While traumatic brain injuries are the leading neurosurgical cause of morbidity and mortality worldwide, studies to determine the relative efficacy of different treatments have proven to be notoriously difficult to conduct. Factors contributing to the difficulty include tremendous variability among patients who are injured (host factors) and equally daunting diversity among individual injuries with respect to mechanisms, pathological consequences, and confounding physiological insults (injury factors). Add to this mix the fact that multiple treatments are administered to most head-injured patients in different ways and in different combinations, and analyzing the effect of any specific intervention becomes almost prohibitively complicated. For these reasons, even large, carefully planned, multicenter prospective trials to assess efficacy of specific treatments in traumatic brain injury have failed to clearly elucidate benefits of individual interventions. Nonetheless, gains have been made in mortality, likely because careful attention to standardization of treatment protocols helps to minimize delayed and preventable damage.1 To overcome some of the classic difficulties in head injury trial design, Sakellaridis and colleagues2 address the question of whether mannitol or hypertonic saline is more effective in treating elevated intracranial pressure (ICP) by utilizing a blinded, alternating treatment protocol. In this paradigm, patients with severe brain injuries who have multiple episodes of elevated ICP are randomized to one treatment or the other, and then the two treatments are alternated with each subsequent episode of pressure elevation. In this way, the investigators were able to compare the effects of each treatment in the same patient over multiple uses, ultimately collecting data on 199 separate events of increased ICP in 29 individual patients. In order to keep the osmotic load of the two different treatments constant, they used 15% saline administered by central intravenous line, which was measured as equivalent to 20% mannitol at the doses used. All other obvious treatment protocols were standardized among the patients, who had a range of ages and injury types, including surgical and nonsurgical lesions. The authors assessed short-term effects of the two hyperosmolar treatments (the magnitude and duration of ICP reduction) as well as overall status of the entire cohort with respect to fluid and electrolyte status, length of stay, and morbidity and mortality rates. They conclude that hypertonic saline and mannitol have equivalent efficacy in treating episodes of elevated ICP. Many other trials in animals and humans comparing various concentrations of hypertonic saline to mannitol have, by and large, reached similar conclusions. This alternating treatment protocol approach to comparative efficacy works well when the treatment indication is episodic and recurrent in a single patient, and when short-term outcomes are easily measurable. However, because all patients received multiple doses of each agent, the design was unable to address the question of whether one agent might be preferable in a specific injury type or in a particular host. Similarly, no conclusions can be drawn about the toxicity of the individual agents nor their effects on more global or long-term outcomes. Intracranial pressure is one of the obvious “surrogate end points” that researchers in the head injury field have attempted to define as targets for clinical trials for many years, to get around the obstacles of most prospective clinical trials utilizing long-term outcome measures. In order for this approach to provide useful data to the clinician, interventions must have an immediate effect on something that can be measured—from blood, radiology parameters, or by some other type of monitoring device— and differences in this quantity should be demonstrated to correlate with long-term functional outcome. A carefully performed alternating treatment protocol as utilized here may be applicable to comparisons between a variety of acute and subacute intervention choices. However, more complete understanding of the effects of head injury interventions will continue to require multiple approaches, including basic science and animal models, trauma registries, and prospective trials with sophisticated data analysis capabilities to help unravel the role played by any one piece of the complex puzzle that is this disease.