Various commercially available and do-it-yourself (DIY) models were used to teach ultrasound-guided i.v. insertion to emergency medicine (EM) residents and medical students. Commercial models decayed with time, but DIY versions were affordable, easy to create, and quickly abandoned. They used a controlled study to subjectively evaluate how well DIY models and commercially created models compared with human tissue, both tactilely and sonographically.

Three models for US-guided i.v. instruction was tested: a commercially available model US training block model, a handmade tofu model, and a homemade gelatin model. All three models were compared to human tissue-guided i.v. insertion in the United States. Participants in the study were EM residents and attendings who had prior experience with US-guided i.v. insertion in actual patients. Participants with varying levels of training and experience with US-guided i.v. insertion subjectively described how each media compared tactilely and sonographically, which model was most similar to live humans overall, and which model was best for instructing learners after practicing peripheral i.v. placement under US guidance using the three media.

The gelatin model had a much better overall score (total of visual and sonographic scores) than the other models, suggesting that the gelatin model was judged to be the most accurate to human anatomy compared to the other models. Homemade alternatives to commercial simulators that are less expensive can be realistic and effective surrogates for learning US-guided peripheral i.v. placement.