during the 3 months prior to intake, 70.41% (n = 295) reported taking an appropriate antianxiety medication during follow-up; 62.05% (n = 260) took an SSRI/SNRI and 34.37% (n = 144) a benzodiazepine at any time during follow-up. Of those with GAD, 5.34% (n = 6) received buspirone and 0.89% (n = 1) received gabapentin during follow-up. Potentially adequate pharmacotherapy (appropriate medication, at an appropriate dose, for at least 8 consecutive weeks) was received by 60.38% (n = 253) of the sample, during follow-up.

Exclusive of having received psychotherapy at intake, 60.38% (n = 253) of patients received psychotherapy, at some point during follow-up. Approximately one-third (n = 142, 33.89%) received CT, 90 (21.48%) received BT, and 152 (36.27%) received potentially adequate psychotherapy (CT and/or BT) during follow-up.

Examined together, 69.21% (n = 290) of participants received any potentially adequate treatment, pharmacotherapy and/or psychotherapy, during the follow-up period.

Predictors of Receiving Treatment during the Follow-Up Period (Table 5)

TABLE 5.

Predictors of receiving potentially adequate anxiety treatment over the 5-years of follow-up (exclusive of intake)

	Odds ratio (95% CI)	Wald χ2	DF	P-value
Model 4. Potentially adequate pharmacotherapy
MDD at intake	1.80 (1.11–2.90)	5.76	1	.0164
Panic w/agoraphobia at intake	2.95 (1.74–4.99)	16.25	1	<.0001
Social phobia at intake	1.67 (1.04–2.69)	4.52	1	.0336
Ethno-racial minority	0.45 (0.24–0.83)	6.58	1	.0103
Income less than $20 K	1.87 (1.18–2.95)	7.21	1	.0073
Model 5. Potentially adequate psychotherapy
PTSD at intake	1.61 (1.03–2.54)	4.30	1	.0382
Global social adjustment	1.53 (1.18–1.98)	10.42	1	.0012
College education	2.20 (1.29–3.76)	8.33	1	.0039
Ethno-racial minority	0.51 (0.27–0.97)	4.23	1	.0398
Income less than $20 K	1.99 (1.23–3.21)	7.88	1	.0050
Model 6. Potentially adequate anxiety treatment
MDD at intake	1.91 (1.14–3.19)	6.12	1	.0134
Panic w/agoraphobia at intake	3.17 (1.77–5.66)	16.41	1	<.0001
Ethno-racial minority	0.40 (0.21–0.77)	7.47	1	.0063
Medicaid/medicare	3.04 (1.64–5.63)	12.41	1	.0004

Results from final forward stepwise model including all variables that entered at P ≤ .05.

Potentially adequate pharmacotherapy

Patients with PDA, social phobia, or MDD at intake, and those earning less than $20,000 per year were more likely to receive adequate pharmacotherapy during follow-up. Those who were ethnic minority members were less likely to receive potentially adequate pharmacotherapy.

Potentially adequate psychotherapy

Patients were more likely to receive potentially adequate psychotherapy during follow-up if they had graduated college, were more impaired, earned less than $20,000, and had PTSD at intake. Those who were ethnic minorities were less likely to receive potentially adequate psychotherapy.

Any potentially adequate treatment

Patients with MDD, PDA, and with medicaid/medicare were more likely to receive potentially adequate anxiety treatment. Ethnic minorities were less likely to receive potentially adequate care.

DISCUSSION

When examined cross-sectionally at intake, few primary care patients with anxiety disorders received potentially adequate treatment. Potentially adequate pharmacotherapy was received by fewer than 20%, psychotherapy by 14% (13% received CT techniques and 5% received BT) and both pharmaco- and psychotherapy by 5% of participants. In total, about a quarter of patients received potentially adequate care.

These rates are somewhat lower than those from primary care samples in western and southern United States^{[25, 26]} and than rates of adequate care for individuals with depression. In a US sample, 38% of those with MDD received minimally adequate care, either from a PCP or specialist, in the past year.^[24] Data from a primary care site with colocated behavioral health found that 53% of those with depression received minimally adequate care in the last 3 months.^[36] Our cross-sectional rates of anxiety treatment adequacy are more comparable to those found for substance disorders, for which approximately 26% of patients may receive minimally adequate care.^[24]

When we followed our participants for up to five more years, substantially more eventually received potentially adequate anxiety treatment. After one year of follow-up, less than half (47%) of patients received potentially adequate care. When examined for up to the 5 years, rates of potentially adequate treatment increased to 60% for pharmacotherapy, 36% for psychotherapy, and 26% for receiving both potentially adequate pharmaco- and psychotherapy. In total, over two-thirds (69%) of these primary care patients with anxiety disorders eventually received potentially adequate treatment.

These increased rates may be due to a number of factors. Statistically, we expect more occurrences when examined over a longer time period. Additionally, findings may reflect “watchful waiting”;^{[27, 28]} PCPs watching for symptoms that do not resolve without treatment before recommending additional care. At intake and follow-up, worse functioning was associated with a higher likelihood of receiving potentially adequate psychotherapy; PCPs may refer patients to CBT only after other treatments have failed to improve their functioning. Lastly, higher rates of potentially adequate treatment over follow-up may indicate a cohort effect. During the follow-up years, direct-to-consumer advertising of psychotropic medications increased, SSRIs/SNRIs increasingly received FDA indications for anxiety disorders, and CBT may have become increasingly accessible in the northeastern United States. However, rates of adequate treatment for each follow-up year (see Table 3) seem to argue against this last possibility; there was no trend for treatment rates to increase over time. Potentially adequate pharmacotherapy increased between intake and year 1, but this is likely because at intake only the last 3 months were examined, whereas at year 1, the last 12 months were examined. Similarly, the cumulative rate of adequate treatment received during follow-up is much larger than that of any individual follow-up year. This suggests that rates of adequate treatment did not increase over time, but rather, when a larger time period is examined, more individuals eventually get adequate care for a portion of time. As CBT is generally a time-limited treatment, with new skills practiced after treatment ends, one is not expected to continue in therapy for several years. For pharmacotherapy however, maintenance treatment is more the norm. Thus, sporadic periods of adequate pharmacotherapy are likely a less positive finding. It is somewhat encouraging, however, that over a period of up to 5 years, nearly two-thirds of primary care patients in our sample eventually received potentially adequate anxiety disorder treatment, at least for some period of time.

Consistent with epidemiological studies of primary care treatment,^[24] PDA was repeatedly a predictor of potentially adequate care. This appears due to the relationship between PDA and adequate pharmacotherapy. PDA may receive better pharmacotherapy than other anxiety disorders in primary care, due to providers’ ability to recognize and treat panic (as more medications have FDA indications for panic than for any other anxiety disorders) and due to the severity of associated impairment with agoraphobia. Similarly, many factors associated with an increased likelihood of receiving potentially adequate treatment appear related to symptom and impairment severity. These include the presence of PDA, MDD, social phobia, or PTSD, more severe global impairment, being unmarried, having medicaid/public insurance, and income less than $20,000. These data are somewhat consistent with findings on prediction of adequate depression treatment in primary care, in which symptom severity and comorbid anxiety have been associated with receiving adequate care.^[36] They are promising when viewed from the perspective of treatment reaching those most in need of care. Higher education however, was consistently associated with the likelihood of receiving potentially adequate psychotherapy. This may reflect access to information about such care, and/or provider biases regarding who is an appropriate candidate for CBT. Continued adaptation and implementation efforts to bring adequate psychotherapy to individuals with lower education levels are needed. Importantly, being a member of an ethnoracial minority group was consistently associated with being less likely to receive potentially adequate care. This finding cannot be explained by low income, which was related to an increased likelihood of potentially adequate care. Mechanisms that may contribute to this disparity, such as provider biases and patients’ beliefs about receiving mental health treatment, need continued research attention.

CONCLUSION

Limitations of the current study include that all recruitment occurred in one geographic area. The US northeast has a higher proportion of trained therapists than many other regions of the country. Thus, CBT rates may be lower in other areas. Research involving sites across the United States, with more diverse samples, would be important to examine the generalizability of our findings as our investigation is not epidemiological. Additionally, our measure of psychosocial treatment did not include treatment techniques with a more recent evidence base, such as mindfulness meditation and acceptance, or the duration and/or the number of psychotherapy sessions. It is likely duration/frequency of treatment could affect the adequacy of psychotherapy, although longer treatment does not necessarily equate to better treatment. Further, our definition of adequate pharmacotherapy was chosen so as to be comparable to other studies and included benzodiazepines as an adequate treatment for all anxiety disorders. This is controversial. If benzodiazepines were not included as adequate treatment for all or some of the anxiety disorders (e.g., GAD), rates of treatment adequacy would likely have been lower. Interpretation is also limited by retrospective, self-report. Participants reported back on treatment since their last interview. Further, two-thirds of primary care patients refused to be screened for our study, and 40% of those screening positive repeatedly no-showed or refused a SCID.^[29] It is possible that individuals receiving quality treatment were more willing to discuss their anxiety, and thus, receipt of appropriate care may be overestimated in this paper. Additionally, follow-up data were collected between 2002 and 2007. Rates of adequate treatment may have increased since that time. For example, growing support for integrated primary care^[37] may have led to improved care.

Nonetheless, the data show the importance of taking a longitudinal view, rather than a single “snap shot” when examining treatment utilization. Our longitudinal data indicate that a much greater proportion of primary care patients with anxiety disorders may receive potentially adequate treatment, including CBT, than has previously been assumed from cross-sectional data. Further, it is encouraging that results suggest those with the greatest impairment are most likely to receive these techniques. As there will likely always be limited availability of CBT therapists to serve the needs of the entire population, it may be that those who are most impaired should get this degree of intervention. Unfortunately, our results show that individuals with less education were less likely to receive potentially adequate psychotherapy and ethno-racial minorities were less likely to receive any potentially adequate care. These are specific areas in which implementation efforts need to be focused. Further, though rates of potentially adequate treatment over follow-up were higher than expected, rates each year were still relatively low, and nearly two-thirds of patients in our sample never received potentially adequate psychotherapy. Thus, our results support efforts toward further dissemination and implementation of CBT for anxiety disorders in real world settings, and given limited resources, further research regarding which patients are most likely to benefit from this level of care.