Teach-Back Boosts Patient Confidence and Adherence, Review Finds

Teach-back, the communication technique in which clinicians ask patients to explain newly received instructions in their own words, has long been promoted as a cornerstone of high-quality clinical communication. A new systematic review and meta-analysis published in the Journal of General Internal Medicine now offers the most rigorous test to date of whether the method actually delivers measurable benefits. The verdict is nuanced: teach-back does not clearly improve what patients know, but it appears to substantially boost their confidence and short-term adherence to health behaviors.

The research team, led by Judith B. Vick of Yale University School of Medicine and Karen M. Goldstein of the Durham VA Health Care System, searched MEDLINE, Embase, and CINAHL Complete for studies published between 2018 and 2026, supplementing these with pre-2018 trials identified in earlier systematic reviews. To be included, studies had to involve adult patients or care partners, deliver teach-back within a single encounter, include a comparator group, and report proximal or intermediate patient outcomes as defined in the team’s conceptual model. Two independent investigators screened every citation at the title/abstract and full-text levels and assessed risk of bias, ultimately synthesizing 18 randomized controlled trials involving 1,985 participants.

Teach-back itself is deceptively simple. Rather than asking patients whether they understood, a question that almost always yields a polite yes, the clinician says something like: I want to make sure I explained this well; can you tell me in your own words what you will do when you get home? If the patient’s restatement reveals gaps, the clinician re-teaches the material and checks again. The technique draws on cognitive science principles, including cognitive load theory, which holds that working memory can be overwhelmed by complex new information, and that active recall helps consolidate learning far better than passive listening.

The review’s first major finding concerned knowledge acquisition. Nine randomized controlled trials assessed whether teach-back improved what patients actually learned, but conceptual inconsistencies across the studies, differences in how knowledge was defined and measured, precluded meta-analysis. Across those trials, the researchers found no clear pattern of effect on knowledge, and they rated the certainty of evidence as very low using the GRADE framework, the Grading of Recommendations Assessment, Development and Evaluation system that weighs risk of bias, inconsistency, indirectness, and imprecision.

The picture brightened for self-efficacy, the psychological construct describing a person’s belief in their own capability to perform behaviors and reach goals. In a meta-analysis of five randomized controlled trials encompassing 416 participants, teach-back interventions produced a large increase in self-efficacy relative to usual care, with a standardized mean difference of 2.40 and a 95 percent confidence interval of 0.37 to 4.44. The certainty of evidence was rated very low, and the confidence interval was strikingly wide, ranging from small to very large effect sizes, with substantial heterogeneity across trials.

Adherence to health behaviors showed a similar pattern. Seven randomized controlled trials with 571 participants assessed whether teach-back improved patients’ self-reported adherence to recommended behaviors such as medication taking, dietary changes, or self-care routines. The pooled standardized mean difference was 1.04, with a 95 percent confidence interval of 0.45 to 1.64, indicating a large effect, and the certainty of evidence was rated low. As with self-efficacy, the confidence intervals were wide and heterogeneity was substantial, meaning the true effect could plausibly fall anywhere from modest to very large.

These findings carry real technical significance for how behavioral interventions are evaluated. Standardized mean differences express effects in units of pooled standard deviation, allowing trials that used different measurement scales to be combined. An SMD of 2.40 for self-efficacy is enormous by clinical research standards, where effects above 0.8 are conventionally labeled large. Yet the very width of the confidence intervals, spanning from 0.37 to 4.44, signals that the evidence base remains thin and variable, a caution the authors emphasize rather than bury.

The self-efficacy result also resonates with a deep theoretical lineage. Albert Bandura’s 1977 unifying theory of behavioral change positioned self-efficacy as the central mechanism through which interventions alter health behavior, and decades of research have confirmed that confidence in one’s ability predicts whether people actually attempt and sustain behaviors like medication adherence or self-management of chronic disease. If teach-back reliably builds that confidence, even without clearly boosting factual knowledge, it may operate through a different pathway than the information-transfer model that has dominated patient education.

Why might teach-back fail to improve knowledge while improving confidence and adherence? One possibility is that the knowledge trials measured learning too inconsistently for any true effect to emerge, with some studies using validated instruments and others using ad hoc quizzes. Another is that teach-back’s real value lies not in transferring more information but in creating a feedback loop: when patients restate instructions, clinicians can correct misunderstandings, and patients experience a sense of mastery and partnership that fuels motivation. The technique may also reduce anxiety and increase surgical cooperation, as one included trial of elderly ophthalmology surgery patients suggested, effects that knowledge scores alone would miss.

The review, funded by the Department of Veterans Affairs Health Systems Research Evidence Synthesis Program, arrives at a moment when health systems, including the Agency for Healthcare Research and Quality and the Institute for Healthcare Improvement, actively promote teach-back in toolkits and training programs. The authors’ conclusion is measured: they did not identify a clear benefit of teach-back on knowledge acquisition, but they did find evidence that teach-back improves self-efficacy and self-reported, short-term adherence to health behaviors. For clinicians, the message is that this simple conversational move, asking patients to teach the instruction back, may matter less for what patients know and more for what they believe they can do, a distinction that could reshape how the field studies and deploys communication techniques in the years ahead.

Subject of Research: Teach-back communication effectiveness in clinical settings

Article Title: Teach-Back in Clinical Communication: A Systematic Review and Meta-analysis

Article References: Vick, J. B., Pace, R., Klimek-Johnson, P., Rushton, S., Roberts, M., Lewinski, A., Neiman, J. H., Chen, D., Joseph, L., Der, T., Gordon, A. M., Jacobs, M., Crossgrove, R., Hoenig, H., Cantrell, S., Sims, C., Namasingh, N., Thornlow, D., Alishahi, A., … Goldstein, K. M. (2026). Teach-Back in Clinical Communication: A Systematic Review and Meta-analysis. Journal of General Internal Medicine. https://doi.org/10.1007/s11606-026-10678-y

Image Credits: AI Generated

DOI: 10.1007/s11606-026-10678-y

Keywords: teach-back, clinical communication, systematic review, meta-analysis, patient education, self-efficacy, health behavior adherence, Journal of General Internal Medicine, Clinical, Communication, Systematic, Review