Guarding the Guidelines: Objectivity in Treatment Recommendations

Treatment guidelines have utility in amassing, distilling, and summarizing comparative treatment evidence for the field at large. Organizations engage panels of experts who volunteer their time to review and critically assess treatment outcome literature for particular diagnoses, then issue recommendations under the organizational imprimatur. For psychotherapies, treatment guidelines grant a seal of approval somewhat analogous to U.S. Food and Drug Administration (FDA) approval of pharmacotherapies (3). Ideally, guidelines based on the results of randomized clinical trials (RCTs) objectively distinguish therapies of demonstrated efficacy from those of uncertain efficacy, providing clinicians with recommendations on what works for whom for a given diagnosis.

Measuring the impact of guidelines is difficult, but their inclusion or exclusion of a particular treatment undoubtedly carries weight and influences clinical and administrative decision making. For example, although the Department of Veterans Affairs (VA) hospital system and Department of Defense (DoD) guidelines state that they are not prescriptive and do not represent policy (4), they tend to guide the funding of particular trainings over others, the development of standardized templates for only certain treatments (e.g., IPT for major depression but not PTSD), and the development of tracking and measurement systems. Inclusion in guidelines may also strengthen clinician arguments for patient insurance reimbursement.

Yet guidelines have limitations. Because of the labor required and the relatively slow accumulation of new research findings, guidelines appear only at lengthy intervals. The last American Psychiatric Association Clinical Practice Guideline for PTSD was published in 2004, the last for major depression in 2010. They thus can fall behind clinical advances.

Moreover, treatment guidelines are unfortunately subject to political influence. As has too often occurred since the fracturing of Freud’s circle, psychotherapy proponents can be partisan rather than objective. Methodology is manipulable. Delimiting inclusion criteria such as the time frame for inclusion of relevant studies can selectively harvest evidence favoring one therapy over others (5). Guidelines may overly depend on meta-analyses, which in turn depend on the quality of their included primary studies (i.e., “garbage in, garbage out”) and lose the fine grain of individual research trials. Their retrospective outlook and time lag to publication mean meta-analyses are invariably dated and ignore subsequent studies of interest (5). Thus, meta-analyses can be helpful but should not constitute the sole data source. Guideline panels may be slanted by treatment allegiance imbalances among the experts. Some panels labor harder than others. One of us (J.C.M.) resigned from the treatment guideline panel of a major psychological organization after a decision emerged that the group would rely on metadata, the already dated conclusions of published systematized reviews and meta-analyses, rather than critiquing the primary literature.

Perhaps because mental health treatment guidelines that bear the aegis of august institutions appear authoritative, pushback has been limited. Parker (6) critiqued the 2004 Australasian depression treatment guidelines on broad methodological and definitional grounds such as range of illness severity, lack of generalizability of patients in RCTs, and vagueness of treatment recommendations. Parker (6) warned of “the risk … of opinion dressed up as fact.” Hoge and colleagues (5) recently published an incisive and devastating critique of the shift in recommendations between the VA/DoD PTSD treatment guideline editions issued in 2017 and those of 2023.

Whereas the 2017 VA/DoD guidelines recommended seven trauma-focused psychotherapies, the 2023 guidelines reduced the number to three: prolonged exposure, cognitive processing therapy, and eye movement desensitization and reprocessing—exposure therapies all. The 2023 guidelines downgraded all non–trauma-focused psychotherapies and decreased the number of recommended pharmacotherapies for PTSD. For IPT, this meant a shift from “weak evidence for” in 2017 to “insufficient evidence to recommend for or against” in 2023, a change made without further explanation. Other downgraded therapies received a similarly cryptic dismissal. These changes did not reflect any weakening of evidence supporting the treatments, but rather, as Hoge et al. (5) charged, a “selective interpretation of evidence and extrapolating recommendations” based on methodological manipulation.

As context, the PTSD research literature supports the efficacy of non–exposure treatments for PTSD, including present-centered therapy and IPT (2, 7), yet the 2023 guideline obscured this. In short, the new guidelines smacked of bias, favoring some treatments over others despite limited and even contrary evidence. They read as narrow-minded, overly restrictive, and oblivious to developments in the field. We share Hoge and colleagues’ (5) disappointment with those guidelines, which seem to excessively promote three older exposure therapies while arbitrarily discounting other treatments.

Originally developed as an affect-focused (7), time-limited, manualized psychotherapy for major depression, IPT has repeatedly demonstrated efficacy in RCTs for various disorders. IPT is included in mood and eating disorder guidelines. Research testing of IPT for PTSD began only in the early 2000s, with a surge of studies in the past decade (2). Lacking evidence, IPT was understandably not included in earlier treatment guidelines. In 2017, however, the VA/DoD guideline listed IPT as a “suggested” treatment—a rank below a recommended treatment—based on “weak evidence” (8). This endorsement may have been overly generous, in that it depended heavily on growing evidence of the efficacy of IPT for treating civilians with PTSD. Veteran and military personnel outcome data were limited to a few small open treatment trials (2).

Since 2017, the case for IPT has strengthened. Thirteen clinical trials, seven of them randomized, now support IPT’s utility for adults with PTSD. Three of those trials found IPT noninferior to prolonged exposure (N=2) or to the selective serotonin reuptake inhibitor sertraline (N=1), which has an FDA indication for PTSD (2). IPT has thus shown equivalent efficacy to “first-line” recommended treatments while outcompeting weaker control conditions. This body of evidence, conducted by multiple independent research groups, more than suffices to demonstrate efficacy and merit inclusion of IPT in clinical treatment guidelines (2).

Regarding the VA/DoD guidelines for veterans, IPT has now been tested in five military trials comprising 187 patients. Although most were small open trials, Shea and colleagues (9) in 2023 found IPT comparable with “gold-standard” prolonged exposure in a two-site VA hospital randomized trial of 115 veterans with PTSD due to war zone trauma. Yet despite bolstered evidence, the VA/DoD recommendation withered. The most recent guidelines downgraded IPT to a noncommittal “neither for nor against” status (4). This is unfortunate: there being no panaceas for PTSD, patients and therapists benefit from having a range of available effective treatments. IPT, as an antidepressant, non–exposure psychotherapy, may have particular benefit for patients with PTSD and comorbid depression and may have lower dropout than exposure-focused treatments (2).

We do not wish to seem partisan in defending IPT. Therapists and researchers conducting other evidence-based approaches might take similar issue with the recent guideline: practitioners of trauma-focused cognitive therapy or skills training in affective and interpersonal regulation (STAIR), for example. Besieged by discriminatory insurance reimbursement, the shift in academic psychiatry from clinical research to neuroscience, and other factors, psychotherapists need to work together rather than fighting among themselves. Because clinicians may not stay current with the research literature, the availability of accurate, objective guidelines can promote dissemination of clinical knowledge and enhance mental health care. Sectarian biases impede that flow. As mental health professionals, we should all endeavor to recognize our biases and rise above them (10). Of the two contradictory Hume epigraphs heading this article, we side with the first.

As Parker (6) noted, “guidelines are difficult to prepare and easy to criticize.” Organizations initiating guidelines should take care to balance their roster of experts, ensuring that no one allegiance predominates. Those experts should check their allegiance at the door, striving to cooperate and collaborate rather than compete, seeking objective readings of a diverse and rocky terrain that no one treatment rules. As Falkenström et al. (11) recommended in 2013 for researchers conducting clinical trials, guideline panelists should acknowledge and seek to control their potential treatment allegiances. Even the appearance of bias discredits what should be an objective document: guidelines, like Caesar’s wife, must be above suspicion. Methodological approaches should reflect broad consensus. Guideline groups should regularly issue preliminary drafts and invite commentary before publishing final guidelines. Guideline panelists should be our Virgils—wise, reasoned, and protective.

After this Viewpoint was accepted for publication, the American Psychological Association released draft PTSD treatment guidelines that repeat the mistakes Hoge and colleagues documented (5). Relying on systematized reviews and meta-analyses rather than using current studies, these guidelines again recommend a few older exposure treatments and deem IPT has “insufficient evidence for the Panel to recommend for or against the following psychological interventions over no intervention ([waitlist]) or [treatment as usual]” (American Psychological Association Guideline Update Panel for the Treatment of PTSD in Adults, 2024, unpublished). The panel does not state how many of the seven IPT randomized trials it included in reaching this judgment.