Analytical pipeline conducts population-wide immunodeficiency screenings
Click Here to Manage Email Alerts
A two-step analytical pipeline facilitated the identification of individuals with primary immunodeficiency and accurately quantified their clinical risk, according to a study published in The Journal of Allergy & Clinical Immunology.
Other health care entities can adapt the freely available system for their own use as well, Nicholas L. Rider, DO, associate professor of pediatrics, division of clinical informatics, Liberty University College of Osteopathic Medicine and the Liberty Mountain Medical Group, and colleagues wrote.
The Software for Primary Immunodeficiency Recognition, Intervention and Tracking (SPIRIT) Analyzer uses an internal, pre-assigned weighting system for more than 1,400 ICD10 claim codes relevant to infectious complications associated with primary immunodeficiency (PI) and inborn errors of immunity (IEI) to assess entire populations of patients and generate risk scores for specific individuals in those populations.
The study involved 427,110 individuals enrolled in the Texas Children’s Health Plan as of Sept. 1, 2019. In the first step of the identification process, the SPIRIT Analyzer assessed this population at 6-month intervals for 30 months beginning in September 2020. Using 12-month blocks of data, the analyzer classified individuals as high risk, low-medium risk or no claim of interest.
The baseline analysis and March 2021 assessments identified 661 patients as high risk, with 49 already known to have PI/IEI. The researchers then assigned data from the other 612 patients to a control group to build a machine learning model. In the second step of this identification process, this model calculated risk ratios, or “PI ratios.” The PI ratios of high-risk patients next were ranked to prioritize their evaluation.
The researchers further assessed high-risk patients by reviewing claims data and electronic health record charts to see if they had a known PI/IEI, if they had already been referred to an immunologist, if they had a secondary risk for infection or unlikely PI/IEI or were candidates for PI/IEI evaluation.
In March 2022, the researchers compared the clinical endpoints of the cohort with their initial risk categories from September 2020. Analysis including PI ratios found 37 individuals who were candidates for clinical immunology evaluations, with 16 of them ultimately referred.
The researchers called this 16-patient total a manageable number for any health system clinical immunology team, driving cost savings by preventing unnecessary evaluations and low-yield testing.
The seven high-risk patients who were referred had significantly higher mean PI ratios than the 329 patients who were not referred (0.34 ± 0.24 vs. 0.27 ± 0.18; P < .001), indicating that this stratification scheme was useful, the researchers wrote.
However, the researchers noted that they do not know the diagnoses for the 16 patients who were referred and that the 21 candidates who were not referred are no longer in their health system.
DiGeorge syndrome was the most common precise PI diagnosis among the high risk and low-medium risk groups. Selective IgA deficiency, which the researchers called a much less complex condition, was the most common precise PI diagnosis in the group with no claim of interest.
Claims per individual rates included 8.4 for the high-risk group, 1 for the group with low-medium risk and 0.06 for the group with no claims of interest, suggesting that the high-risk patients may be more complex and have more data available for analysis for a clearer risk picture, according to the researchers. Or, they continued, high-risk patients may be at a more advanced state of evaluation.
Despite the differences in evaluations, the researchers noted, high-risk patients who were not directed toward clinical immunology evaluations will continue to be analyzed over time and still may be referred for evaluations. Also, the researchers wrote, all individuals with risk will be monitored longitudinally.
There is a need for additional work to prove this pipeline’s analytical utility in other subpopulations and across a larger segment of the general population, the researchers wrote. But health systems already can implement the SPIRIT Analyzer tool inexpensively to facilitate the early recognition of patients with PI/IEI, the researchers continued, preventing adverse clinical outcomes, containing costs and facilitating best-practice care.
Perspective
Back to Top
Nicholas L. Hartog, MD, FAAAAI, FACAAI
This is fantastic work by Rider and colleagues. Like many rare diseases, the delay to diagnosis is a significant issue in primary immunodeficiency. The years it takes to make a diagnosis of primary immunodeficiency leads to increased patient morbidity, mortality and increased cost among other things.
Any step to assist the diagnostic process and get a patient with a high chance of having a primary immunodeficiency in front of an immunologist quicker is a great thing for these patients. As immunologists, we chronically struggle with where these patients are “hiding.”
There is no screening for most primary immunodeficiencies, outside of newborn screening for severe combined immunodeficiency. All other patients must have a clinical suspicion prior to workup being started. Things are often missed for years before we see a patient, and we see this way too often.
This is a very important step in getting electronic medical records to work for us and the patients instead of acting only as a billing software. We have the ability to use this kind of system to identify patients at high risk for rare diseases. This is important proof it can be done.
I would like to see follow-up with diagnosis of the high-risk patients and if a genetic or other diagnosis was made. In addition, I would like to see follow-up on what treatment was started, how long from onset of symptoms it was until they were identified, and so on. This would help us to come up with cost saving analysis.
I also would expect this system will help with earlier diagnosis and likely will save a significant amount of money. In addition, it should be employed at a variety of health systems so it can be further validated in diverse populations.
Collapse
Read more about
Click Here to Manage Email Alerts