Daktari Diagnostics has a big mission: to improve point-of-care diagnostics for patients in resource-poor markets. By using Quantrix to measure outcomes of its new medical diagnostic tool against the gold standard of laboratory instruments, the company is pinpointing the variances it must address in order to deliver consistently accurate diagnostic results.
Cambridge, MA-based Daktari’s first product is a CD4 cell counter, which rapidly analyzes blood samples to measure a specific type of white blood cell used in monitoring people with HIV. The Daktari CD4 system is being extensively field-tested in Africa. Existing diagnostic protocols test blood samples on a laboratory-based device that requires a highly trained operator and a multi-day turnaround for results. Patients in emerging markets like Africa often access healthcare services via local clinics and district hospitals, and it can be difficult to require them to make multiple trips just to get their test results days later. The Daktari device changes the equation by testing the blood samples on the spot, delivering results within 15 minutes.
“Our goal is to get the results of our device within 10 percent of the clinical values of the gold standard of the lab-based device,” says Fred Farber, Chief Technology Officer, Daktari. The Daktari device analyzes more than 10,000 unique measurements per sample. In addition, the study personnel who are processing the samples also track temperature, humidity, time between sample collection and processing, and other environmental and operator data. Any one piece of data – or combination thereof – can explain a variance from the gold standard result. Farber uses Quantrix to analyze those data points.
“Why did our device give us a ‘wrong’ answer? Why did it give us a ‘right’ answer? I need to identify the differences so that we can fine-tune our device,” says Farber. “I use Quantrix as a sophisticated cause-analysis tool to separate runs using both its computational and visualization techniques.”
Farber uses DataLink, the Quantrix data import module, to enter the 10,000 data points collected by the device for each sample, or 'run', as a separate dimension within the Quantrix model. In addition, each batch of user-collected data is imported; all data is tagged by sample ID. Where Quantrix is unique from other spreadsheet-based tools is its ability to model the data in multiple dimensions without creating new formulas for each run. Farber creates queries to calculate and graph the runs that are any chosen percentage outside of the gold standard, and can display them next to those that are within the range. “Quantrix allows me to look at each run as a different flavor of the same element, and process them as dimensions,” says Farber. “In Excel, each of those runs would be a discrete entity, and would require its own formulas.”
The Quantrix approach enables Daktari to rapidly process large volumes of data using a variety of scenarios. For example, if Farber wants to calculate the impact of room temperature or the time to process blood on the quality of results, he can drag and drop the formula into the matrix. “Rather than set up different calculations for each variation of a run, I can create a run as a dimensional array and drop it into different areas of the model,” says Farber. “Quantrix thinks in as many dimensions as you can set up, and you can take those dimensions and use them in a variety of ways. That makes it different from Excel.”
Another feature of Quantrix that Farber relies upon to analyze variances is its ability to select from multiple formulas in a model. He can create formulas to find runs within the minimum, maximum, and median variances from the gold standard, as well as those for any other percentage variation, and choose which ones to use to graph results. “I’m trying to see what is changing,” says Farber. “In Quantrix, I can leave all of my formulas in place and it is super simple to go back and forth.”
For Daktari, Quantrix is an important tool in its goal to improve healthcare. “Everything we’re doing is to improve point-of-care diagnostics,” says Farber. “Quantrix helps us with the root cause analysis to find out how to improve our technology.”