While not the most complete drug reference available, I do believe the Lexi-comp databases are the best suited for pharmacists. The information is laid out in a concise, user friendly manner. There is so much information to chose from that it must be difficult for companies like Lexi-comp to design resources that don’t overwhelm the user. I would like to see better information on certain topics; compatibility, extemporaneous preparations, off label uses, etc.

I’m not surprised by your comment regarding Micromedex. I’ve never been a big fan and don’t trust the information beyond the basics. I also find it interesting that these resources demonstrate a 4% error rate across the board. When you consider the amount of information contained in these databases that can add up in a hurry. It would be interesting to know if any of these errors have ever led to an adverse drug event.

One assessed online drug info databases (BMC Med Inform Decis Mak full-text at: http://www.biomedcentral.com/content/pdf/1472-6947-7-7.pdf) and found Clinical Pharmacology performed best, barely edging Micromedex. An interesting aspect that I think adds some current relevance is that our findings almost mirrored those of Amy Peak’s at Butler – despite both studies using different approaches and being conducted (unwittingly) concurrently. I think those two studies were also the first to make people ask the same question you proposed – why do most hospitals automatically buy Micromedex?

The other study extended the previous by evaluating online drug info databases’ PDA counterparts (Pharmacotherapy http://www.atypon-link.com/PPI/doi/abs/10.1592/phco.27.12.1651). As you suggested in your post – Lexi performed the best in the PDA category. I also think it is notable that Lexi’s PDA products have fared well in other specialties since then, including an October 2009 article on nursing-specific databases where it tied for best performer (Int J Med Inform http://bit.ly/4AjHgX).

A final takeaway from the literature on this topic is the room for improvement in these tools – highlighted by the most recent (Oct 2009) study which demonstrated a 4% error rate across all databases.