Human Simulation Can Improve Clinical Trial Design
Simulated Patients and Clinical Trials
Computer-controlled simulated patients can improve the design of clinical trials before any real patients are even enrolled, Duke University Medical Center researchers have found. Their new finding follows an earlier Duke study that such simulated patients can improve the performance of clinical trial coordinators. Such patient simulators are basically lifelike computer-controlled mannequins that can mimic the physiological responses of living patients to procedures.
Improving design of clinical trials before they start is important, the researchers said, because it is commonly known that data collected during the early stages of a trial may not be as accurate or reliable as data collected later in the trial, since there can be a learning curve for health care workers. Just as importantly, the researchers continued, clinical trials whose protocols are tested before implementation can lead to fewer unwanted events, such as adverse patient reactions, thereby making the trial safer for patients.
Patient simulators can be used to improve clinical trials because they can be programmed to react physiologically to different scenarios, such as the administration of drugs or the placement of intravenous lines. The researchers can control the reactions of the simulated patients to enable learners to experience a variety of situations that may occur.
"We believe that all clinical trials should have their procedures and protocols evaluated before patients are enrolled," said Melanie Wright, Ph.D., a human factors specialist in the Duke University Human Simulation and Patient Safety Center. "In the case of complex trials, this is especially important and represents an area where human simulation can play a role in ensuring patients' safety and the reliability of the data collected."
The results of Wright's study were published in the December, 2005, issue of the journal Contemporary Clinical Trials. The research was supported by the National Institute of Neurological Disorders and Stroke and the Office of Research Integrity.
"Human patient simulation can be used to evaluate clinical trial protocols in the form of high-fidelity walk-throughs of different scenarios," Wright said. "Additionally, patient simulation offers the opportunity to rehearse clinical trial scenarios to evaluate not only the procedures, but the data collection tools used to support the trial.
"In our recent experience, simple observations by the trial's principal investigator and simulation operators, when coupled with the reports by those who actually participated in the simulation, resulted in substantive improvements in the protocol and procedures," she continued.
For their analysis, the Duke team evaluated a draft protocol of a clinical trial involving administering anesthetics during an open heart procedure. The protocol required constant monitoring of the patients' vital signs, as well as interacting with other members of the surgical and anesthesia teams. A total of 48 trial coordinators and ten data monitors, as well as the principal investigator, went through a full day of training on the simulator.
After observing the training sessions of the draft protocol, the researchers found that one potential procedural flaw might lead to patients receiving an unnecessary double-dosing of the drug. Since the trial involved the use of a heart-lung machine to take over the function of the heart during surgery, the researchers found that to the drug might be administered both directly to the patient, as well as introduced to bypass circuit, depending on the timing of events.
Additionally, the researchers determined the optimal time for study coordinators to discuss medical management of the patient with health care providers in the post-operative intensive care unit. Walk-throughs of the protocol identified specific opportunities when the trial coordinators would not be busy with other tasks associated with the trial.
"Another major change to the protocol was a decision by the trial sponsor to include all coordinators in the simulation training," Wright explained. "The original plans called for only the monitors to receive training; however, after participating in the simulation training, they decided that it would be beneficial for the coordinators to be trained as well."
Finally, the experience led to changes in the paperwork and checklists that trial coordinators follow to ensure that the trial is being conducted in the same manner at all participating sites.
"We believe that use of human simulation can be cost-effective for sponsors of clinical trials," Wright said. "There are up-front costs associated with running the simulations, but the potential savings of not losing data collected early in the trial and the need to reprint or reproduce trial materials should offset those costs."
She added that while full-scale simulation can be effective in spotting potential problems in complex clinical trials, researchers can also design simpler simulations for simpler or more straightforward trials.
The clinical trial involved in the study was supported by the Global Perioperative Research Organization, a joint effort of the International Anesthesia Research Society, the Duke Clinical Research Institute and Duke's Department of Anesthesiology to improve outcomes of patients during surgery. The Simulation Center is a joint effort of Duke's School of Medicine, School of Nursing and Department of Anesthesiology.
Other Duke researchers involved were team members were Jeffrey Taekman, M.D., Eugene Hobbs, Mark Newman, M.D., and Mark Stafford-Smith, M.D.