App breaks down language barriers for patients, doctors

Jerrit Tan, a first-generation Chinese-American graduate of the University of Michigan, used to skip school to act as a translator for his immigrant parents and extended family when they visited a doctor.

“All my grandparents and uncles and aunts only spoke Chinese,” he recalled. “I saw how hard it was for people to get access to healthcare.”

That experience helps explain why, after helping to build Google's first New York City sales team, he quit to develop a smartphone app that provides medical interpretation services. Canopy Apps' first product provides access to translations of medical phrases in 15 languages, including Spanish, Arabic and Chinese.

A free version of the app has already been downloaded 80,000 times. Paying clients include Duke University Health System, UNC Health Care, the Visiting Nurse Service of New York and the University of Pennsylvania Health System, whose physicians and staff can use the advanced version.

Caring for the millions of Americans who speak little or no English poses huge challenges for healthcare providers. A 2012 report from the Agency for Healthcare Research and Quality found patients with limited English proficiency experience longer hospital lengths of stay and greater risks of infections, falls, pressure ulcers and readmissions, compared with English-speaking patients with similar conditions.

Hospitals are required to use interpreters when non-English-speaking patients arrive at their doorsteps. But in many locations, interpreters can be difficult to reach. Telephone services often have poor audio quality or present a logistical nightmare in a physician's office or a hospital room.

To create its smartphone-based solution, New York City-based Canopy assembled a library of translated medical phrases. The user pages through a number of categories to find the appropriate phrase, which is then read in the selected language by a digitally recorded voice. When an interpreter is needed, the app connects providers to the hospital's on-call interpretation service through the smartphone.

Dr. Raviraj Patel, an emergency medicine physician with Dallas Methodist Hospital, signed up for the Canopy app and service after growing frustrated with the three-way headsets that connected him, the patient and a remote translation service. The sound quality was poor, and the three-way phone conferences often resulted in long, convoluted conversations that bogged down operations in the busy emergency department. He sometimes had to wait for one of the few available headsets.

But the Canopy app's medical dictionary proved ideal for the quicker, more rudimentary conversations needed to obtain basic information. The availability of prompt access to phone interpreters served as a backup. Although Patel speaks Spanish—the first language of many ED patients at the Dallas facility—he said the app is helpful if he forgets the Spanish equivalent for a particular medical phrase.

“Sometimes medical terminology eludes you,” he said. “A simple thing like 'gall bladder' or 'urinary bladder' can be at your fingertips and (it) can help you pull up those words in multiple languages.”

Tan pitches his service to hospitals based not only on the interpretation app, but also the device's ability to rate the quality of existing interpreters. “As a hospital exec, you can say, 'Our interpreters are low quality. That's going to cost us a lot of money when they make a mistake and we get sued,' ” he said. It also helps administrators track how many foreign-born patients arrive at their hospital, when they arrive and what languages they speak. That data enable the hospital to schedule interpreters more efficiently.

Canopy Apps launched in 2010, and has received $2 million in grants from the National Institutes of Health and a roughly $100,000 investment from technology incubator Pilot Health Tech NYC, an organization funded by the city of New York.

Canopy faces some tough competition, including its CEO's former employer. Many doctors rely on Google Translate, which uses machine learning to automatically translate phrases. But a December 2014 study in the BMJ journal testing Google Translate in various European languages found only a 57.7% accuracy rate on a number of medical phrases.

Tan and his Canopy team hope they'll eventually be large enough to undergo similar scientific scrutiny.

Follow Darius Tahir on Twitter: @dariustahir