Here you will find a detailed algorithm for treating depression in adults. It includes major depression, dysthymic depression, delusional depression and bipolar illness in a depressed phase.
The most recent text versions (abbreviated compared to this Internet version) are available from the following references: For non-bipolar depression, see Osser DN and Patterson RD. Algorithms for the pharmacotherapy of depression, parts one and two. Directions in Psychiatry 1998;18:303-334. For bipolar depression, see Dantzer A, Osser DN. Algorithms for the pharmacotherapy of acute depression in patients with bipolar disorder. Psychiatric Annals 1999;29(5):285-300.
Historical Note: Over the past decade, the authors have been producing algorithms for dealing with a variety of clinical problems in psychopharmacology. The first algorithm for the use of medication in depression was published in 1988,1 and it was updated in 1993.2 Then, in 1996, the first computerized version was prepared in HTML as web pages on the Internet.3 These algorithms were a synthesis of relevant published literature, practice guidelines, and expert consensus opinion, blended with additional informally-derived expert consensus and clinical experience as necessary to address a wide range of common clinical problems for which evidence-based data was unavailable. Revision and expansion of the Internet algorithm occurred in 1997 when it was presented in a software media session and in a symposium at the annual meeting of the American Psychiatric Association in San Diego. It was updated again in 1998 and 1999 to coincide with publication of the text versions noted above. Further updates have been made up to the date on the title page.
A Few General Remarks: Although the focus of the algorithm is on pharmacotherapy, this does not imply that pharmacotherapy is recommended for all patients with depression. Psychotherapy is effective as the sole treatment for many patients, and some patient require both. A treatment algorithm that includes all psychosocial interventions would be an even more difficult undertaking given the relative paucity of evidence available.(4) This algorithm focuses on the choice of pharmacotherapy when a decision has been made to consider medication treatment.
The algorithm is formatted in the computer in the style of a clinical psychopharmacology consultation. The computer functions, in effect, as a virtual consultant as it asks the clinician a series of questions about a hypothetical patient. Depending on the answers, various suggestions for further workup of that patient or recommendations for treatments to consider are made. The authors have been seeking to capture the process of how the mind of an expert psychopharmacologist evaluates and decides how to select the best medication for a particular patient (i.e. - the "art" of prescribing).
The entrance point to the algorithm is when the clinician or clinical team has screened the patient for medical causes of depression for which specific treatments are appropriate (e.g. - endocrine and sleep disorders, drug side effects or interactions, or depressive problems that are part of ongoing, active substance abuse/dependence, or recent withdrawal from substances). The recommendations are thought to apply best to patients who have been abstinent from drugs or alcohol for at least one week.(5) Although one might consider pharmacotherapy for a patient still actively abusing substances, an algorithm for that situation is not proposed in this program.
The Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, published by the American Psychiatric Association (DSM-4),(6) contains a comprehensive list of possible medical causes of depression which is useful to review. Active awareness of the patient's problem list on Axis III will also be useful in selecting the safest approach from among alternative treatments recommended in the algorithm. Dosaging strategies may also differ from the generic strategy recommended in some of these situations.
The patient's current drug regimen should also be noted carefully. Again, the selection of treatment and dosing strategy may be affected by the knowledge of drugs that the patient is already taking.
Diagnosis of Depression Subtype:
The use of this algorithm is predicated upon the clinician making a criteria-based DSM-4 diagnosis. Since this is an evidence-based as well as consensus-informed algorithm, it is important to note that almost all the published research studies in modern psychopharmacology (over the past 15 years) utilize DSM criteria as the basis for entering patients into the studies. Hence, if the clinician uses more vague or alternative criteria, this may reduce the validity of the recommendations to be found in this algorithm. A central premise of the approach to treatment outlined in this program is the primary importance of treating diagnoses, not symptoms. The vast majority of the evidence basis for psychopharmacology practice is comprised of information about treating patients with particular diagnoses, not symptoms independent of diagnosis. Many treatment errors arise from treating symptoms without identifying their cause. It appears to be one of the primary causes of polypharmacy and resultant toxicity from drug interactions and compounded side effects.
In some patient populations it may be more difficult to use the DSM criteria for depressive disorders. For example, geriatric patients with medical illnesses may have psychomotor retardation and anergia secondary to non-depressive causes. The DSM criteria may overdiagnose depression in such patients. However, it is considered desirable to err on the side of overdiagnosis considering the importance of identifying and treating depression in this population.7 Use of the DSM may lead to underdiagnosis in other situations: for example, it has been proposed that some patients have "masked depression" presenting as a somatoform disorder. There is little evidential literature suggesting how to treat this kind of depression, but like "subsyndromal depressive symptoms"8 (for which there also have been no placebo controlled studies), it may be assumed that masked depression may be associated with significant impairment. These patients will be diagnosed "Depression Not Otherwise Specified" if they have no other primary Axis I problem in the DSM-4, and the algorithm will propose treatment sequences for these patients.
In summary, although psychiatric diagnosis is a longitudinal process in which repeated clinical observations of symptoms, course, and response to treatment occur over time, a reasonable starting point for the algorithm is a criteria-based assessment adjusted as needed by clinical judgement. With each sequential trial of treatment that fails to produce the expected result, however, a thorough review of the diagnosis should be repeated. For example, a presumed non-psychotic depression may be psychotic, or a unipolar depression may be bipolar. Patients are often reluctant to reveal some of their symptoms or even flatly deny them on direct questioning for a long time. The medical differential should also be reconsidered at each step.
Often despite a thorough evaluation the clinician remains unsure about the exact DSM diagnosis or subtype diagnosis. If more than one diagnosis is under consideration, the algorithm for each option should be explored and the recommendation options compared. If there are overlapping recommendations, those might be the treatments-of-choice that cover the diagnostic ambiguity.
It should be noted that the patient may have comorbidity on Axis I or II of the DSM-4. Unless that comorbidity is considered the primary problem, in which case first-line treatment should be directed toward that disorder, these algorithms propose that initial pharmacotherapy (to be described) is generally appropriate regardless of secondary comorbidity, with some exceptions that will be described. Initial pharmacotherapy is defined as at least two adequate trials (as specified in the algorithm) of pharmacotherapy, or at least two reasonable attempts to provide adequate trials. Following this, the patient will meet criteria for a treatment-resistant depression if response is still unsatisfactory. Unsatisfactory is defined as when there are still residual subsyndromal depressive symptoms. Before proceeding further, Axis I and II comorbidity (that up to now appeared to be of secondary importance) will be taken into account and then further recommendations will derive from this more comprehensive assessment.
Finally, one more comment on the structure of the recommendations: under each diagnosis, there will be recommendations after consideration of a variety of contingencies. A first-line recommendation is indicated whenever a rational basis for such a preference can be formulated (current as of this writing but subject to change at any time). Frequently, other choices of almost comparable merit exist and these are provided, and the advantages and disadvantages of each are discussed. Particular circumstances of the individual patient may make one of the other options clearly first-line for that patient. However, the patient's preference is also important, and as part of the informed consent process the physician is expected to present the significant alternatives to the patient. Note, however, that the emphasis placed on the alternatives by the physician may have considerable (and sometimes undue) influence on the patient's thinking. As one example, consider sexual side effects of antidepressants. In order for patients to make an informed decision on whether to begin an SSRI, the discussion should describe sexual side effects and mention that there are some antidepressants that cause them less frequently. If the physician does not mention the alternatives, patients will not have the opportunity to to make a choice based on their own preferences and knowledge of the relevant facts. In the development of these algorithms, the authors have continuously tried to make explicit the relevant facts regarding the preferences indicated, so that users of these algorithms can make objective evaluations of what would be best for them.
References for Introduction:
(1)Osser DN. Treatment resistant problems: depression and dysphoria. In: Tupin JP, Shader RI, eds. Clinical handbook of psychopharmacology. 2 ed. New York: Jason Aronson; 1988:269-328.
(2)Osser DN. A systematic approach to the classification and pharmacotherapy of nonpsychotic major depression and dysthymia. Journal of Clinical Psychopharmacology. 1993;13:133-144.
(3)Osser DN. Algorithm for the pharmacotherapy of depression. . 1.0 ed. Lexington, Massachusetts: Mental Health Connections, Inc.; 1996.
(4)Persons JB. Indications for psychotherapy in the treatment of depression. Psychiatric Annals. 1998;28:80-83.
(5) Mason BJ, Kocsis JH, Ritvo EC, Cutler RB. A double-blind, placebo-controlled trial of desipramine for primary alcohol dependency stratified in the presence or absence of major depression. Journal of the American Medical Association. 1996;275:761-767.
(6)Diagnostic and statistical manual of mental disorders,. . 4 ed. Washington, D.C.: American Psychiatric Association; 1994.
(7)Chochinov HM, Wilson KG, Enns M, Lander S. Prevalence of depression in the terminally ill: effects of diagnostic criteria and symptom threshold judgments. American Journal of Psychiatry. 1994;151:537-540.
(8)Judd LL. The role and clinical significance of subsyndromal depressive symptoms in unipolar major depressive disorder. Journal of Affective Disorders. 1997;45:5-18.
Many of the questions and recommendations have a clinical confidence rating icon in their windows. These give an indication of the scientific strength of data supporting the use of the recommendations and distinctions made by the questions. Confidence ratings have three levels:
Your feedback and your clinical experience are most valuable to help us refine this algorithm.
The "Feedback" button takes you to our automatic system for sending observations to us. Recent feedback of interest to users is posted regularly in the "Users Comments" section.
Our US Mail address is:
Dr. David N. Osser
% Mental Health Connections, Inc.
21 Blossom St.
Lexington, MA 02173
E-mail: mhc@mhc.com
This program is designed to supplement and enhance the expertise of the psychiatrist. However, management of any individual patient is the responsibility of the prescribing physician. Recommendations from this program should only be used in the context of your overall knowledge of the patient.
End of Introduction.
Picture Credit: The "Blue Lady" painting is owned by NARSAD and used with permission.
12/31/03 NOTE: Version 4 of this program contains more up-to-date information. It is available at www.mhc.com/Algorithms.
Here we list updates and additions to this program. They are listed by date, with the most recent first. When done, go back to title page menu.
January 10, 2002. As the FDA's new black box warning on nefazodone (Serzone), issued last month, is being evaluated by experts and risk managers, a consensus seems to be building that this antidepressant should be reserved for patients who have not tolerated or not responded to other antidepressants. This is unfortunate, in view of its low incidence of sexual side effects and other advantages. Nevertheless there is clearly a small but significant potential for severe liver injury. Therefore, as of this date, we have removed nefazodone as one of the first-line recommendations for major depression and dysthymia. It is suggested in selected areas of the algorithm as an alternative to the other first-line recommendations, assuming that a full discussion of the risks and benefits has been held with the patient and documented. Further information about the hepatotoxicity may be found in the section on dosing and using nefazodone.
December 14, 2001. Update in question on selection of initial antidepressant when the patient is about to receive pharmacotherapy for the first time: should St. John's wort or SAMe be considered? and...are there gender differences in the response rate to different antidepressants? Update on merits of various augmentation strategies (e.g., pindolol, bupropion) to consider if the patient has not responded satisfactorily to two or three antidepressant monotherapy trials.
August 7, 2000 Significant change in algorithm sequence in the early phases. Bupropion SR is elevated to one of the first-line options. Venlafaxine and mirtazapine are made equivalent to tricyclics as recommendations for patients resistant to first-line agents. Sequential monotherapies with these are equally recommended to the SSRI-tricyclic combination strategy. Added new information on procedures for switching from one antidepressant to another. Added section on antidepressant side effect management. Reviewed the American Psychiatric Association Practice Guidelines for the Treatment of Major Depression, April 2000 update, and ensured that the information in this program is consistent with their recommendations, whenever possible. There were a few significant differences in recommendations, and we made sure that the evidence-basis for our different approach in those instances was clearly indicated.
February 14, 2000 Updated the link to PubMed's new site. Further refinements of the treatment trial reporting process.
December 12, 1999 Refinements made to the recommendation numbering scheme. These are the numbers used to report drug trials. All recommendations for patients with bipolar depression now will be reported as with or without psychosis. For example recommendation #40 is without psychosis; #40P is the one with psychosis.
November 4, 1999 Many refinements of the drug trial reporting process have been made. Clinician and patient versions of a Patient Functioning Questionnaire have been added.
October 14, 1999 A few changes in the text.
October 3, 1999 The drug trial reporting form has been improved and recommendations have been numbered so the user can easily identify the appropriate recommendation for reporting.
May, 1999: Extensive revision and update of the section on treating the various forms of bipolar depression. See flowchart for summary of changes (select bipolar depression as the diagnosis), or begin a consultation and select bipolar depression as the diagnosis.
May, 1999: General update to the non-bipolar sections of the algorithm, including revisions of drug costs, cytochrome interactions, inclusion of citalopram in relevant areas, procedures for adequate trials of antidepressants, author biography.
May 16, 1998: Updates to the algorithm for psychotic depression. ECT as a first line treatment is narrowed to the situation when the patient is severely ill. New reports with using olanzapine alone, and certain SSRI's alone (fluvoxamine, sertraline) in psychotic depression are acknowledged and included as options to consider, especially if more study confirms efficacy. See flowchart for summary of changes.
May 16, 1998: General update to the non-bipolar sections of the algorithm. About 25 new references added, new alternatives to consider at various steps, additional points in support of various recommendations, downgrading of alternatives not as favored at this point.
May 16, 1998: Update of "About the author." It now includes a financial disclosure statement.
July 27, 1997: Minor corrections, plus another spot was added where mirtazapine is mentioned as an option. This is in non-melancholic-non-severe depression when there have been adequate trial(s) of SSRI's/nefazodone.
May 12, 1997: In the Places to Go menu, we have added an algorithm for evaluation and management of frequent relapses. See it under "Barriers: frequent relapse..." or click here to see it now. Then return here.
May 12, 1997: In the section on treating depression meeting criteria for the Atypical Features Specifier of DSM-4, we added phototherapy as an option. Click here to see it now. Answer the question "yes" and check the recommendation if the patient has not had an MAOI trial. Then return here.
May 12, 1997: We have added discussions of mirtazapine (Remeron) at appropriate places in the algorithm. Although in no instance is it the primary recommendation, it may be an option for certain treatment-resistant patients. Return here after reading the relevant text.
May 12, 1997: Role of bupropion as augmentation for partial response to SSRI antidepressants: Click here, choose "no" answer to question about tricyclic usage, and read recommendation text. Then return here.
April 7, 1997: In the Places to Go menu, under "Adequate Trial of SSRI" there are some additional observations from the "expert consensus" panel regarding dosage strategies in bipolar (as opposed to non-bipolar) depressive episodes. Click here to go to it. Then return here.
March 27, 1997: In the Places to Go menu, there is a new algorithm for evaluation and management of noncompliance. See it under "Barriers: noncompliance..." or click here to see it now. Then return here.
March 27, 1997: In the Places to Go menu, there is an interesting quote from the writings of Feodor Mikhailovich Dostoevsky. To read it now, click here. Then return here.
March 27, 1997: In the opening screen to begin a consultation, there is a new caution requiring that the clinician employ criteria-based DSM-IV diagnoses before using this algorithm. The recommendations, which are substantially based on research which employs DSM diagnoses, will be much less accurate if alternative or more vague diagnostic criteria are used.