Importance Of Defining Use Cases

Defining the use case for a PIM system is important because it determines the personnel, infrastructure, and other resources needed and the privacy and security measures required for implementation, including any need for data-sharing agreements or a linkage honest broker. Use cases for a single site’s data, for instance, have many fewer requirements than multi-source use cases, especially when personally identifiable information (PII) is involved, which can exponentially increase the requirements.

The specified use case and its requirements will also influence subsequent decisions about how the data are handled and assessed. For example, accuracy is important to all use cases, but the particular use case will help determine the relative importance of the metrics used to assess the accuracy of the matching process.

In assessing accuracy, some use cases might be more tolerant of false positives (non-matches incorrectly identified as matches), while others might be willing to accept more false negatives (linkages identified as non-matches that, in fact, are matches). Clinical care use cases are most concerned about eliminating false positives since incorrectly merging two different patients' data could lead to serious adverse consequences. However, public health use cases are often most concerned with eliminating false negatives. In public health, one might say, for instance, that it’s more important to identify all people who may be asthmatic at the risk of including some non-asthmatics versus missing some people who are actually asthmatic. Also, public health use cases must be concerned about scalability, not just accuracy, which will determine the statistical methods used.

In addition, specifying why being able to uniquely identify a patient is important should be part of defining the use case. For example, identifying a unique patient across multiple records or multiple sites may be important to ensure all information about that patient is considered in providing informed care or to avoid duplicate testing. Or, if the use case involves deduplication (described in next section), the reason may be to support patient care, or it could be to ensure accuracy and avoid overreporting in research or public health surveillance.

More broadly, having a comprehensive picture of the universe of PIM use cases and their key determinants supports systematic, well-informed decision-making and effective planning, whether at the site, regional, state, national, or international level.