Prediction of High Acute Care Utilization in Children with Sickle Cell Disease

Background: A small percentage of patients with sickle cell disease (SCD) account for a high percentage of acute care utilization and painful vaso-occlusive episodes (VOEs) requiring hospitalization. Chronic pain may develop in patients with SCD who experience frequent VOEs. Little is known about the mechanisms mediating the development of chronic pain in SCD, and there are no reliable clinical predictors for the development of a chronic pain phenotype. We aimed to determine the factors associated with the development of chronic pain phenotype characterized by high acute care utilization, in children and adolescents with SCD.

Methods: We performed a single-institution, retrospective case-control study of all patients aged 10-25 years old who received SCD care at our institution from 2016-2020. Patients were identified using ICD-10 codes for SCD. For patients meeting inclusion/exclusion criteria, we extracted demographic, laboratory, and opioid prescription data. ED visits and hospitalization encounters during which parenteral opioids were administered were identified in the electronic medical record. Distribution of acute care encounters among all included patients was used to establish cases who met criteria for high acute care utilization, defined as > 85^th percentile for all patients. Laboratory data were collected from the index year, defined as the first year a case met criteria for high acute care utilization, or from the matched year within 1-2 year window, for controls, as well as from the year prior. Chi-square and student t test were used to compare categorical and continuous variables, respectively, between cases and controls using SAS (Cary, NC). ROC curves were created to determine the discriminatory ability of using opioid prescription data to identify patients meeting criteria for high acute care utilization.

Results: A total of 200 patients with SCD (all genotypes) met inclusion/exclusion criteria for the study. The 85^thpercentile cut point corresponded to ≥4 acute care encounters in one year. Using this definition, 36/200 (18%) patients who were considered "high acute care utilizers" and had ≥4 acute care visits during any 1-year period during 2016-2020 were chosen as cases. Matching 2:1 by genotype and calendar year, 72 "low-utilizers" were randomly selected as controls among the 127 patients who had ≤1 acute care visit during each year of inclusion. Age, sex, and hydroxyurea use did not differ among cases vs controls. We found a higher proportion of patients with Medicaid insurance (83% vs 62%, p=0.03), as well as asthma (33% vs 11%, p=0.008) and anxiety or depression (39% vs 2%, p<0.001), among cases vs controls. Steady-state laboratory variables did not differ significantly among cases vs controls. Total opioid prescriptions were significantly higher in both index year [7.7 (95% CI: 5.2-10.2) vs 1.3 (0.9-1.7), p<0.001] and in the year prior [4.6 (95% CI: 1.9-7.2) vs 1.0 (0.7-1.3), p=0.01] among cases vs controls. Total morphine equivalents during the index year utilizers [3540 (95% CI: 1010-6060) vs 554 (355-753), p=0.02] and in the year prior [1386 mg (95% CI: 597-2170) vs 448 mg (249-648), p=0.01] were significantly higher among high acute care utilizers. ROC curves for average total opioid prescriptions and total morphine equivalents prescribed in the year prior to meeting criteria for high acute care utilization demonstrated AUCs of 0.80 and 0.81, respectively. Using these ROC curves, a cutpoint of ≥2 opioid prescriptions in the prior year was associated with a maximal sensitivity of 75% and a specificity of 76%. A cutpoint of ≥400mg of total morphine equivalents in the year prior was associated with a maximal sensitivity of 81% and a maximal specificity of 71%.

Conclusions: We found no relationship between laboratory variables and high acute care utilization in children and young adults with SCD. Demographic and clinical factors associated with high acute care utilization include Medicaid insurance status, asthma and mood disorders as well as number of opioid prescriptions in the year prior to meeting our criterion for high acute care utilization. Future studies should be conducted to prospectively identify predictors of high acute care utilization and the development of chronic pain in SCD. The role of opioids for the treatment of chronic pain in SCD and the relationship of opioid to the development of high acute care utilization and chronic pain should also be further studied.

Liem:BlueBird Bio: Research Funding; GBT: Research Funding; NHLBI: Research Funding; Editas: Research Funding.