Reading time: 3 minutes, 57 seconds (991 words)
“If we only ask whether interventions help, then we risk missing when they harm…. Safer interventions require the same scientific attention to harms as we already give to benefits.”
That was the conclusion to Dr. Lindsey Gilling’s presentation to the International Forensic Psychiatry Lecture Series.
She provided recommendations for forensic mental health practitioners, researchers, funding agencies, and journal publishers to expand the science of effectiveness to include the science of safety.
How and When Negative Effects Happen
“Having good intentions doesn’t immunize us from unintended consequences of the work that we do.” Dr. Gilling stressed that therapy does work, but it does not work for everyone.
She emphasized that negative effects rarely reflect a single point of failure but a dynamic interaction of patient, therapist, treatment, and system factors.
The risk of having negative effects rises when the patient has higher symptom severity and complexity, as well as personality, interpersonal, and relational difficulties, such as attachment and alliance difficulties. Previous negative experiences in therapy and trauma also increase risk.
Therapist-related risk factors include weak skills in building and repairing therapeutic relationships; not responding to patients’ needs; and limited training and experience in the particular therapy. Other risk factors include stress, burnout, or countertransference difficulties, as well as low empathy, warmth, or interpersonal effectiveness.
Risk can also rise when treatments are a mismatch for the patient, including overtreatment of low-risk patients and undertreatment of high-risk patients. Commonly used in forensic mental health settings, group-based treatments also increase risk if shaming, exclusion, or reinforcing pro-offending attitudes are not well managed.
Risk also increases when the organizational culture prioritizes risk management at the expense of relational work. Limited time, high caseloads, and restricted treatment options can also increase the risk of harm, particularly when they lead to high staff turnover and burnout.
Research Evidence on Harms
Most research on psychological treatments focuses only on their effectiveness. Few report harm. That means the evidence base is incomplete, which creates bias and uncertainty.
Unwanted events include adverse events (AEs) and serious adverse events. These can be unrelated to therapy, such as the natural course of the mental disorder or life stressors, or negative treatment effects, such as side effects of medications or malpractice. Unwanted events can include some therapeutic discomfort, such as guilt or shame. Because of that confusion in terminology, unwanted events can happen 30 to 60 per cent of the time. Deterioration has been found to be between five and 10 per cent, with lasting harm occurring five per cent of the time.
Consolidated Standards of Reporting Trials
Many researchers are familiar with the Consolidated Standards of Reporting Trials (CONSORT) statement on how to report randomized controlled trials (RCTs), a gold standard for evaluating interventions.
However, researchers’ reporting of social and psychological intervention RCTs were often poor, with missing information. Traditionally, RCTs reported on benefits more than harms. AEs were often poorly defined, inconsistently measured, and selectively reported. This limits informed clinical decision-making and patient safety.
So, the CONSORT group created CONSORT-SPI 2018 (social and psychological interventions). While the checklist is the same, the supporting documents show how the checklist fits into these studies and provide guidance on reporting those commonly missing pieces.
The 2004 version of the CONSORT Harms included 10 recommendations for reporting harms, including definitions, methods for collecting data, severity, withdrawal, and analysis. The 2022 revision added three more items, with harms being integrated into the main CONSORT framework.
Current Research
Dr. Gilling and her colleagues are doing a systematic review of English-language research papers published after July 2018 to determine whether harm reporting improved. Using an existing 21-item checklist developed from the 2004 CONSORT Harms statement, the researchers analyzed 37 RCTs of psychological interventions for adults detained in prisons or forensic hospitals around the world. Only 5.4 per cent had moderate harms reporting. All others had very low harms reporting.
One of the barriers to high-quality reporting could be difficulty detecting negative effects and determining whether they are attributable to the intervention. The psychological intervention research field still prioritizes benefits over risks, and journals have yet to enforce reporting standards.
Potential Solutions
Dr. Gilling encouraged forensic mental health practitioners to build psychologically safe cultures where concerns can be openly discussed. This starts by transparently discussing potential risks and benefits during informed consent and explicitly asking patients about negative effects and deterioration throughout treatment. This requires practitioners to distinguish between therapeutic discomfort and harmful effects.
Practitioners should design interventions that are responsive and person-centred, prioritize patient safety, and continuously monitor outcomes, not only when the person completes a program.
She suggests researchers should:
- Clearly define expected AEs, serious adverse events, and negative outcomes in protocols;
- Incorporate AEs into informed consent procedures (“Participants can’t provide truly informed consent if we fail to tell them about the possible negative or adverse outcomes associated with the treatment”);
- Systematically monitor harms throughout trials;
- Transparently report deterioration and AEs, even when the cause is uncertain;
- Use standardized reporting guidelines, such as CONSORT Harms, including null and negative findings; and
- Report not only on effectiveness but also on who benefits, who deteriorates, why negative effects emerge, and how risks can be reduced.
Dr. Gilling recommends journal publishers endorse and enforce CONSORT and CONSORT Harms guidance, require authors to state when harms were not monitored, and encourage publication of null and negative findings.
Research grant providers should:
- Require potential harms be monitored in the trials they fund;
- Require prespecified AE protocols and reporting plans;
- Prioritize transparency, replication, and patient safety; and
- Support methodological research on harms in psychotherapy.
All of this work would help build a science of safety, not only a science of effectiveness.
For more details, watch Dr. Gilling’s May 7, 2026, lecture, “The Gap Between Intent and Impact: Understanding Negative Effects of Psychological Interventions in Forensic Settings.”
Written by Virginia St-Denis and approved by Dr. Lindsey Gilling.
