Address Verification is a Life-saving Undertaking During the COVID-19 Pandemic
05/28/2020 by Juli Yaguda
As the COVID-19 crisis has gone on and the world is beginning to re-emerge into a reality where exposure risks increase, sharing accurate information between institutions has grown from crucial to life-saving. Against a backdrop of rising infection rates, health care professionals need to be able to identify those who may have been exposed to a carrier of the virus. This person-to-person effort, called contact tracing, is a proven means of slowing a pandemic’s spread and gathering data to help science learn more about the virus.
While contract tracing can take on many forms and can even leverage cell phone data to establish concrete longitude and latitude coordinates for areas at-risk of exposure, it doesn’t tell the whole story. For example, not everyone’s cell phone is always on, and older or poorer populations may not even have one. Plus, privacy concerns can prevent the gathering of location data. Amid such challenges, contract tracing needs the basics first.
Through contact tracing, hospital and public health staff interview a patient one-on-one to establish a list of those they may have interacted with when they were most infectious. From there, health care workers must contact these at-risk individuals as soon as possible to warn them of their exposure and, hopefully, minimize their future interactions. As more people become aware of their potential infected status, they avoid interactions with others to prevent further infections.
A massive undertaking like this requires an equally massive set of resources in the form of technology, people, and processes to be effective.
Acquiring personal data from a patient at the time of their admission to a hospital is a relatively straightforward task. However, medical staff regularly find that progressing to the next level of tracing their recent history introduces multiple complications.
When attempting to slow the spread of a contagious virus, every moment is critical. As patients struggle to remember key details for every person they’ve encountered, health care professionals can save precious time if they can resolve address information into an organized, standardized format. Effective address verification is vital to drawing the most useful information for contact tracing.
In cases like these, even partially remembered details can be useful in drawing a more complete picture of a virus hotspot. At first, that area includes only a specific patient and their location history. But that area then expands to who they have come in contact with and, from there, every person those individuals may have contacted as well.
Unless health care workers can resolve these incomplete details into a usable format to establish accurate geolocations, the only thing they may truly know is who is infected today.
In an effort to keep pace with the spread of the coronavirus (COVID-19) pandemic, health officials estimate a need for some 300,000 additional health care workers to interview infected patients and those with whom contact has been made. These interviews can yield crucial details in contact tracing to slow the infection rate.
Under these sorts of demanding conditions, training windows for those recording this data grow shorter. And with that, the potential for inconsistent data increases as well. Depending on available resources at a healthcare facility, health care workers can key information into a spreadsheet or take it down with pencil and paper. But in whatever format, there can be little quality control at the time of data entry. As a result, the work to identify potential viral hot spots becomes that much harder.
Data management software plays a key role in alleviating incorrect or incomplete data. Zencos has been working in the data management field for over 20 years incorporating the power of SAS Software. SAS software can process a set of unclear or incomplete addresses and run them through a basic data cleaning process that includes standardization with match-codes and an embedded knowledge base of standard definitions. From there, SAS data management software can ensure proper address formatting to ensure the data is ready for verification and validation.
Here, a data pack works within SAS to verify each address against U.S. Postal standards to find potential matches. If the data does not resolve, the software issues an alert that more input is required. Then, once the user has filled in the needed blanks, the end result is a clean set of street addresses in a standardized format consistent with existing postal data.
At this stage, health care workers can group the information then forward it to the Centers for Disease Control, for example, for network analysis. However, if critical address information for those at risk of infection isn’t clean and standardized, it may as well not exist.
With contact tracing at a premium, new resources are in development to slow the spread of coronavirus. This spring, Google announced it was producing a new app with Apple to track the locations of Android and iPhone users, and Microsoft announced a similar effort. Using Bluetooth signals, these apps aim to track mobile devices to determine who may have been exposed to a confirmed case of COVID-19.
But these technologically advanced efforts can’t fully perform the job of contact tracing. Even setting aside privacy concerns, tracking mobile phone locations doesn’t cast a wide enough net to control an outbreak. For instance, not everyone in a given population may have the right mobile device to generate tracking information. And without an in-person interview to explain the context of someone’s interaction, more false positives may result.
According to public health experts, contact tracing through person-to-person interviews is the most reliable strategy in slowing a pandemic’s spread.
From the format of a given person’s name or their street name, so much ambiguity can surround address information. In most cases, a human being would be able to validate similar-looking locations into a single record. But at the scale of millions of address records, the practical work behind the verification effort quickly grows unreasonable.
As the volume of records increases, a machine-learning system can prepare address data as it’s gathered so it is clean, organized, and available within a short period of time. Software can also track the residential history at a given address to confirm whether a given person has recently moved.
The coronavirus pandemic has revealed the varying ways institutions have been caught unprepared to manage a crisis of this magnitude. There’s still no substitute for person-to-person research to gather the information needed to stem the pandemic’s progress. Technology remains a key component of effectively processing all that data.
But unless that data is cleaned and prepared before being analyzed, the information can stall or even mislead when it’s needed most. And now is when it is needed most.