Characteristics of Patients Who Attended Behavioral Health Services After Primary Care Referral With Referral Management Support
Abstract
Objective:
This study examined whether documented disparities in access to behavioral health specialty care persisted in a novel integrated primary care model situated in a large health system when triage and referral management supports were provided by a centralized resource center for patients with behavioral health needs.
Methods:
Patients triaged and referred to specialty behavioral health care who did or did not attend a specialty care visit (N=1,450) were compared in terms of various demographic and clinical characteristics by using binary logistic regression.
Results:
Among patients with attendance data, financially unstable individuals were more likely than financially stable counterparts to miss their first appointment with a specialty behavioral health provider after referral from primary care. Previously documented attendance disparities based on race, ethnicity, and illness severity were not observed.
Conclusions:
These findings can inform targeted strategies to increase attendance among patients with financial insecurity and reduce disparities in outpatient behavioral health services.
Highlights
•
Financial security was associated with better attendance at the first specialty behavioral health appointment after referral from primary care.
•
Patients’ predictions of whether they would attend their first specialty appointment corresponded with their actual attendance.
•
Previously documented disparities in access to care, such as those based on race, ethnicity, and mental illness severity, were not observed in this novel triage and referral model.
Access to specialty behavioral health care remains a challenge in the United States. Most sources report that, among individuals referred to specialty behavioral health care from primary care, 60% to 70% attend the initial appointment with the provider to whom they were referred (1), a lower rate than in other specialties (63% to 83%) (2). Poor attendance at specialty behavioral health appointments has been linked to personal consequences, such as increased risk of psychiatric hospitalization (3), as well as systems-level consequences, such as underutilization of resources and treatment delay for other individuals seeking care (1). Non-White Hispanic/Latinx individuals experience lower behavioral health referral and utilization rates, compared with non-Hispanic/Latinx Whites (4, 5). Findings on associations between attendance and age (1, 6), gender (2, 6), overall health status (5, 7), and mental illness severity (7, 8) have been mixed. Patients with depression are significantly more likely to attend behavioral health visits when appointments are made for them, whereas patients without depression are less likely to respond to such administrative changes (2, 6). There is sparse literature on the effects of financial stability on appointment attendance, although lower socioeconomic status is associated with lower rates of behavioral health care utilization (4).
The collaborative care model (CoCM), the leading model for integrating behavioral health treatments in primary care (9, 10), has been shown to reduce attendance disparities related to race and ethnicity (11). Key components of the CoCM model include a core team consisting of a primary care provider, mental health provider or care manager, and consulting psychiatrist; use of evidence-based practices and measurement-based care; and a registry to track patients. Implementation of the CoCM was initiated by the University of Pennsylvania Health System (Penn Medicine), a large and diverse health system, in January 2018. This primary care system, referred to as Penn Integrated Care, includes a novel resource center that provides key triage and referral management activities for patients with more complex needs.
Although CoCM programs can improve access to care, CoCM is not an appropriate referral for all patients, particularly those with more severe or persistent behavioral health needs (12). In the Penn Integrated Care model, primary care providers are encouraged to refer patients with any behavioral health symptom or condition, regardless of severity, to the resource center. Intake coordinators assess the patient by telephone using validated measures, such as the Patient Health Questionnaire–9 (PHQ-9) (13), and triage them to the most appropriate level of care using an established algorithm (14). Patients may be triaged to CoCM in the primary care practice or to a variety of community settings or providers, such as an outpatient therapist or psychiatrist or substance use disorder treatment. To facilitate access to specialty behavioral health care in the community (12), the resource center helps patients schedule their appointment, follows up with those referred to specialty care by phone (typically 1 to 2 weeks after the appointment) to ascertain attendance, and offers to help reschedule or find a new provider if needed. At the time of referral, patients are also asked to provide their self-rated likelihood of attendance by reporting on a 9-point Likert scale the likelihood that they will attend their outpatient visit.
Preliminary research by our team indicated that 88% (N=845 of 961) of patients screened by the resource center and referred to CoCM in a Penn Integrated Care primary care practice attended one or more appointments with a Penn Integrated Care mental health provider (15). However, attendance at specialty behavioral health appointments after referral in this population has yet to be investigated. Previous research has largely relied on either the primary care or specialty office to determine if patients attend a behavioral health appointment after referral.
In this study, we identified patient factors associated with attendance of the first specialty behavioral health visit after referral from primary care and triage by the resource center. We hypothesized that disparities related to patient sociodemographic and clinical factors would be observed, consistent with the literature, such that non-White Hispanic/Latinx patients (4, 5) and patients who self-report financial limitations (4) would have lower rates of attendance, compared with non-Hispanic/Latinx White patients who report financial security. Because the Penn Integrated Care resource center provides patient support services that have been shown to bolster attendance rates for patients with depression, we hypothesized that there would be minimal disparities based on depression severity and endorsement of suicidal ideation (8). Referral management supports like those provided by the Penn Integrated Care resource center are uniquely positioned to implement interventions to reduce disparities in access to care.
Methods
This project was reviewed by the University of Pennsylvania Institutional Review Board and determined to be exempt. Data were extracted from the Penn Integrated Care resource center Behavioral Health Laboratory software (12), including structured interview questions and validated symptom measures for all patient encounters between January 2018 and February 2020. For patients with multiple referrals to the resource center, we used the first encounter.
To ascertain any disparities in the triage and referral process itself, we first used Pearson chi-square tests and two-tailed independent samples t tests to compare patients triaged to CoCM (N=2,108) with those triaged to specialty care (N=4,182) on six variables of interest: self-reported race, ethnicity, financial stability, and overall health status; severity of depression (PHQ-9 score); and endorsement of recurrent suicidal thoughts (PHQ-9 question 9). Previous literature has shown these variables to contribute to attendance disparities (4, 5, 7, 8).
To determine if differences existed between patients who did or did not complete a follow-up call to determine attendance, we compared the 1,450 individuals who were reached with individuals who were referred to specialty care but could not be reached or did not have a completed follow-up call because of administrative reasons (N=1,590). Follow-up calls may not have been completed either because the resource center was unable to reach the patient or because follow-up calls were not attempted. Follow-up calls were not attempted when referral volumes were high or the resource center was short staffed.
For our main analysis, we compared patients who did or did not attend their first specialty visit after referral (N=1,450). We examined descriptive statistics for patient sociodemographic and clinical variables, then used binary logistic regression to identify variables that predicted attendance. To correct for multiple comparisons, we used a Bonferroni-adjusted p value of 0.008.
The following categorical variables were dichotomized for analysis: race (White versus non-White), financial security (“are comfortable” versus “have just enough to get along” or “can’t make ends meet”), good health status (“excellent,” “very good,” or “good” versus “fair” or “poor”), and endorsement of recurrent suicidal thoughts (“not at all” versus “several days,” “more than half the days,” or “nearly every day”). Because the resource center follows an algorithm to determine which assessment tools to administer, not all patients were asked all questions.
Results
Individuals referred to specialty care (N=4,182) were more likely to have more serious illness than those triaged to CoCM (N=2,108), as evidenced by worse overall health status (χ2=5.43, df=1, p=0.020) and higher PHQ-9 scores (t=18.80, df=5,589, p<0.001).
Table 1 reports descriptive statistics for patients triaged to specialty behavioral health care by the resource center. The majority identified as female, of non-White race, and of non-Hispanic/Latinx ethnicity and reported financial insecurity. Of the 1,450 individuals who were able to be reached by the resource center to confirm attendance of their first specialty visit, 65% (N=947) attended their visit.
| Total | Attended first visit | Did not attend | ||||
|---|---|---|---|---|---|---|
| (N=1,450) | (N=947) | first visit (N=503) | ||||
| Characteristic | N | %a | N | %a | N | %a |
| Age (M±SD years) | 41.9±14.7 | 41.5±14.4 | 42.6±15.1 | |||
| Gender | ||||||
| Male | 401 | 28 | 276 | 29 | 125 | 25 |
| Female | 1,049 | 72 | 671 | 71 | 378 | 75 |
| Raceb | ||||||
| Asian/Pacific Islander | 40 | 3 | 23 | 3 | 17 | 3 |
| Black/African American | 797 | 55 | 502 | 54 | 295 | 59 |
| Native American/Alaskan Native | 5 | 0.3 | 3 | 0.3 | 2 | 0.4 |
| White | 454 | 32 | 314 | 34 | 140 | 28 |
| Other or more than one race | 118 | 8 | 81 | 9 | 37 | 7 |
| Prefer not to disclose | 24 | 2 | 14 | 2 | 10 | 2 |
| Ethnicityb | ||||||
| Hispanic/Latinx | 99 | 7 | 62 | 7 | 37 | 7 |
| Non-Hispanic/Latinx | 1,320 | 92 | 859 | 92 | 461 | 92 |
| Prefer not to disclose | 19 | 1 | 16 | 2 | 3 | 0.6 |
| Married or partneredc | 414 | 29 | 273 | 29 | 141 | 28 |
| Employedc | 907 | 63 | 593 | 63 | 314 | 63 |
| Financial situationb | ||||||
| Are comfortable | 491 | 34 | 351 | 38 | 140 | 28 |
| Have just enough to get along | 589 | 41 | 369 | 39 | 220 | 44 |
| Can’t make ends meet | 334 | 23 | 203 | 22 | 131 | 26 |
| Prefer not to disclose | 24 | 2 | 14 | 2 | 10 | 2 |
| General healthb | ||||||
| Excellent | 70 | 5 | 50 | 5 | 20 | 4 |
| Very good | 200 | 14 | 137 | 15 | 63 | 13 |
| Good | 461 | 32 | 310 | 33 | 151 | 30 |
| Fair | 489 | 34 | 303 | 32 | 186 | 37 |
| Poor | 218 | 15 | 137 | 15 | 81 | 16 |
| Behavioral health assessmentd | ||||||
| Current psychotic symptomse | 81 | 6 | 53 | 6 | 28 | 6 |
| Active maniae | 60 | 4 | 40 | 4 | 20 | 4 |
| PHQ-9 score (M±SD)f | 12.6±6.2 | 12.8±6.3 | 12.3±6.2 | |||
| Score of 1 or higher on PHQ-9 item 9g | 292 | 20 | 207 | 22 | 85 | 17 |
| GAD-7 score (M±SD)h | 11.9±5.7 | 12.0±5.6 | 11.7±5.9 | |||
| Self-rated likelihood of attendance (M±SD)i | 7.7±1.6 | 7.9±1.5 | 7.4±1.9 | |||
a
Percentages are rounded to the nearest whole number unless less than 1%.
b
N=1,438 for all patients, 937 for those who attended the first visit, and 501 for those who did not attend the first visit.
c
N=1,438 for all patients.
d
Assessment instruments were administered by using the Behavioral Health Laboratory algorithm. Sample sizes vary by item because not all items or instruments were administered to all individuals.
e
N=1,435 for all patients, 934 for those who attended the first visit, and 501 for those who did not attend the first visit.
f
Scores on the Patient Health Questionnaire–9 (PHQ-9) range from 0 to 27, with higher scores indicating greater severity of depression.
g
“Over the last two (2) weeks, how often have you been bothered by any of the following problems? Thoughts that you would be better off dead or of hurting yourself in some way.” Possible responses are 0, not at all; 1, several days; 2, more than half the days; 3, nearly every day.
h
Scores on the Generalized Anxiety Disorder–7 (GAD-7) scale range 0 to 21, with higher scores indicting greater severity of anxiety. N=1,412 for all patients, 911 for those who attended the first visit, and 501 for those who did not attend the first visit.
i
Item rated on a 9-point Likert scale ranging from 0, not likely to attend, to 9, very likely to attend. N=495 for all patients, 352 for those who attended the first visit, and 143 for those who did not attend the first visit.
Of patients referred to specialty care (N=1,450), we found that those with attendance data were more likely to endorse financial security (χ2=10.49, df=1, p=0.001) and had higher PHQ-9 scores (t=−2.76, df=3,003, p=0.006). There were no significant differences for gender, race, ethnicity, health status, suicidal thoughts, or self-rated likelihood of attendance.
For our main analysis, we conducted a binary logistic regression analysis with attendance at the first specialty behavioral health appointment after referral (confirmed by a yes-or-no response) as the dependent variable, and entered gender, race, ethnicity, health status, PHQ-9 score, suicidal thoughts, and financial security as independent variables in the model. Individuals reporting financially security were 1.5 times (odds ratio=1.5, 95% confidence interval=1.17–1.98) more likely to attend their appointment (B=0.42±0.14, p=0.002) than those who endorsed financial insecurity. Differences were nonsignificant for all other variables.
We did not include self-rated likelihood of attendance in the regression analysis because the question was asked of only approximately one-third of the sample (N=495). We conducted an exploratory t test that demonstrated that those who attended their visit had significantly higher self-rated likelihood of attendance (t=−2.72, df=223, p=0.007) than nonattenders.
Discussion
Financial security was the only significant difference observed between patients who did or did not attend their initial appointment with a specialty behavioral health provider after referral from primary care and triage by the Penn Integrated Care resource center. Individuals who endorsed financially security were 1.5 times more likely to attend their first appointment than those endorsing financial insecurity.
This is the first study to our knowledge to assess the relationship between financial security and patient attendance at specialty behavioral health appointments after referral from primary care. Our findings are corroborated by research showing that lower socioeconomic status is linked to lower rates of behavioral health care utilization (4) and underscore the need for supports for patients with limited finances, such as the provision of subsidized care or transportation vouchers. Finances, socioeconomic status, race, ethnicity, and illness severity are particularly correlated with one another in the United States. Future studies that are sufficiently diverse and powered to explore the relative contribution of these factors are needed to determine whether the attendance disparity we observed is based solely on financial security.
Our analyses showed no significant difference in specialty care attendance by race, ethnicity, overall health status, or severity of depressive symptoms, despite disparities documented in the literature. One possible explanation is that the Penn Integrated Care resource center model, which comprises CoCM practices supported by a centralized triage and referral management service, reduces disparities in access to specialty behavioral health care. Without a control group, we cannot make this conclusion; however, the Penn Integrated Care resource center facilitates scheduling, which has been shown to increase attendance (2, 6), and uses validated assessment tools and decision support software to ensure patients are connected with the most appropriate care. Previous research has shown that CoCM reduces outcome disparities related to race and ethnicity (11). Future studies that investigate the added value of triage and referral management supports in CoCM would be interesting.
The attendance rate observed in this sample (65%) is consistent with rates reported in the literature (60% to 70% attendance in prior studies) (1). Patients with attendance data were more likely to be financially secure and to have a higher PHQ-9 score. The finding of financial security makes sense, given that those who are financially secure are more likely to have stable housing and consistent telephone access required to be contacted by the resource center. More research is needed to investigate the link between depression severity and ability to be contacted, and researchers should be mindful that those who are financially insecure may have less complete data.
Self-rated likelihood of attendance was significantly associated with actual attendance. This finding illustrates patients’ awareness of personal barriers and the multifactorial decision making involved in initiating specialty behavioral health care (8). Brief intervention strategies, such as motivational interviewing or problem solving around structural barriers, may be useful for patients who indicate low likelihood of attendance.
Our study is uniquely situated within the context of a novel behavioral health assessment, triage, and referral management entity within a large and diverse CoCM program. A limitation is that our findings are based on data for patients who were able to be reached by the resource center to confirm attendance. Future studies that examine attendance by type of outpatient program or provider referred to or that take into account patient out-of-pocket costs would be interesting. Navigating specialty behavioral health care is complex; qualitative studies examining patients’ decision-making processes and experiences surrounding psychiatric referral attendance would add to the existing literature.
Conclusions
The literature has shown that poor attendance of specialty behavioral health appointments is associated with a higher risk of psychiatric hospitalization and inefficient utilization of resources on a systems level. We found that self-reported financial instability plays a significant role in explaining nonattendance of the first specialty behavioral health appointment after referral from primary care. Improved strategies to ensure equitable access to care and to provide additional support to patients with limited financial resources are needed. Previously documented disparities based on race, ethnicity, and illness severity were not observed in our sample; further studies are needed to determine whether the additional supports provided by a centralized resource center can bridge disparities in behavioral health care access.
References
1.
de Figueiredo JM, Boerstler H, Doros G: Failure of high-risk minority patients to show up for outpatient psychiatric treatment. Int J Ment Health 2009; 39:91–105.
2.
Forrest CB, Shadmi E, Nutting PA, et al: Specialty referral completion among primary care patients: results from the ASPN Referral Study. Ann Fam Med 2007; 5:361–367
3.
Grinshpoon A, Lerner Y, Hornik-Lurie T, et al: Post-discharge contact with mental health clinics and psychiatric readmission: a 6-month follow-up study. Isr J Psychiatry Relat Sci 2011; 48:262–267
4.
Narain K, Xu H, Azocar F, et al: Racial/ethnic disparities in specialty behavioral health care treatment patterns and expenditures among commercially insured patients in managed behavioral health care plans. Health Serv Res 2019; 54:575–585
5.
Jones AL, Cochran SD, Leibowitz A, et al: Racial, ethnic, and nativity differences in mental health visits to primary care and specialty mental health providers: analysis of the Medical Expenditures Panel Survey, 2010–2015. Healthcare 2018; 6:29
6.
Craig TJ, Huffine CL, Brooks M: Completion of referral to psychiatric services by inner city residents. Arch Gen Psychiatry 1974; 31:353–357
7.
Killaspy H, Banerjee S, King M, et al: Prospective controlled study of psychiatric out-patient non-attendance. Characteristics and outcome. Br J Psychiatry 2000; 176:160–165
8.
Grunebaum M, Luber P, Callahan M, et al: Predictors of missed appointments for psychiatric consultations in a primary care clinic. Psychiatr Serv 1996; 47:848–852
9.
Archer J, Bower P, Gilbody S, et al: Collaborative care for depression and anxiety problems. Cochrane Database Syst Rev 2012; 10:CD006525
10.
Miller CJ, Grogan-Kaylor A, Perron BE, et al: Collaborative chronic care models for mental health conditions: cumulative meta-analysis and metaregression to guide future research and implementation. Med Care 2013; 51:922–930
11.
Angstman KB, Phelan S, Myszkowski MR, et al: Minority primary care patients with depression: outcome disparities improve with collaborative care management. Med Care 2015; 53:32–37
12.
Krahn DD, Bartels SJ, Coakley E, et al: PRISM-E: comparison of integrated care and enhanced specialty referral models in depression outcomes. Psychiatr Serv 2006; 57:946–953
13.
Kroenke K, Spitzer RL, Williams JBW: The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med 2001; 16:606–613
14.
Tew J, Klaus J, Oslin DW: The Behavioral Health Laboratory: building a stronger foundation for the patient-centered medical home. Fam Syst Health 2010; 28:130–145
15.
Wolk CB, Last BS, Livesey C, et al: Addressing common challenges in the implementation of collaborative care for mental health: the Penn Integrated Care program. Ann Fam Med 2021; 19:148–156
Information & Authors
Information
Published In
History
Received: 15 July 2020
Revision received: 7 November 2020
Revision received: 20 January 2021
Accepted: 25 February 2021
Published online: 2 June 2021
Published in print: December 01, 2021
Keywords
Authors
Competing Interests
The authors thank their partners in the Penn Integrated Care program as well as Cecilia Livesey, M.D., and Matthew Press, M.D., for providing feedback on the manuscript.
Competing Interests
The authors report no financial relationships with commercial interests.
Metrics & Citations
Metrics
Citations
Export Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.
For more information or tips please see 'Downloading to a citation manager' in the Help menu.
View Options
View options
PDF/EPUB
View PDF/EPUBLogin options
Already a subscriber? Access your subscription through your login credentials or your institution for full access to this article.
Personal login Institutional Login Open Athens loginNot a subscriber?
PsychiatryOnline subscription options offer access to the DSM-5-TR® library, books, journals, CME, and patient resources. This all-in-one virtual library provides psychiatrists and mental health professionals with key resources for diagnosis, treatment, research, and professional development.
Need more help? PsychiatryOnline Customer Service may be reached by emailing [email protected] or by calling 800-368-5777 (in the U.S.) or 703-907-7322 (outside the U.S.).