Hardware costs continue to plummet, and while that may be bad news for some device makers, it does bode well for developing countries where access to important sensory gadgets has public health implications.

A group at Harvard led by chemist George Whitesides published a paper in PNAS last week on how to build an inexpensive multi-purpose sensor. 

The idea is to make one low-cost device to perform a wide variety of chemical analysis jobs, such as measuring diabetics' glucose levels, the presence of malaria antibodies in the blood, and the concentration of electrolytes in urine or heavy metals in drinking water. Farmers, soldiers, government officials and healthcare workers all need to know such data points but the sensors are often contained in separate devices, or the tests must be sent off separately for lab results.

"Here you can do it on the spot," lead researcher Alex Nemiroski says. "Primarily what we've innovated on it is the hardware."

The group looked at standard sensors and tried to strip out every conceivable bell and whistle, not to mention recalibrated accuracy for clinical work, not necessarily more rigorous academic research. After all is said and done Nemiroski estimates the cost of the device at about $25 and "We're sure that price can be pushed down a lot more." They designed the device to perform a variety of electrochemical analytic techniques -- chronoamperometry, cyclic voltammetry, differential pulse voltammetry, square wave voltammetry, and potentiometry.

However, "We don't want this to remain an academic exercise," he says. The group has partnered with non-profit Diagnostic For All, which creates testing devices for developing countries, to begin the process of making these prototypes at scale.

Detecting large molecules like protein antibodies requires very different sensory pulses than when determining electrolytes -- perhaps counterintuitively, Nemironski says it's actually trickier to determine the presence of the former rather than the latter. However he notes that they have developed specificity to determine individual concentrations of cadmium, iron, and zinc.

"There's still work to be done on the biochemical side to make these tests," he says.

The group provides a list of ingredients and instructions in their paper for those that want to build the device themselves. It's not at the level where "anyone" can do it -- assembly requires soldering and basic programming knowledge to tell the device what kind of test to run. Nemironski says his group will soon release a library of code for the various tests so people don't have to program their own.

Already members of Nemironski's team have traveled to India to begin field trials and assess setting up a database that manages patient records. Getting that information out of the phone isn't a simple task. The Internet can be tough to reach in remote locations, cell networks change data standards and phones have varying capabilities. So the device transmits data over voice.

"To reach those people we need to reach all generations of technology, so we have to develop the IT strategy," he says.