The Development Group for the Standardized Treatment Guideline for Depression was formed as a multidisciplinary team consisting psychiatrists, administrative researchers, clinical psychologists, systematic review experts, and specialists in preventive medicine. Throughout the process of developing the clinical practice guideline, members of the group carried out ongoing comparative assessment and evaluation of all results while performing the investigations and procedures required in order to base the findings on scientific methodology (8).

The Development Group investigated pre-existing domestic and international treatment guidelines, as well as trends in the management of depression among Korean psychiatrists. Subsequently, the Development Group analyzed the results of previous investigations (10-50) and developed a non-pharmacological treatment guideline for depression. The Development Group also assessed the domestic status of depression treatments in Korea. The 'Evidence-Based, Non-Pharmacological Treatment Guideline for Depression in Korea' was indicated at Korean adult patients with an initial diagnosis of mild or moderate depression by DSM-IV (9).

Common self-care methods for depression include bibliotherapy, computer-based treatment, dietary supplements (St. John's wort, S-adenosylmethionine, selenium, vitamin B, C, D, folic acid, Ginkgo biloba, glutamine, tyrosine, natural progesterone, oriental medicine, caffeine, alcohol, omega-3 fatty acids, and others), acupuncture therapy, light therapy, anion therapy, massage therapy, exercise therapy, relaxation therapy, music therapy, hypnotherapy, yoga, meditation, and aromatherapy (51). Among these, the literature evidence was the most convincing for St. John's wort, exercise therapy, bibliotherapy, and light therapy during the winter. In the context of domestic situation in Korea, exercise therapy and bibliotherapy were considered to be worthy of recommendation as first stage of treatments for mild or moderate depression. St. John's wort was not included in this guideline, since it is classified as not a Over-the-Counter (OTC) drug but a prescription medication in Korea. Light therapy was also excluded, because it is not widely clinically-available in Korea. In addition to self-care methods, this guideline addressed psychotherapy methods. For any form of psychotherapy, it is important to consider whether a given method has independent and well-established therapeutic features that justify the title of 'psychotherapy' because an unverified psychotherapy is nothing but a placebo treatment. According to Wampold et al. (52), psychotherapy should have the following key elements to be distinguished from general clinical management or a placebo treatment. First, the therapy is mediated by a trained specialist. Second, the therapist can carry out treatments tailored to the needs of each patient. Third, there should be a unique and clear psychotherapeutic principle that distinguishes it from other forms of psychotherapy. From this viewpoint, only a few kinds of psychotherapy are comparable to pharmacological treatments or placebo controls. These are cognitive-behavioral therapy (CBT), short-term psychodynamic supportive psychotherapy (SPSP), interpersonal psychotherapy (IPT), problem-solving treatment, and marital therapy. Among these, the guideline evaluated three psychotherapies, namely CBT, SPSP, and IPT because these three methods have been supported by theoretical and clinical evidence. In addition, therapeutic and theoretical evidences suggest that these three psychotherapies could be applied to different groups of depressed patients (53). On one hand, both CBT and IPT follow structured procedures with limited numbers of therapy sessions. However, CBT deals with the association between negative emotions and thoughts, while IPT deals with the association between negative emotions and corresponding life-events (mostly, interpersonal). On the other hand, psychodynamic psychotherapy deals with transference relationships, and its treatment principles are based on earlier theories of psychotherapy. Psychodynamic approaches have been claimed by a number of psychotherapists to be effective, but the classical psychoanalytic treatment is not suitable in the management of depression because the treatment is ultimately aimed at self-understanding and personality change. Especially, SPSP is defined as 16-session psychotherapy which focus on the affective, behavioral, and cognitive aspects of interpersonal or intrapersonal relationships (54). As a short term psychodynamic psychotherapy, SPSP recognizes the therapeutic utility of the transference relationship but does not interpret this relationship to the patient. Thus, among a number of self-care or psychotherapeutic methods, the Development Group regarded exercise therapy and bibliotherapy as self-care procedures available for mild or moderate depression, and also considered CBT, SPSP and IPT as available psychotherapeutic methods (8). Hence, our key questions referred to exercise therapy, bibliotherapy, CBT, SPSP, and IPT.

These key questions were designed to elicit accurate and correct answer, usually by employing the "PICO" method. The basic elements of a key question were P, I, C, and O, where P represents patients or the corresponding problems (patient population), I represents an intervention such as diagnostic evaluation, prognostic factors, and treatments (interventions), C represents an alternative intervention with which the former is to be compared (comparison), and O represents clinical outcome (outcome) (55).

Within the scope expressed in the Core Standard Ideal (COSI), the following literature sources were searched: PubMed, EMBASE, Cochrane CENTRAL, Korea Med, KMbase, RICH, and the National Assembly Library. Since systematic reviews and meta-analyses did not include Korean publications, the administrative researchers were obliged to search the literature manually and extract the contents of the relevant journals to access and evaluate original publications.

Literature search was carried out in several steps, including the development of the search strategy itself. First, the researcher who was responsible for selection of key questions converted the questions into the PICO format (55) and forwarded them to the literature search team. The team extracted preliminary search words and sent them back to the researcher in charge. Second, the team searched the database using the selected search words, listed the initial basic search results and abstracts, and sent this data back to the researcher in charge. Then, the researcher read and judged each abstract and created a list of suitable studies, given the key question being asked. The literature search team obtained the original publications in the list. Third, the researcher in charge received the original publications from the search team, read them carefully, and selected the publications to be used in the systematic review. Finally, the search team searched for published findings of randomized controlled trials (RCTs) conducted after the publication of the most recently selected systematic reviews.

For evaluation of each study, the Development Group used the Scottish Intercollegiate Guidelines Network (SIGN) grading system (56). Meta-analyses and RCT were allocated as top level evidence; un-randomized clinical research and observational research as mid level evidence; and, expert opinion and case reports were allocated as low level evidence. The SIGN systematic review evaluation method was generally used to judge the following criteria: first, if the research questions were carefully chosen and well-focused; second, if there was a methodological description; third, if the literature search was comprehensive enough to discover all the appropriate findings; and fourth, if all the research findings were homogeneous enough to be combined together. Publications in languages other than Korean or English were excluded from the scope of this search. The Development Group had been trained by an expert of evidence-based medicine in evaluating the appropriate literature with experts in methodology, and such efforts were continued to unify the evaluation criteria used by all researchers.

Evidence levels used are summarized in Table 1 (57). These were prepared by systematically summarizing relative research findings to allow readers to compare the results in a single view. In general, the tables included research type and quality, research type, arbitration, confidence intervals, and results.

The Development Group classified grades of recommendation into 4 levels, A, B, C, and Good Practice Point (GPP), as indicated in Table 2. GPP is a recommendation grade with an evidence level of 3 or 4 (57). The development of recommendations consisted of synthesizing and analyzing all the data gathered for each key question. However, if findings were diverse or controversial on a certain issue, the evidence was not consistent, most of the claims underlying the evidence were inadequate or poor, or if the evidence had a high evidence level but were of a lower clinical applicability, careful consideration were given to the recommendation grade. As a consequence, expert opinion sometimes influenced recommendation grades, and official agreement amongst all the researchers were used in place of evaluations from a single researcher.

The Development Group requested external experts to review the draft of the guideline for its feasibility and availability, and the review result was considered for during the revision process after a subsequent internal discussion. The 'Evidence-Based, Non-Pharmacological Treatment Guideline for Depression in Korea' was certified at the Annual Meeting of the Korean Neuropsychiatric Association in April, 2010.