Developing standardized sharing of clinical data as a method to assess the longitudinal retention of HIV treatment.

Methods

Summary

Clinical data will be collected and organized in a structured format in a data warehouse. Following the collection of the data, it will be standardized into a data model amenable to data analytics mainly through SQL queries of the data base to interrogate for specific diagnoses or procedures. 

Another critical activity will be the anonymization of the data using the privacy preserving record linkage software, this anonymization in tandem with the application of pseudo identifiers will form the basis of the privacy protections for patient participants. 

Challenges

One challenge I anticipate will be onboarding willing clinics or health systems - there are various factors that might prevent clinics or health systems from sharing their clinical data - some might include, lack of the required infrastructure to digitize the clinical data, another could be lack of appropriate clinical data warehousing applications or there might not exist appropriate governance structures to allow the sharing of the data. To resolve these challenges, we intend to work with each health system to identify any hinderances and resolve them using the project funding requested: 

1. In the case of lack of infrastructure, we will utilize some of the project funds to acquire affordable computers with sufficient storage space to hold the clinic data - the project does not require very high performing computers but rather reliable storage. Since we intend to partner with 6 clinics, the funds requested for equipment are sufficient to purchase sufficient disk storage to integrate in existing equipment. Otherwise, we will consider the use of cloud infrastructure for cold data storage as a cost saving measure.
2. We will work with clinics that dont have clinical data warehousing applications to adopt the use of open source clinical data warehousing applications. This is an affordable approach that will lead to the rapid digitization of the underlying clinical data.
3. Where governance structures are lacking,  the Data Protection and Privacy Act of 2019 will form the basis of formulating the governance of data collection and transmission -- the underlying principal being the protection of the patient privacy and the security of their sensitive information.

Pre Analysis Plan

We anticipate receipt of clinical data in the form of diagnoses, procedures, encounters and other typical clinical data. These data will  be normalized into a standard format or data modal amenable for structured query language (SQL) querying and stored in a data base. Organization of the data into this standard format will thereby render it available for data analytics to answer questions of relevance to health departments and other stakeholders.

For example a health department in a given county or district might be interested in the prevalence of a given condition - say measles or typhoid; with the data appropriately stored and rendered available for data analytics, it will be possible to query the existing data and create a report for the health department in a fairly rapid manner. 

The strength of this approach will depend on access to clinical data of as much of the patient population. Therefore, we anticipate that the significance of the conclusions will impacted by the data set we have access to.