ResearchPod Summary
While the medical and mental health professions widely embrace the principle of primum non nocere ("first, do no harm") and ethical codes mandate avoiding client harm, the field of clinical psychology has historically paid remarkably little attention to potentially harmful therapies (PHTs). Major psychotherapy handbooks and guidelines often dedicate negligible space to adverse effects, contrasting sharply with psychiatry, where adverse drug reactions receive intense scientific and regulatory scrutiny. Because psychology lacks a formal regulatory body equivalent to the Food and Drug Administration, the responsibility of monitoring, identifying, and avoiding unsafe psychological interventions falls entirely upon the profession itself.
Traditional arguments for psychotherapy-induced deterioration typically rely on two lines of evidence: outcome studies estimating that 3% to 10% of clients deteriorate, and meta-analyses revealing negative effect sizes in a minority of studies. However, both metrics are flawed. Deterioration percentages can overestimate harm by misattributing natural worsening to therapy (the post hoc ergo propter hoc fallacy) or underestimate harm by missing clients whose recovery was merely slowed. Methodological obstacles further complicate the identification of PHTs, including treatments that increase outcome variance while leaving group means unchanged, side effects that manifest in some symptom domains but not others, multidimensional forms of harm, and negative impacts on clients' friends or relatives.
Recognizing and researching PHTs carries profound implications for clinical science and practice. First, awareness of harmful interventions helps practitioners fulfill ethical obligations to minimize client risk and allows them to guide clients away from dangerous adjunctive treatments. Second, studying the mechanisms of client deterioration provides vital insights into the therapeutic process, highlighting processes like premature termination of exposure or maladaptive modeling that can cause roadblocks even in generally effective treatments. The field must prioritize identifying PHTs, require researchers to report full score ranges and drop-out data, and conduct independent replications to safeguard mental health consumers.
Alex: Welcome to another episode of ResearchPod. Today we're looking at a paper by Scott Lilienfeld that examines psychological treatments that inadvertently cause harm.
Sam: So the paper is asking why medical treatments are heavily monitored for side effects, while mental health care has historically paid much less attention to treatments that might make patients worse?
Alex: Exactly. While we have extensive lists of treatments that work, the field has largely neglected systematic tracking of hazardous psychological interventions.
Sam: And the core issue is that people assume all counseling is helpful by default, so nobody's actively watching for when it backfires?
Alex: Precisely. When someone seeks help after a severe trauma, they might receive a mandatory counseling session that forces emotional processing too quickly—and that can actually worsen their long-term symptoms.
Sam: That feels counterintuitive. What do researchers call these negative outcomes caused by the treatment itself?
Alex: There's a term for when a treatment causes unintended distress: an iatrogenic effect. In plain language, it just means the treatment made the problem worse—like a medicine with a severe side effect.
Sam: But unlike prescription drugs, therapy doesn't have an equivalent to the FDA tracking safety problems.
Alex: That's the core vulnerability. Psychology has to police itself, and for a long time, it didn't prioritize tracking negative outcomes.
Sam: Why was there such a blind spot?
Alex: For decades, researchers were focused on proving that psychotherapy worked at all compared to doing nothing. Because early studies showed positive averages, the field assumed therapy was uniformly beneficial. But averages can hide a lot. Studies consistently suggest that a meaningful minority of clients—somewhere in the range of five to ten percent—get worse after starting psychotherapy.
Sam: How do researchers even measure whether a treatment caused the worsening, rather than the patient's condition just naturally declining?
AI-generated third-party summary by ResearchPod. Not official content or an endorsement by the paper authors or affiliated organizations.
Alex: To know if a treatment caused harm, you have to compare patients who received the intervention against a control group that received nothing. Without that comparison, you can't tell if the decline was due to therapy or the natural course of the illness.
Sam: And I imagine that creates real tension for practitioners who want to help but might be using techniques that carry hidden risks.
Alex: That tension is sometimes called the scientist-practitioner gap. Research flags certain approaches as risky, but those warnings don't always reach the clinicians actually using them.
Sam: So what are some specific examples of these risky approaches?
Alex: One classic example is critical incident stress debriefing—a single-session group intervention given to people right after a traumatic event. The idea sounds reasonable: get people talking immediately. But it forces them to relive intense emotional details before their minds have had any chance to naturally process what happened, which can disrupt normal coping mechanisms rather than support them.
Sam: So instead of helping, it can re-traumatize people.
Alex: That's what controlled evaluations suggest. And it's a useful illustration of a broader pattern—a technique that feels intuitively helpful but produces the opposite effect when tested carefully.
Sam: Are there other techniques on this list that researchers have evaluated?
Alex: Yes. Another prominent example involves juvenile delinquency programs that try to frighten at-risk teenagers by exposing them to prison environments. The logic seems sound—show kids the harsh reality, and they'll change course. But controlled evaluations found these programs actually increased the odds of future offenses compared to control groups.
Sam: Why would that happen?
Alex: The mechanism is something researchers call peer deviancy training. When troubled teenagers are grouped together in an intense setting—even one meant to be a deterrent—they inadvertently reward and reinforce each other's antisocial attitudes. The group dynamic normalizes the very behavior the program was trying to prevent.
Sam: So the intervention itself creates the social feedback loop that makes things worse.
Alex: Exactly. And that same dynamic appears in other group-based programs for at-risk youth. Structure matters enormously. Without it, peer influence can run in the wrong direction.
Sam: What about therapies that deal with communication disorders, like autism?
Alex: There's a technique where a helper physically supports an autistic person's hand to type on a keyboard, based on the assumption that the main barrier is physical control rather than cognitive processing. Controlled tests have consistently shown that the messages are unknowingly generated by the helper guiding the hands—not by the person being supported.
Sam: So the helper is inadvertently projecting their own thoughts onto the board.
Alex: Precisely. And that dynamic has led to serious consequences—including uncorroborated allegations of abuse against family members based on messages that weren't actually coming from the person they appeared to come from.
Sam: That's a severe outcome. Are there physical interventions that carry similar risks?
Alex: Certain attachment therapies for children—involving intensive holding or rebirthing simulations—attempt to release unprocessed distress by physically restraining or wrapping patients. Because these methods lack controlled trial support and carry extreme physical risks, they sit at the most serious end of the harm classification. There are documented cases of suffocation.
Sam: And memory work? I know some therapies try to recover forgotten events from childhood.
Alex: Suggestive methods like hypnosis and repeated prompting can inadvertently construct entirely false memories of trauma in some individuals. The technique creates the memory rather than uncovering a real one. And those false memories have been linked to severe psychological distress and significant family disruption.
Sam: It keeps coming back to the same underlying problem—good intentions, but no systematic check on what's actually happening to patients.
Alex: And that's where the paper's practical recommendations come in. One of the more well-evaluated tools is a session-by-session feedback system. Clients fill out standardized questionnaires at every appointment, and clinicians receive a prompt warning if a client is failing to make expected progress. If the scores show someone getting worse, the therapist gets an alert to change course before real damage accumulates.
Sam: Like an early warning system built into the therapy itself.
Alex: Exactly. Studies suggest this kind of feedback loop is associated with meaningful reductions in symptom deterioration. It's a relatively simple mechanism, but it addresses the core problem—therapists often don't know a treatment is backfiring until significant harm has already occurred.
Sam: There's also a broader institutional piece, isn't there? It's not just about individual clinicians.
Alex: Right. Lilienfeld argues that major professional organizations need to stop awarding continuing-education credits for unvalidated methods. If a technique hasn't been tested and carries known risks, the profession shouldn't be certifying people to use it. Identifying and discouraging harmful treatments, he suggests, should take precedence over cataloging beneficial ones—because safeguarding clients from demonstrable harm is the primary ethical obligation.
Sam: Though the paper does acknowledge limits in the current evidence base?
Alex: It does. Many of the studies rely on case reports or quasi-experimental designs rather than randomized controlled trials, which makes it difficult to calculate precise harm rates. The mechanisms are often clear in theory, but nailing down exact figures requires more controlled research.
Sam: So the honest position is: we know enough to act, but not enough to be complacent about the research itself.
Alex: That's a fair summary. Good intentions aren't sufficient. Clinicians need systematic awareness—not just of what works, but of what causes harm. That's the core message of the paper. Thanks for listening to ResearchPod.