Barriers to Implementation of Evidence-Based Stroke Prevention in Sickle Cell Disease: A Preliminary Study

Introduction: Evidence-based guidelines for stroke prevention in sickle cell disease (SCD) strongly recommend chronic red cell transfusion therapy (CRCT) for patients at high risk for stroke. Patients are defined as high risk if they have had an abnormal transcranial Doppler (TCD) screen or have already suffered an initial stroke. Despite these guidelines, CRCT is likely underutilized. The barriers to implementation for CRCT are poorly understood but are likely due to a combination of patient and health-care-access variables. This study was performed to assess the implementation of CRCT in patients with SCD at high risk for stroke treated at a large, Southeastern hospital. It was hypothesized that age, provider type, and distance from comprehensive care would be associated with CRCT status. Additional factors, including indication for CRCT and insurance status, were also examined as exploratory variables.

Methods: A retrospective chart review was used to identify patients defined as high risk for stroke, their current CRCT status as well as additional demographics. High-risk patients were identified using current clinic rosters and previous stroke prevention trial databases from January 1, 2000 to December 31, 2014. Information collected included patient's current age, gender, indication for CRCT, last known zip code, insurance status, and provider type at last known encounter (SCD versus primary care or ED only). Demographics were compared between patients maintained on CRCT, patients still receiving care but not on CRCT, and patients who had been lost-to-follow-up (LTFU, treatment status unknown) using one-way ANOVA and Chi-square tests or Fisher's Exact tests. Age was viewed as a potential confounding variable for some of the characteristics (i.e., provider type, insurance status). Logistic regression was also used to examine associations between specific characteristics and CRCT status after adjusting for age.

Results: The patient population included 143 patients with SCD at high risk for stroke. At the time of the assessment, 13 patients were deceased. Thus, the remaining 130 patients were used for subsequent analysis (current age: M = 23.2, SD = 8.1; 68 male, 62 female). Of those living, 52% (68/130) were being treated with CRCT, 24% (31/130) were still being seen but were not receiving CRCT, and 24% (31/130) were LTFU. Older patients were more likely to be off CRCT or LTFU (p < 0.0001). Over 91% of patients on CRCT had a provider with training in SCD whereas 55% of patients not on CRCT had a provider without SCD training (p < 0.0001), and this effect remained significant after adjusting for age. Distance from comprehensive care did not predict CRCT status, but was related to those LTFU (p = 0.005). Of the patients who remained on CRCT treatment, indication for treatment was history of stroke for 62% of patients versus 48% of patients who were not being treated and 77% of those LTFU (p = 0.020).

Conclusions: Current NHLBI guidelines strongly recommend CRCT for patients with SCD at high risk for stroke; however, our study suggests that there has been poor implementation of this guideline. Despite the improved capacity to screen for patients at risk for stroke using TCD (per the STOP study), patients are not maintained on this therapy regardless of the lack of a safe stopping point. In our hospital, close to half of patients were either not receiving CRCT for stroke prevention or had been LTFU. In terms of patient and health-care characteristics, adult patients without a SCD specialty provider were less likely to be on CRCT for stroke prevention even if they continued to receive care in the hospital system. Indication for treatment was also associated with CRCT status, such that those with a history of prior stroke were more likely to be on treatment versus those with other neurological indications (such as abnormal TCD). The results were also notable for a distinct group of individuals at high risk for stroke who were LTFU. This particular patient group was mostly comprised of adult patients with a history of prior stroke who resided far from specialty care and may never have been transitioned into adult care. To further assess the barriers to dissemination and implementation of CRCT and further explore our preliminary results, we are planning a multi-site investigation, including assessment of patient-reported concerns and evaluation of different types of care environments for SCD.