Clinical Results for Patients with Major Depressive Disorder in the Texas Medication Algorithm Project
Total Page:16
File Type:pdf, Size:1020Kb
ORIGINAL ARTICLE Clinical Results for Patients With Major Depressive Disorder in the Texas Medication Algorithm Project Madhukar H. Trivedi, MD; A. John Rush, MD; M. Lynn Crismon, PharmD; T. Michael Kashner, PhD, JD, MPH; Marcia G. Toprac, PhD; Thomas J. Carmody, PhD; Tracie Key, BSN, RN; Melanie M. Biggs, PhD; Kathy Shores-Wilson, PhD; Bradley Witte, BA; Trisha Suppes, MD, PhD; Alexander L. Miller, MD; Kenneth Z. Altshuler, MD; Steven P. Shon, MD Context: The Texas Medication Algorithm Project is an Psychiatric Rating Scale total score was higher than the evaluation of an algorithm-based disease management median for that clinic’s routine quarterly evaluation of program for the treatment of the self-declared persis- each patient. tently and seriously mentally ill in the public mental health sector. Main Outcome Measures: Primary outcomes in- cluded (1) symptoms measured by the 30-item Inven- Objective: To present clinical outcomes for patients with tory of Depressive Symptomatology–Clinician-Rated scale major depressive disorder (MDD) during 12-month al- (IDS-C30) and (2) function measured by the Mental Health gorithm-guided treatment (ALGO) compared with treat- Summary score of the Medical Outcomes Study 12-item ment as usual (TAU). Short-Form Health Survey (SF-12) obtained every 3 months. A secondary outcome was the 30-item Inven- Design: Effectiveness, intent-to-treat, prospective trial tory of Depressive Symptomatology–Self-Report scale comparing patient outcomes in clinics offering ALGO with (IDS-SR30). matched clinics offering TAU. Results: All patients improved during the study Setting: Four ALGO clinics, 6 TAU clinics, and 4 clin- (PϽ.001), but ALGO patients had significantly greater ics that offer TAU to patients with MDD but provide symptom reduction on both the IDS-C30 and IDS-SR30 ALGO for schizophrenia or bipolar disorder. compared with TAU. ALGO was also associated with sig- nificantly greater improvement in the SF-12 mental health Patients: Male and female outpatients with a clinical score (P=.046) than TAU. diagnosis of MDD (psychotic or nonpsychotic) were di- vided into ALGO and TAU groups. The ALGO group in- Conclusion: The ALGO intervention package during 1 cluded patients who required an antidepressant medi- year was superior to TAU for patients with MDD based cation change or were starting antidepressant therapy. on clinician-rated and self-reported symptoms and over- The TAU group initially met the same criteria, but be- all mental functioning. cause medication changes were made less frequently in the TAU group, patients were also recruited if their Brief Arch Gen Psychiatry. 2004;61:669-680 AJOR DEPRESSIVE DISOR- medical conditions (eg, cardiac heart dis- der (MDD) is a preva- ease,9-11 myocardial infarction,12-14 chronic lent, serious, debilitat- pain,15 diabetes,16 and asthma).17,18 The di- ing illness that affects rect monetary cost of treatment, com- 7% to 12% of men and bined with the indirect costs from lost pro- M20% to 25% of women in their lifetime.1,2 ductivity, are substantial19-21 and have been The course of MDD is typically chronic or estimated to be between $44 and $53 bil- recurrent.3 From 10% to 30% of patients lion per year.22-25 have major depressive episodes that last Despite the high prevalence of MDD loner than 2 years, and 20% to 30% have and the wide availability of effective treat- MDD superimposed on dysthymic disor- ments, undertreatment is common.8,26,27 der (double depression).4-6 Major depres- The aim of treatment is symptomatic re- sive disorder accounts for up to 60% of mission and functional recovery28 with psychiatric hospitalizations, and 8% to 15% continuation treatment to prevent re- of these patients commit suicide.7,8 Fur- lapse.29-32 Symptomatic improvement (re- Author affiliations are listed at thermore, depression worsens the mor- sponse) is distinguished from remission the end of this article. bidity and mortality of several general (ie, minimal or no symptoms), because re- (REPRINTED) ARCH GEN PSYCHIATRY/ VOL 61, JULY 2004 WWW.ARCHGENPSYCHIATRY.COM 669 ©2004 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 mission, in contrast to a response with residual symp- To increase the probability of appropriate algorithm imple- toms, is associated with better functioning33,34 and a bet- mentation, an extensive provider support system with ad- ter prognosis.35-39 ditional personnel funded by research moneys was used. Most randomized controlled efficacy trials typi- cally have engaged symptomatic volunteers with mini- mal concurrent psychiatric or general medical illnesses METHODS and minimal levels of treatment resistance. Conse- quently, findings from these studies may not generalize STUDY DESIGN to self-declared patients seen in clinical psychiatric prac- This study is an effectiveness, intent-to-treat, prospective trial that tice. Moreover, few studies define how to treat those with compares patient outcomes in clinics offering ALGO with matched an unsatisfactory clinical response to the initial treat- clinics offering TAU. Clinics were prematched based on mental ment or compare the benefits of different medication op- health and mental retardation authority and urban status. Evalu- tions given sequentially.40 These efficacy trials indicate able patients were postmatched based on symptoms (30-item In- that approximately 35% of participants achieve remis- ventory of Depressive Symptomatology–Clinician-Rated scale [IDS- sion in 6 to 8 weeks,29 although higher remission rates C30] and 30-item Inventory of Depressive Symptomatology–Self- 41,42 Report scale [IDS-SR30] scores) and length of illness (see Rush et are found in longer treatment trials. In the longer term, 66 10% to 30% of patients who do not respond or enter re- al for a detailed review of the rationale and design). This mul- tisite study evaluated the clinical benefits of ALGO provided mission quickly subsequently develop depressive re- in 4 clinics compared with 6 clinics that offered TAU lapses during the ensuing 4 to 6 months despite contin- (TAUnonALGO) and an additional 4 clinics that offered TAU to 43 ued pharmacotherapy. Because no one treatment is a patients with MDD but also provided ALGO for either schizo- panacea, clinicians often use a sequence of treatment steps phrenia or bipolar disorder (TAUinALGO). Physicians from all (either monotherapies or combinations) to increase the 14 clinics had access to the same medications. The TAUinALGO likelihood of response or remission. Recent efforts have clinics were intended to assess the effect of an “algorithm cul- aimed to define guidelines or algorithms for the appli- ture” associated with the implementation of any of these algo- cation of pharmacotherapeutic options for MDD.44-50 De- rithms on treatment practices. If no differences between TAU cision tree–based algorithms hold the promise of in- groups were found, they could be combined for comparison with creased consistency of treatment across practitioners, ALGO. Randomizing patients among physicians and clinics would require health care providers to ignore their algorithm training which in turn should lead to better clinical outcomes and and consultation interventions when treating control patients. To more efficient use of health care resources. Algorithm- randomize by health care providers within the same clinic risked guided treatment provides a basis for improving the qual- the “water cooler” effects (ie, algorithm physicians would talk to ity of treatment in both the public and private sectors. and affect the practice of TAU physicians). Furthermore, physi- To our knowledge, the present study is the first con- cians in a clinic typically cross-covered for each other, further lim- trolled trial to evaluate algorithm-based treatment of de- iting feasibility. pression in a public sector population treated by psy- The primary aim was to assess whether ALGO produced chiatrists. One open trial51 of the impact of algorithm- better clinical outcomes in terms of either an earlier onset and/or driven treatment on symptomatic outcomes in a a greater overall effect during a 1-year treatment period. We psychiatric (inpatient) population showed effectiveness hypothesized that ALGO would produce (1) a faster and more robust improvement in symptoms, (2) better functioning, and for an algorithm in the inpatient setting but lacked a con- (3) a lower side effect burden than TAU. trol condition. The study was conducted in accordance with interna- 51-65 A series of studies, including those conducted tional guidelines for good clinical practice and the Declara- by Katon et al,52,64 have evaluated clinical outcomes fol- tion of Helsinki and approved by the institutional review boards lowing the use of Agency for Health Care Policy and Re- at The University of Texas Southwestern Medical Center at Dal- search–based, guideline-driven treatments (see the “Com- las and The University of Texas at Austin. On study entry, symp- ment” section). Katon et al52 conducted a randomized toms, function, quality of life, side effect severity and burden, controlled trial of a guideline-driven intervention vs usual and health care service utilization and treatment costs were evalu- care in the treatment of patients with major (n=91) or ated at baseline and quarterly for at least a 12-month period minor depression (n=126) in a primary care setting. For for all available participants. major but not for minor depression, the intervention was ALGORITHM INTERVENTION associated with greater adherence to adequate medica- tion doses and more favorable ratings of antidepressant ALGO included 2 consensus-driven, medication management medications benefit, as well as higher ratings of the qual- algorithms (one each for psychotic and nonpsychotic forms of ity of care and better symptomatic outcomes. Most other MDD)48 and expert consultation (offered on biweekly telecon- trials conducted among primary care settings evaluated ference) and on-site clinical support from clinical coordina- broadly defined guideline-driven principles (eg, Did the tors and a patient and family education program provided by 67 patient complete the acute-phase trial or not? Was the the clinical coordinators.