Patients who obtained authorization but did not get initial mental health treatment needed treatment as much as or more than patients who presented for care.
Objective: We sought to determine what demographic and clinical factors are associated with receipt of initial mental health treatment.
Study Design and Methods: A total of 1177 patients completed structured clinical interviews (Michigan Screening for Treatment and Research Triage) when they called to authorize mental health benefits. Measures included age, sex, alcohol use, drug use, anxiety, depression, medical history, behavioral health treatment history, psychosocial stressors, functioning, and suicidality. Multivariate analyses determined the association between these variables and a behavioral health claim within 90 days of the interview.
Results: Among those completing interviews, 85% attended initial mental health treatment. Factors significantly associated with increased odds of treatment initiation were good self-rated health (odds ratio [OR] = 1.70; 95% confidence interval [CI] = 1.15, 2.50), support of family or friends (OR = 1.71; 95% CI = 1.11, 2.65), previous outpatient mental health visits (OR = 1.56; 95% CI = 1.11, 2.19), and recent alcohol use (OR = 1.41; 95% CI = 1.00, 1.97). Factors associated with decreased odds of treatment initiation were recent period of total disability (OR = 0.62; 95% CI = 0.45, 0.87), any previous suicide attempt (OR = 0.56; 95% CI = 0.36, 0.87), 6 or more physician visits for medical reasons this year (OR = 0.64; 95% CI = 0.44, 0.92), and legal problems (OR = 0.31; 95% CI = 0.16, 0.61). In multivariate analyses, family support, history of medical visits, and recent alcohol use were no longer significant predictors.
: Most individuals in this insured population who completed an initial telephone assessment had an initial behavioral health claim. However, patients with greater health or social service needs were at higher risk for not obtaining treatment, suggesting the need for greater outreach and attention by providers and insurers.
(Am J Manag Care. 2009;15(2):105-112)
In this analysis of 1177 patients who completed a structured clinical screening interview (Michigan Screening for Treatment and Research Triage) to obtain authorization for behavioral health benefits, we found that:
Although mental health services are recommended for most common mental disorders, epidemiologic studies show that the majority of those suffering from mental illness are not receiving treatment.1,2 Factors associated with limited access to care, such as living in a rural area, low average income, and lack of insurance, may lead to the underutilization of mental healthcare services.2 Research demonstrates that “no-show” patients are more likely to be younger, male, and unmarried; to have children; and to have less mental healthcare experience.
In addition to clinicians remaining alert for symptoms of mental health disorders, using formal screening tools to identify patients who manifest symptoms, and diagnosing those who meet criteria, patients must recognize a need for treatment, make the initial contact, attend the first visit, and adhere to a recommended course of action before they can benefit from treatment.3 Patients who have made initial contact to receive mental health services have progressed in this treatment-seeking pathway; however, many do not take the next step in obtaining needed care—completing an initial behavioral health visit. Prior studies have reported that among patients attempting to access care in community mental health settings, approximately 25%-40% do not show up for their first appointment.4-7 These prior studies focused largely on community mental health settings, often serving a predominately seriously mentally ill population. Frequently, these studies also were limited by small sample sizes and lack of key patient information, such as symptom burden, family support, and current levels of functioning, which bear on the need for mental health treatment.
In this study, we used data from a standardized telephone intake interview routinely conducted by a managed behavioral healthcare organization affiliated with a university medical center. The standardized interview, which was used as a component of the organization’s routine triage and referral activity, included detailed information on patients’ mental health symptoms, home and work environments, suicidal ideation, and substance use, allowing us to examine demographic and clinical factors associated with patients’ initial behavioral health treatment. Identifying factors that predict failure to follow up with behavioral health treatment may lead to targeted interventions to improve engagement in care.
The study population consisted of patients initially assessed for treatment by a managed behavioral healthcare organization associated with a large Midwestern university between January 1, 2003, and March 31, 2006. Patients routinely completed a telephone-based assessment interview (Michigan Screening for Treatment and Research Triage [M-START]) when they called to request authorization for services. All patients who called to authorize their benefits were approved to initiate care. The care management system tracks interview dates, and assessments are not completed more than once in a 1-year period. Furthermore, all participants were continuously enrolled for the duration of the study, and no participants completed more than 1 interview.
We examined whether patients had a behavioral health claim within the 90 days after their M-START interview. Although wait times may be associated with not showing up for treatment,4-6 we chose a 90-day window for 2 reasons: (1) this time window is a standard metric used by the National Committee for Quality Assurance to evaluate effectiveness of acute treatment for depression among managed care organizations8 and (2) this time window was sufficient for all patients to be able to be seen by the behavioral health provider (ie, all members should have been able to be seen within 90 days if they chose to do so).
The M-START interview is a comprehensive telephonebased, computer-assisted structured interview, consisting of triage and need-relevant measures of psychiatric symptoms that allow for referral to appropriate specialties as needed, health and functional assessments, and psychosocial supports. M-START is administered by trained clinical social workers and takes approximately 10-15 minutes to complete. The interview includes stem questions from validated and reliable clinical instruments (eg, the Patient Health Questionnaire [PHQ-9]9 for depression, the Structured Clinical Interview for DSM-III-R [SCID]10 for mood and anxiety disorders, the CAGE11 for alcohol use). Additional items were generated from the expert opinion of 23 clinicians and researchers. The Delphi method12 was used to ask the experts to narrow down an initial list of potential items into a brief interview designed to identify patient treatment needs by indicating their rankings of importance of individual items.
In particular, patients are asked about their age, alcohol use, anxiety, behavioral health treatment history, depression symptom severity (using PHQ-99), drug use, sex, psychosocial issues in their environment (eg, family and relationships, significant loss, legal problems), medical history, functional impairment (eg, inability to work or carry out normal activities), and suicidality.
Patient M-START interview data then were linked to outpatient behavioral health claims from the managed care organization that administered the interviews. Behavioral health claims were identified based on Current Procedural Terminology codes 90801-90911.13
We first conducted bivariate analyses using t tests (for continuous measures) and Χ2 tests (for categorical or dichotomous measures), comparing patients who had a behavioral health claim within 90 days of their M-START interview with those who did not have a claim during that time period. As we conducted 48 bivariate tests, we used the false discovery rate14 to adjust for multiple comparisons. In this case, our adjusted P values for significance were ([n + 1]/[n*2])*0.05 = (49/96)*0.05 = 0.026. All predictors identified as potential significant predictors of behavioral health claims at the P ≤.026 level were used in multivariable analyses. Next, we used stepwise logistic regression analysis to evaluate which factors influenced whether a patient had a behavioral health claim within the 90-day period. We removed predictors that were no longer significant in the full model, with the exception of age, sex, and PHQ-9 score, which were left in regardless of significance because of their relevance as demographic and clinical factors associated with behavioral health treatment.
Between January 1, 2003, and March 31, 2006, 1177 patients completed an M-START interview. Of those, 1006 (85%) had an outpatient behavioral health claim within 90 days after their M-START interview. Baseline characteristics of the population completing interviews are presented in Table 1.
In univariate analyses, the following characteristics were associated with increased odds of behavioral treatment initiation: good self-rated health (odds ratio [OR] = 1.70; 95% confidence interval [CI] = 1.15, 2.50), having support of family or friends (OR = 1.71; 95% CI = 1.11, 2.65), having made any outpatient mental health treatment visits in the past 2 years (OR = 1.56; 95% CI = 1.11, 2.19), and having consumed alcohol in the past 3 months (OR = 1.41; 95% CI = 1.00, 1.97). Conversely, the following characteristics were associated with decreased odds of treatment initiation: functional impairment (being totally unable to work or carry out normal activities for any days in the past 30 days [OR = 0.62; 95% CI = 0.45, 0.87]), having ever made a suicide attempt (OR = 0.56; 95% CI = 0.36, 0.87), having had 6 or more doctor visits for medical reasons this year (OR = 0.64; 95% CI = 0.44, 0.92), and having legal problems (OR = 0.31; 95% CI = 0.16, 0.61).
In multivariate analyses (presented in Table 2, model 3), factors that continued to increase the odds of treatment initiation included good selfrated health (OR = 1.55; 95% CI = 1.01, 2.37) and having any mental health outpatient visits in the last 2 years (OR = 1.82; 95% CI = 1.27, 2.60). Alternatively, factors that decreased the odds of behavioral treatment initiation included functional impairment (being totally unable to work or carry out normal activities for any days in the past 30 days [OR = 0.65; 95% CI = 0.45, 0.94]), having ever made a suicide attempt (OR = 0.61; 95% CI = 0.38, 0.98), and having legal problems (OR = 0.30; 95% CI = 0.15, 0.61). Family support was no longer a significant predictor in fully adjusted analyses. In fully adjusted models containing both medical visits and self-rated health, neither was a significant predictor as a result of the strong relationship between the 2 factors (Χ2 = 266.3, P <.0001, or in correlational terms, r = -0.48 with P = .032). For this reason and given the ubiquity of the selfrated health measure in research and clinical settings, the “medical visits” measure was dropped from the final model (see Table 2).
This study presents new findings regarding demographic and clinical factors that are associated with showing up for initial outpatient behavioral health treatment among privately insured individuals. In this population, the large majority of the patients who called to initiate treatment subsequently had a behavioral health claim. This finding differs from those of previous studies that have reported higher no-show rates, although most of these studies were conducted in community mental health populations.4-7
Interestingly, our follow-up rates were similar to those found in a study conducted in a community mental health center in the Netherlands. That study also found long waiting times, lack of motivation, and resolution of the problem as the primary reasons why patients did not follow through with initial treatment. The authors concluded that measures to reduce the no-show rate may be counterproductive by discouraging active coping.15
Unfortunately, in the present study, patients who did not have a behavioral health claim appeared to be those who were most functionally impaired. Although one would expect that patients with more psychiatric illnesses or greater symptom severity would be in greater need of treatment, the presence of psychiatric symptoms from a variety of disorders (depression, anxiety, panic, posttraumatic stress disorder, obsessivecompulsive disorder) and the severity of depressive symptoms as measured by PHQ-9 did not appear to affect the likelihood of following through with a behavioral health visit. More troubling, patients with poor selfrated health, difficulty functioning at work and in other activities, past suicide attempts, or with legal problems were less likely to initiate behavioral healthcare. Conversely, patients who had prior experience with the mental health system were more likely to engage in care.
Our findings suggest that health plans and providers cannot assume that patients who initiate contact for accessing mental health services but then fail to show for their initial treatment are individuals who have less need of care—indeed, these individuals may have greater need for treatment. Not only are no-shows detrimental from the perspective of patient health and well-being, but they also are disruptive to providers and overall workflow.16,17 As a result, there may be an institutional incentive for insurers to ensure that patients follow through with care. Behavioral health organizations and provider clinics may need to inquire about patients’ functional capacity at the time of mental health intake and assist those who report poor health or functional compromise in following through with care. Reminder calls prior to appointments are one demonstrated way that healthcare plans have decreased no-shows.18-20 Healthcare organizations also may need to regularly call no-show patients to further ascertain patient need and encourage follow-up for compromised individuals as part of a “best practice” of dealing with noncompliance with treatment.21 Furthermore, care management programs such as depression disease management programs have been demonstrated to improve both coordination of care among providers and patient adherence and outcomes.22
We note that initiating care is an essential but not sufficient condition for successful mental health treatment. Thus, future research is needed regarding demographic, clinical, and treatment predictors of continued treatment and of adequate treatment trials.
Although this study used a comprehensive, structured interview to provide new information on possible correlates of behavioral health treatment initiation, several limitations must be noted. The population included in our study had a high follow-up rate with behavioral health treatment (85%), which may be higher than expected based on previous research for several reasons. This population already was more inclined than the general population to follow through with care, as these patients had already taken the first steps to initiate a phone call and seek to authorize mental health benefits. Furthermore, this population represented insured individuals who perhaps had higher rates of employment, income, and overall socioeconomic status, factors that also may be associated with follow-through with treatment. Prior research with a university employee population indicates that even when insurance benefits and access to care are constant and easily attainable, whites and those with high incomes consume more mental health benefits than other people do.23
Another concern is that we can’t say anything from our data about the quality of the treatment a patient received, nor do we have any follow-up data on severity of emotional distress or changes in the course of illness and outcome over time. We do not have any claims data on possible prior behavioral health treatment (eg, specific providers, inpatient stays, psychiatric emergency room visits) other than patients’ responses to the question regarding whether they had previously had outpatient behavioral healthcare in the past 2 years. We also note that our focus was on specialty behavioral health treatment and not on potential treatment for emotional distress in primary care settings, as the M-START interview is specifically tailored to patients seeking to use behavioral health benefits.
To know more about a patient’s prior and current illness and treatment trajectory over time, we would need more data on both frequency and type of behavioral health treatment, as well as type of provider, medication adherence, and a follow-up period longer than 90 days. This information would perhaps provide greater insight into why a patient may not end up having a behavioral health claim after calling to initiate benefits. However, our focus here was not on quality of care and overall treatment outcomes, but rather on trying to identify who does and does not come in for that first visit, and what factors may be associated with the likelihood of attendance. Further research including a qualitative component might help identify reasons why patients do not follow through with treatment after calling to authorize visits.
Finally, there can be questions about the generalizability of findings from 1 interview and 1 managed care plan, with patients seeking care in an academic medical setting. Patients who have other types of healthcare, providers, and locations may have a different likelihood of and characteristics associated with behavioral health treatment initiation. Furthermore, there are limitations associated with using patient self-report data (eg, the data from the M-START interview) in assessing the true underlying nature of patients’ disease. However, despite these limitations, this study provided a unique opportunity to examine these issues by pulling together data from both an intake screen and from behavioral health claims to identify treatment initiation.
It is heartening that most patients who completed an initial telephone assessment went on to have a behavioral health claim within 90 days after their interview. However, patients who may have had greater health or social service needs were less likely to have initiated treatment, whereas those patients with better health and experience with the mental health treatment system were more likely to initiate treatment. This study extended prior work on the process of treatment engagement by examining a large number of demographic, social, and clinical characteristics associated with treatment initiation after assessment. Our findings suggest that patients experiencing the greatest need for services may be least likely to initiate treatment, necessitating greater outreach and attention by providers.
The authors would like to acknowledge the contributions of Thomas Spafford and Michelle Kaston, BS, to earlier drafts of this manuscript.
Author Affiliations: From the Health Services Research and Development Center of Excellence (KZ, DEW, MV), Department of Veterans Affairs, Ann Arbor, MI; the Department of Psychiatry (KZ, PNP, RJM, JSK, HW, DJD, MMB, MV), University of Michigan, Ann Arbor; and the Blue Care Network of Michigan (DJD), Ann Arbor.
Funding Source: This publication was made possible by the University of Michigan’s Department of Psychiatry and Venture Investment Funds, the Department of Veterans Affairs HSR&D’s RCD 98-350 Career Development Award, and Coordinating Mental Health Services in Integrated Delivery Systems (principal investigator, Marcia Valenstein, MD).
Author Disclosure: The authors (KZ, PNP, RJM, JSK, HW, DEW, DJD, MMB, MV) report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.
Authorship Information: Concept and design (KZ, PNP, RJM, DJD, MV); acquisition of data (KZ, JSK, HW, MV); analysis and interpretation of data (KZ, RJM, DEW, DJD, MMB, MV); drafting of the manuscript (KZ, PNP, HW, DJD); critical revision of the manuscript for important intellectual content (KZ, PNP, DJD, MMB, MV); statistical analysis (KZ, RJM, DEW); provision of study materials or patients (JSK); obtaining funding (MV); and administrative, technical, or logistic support (JSK, HW, DEW, DJD, MMB).
Address correspondence to: Kara Zivin, PhD, Department of Psychiatry, University of Michigan, 4250 Plymouth Rd, Box 5765, Ann Arbor, MI 48109. E-mail: firstname.lastname@example.org.
1. American Psychiatric Association, Work Group on Major Depressive Disorder. Practice Guideline for the Treatment of Patients with Major Depression. 2nd ed. Arlington, VA: American Psychiatric Association; 2006.
2. Wang PS, Lane M, Olfson M, Pincus HA, Wells KB, Kessler RC. Twelve-month use of mental health services in the United States: results from the National Comorbidity Survey Replication. Arch Gen Psychiatry. 2005;62(6):629-640.
3. Donabedian A. Aspects of Medical Care Administration: Specifying Requirements for Health Care. Cambridge, MA: Harvard University Press; 1973.
4. Folkins C, Hersch P, Dahlen D. Waiting time and no-show rate in a community mental health center. Am J Community Psychol. 1980;8(1):121-123.
5. Orme DR, Boswell D. The pre-intake drop-out at a community mental health center. Community Ment Health J. 1991;27(5):375-379.
6. Barton AK. Following up on aftercare: show versus no-show rates in North Carolina. Hosp Community Psychiatry. 1977;28(7):545-546.
7. Kruse GR, Rohland BM, Wu X. Factors associated with missed first appointments at a psychiatric clinic. Psychiatric Serv. 2002;53(9):1173-1176.
8. National Committee for Quality Assurance. Antidepressant Medication Management (Effective Acute Treatment Phase). Washington, DC: National Committee for Quality Assurance; 2003.
9. Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. 2001;16(9):606-613.
10. Spitzer RL, Williams JB, Gibbon M, First MB. The Structured Clinical Interview for DSM-III-R (SCID), I: history, rationale, and description. Arch Gen Psychiatry. 1992;49(8):624-629.
11. Mayfield D, McLeod G, Hall P. The CAGE questionnaire: validation of a new alcoholism screening instrument. Am J Psychiatry. 1974;131(10):1121-1123.
12. Helmer O, Rescher N. On the epistemology of the inexact sciences. Management Science. 1959;6(1):25-52.
13. American Medical Association. Psychiatry. In: Current Procedural Terminology (CPT) 2003 Standard Edition. Chicago, IL: AMA Press; 2002.
14. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J Royal Stat Soc B. 1995;57(1):289-300.
15. Peeters FP, Bayer H. ‘No-show’ for initial screening at a community mental health centre: rate, reasons and further help-seeking. Soc Psychiatry Psychiatr Epidemiol. 1999;34(6):323-327.
16. Moore CG, Wilson-Witherspoon P, Probst JC. Time and money: effects of no-shows at a family practice residency clinic. Fam Med. 2001;33(7):522-527.
17. Lacy NL, Paulman A, Reuter MD, Lovejoy B. Why we don’t come: patient perceptions on no-shows. Ann Fam Med. 2004;2(6):541-545.
18. Gariti P, Alterman AI, Holub-Beyer E, Volpicelli JR, Prentice N, O’Brien CP. Effects of an appointment reminder call on patient show rates. J Subst Abuse Treat. 1995;12(3):207-212.
19. Shepard DS, Moseley TA 3rd. Mailed versus telephoned appointment reminders to reduce broken appointments in a hospital outpatient department. Med Care. 1976;14(3):268-273.
20. Gates SJ, Colborn DK. Lowering appointment failures in a neighborhood health center. Med Care. 1976;14(3):263-267.
21. Cruz M, Cruz RF, McEldoon W. Best practice for managing noncompliance with psychiatric appointments in community-based care. Psychiatr Serv. 2001;52(11):1443-1445.
22. Katon W, Von Korff M, Lin E, et al. Collaborative management to achieve depression treatment guidelines. J Clin Psychiatry. 1997;58(suppl 1):20-23.
23. Richman BD. Insurance expansions: do they hurt those they are designed to help? Health Aff (Millwood). 2007;26(5):1345-1357.