Distributed data processing for public health surveillance

Lazarus, Ross; Yih, Katherine; Platt, Richard

Although the data being transferred to the datacenter is arguably not identifiable PHI because of the high level of aggregation, we use the Public Health Information Network Messaging System [8] (PHINMS), a freely available, secure, data transfer software suite developed by the CDC, to transfer aggregate data. A PHINMS server operates at the datacenter and each data provider operates a PHINMS client, using a security certificate supplied by the datacenter for encryption and authentication. PHINMS allows fully automated operation at both the datacenter and at each data provider. PHINMS communicates over an encrypted channel and usually requires no special modification to the data provider firewall, since it is only ever initiated by an outgoing request (the data provider always initiates the transfer of new data) and uses the same firewall port and protocol (SSL on port 443) as commercially encrypted services such as internet banking. PHINMS is reasonably robust to temporary connectivity problems, as it will try to resend all messages in the queue until they are delivered. Data transmission is one of the least problematic aspects of maintaining this system. We provide automatic installation software and it runs more or less instantaneously and transparently, without intervention in our experience. No training is needed as the process is fully automated.