On Oct. 13, 2008 Hospira announced that it had acquired the EndoTool business from MD Scientific. (Press release) The EndoTool glucose management system is software used to determine optimal insulin dosages to help  establish and maintain glycemic control. Target markets for the product include critical care and surgery, as well as lower acuity areas on hospitals. Hospitals are also considering use EndoTool in Labor and Delivery. The product was launched 18 months ago by MD Scientific, and seen increadible adoption (60 hospitals currently). The product won't be "relaunched" under the Hospira brand. You can read the publicly available FDA 510k stuff here.

Software designed to support the application of clinical protocols has been in the works from various vendors. Patient monitoring examples include Philips Protocol Watch, soft-launched back in February 2007, and . These applications automate what are otherwise onerous manual calculations with data acquired from medical devices and integrated with data from other information systems. This is workflow automation of the most important kind, diagnosis and therapy delivery. These applications are typically regulated as Class II medical devices.

Last week I spoke with Philip Settimi, MD, vice president of global strategic marketing for Hospira. According to Settimi, "EndoTool replaces spreadsheets of physician preferences and worksheets full of manual calculations for managing patient glucose levels." Such manual methods are obviously inefficient, but also susceptible to human error. This approach provides an effective tool to impose a controlled and centralized tool for managing tight glycemic control (TGC). Endo Tool comes with a specific protocol based on sophisticated algorithms to support glucose management. The key: taking all that complexity (the calcs) at the point of care and automating those 33 different non-linear equations.

In a typical use model, the end user selects the patient to establish patient context, enters one or two blood glucose data points, the software recommends dosing and when to test glucose again. The system, after a couple data points, creates a mathematical model of the patient's physiology (physiological curve) and allows for the more effective and more rapid zeroing of the patient into the ideal zone. Thus the system constantly adjusting up and down the physiological curve.

It is easy to automate part of a workflow process with some computer software. To fully automate a workflow beyond calculations - to include patient demographics, orders, charting and reporting - that requires integration with bedside medical devices (sometimes more than one) and other information systems. Surprisingly, EndoTool already has interfaces with ADT (admission, discharge and transfer) for patient demographics. Currently the user has to manually enter glucometer test results.

System architecture is always a point of interest in products like this. EndoTool works on any standalone or networked PC in hospital and uses existing enterprise networks. In network configurations, clients communicate with a server. Unlike a product that automates a paper process, this system scales up to support a large patient population, and interfaces.The server runs a centralized database. Client applications can be installed remotely for lower operating costs. Hospira provides on site installation and training.

Moving the software to the pump is also a possibility. Settemi noted that he sees value in integrating EndoTool with MedNet and possibly the pumps themselves. "The notion of having decision support technology like this, embedded in the pump, is a potential productivity and patient safety enhancement," said Settimi. One big caveat is the resulting user experience. Hospira's new Symbiq pump's large touch screen is well suited for an application like this. He expects to see a hybrid model like patient monitoring, where the user interface is on the device at bedside with some or all of the software running on a server somewhere on the network.

Infusion pumps are used to deliver many high risk meds; many of these are titrated and could benefit from both automated workflow for managing the titration, and medical device interoperability between the pump and protocol software.  Settimi noted that Hospira sees an opportunity to offer a more expanded menu of protocols in the future. The system manages a broad range of patients with their current protocol: peds, adult diabetic, non diabetic, DKA, etc. There's an opportunity from a technology perspective, yet from a user perspective there's a big benefit in a common approach. Hospira believes that Endo Tool represents a technology platform, beyond TGC, to include narcotics, coagulation therapy, and a broad array of infusion therapy administration.

I mentioned Philips' Protocol Watch earlier as another example of automating protocols used in patient care. Another example is Visensia from OBS Medical. Visensia applies a proprietary algorithm for analyzing the physiological parameters from a patient monitor to identify patients with a deteriorating clinical condition. Unlike TGC protocols that are peer reviewed and published, the algorithm in Visensia is a trade secret - sort of like a black box.

Vendors and entrepreneurs are working on lots of different systems to support protocols of various types. Many of these vendors are approaching the problem from a specific diagnosis or other clinical domain. The protocol automation market will probably evolve like many others, where various solutions come to market targeting specific clinical applications. Over time, solutions will broaden in scope in an effort to apply their architecture to a broader array of tasks. Eventually there will emerge generalized protocol support systems that are used throughout a hospital.

Some protocol solutions will come from vendors, looking to extend their proprietary end-to-end solutions. Many new protocol systems will also come from software companies. Who knows, maybe even some existing health care IT vendors will bite the FDA bullet and enter the protocol automation market with their own regulated "medical devices." In any event, this will be an interesting market segment to watch, and one with significant clinical value and impact on patient safety.