Dr. Gaynes is associate professor of psychiatry in the Department of Psychiatry at the University of North Carolina School of Medicine in Chapel Hill.

Dr. Davis is associate professor of psychiatry in the Department of Psychiatry at the University of Alabama at Birmingham and the director of research at the Tuscaloosa Veteran’s Affairs Medical Center in Alabama.

Dr. Rush is professor and vice chair for research, and Dr. Trivedi is professor of psychiatry in the Department of Psychiatry at the University of Texas Southwestern Medical Center in Dallas.

Dr. Fava is professor of psychiatry and associate chief for clinical research of psychiatry at Massachusetts General Hospital in Boston.

Dr. Wisniewski is associate professor in the Epidemiology Data Center at the University of Pittsburgh in Pennsylvania.

Disclosure: Drs. Gaynes, Davis, and Wisniewski report no affiliation with or financial interest in any organization that might pose a conflict of interest. Dr. Rush has been a consultant to Bristol-Myers Squibb, Cyberonics, Eli Lilly, Forest, GlaxoSmithKline, and Organon; has been on the speaker’s bureaus of Bristol-Myers Squibb, Cyberonics, Eli Lilly, Forest, GlaxoSmithKline, and Wyeth; and has received grant/research support from the National Institute of Mental Health (NIMH), the Robert Wood Johnson Foundation, and the Stanley Medical Research Institute. Dr. Trivedi has served on the speaker’s bureaus of Bristol-Myers Squibb, Eli Lilly, Forest, Janssen, Organon, Pharmacia & UpJohn, Solvay, and Wyeth; and has received research grants from Abbott, Bayer, Bristol-Myers Squibb, Eli Lilly, Forest, GlaxoSmithKline, Janssen, Johnson & Johnson, NIMH, MeadJohnson, National Alliance for Research on Schizophrenia and Depression, Organon, Parke-Davis, Pfizer, Pharmacia & UpJohn, Solvay, and Wyeth. Dr. Fava has received research support and/or honoraria from Abbott, Aspect Medical Systems, Bayer, Bristol-Myers Squibb, Cephalon, Compellis, Cypress, Dov, Eli Lilly, Forest, GlaxoSmithkline, Janssen, Johnson & Johnson, Knoll, Lichtwer Pharma GmbH, Lorex, Lundbeck, Novartis, Organon, Pharmavite, Pfizer, Roche, Sanofi/Synthelabo, Solvay, Somerset, and Wyeth.

Funding/support: This project was funded in whole or in part with Federal funds from the NIMH, National Institutes of Health, grant no. N01MH90003 awarded to Dr. Rush; and the NIMH K23 Career Development Award, grant no. MH01951-03, awarded to Dr. Gaynes. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government.

Acknowledgment: The authors would like to thank all the STAR*D investigators for their help in making this large and complex multicenter study possible. They also acknowledge the administrative support of Veteran’s Affairs Research and Development at participating VA Medical Centers.

Please direct all correspondence to: Bradley N. Gaynes, MD, MPH, CB #7160, Department of Psychiatry, University of North Carolina School of Medicine, Chapel Hill, NC 27599-7160; Tel: 919-966-8028; Fax: 919-843-4370; E-mail: bgaynes@med.unc.edu.


Focus points

•Nearly all antidepressant clinical trials use strict exclusion criteria that severely limit their applicability to the majority of depressed patients seen by physicians.

•Even within clinical trials using these less representative groups, less than half of patients with nonpsychotic major depressive disorder will reach remission following an 8–12 week initial antidepressant treatment attempt.

•There is an extremely limited evidence base for how to treat depressive illness after the first attempt has failed.

•The Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study, employing some unique methodologies and allowing four possible sequential levels of treatment, was designed to address some of the deficiencies in the evidence base.

•Enrollment is complete, and initial results for patients failing a first antidepressant trial are expected to be available in June 2005.



Major depressive disorder is expected to become the second leading cause of disability worldwide by the year 2020. A large proportion of this disability can be attributed to the large number of patients (>70%) who do not achieve sustained remission following initial treatment. Presently, the evidence available to guide management for treatment-resistant depression (TRD) is quite limited. The Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study has been designed to evaluate treatment options to improve clinical outcomes for depressed outpatients who do not achieve a satisfactory outcome after one or more treatment attempts. STAR*D helps to define the best “next-step” treatments for patients with TRD, and it will provide descriptive information on the longer-term benefits of successful treatments, as well as the side-effect burden and economic costs associated with different treatments. The study has enrolled over 4,000 patients recruited from primary care and psychiatric settings in the public and private sectors. Preliminary baseline findings are currently being reported. Final results regarding preferred treatment options are expected by May 2006.


Major depressive disorder (MDD) is expected to become the second most disabling medical disease by the year 2020,1 largely due to its high prevalence, the recurrent and chronic nature of the condition, and life-shortening impact caused by both suicide and increased mortality of concomitant general medical conditions. Yet, available rigorous scientific evidence provides only limited guidance on how to successfully manage patients with MDDs, most of whom do not remit after an initial treatment attempt. The majority of randomized controlled trials (RCTs) exclude patients with most concurrent general medical and psychiatric conditions. Even for those relatively uncomplicated cases, only approximately 50% of outpatients with nonpsychotic MDD initially treated with a single evidence-based medication or psychotherapy will experience a clinically significant symptom improvement during the 8–12 weeks of acute phase of treatment.2-6 Even more disturbing, only 20% to 35% of patients will reach symptom remission,5 the aim of treatment for MDD.4,7 Furthermore, the proportion of those responding who later relapse in the continuation phase is also substantial; after 1 year, 37% of successfully treated primary care patients experience recurrence of a major depressive episode,8 a proportion quite similar to those in specialty mental health settings.9,10

For depressed patients, the lack of satisfactory sustained clinical response or remission has severe consequences. While mood disorders as a whole are markedly disabling,11 chronic depressive illnesses in particular appear to be associated with substantial and persistent impairment in physical, social, and general health functioning.12 Persistent and severe depressive symptoms increase the risk of chronic depressive illness,13 disability,14 suicide,15 and the development of general medical conditions.16 Furthermore, initial nonresponse appears to be associated with a subsequently harder-to-treat course of depression. The likelihood of a placebo response to treatment or spontaneous remission substantially decreases after an initial antidepressant failure.17 Progressively more complicated or potent interventions are often required as the patient’s history of antidepressant failure(s) grows.18 Additionally, failed treatments can decrease the likelihood of a patient pursuing subsequent treatment.

The substantial public health significance of nonresponding, persistent depressive illness across all settings is reflected by the considerable utilization of healthcare visits and tremendous monetary costs ($43 billion dollars annually, with $17 billion of that resulting from lost work days).19 How to successfully manage patients with treatment-resistant depression (TRD), that is, those who do not achieve a sustained remission with one or more adequate trials of treatment, is a major public health challenge that spurred the National Institutes of Mental Health to request proposals
to evaluate various alternative treatments for TRD.20

RCTs that compare two or more management strategies for depressions that have not remitted after an initial antidepressant treatment are rare. While nonrandomized studies have begun to address this point (eg, Thase, and colleagues21), no controlled studies have compared at least two potentially active augmenting treatments or at least two potentially active switch treatments. For example, whether switching within class (eg, from one serotonin reuptake inhibitor (SRI) to another) or switching between classes (eg, from an SRI to a norepinephrine/dopamine reuptake inhibitor) has not been studied in a randomized controlled fashion. Consequently, clinical practice guidelines that suggest different second or third treatment steps rely largely on clinical consensus or on open uncontrolled trials rather than on a clear evidence base.17,22,23 Indeed, most practice guidelines2,4,7 and treatment algorithms22 do not recommend specific “next step” treatments or specific treatment sequences.

In addition to the minimal evidence by which to recommend the next treatment approach after initial failure, the evidence focusing on the treatment of patients who have failed two or more trials is even more limited, and is overall insufficient to guide clinical practice. Where psychotherapy fits in is also unknown, though a recent report (AF Schatzberg, et al. unpublished data, 2004) suggests that cognitive-behavioral analytic system psychotherapy results in a 50% response rate, even after an initial failure to achieve a response. Finally, whether the available clinical trial evidence actually applies to representative patients seen by most psychiatrists and primary care physicians, given the strict inclusion/exclusion criteria typically used in most efficacy trials, remains unknown.

The Sequenced Treatment Alternatives to Relieve Depression Study (STAR*D), which is the largest prospective clinical trial of major depressive disorder ever conducted, was designed to address some of these deficiencies in the evidence base.24,25 This article briefly describes the aims, design, and methods of STAR*D (www.star-d.org).

Aims and Designs of STAR*D

The design of the STAR*D was guided by a number of key considerations. First, the aim of treatment for depression is full remission of symptoms—not merely achieving a response with residual symptoms, which are associated with a substantially higher risk of recurrent depressive illness.26 Second, less than half of depressions will remit when treated with 8–12 weeks of antidepressant medication or psychotherapy. Third, a reported history of treatment response is difficult to confirm because of questions about the accuracy of self-report information and the difficulty confirming whether prior trials were adequate in dose and duration. Fourth, many treatment options are available, but it is not clear which course to pursue after a previous treatment has failed. Finally, results must be generalizable to the majority of patients seen in actual practice, ie, patients seen in psychiatric and primary care settings with medical and psychiatric comorbidities.

Guided by the above rationale, the overall aim of STAR*D is to develop and evaluate treatment strategies to improve clinical outcomes for patients with TRD, with a cohort of treatment-resistant patients created prospectively from a pool of those with a current MDD episode. STAR*D has the following objectives:

(1) To determine prospectively which of a number of treatments are the most effective “next-step” treatments for patients with MDD which does not remit following initial and, if needed, subsequent treatments.

(2) To compare relative efficacy and patients’ acceptance of different treatment strategies (eg, augmentation versus switching).

(3) To determine the longer-term benefits of successful strategies.

(4) To compare side-effect burden and economic costs of different treatments.

(5) To further develop and evaluate clinical outcome measures for use in real world practice.


Participating clinical sites include non-academic and non-efficacy research venues, ie, settings that are representative of primary and specialty care practices.  Clinical study sites were selected based on the availability of a large number of depressed outpatients and minority patients, clinicians, and administrative support able to assist the project. These sites largely include practices not typically engaged in efficacy RCTs of depressive illness. Participating clinicians are those who routinely care for patients. Patients enrolled have nonpsychotic MDD of at least a moderate severity who are representative of most MDD outpatients in that they are allowed to have concurrent psychiatric and general medical conditions as long as these conditions do not contraindicate the possible medications required by protocol. To ensure a sample representative of those who present to clinic settings for treatment, patients must either already be in treatment or about to begin treatment in the participating clinic; no advertisements are allowed for patient recruitment. To ensure sufficient power to address our primary objectives, STAR*D goal for recruitment was 4,000 patients (see Rush and colleagues25 for description of sample size calculation). At study initiation, it was expected that approximately half (n=2,000) of these patients would experience an unsatisfactory response following an adequate trial in the first level of treatment, ie, they would meet the criteria for stage I treatment resistance.18

Eligibility criteria for the study are broadly defined and inclusive. Patients must have nonpsychotic MDD identified by the clinician and confirmed by a symptom checklist of the nine depressive symptoms per the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition,27 and for which antidepressant treatment is recommended. Patients must be 18–75 years of age; score at least 14 on the 17-item version of the Hamilton Rating Scale for Depression [HAM-D17]28; and not have a diagnosis of bipolar disorder or obsessive-compulsive disorder (which would require a different treatment strategy), or an eating disorder or seizure disorder (each of which would preclude bupropion treatment in the second level of treatment).

Organizational Structure

The STAR*D infrastructure includes the National Coordinating Center in Dallas (NCC), the Data Coordinating Center in Pittsburgh (DCC), and 14 regional centers (RCs), each overseeing the implementation at 2–4 clinical sites providing primary care (PC) or specialty care in either the public or private sectors. At least three clinicians participate from each site. Thirteen of the 14 Regional Centers have at least one PC site, and six have two PC sites. Nearly half of the clinical sites (19 of 41) are PC settings. Clinical sites were identified based on multiple factors including the availability of depressed outpatients, clinicians, administrative support, and minority populations. The institutional review boards at the NCC, the DCC, and each RC approved the study protocol.


Specific treatments available at each level were selected to balance safety, common practice, and pharmacologic reasoning, while empirically testing options that could be implemented in primary care (levels 1 and 2) or specialty care (levels 1–4). Citalopram (CIT) was selected to represent the selective serotonin reuptake inhibitor (SSRI) class, based on its minimal drug-drug interactions, minimal discontinuation-related symptoms, and relatively shorter half-life. The latter benefit allows one to switch to a different medication and more confidently study the effects of the new medication rather than observing the overlapping effects of two medications in the first weeks of the next level.

All eligible participants who provided written informed consent in accordance with the Declaration of Helsinki are enrolled into level 1 of STAR*D for a 12-week treatment trial with CIT. Participants who attain remission (as defined by a score of ≤5 on the 16-item Quick Inventory of Depressive Symptomatology-Clinician Rating (QIDS-CR-16)29,30 by the end of level 1 then enter a 12-month, naturalistic follow-up phase (Figure).31

Those patients who do not obtain remission with CIT may enter a series of open-label RCTs (with masked outcome evaluations), each of which can last up to 12 weeks (Table).25 Clinicians are guided by evidence-based algorithms based in large part on experience from the Texas Medication Algorithm Project32 that depend on depressive symptom severity (as measured by the QIDS-CR-16), side-effect ratings (as measured by the Frequency and Intensity of Side Effects [FISER] and the Global Rating of Side Effect Burden [GRSEB],26 and length of treatment. As a function of these three variables, the algorithm recommends that physicians increase antidepressant medication dose, continue current dose, decrease dose, or move on to the next level. The first RCT, level 2, compares the effectiveness of seven separate treatments that involve four general strategies. Patients may switch to a different antidepressant (three options: venlafaxine-XR [VEN-XR], sertraline [SER], or bupropion-SR [BUP-SR]); switch to cognitive therapy (CT); augment the CIT with a second medication (two options: BUP-SR or buspirone [BUS]), or augment the CIT with CT. This design feature allows STAR*D to both assess the effectiveness of different treatment strategies (switch versus augment) and to compare the effectiveness of individual treatment alternatives within and across these two strategies. For example, the comparative effectiveness of switching within antidepressant class (ie, to SER) can be compared to switching to a different class (ie, to BUP-SR). Those patients who achieve remission in level 2 may enter the 12-month naturalistic follow-up phase.

Participants who do not achieve remission with CT, whether CT is part of a switch or augment strategy, are eligible to enter level 2A, an RCT comparing the effectiveness of two antidepressant options (BUP-SR or VEN-XR). Level 2A also ensures that patients entering level 3 have been treatment resistant to two separate antidepressant medication trials. As with earlier levels, those who obtain remission enter the 12-month follow-up phase, while those not in remission are eligible to enter level 3.

Level 3 is an RCT comparing the effectiveness of two switch options (mirtazapine [MIRT] or nortriptyline [NOR]) and two augment options (lithium33 or thyroid hormone [THY]). Those obtaining remission enter into the 12-month follow-up phase, while those without such a response are eligible to enter level 4, an RCT comparing the effectiveness of the monoamine oxidase inhibitor tranylcypromine and the combination of VEN-XR and MIRT. Those obtaining remission enter into the 12-month follow-up phase, while those with an unsatisfactory response leave the study and are managed as clinically indicated.


Random treatment assignment occurs in all treatment levels after level 1. A unique feature of the protocol is its attempt to consider patient preference within the constraints of an RCT by employing an equipoise stratified randomized design34 at levels 2 and 3. This approach attempts to strike a balance between two standard designs in multiple treatment option studies, the completely randomized design and the “clinician’s choice” design. Completely randomized designs, in which patients are assigned at random to an available option, require both patient and clinician to find each option acceptable. With clinician’s choice, patients are randomized to a small number of broad strategies with the specific choice being left up to the clinician. This latter design allows some degree of physician and patient preference, but forfeits the ability to make randomization-based comparisons of particular options. The equipoise stratified design attempts to merge the benefits while avoiding the disadvantages of these two designs.34 Where there are at least two possible strategies (eg, switch to a different treatment or augment with a different treatment), patients may choose to exclude certain strategies as long as there are at least two possible options to which one might be randomized. For example, at level 2, a patient may select only the switch medication strategy (three options) or only the augment medication option (two options), but they cannot select only the switch to CT option (one option). They could, however, select both the augment with CT and switch to CT strategies (a total of two options), or they may choose both the augment and switch strategies, but may exclude CT within either or both treatment strategy groups. For a description of the rationale behind the selection of the specific treatment strategies, please see Fava and colleagues.25

Implementation of the Protocol

Clinical research Coordinators (CRCs) located at the clinical sites play a critical role in ensuring the proper implementation of the protocol. The CRCs, all of whom have at least a Bachelor’s level of training or 2 years of healthcare-related education, are trained and certified in executing the treatment protocol and in data collection procedures (screening, application of eligibility criteria, and collection of other study-related data items). The CRCs ensure that treating clinicians properly implement the treatment, and they make recommendations on dose adjustment based on depressive severity and medication side effects based on a clinician procedure manual. These guidelines are not mandatory, but clinicians are contacted should their practice deviate meaningfully from these recommendations. These guidelines are designed to ensure that patients receive an adequate treatment trial within each level. CRCs collaborate closely with both patient and clinicians, administer some of the clinician assessments, make sure all self-rated instruments are completed, and work as a coordinator between the clinical sites, the RCs, the DCC, and the NCC.

Physicians providing pharmacologic treatments and psychotherapists providing CT were trained in the protocol and treatment procedures according to a Clinical Procedures Manual. CRCs ensure protocol implementation and complete a Clinical Record Form at each visit.

CT psychotherapists were additionally trained and certified by expert cognitive therapists from the University of Pittsburgh School of Medicine to ensure mastery of CT skills.35 in press Once certified, therapists had ongoing review of study sessions. Should therapists experience problems adhering to CT techniques, additional supervision was provided; those unable to adhere to CT methods were not allowed to continue to see study patients.

Data Collection

Clinical Data: Baseline

At screening, once consent is obtained, CRCs collect information on clinical and demographic features, eligibility (inclusion/exclusion criteria), depressive severity measures (including the HAM-D1729 and the QIDS-CR1630,31); and the Cumulative Illness Rating Scale to assess general medical comorbidity.36 Patients complete self-report forms to assess depressive severity (QIDS-SR-1630,31) and to screen for concurrent psychiatric diagnoses (using the Psychiatric Diagnostic Screening Questionnaire37).

Clinical Data: Course of Treatment

At each clinical treatment visit within each level (2 weeks, 4 weeks, 6 weeks, 9 weeks, and 12 weeks after level entry), CRCs administer the QIDS-C-16 while patients provide the QIDS-SR-16 and medication side effects (the Patient-Rated Inventory of Side Effects [PRISE], the FISER, and GRSEB) (form available at www.star-d.org). This clinical information is used by the clinicians to implement the treatment protocol (eg, they dictate when to raise or hold the dose). Clinicians fill out a Clinical Global Impression of Improvement38 and complete a Clinical Report Form at the conclusion of each visit.

Research Outcomes

The primary outcome measure is the HAM-D17. Beginning at entry into the study, data are collected by telephone interviews with trained Research Outcomes Assessors (ROAs) who are blind to CRC-collected information, treatment, and treatment settings. At entry and exit from each treatment level and at months 3, 6, 9, and 12 of follow-up, ROAs administer a HAM-D17 and a 30-item Inventory of Depressive Symptomatology-Clinician-rated.39 In addition, a telephone-based Interactive Voice Response system40 collects information on a number of domains, including depressive severity (using the 16-item Quick Inventory of Depressive Symptomatology-Self-Report30), function (including the 12-item Short-form Health Survey41 and the Work and Social Adjustment Scale42), and utilization and costs (using a 15-item Utilization and Cost Patient Questionnaire).


The study should be able to address the following questions:

(1) What is the rate of remission (defined as HAM-D17 ≤7 by the blinded ROA) with adequate treatment with a representative SSRI?

(2) Which of several possible treatments following an unsuccessful initial trial with an SSRI is best in producing remission?

(3) Which of several treatments are to be recommended after a second medication attempt, or even a third medication trial, does not produce remission?

Secondary analyses will identify whether specific baseline features are associated with remission or intolerance.

What to Expect From Analyses of the STAR*D Study

Enrollment is now complete with 4,041 enrollees. Initial baseline results characterizing the patients’ demographic and clinical features of participants are currently being reported. Initial trial results with level 2 treatments are expected by June 2005, while final results regarding preferred treatment options are expected by May 2006.

Results of level 1 will provide the most accurate measure of times to and rates of remission in outpatients managed appropriately in representative primary and mental healthcare settings who were initially treated with an adequate trial of an SSRI. Subsequent results (levels 2, 3, and 4) will provide the first available RCT evidence of which treatments to consider after an initial SSRI does not successfully produce remission. Other key results from this study will include a comparison of relative efficacy and patients’ acceptance of different treatment strategies; a determination of longer-term benefits of successful strategies; a comparison of side-effect burden and economic costs of different treatments; and an evaluation of clinical outcome measures in real-world practice. These results will substantially enhance the evidence base guiding the clinical management of TRD.

Future Directions

Additional key management questions not addressed by the study design and protocol remain to be answered. For example, the role of CT as an initial treatment choice (rather than being a possibility if a first antidepressant attempt does not work) remains unclear. In addition, the role of newer strategies for TRD, such as augmentation with an atypical antipsychotic, as well as some strategies already being used, such as electroconvulsive therapy, were not tested in the current STAR*D protocol. Furthermore, how best to address TRD in the continuation and maintenance phases of treatment (after the acute phase of treatment has ended) remains to be delineated. The rationale that guided the design of STAR*D—the focus on remission, the use of evidence-based medication and psychotherapy treatments, the sequential management approach of what to do next if an earlier treatment does not work, and the emphasis on representative patients and clinicians to improve the generalizability of results—can be successfully used to address these and other clinically relevant research questions about TRD in the future. PP


1. Murray CJ, Lopez AD. Alternative projections of mortality and disability by cause 1990-2020: Global Burden of Disease Study. Lancet. 1997;349(9064):1498-1504.

2. Depression guideline panel. Depression in primary care: Diagnosis and Detection. Vol 1. AHCPR publication No. 93-0550. Rockville, MD: US Department of Health and Human Services, Public Health Service, Agency for Health Care Policy and Research; 1993.

3. Frank E, Karp J, Rush A. Efficacy of treatments for major depression. Psychopharmacol Bull. 1993;29:457-75.

4. Depression guideline panel. Depression in Primary Care: Treatment of Major Depression. Vol 2. AHCPR publication No. 93-0550. Rockville, MD: US Department of Health and Human Services, Public Health Service, Agency for Health Care Policy and Research; 1993.

5. Fava M, Davidson KG. Definition and epidemiology of treatment-resistant depression. Psychiatr Clin North Am. 1996;19(2):179-200.

6. Jarrett RB, Rush A. Short-term psychotherapy of depressive disorders: Current status and future directions. Psychiatry: Interpers Biol Process. 1994;57:115-132.

7. American Psychiatric Association. Practice guideline for the treatment of patients with major depression (revision). Am J Psychiatry. 2000;157(Suppl4):1-45.

8. Lin EH, Katon WJ, VonKorff M, et al. Relapse of depression in primary care. Rate and clinical predictors [see comments]. Arch Fam Med. 1998;7(5):443-449.

9. Kanai T, Takeuchi H, Furukawa TA, et al. Time to recurrence after recovery from major depressive episodes and its predictors. Psychol Med. Jul 2003;33(5):839-845.

10. Hart AB, Craighead WE, Craighead LW. Predicting recurrence of major depressive disorder in young adults: a prospective study. J Abnorm Psychol. 2001;110(4):633-643.

11. Wells K, Stewart A, Hays R, et al. The functioning and well-being of depressed patients: results from the Medical Outcomes Study. JAMA. 1989;262:914-919.

12. Hays RD, Wells KB, Sherbourne CD, et al. Functioning and well-being outcomes of patients with depression compared with chronic general medical illnesses. Arch Gen Psychiatry. 1995;52(1):11-19.

13. NIMH/NIH consensus development conference statement: mood disorders. Am J Psychiatry. 1985;142:469-476.

14. Broadhead WE, Blazer DG, George LK, et al. Depression, disability days, and days lost from work in a prospective epidemiologic survey. JAMA. 1990;264(19):2524-2528.

15. Angst J, Angst F, Stassen HH. Suicide risk in patients with major depressive disorder. J Clin Psychiatry. 1999;60(Suppl 2):57-62; discussion 75-56, 113-116.

16. Krishnan KRR, Delong M, Kraemer H, et al. Comorbidity of depression with other medical diseases in the elderly. Biol Psychiatry. 2002;52(6):559-588.

17. Thase M, Rush A. Treatment resistant depression. In: Bloom F, Kupfer D, eds. Psychopharmacology: the Fourth Generation of Progress. New York, NY: Raven Press; 1995:1081-1097.

18. Thase ME, Rush AJ. When at first you don’t succeed: sequential strategies for antidepressant nonresponders. J Clin Psychiatry. 1997;58(Suppl 13):23-29.

19. Greenberg PE, Stiglin LE, Finkelstein SN, Berndt ER. The economic burden of depression in 1990 [see comments]. J Clin Psychiatry. 1993;54(11):405-418.

20. McCombs J, Nichol M. Pharmacy-enforced outpatient drug treatment protocols: A case study of Medi-Cal restrictions for cefaclor. Ann Pharmacother. 1993;27:155-161.

21. Thase ME, Rush AJ, Howland RH, et al. Double-blind switch study of imipramine or sertraline treatment of antidepressant-resistant chronic depression.[comment]. Arch Gen Psychiatry. 2002;59(3):233-239.

22. Crismon M, Trivedi M, Pigott T, et al. Texas Consensus Conference Panel: The Texas Medication Algorithm Project. Report of the Texas Consensus Conference Panel on medication treatment of major depressive disorder. J Clin Psychiatry. 1999;60:142-156.

23. Fava M, Uebelacker LA, Alpert JE, et al. Major depressive subtypes and treatment response. Biol Psychiatry. 1997;42(7):568-576.

24. Fava M, Rush A, Trivedi M, et al. Background and rationale for the sequenced treatment alternatives to relieve depression (STAR*D) study. Psychiatr Clin North Am. 2003;26(2):457-494.

25. Rush A, Fava M, Wisniewski S, et al. Sequenced Treatment Alternatives to Relieve Depression (STAR*D): Rationale and Design. Control Clin Trials. 2004;25(1):119-142.

26. Judd LL, Akiskal HS, Maser JD, et al. Major depressive disorder: a prospective study of residual subthreshold depressive symptoms as predictor of rapid relapse. J Affect Disord. 1998;50(2-3):97-108.

27. Diagnostic and Statistical Manual of Mental Disorders. 4th ed rev. Washington, DC: American Psychiatric Association; 2000.

28. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiat. 1961;23:56-61.

29. Rush AJ, Trivedi MH, Ibrahim HM, et al. The 16-Item quick inventory of depressive symptomatology (QIDS), clinician rating (QIDS-C), and self-report (QIDS-SR): a psychometric evaluation in patients with chronic major depression. Biol Psychiatry. 2003;54(5):573-583.

30. Trivedi MH, Rush AJ, Ibrahim HM, et al. The Inventory of Depressive Symptomatology, Clinician Rating (IDS-C) and Self-Report (IDS-SR), and the Quick Inventory of Depressive Symptomatology, Clinician Rating (QIDS-C) and Self-Report (QIDS-SR) in public sector patients with mood disorders: a psychometric evaluation. Psychol Med. 2004;34(1):73-82.

31. STAR*D Internet site. Overview slides. Available at: http://www.edc.gsph.pitt.edu/stard. Accessed: November 18, 2004.

32. Trivedi MH, Rush AJ, Crismon ML, et al. Clinical Results for Patients With Major Depressive Disorder in the Texas Medication Algorithm Project. Arch Gen Psychiatry. 2004;61(7):669-680.

33. Holcombe RF, Jacobson J,, Li A, et al. Inclusion of black Americans in oncology clinical trials: the Louisiana State University Medical Center experience. Am J Clin Oncol. 1999;22(1):18-21.

34. Lavori P, Rush A, Wisniewski S, et al. Strengthening clinical effectiveness trials: equipoise-stratified randomization. Biol Psychiatry. 2001;50(10):792-801.

35. Friedman EF, Thase ME, Biggs MM, et al. The implementation of cognitive therapy in STAR*D. Cog Ther Res. 2005. In press.

36. Linn BS, Linn MW, Gurel L. Cumulative illness rating scale. J Am Geriatr Soc. 1968;16(5):622-626.

37. Zimmerman M, Mattia J. The reliability and validity of a screening questionnaire for 13 DSM-IV Axis I disorders (the Psychiatric Diagnostic Screening Questionnaire) in psychiatric outpatients. J Clin Psychiatry. 1999;60:677-683.

38. Guy W. ECDEU Assessment Manual for Psychopharmacology- Revised (US Department of Health, Education, and Welfare Publ No ADM 76-338), 1976; Rockville, MD.

39. Rush AJ, Gullion CM, Basco MR, et al. The Inventory of Depressive Symptomatology (IDS): psychometric properties. Psychol Med. 1996;26(3):477-486.

40. Kobak KA, Greist JH, Jefferson JW, et al. Computerized assessment of depression and anxiety over the telephone using interactive voice response. MD Comput. 1999;16(3):64-68.

41. Ware J, Kosinski M, Keller S. A 12-item Short-form Health Survey: construction of scales and preliminary tests of reliability and validity. Med Care. 1996;32:220-233.

42. Mundt JC, Marks IM, Shear MK, et al. The Work and Social Adjustment Scale: a simple measure of impairment in functioning. Br J Psychiatry. 2002;180:461-464.