Hydroxyurea use in patients with sickle cell disease in a Medicaid population

This is the authors' abstract. We don't add key points for this paper.

The purpose of this study was to examine adoption and utilization of hydroxyurea (HU) in a Medicaid population of adults with sickle cell disease (SCD). A retrospective cohort study of SCD patients aged 16–64 years, enrolled in Florida Medicaid program during a 5-year period from 2001 to 2005, was performed. HU adoption was determined by the presence of at least one HU pharmacy claim using National Drug Codes. Of 2301 adult patients with SCD, nearly 17% of the cohort (n = 384) had at least one pharmacy claim for HU. Furthermore, 33% (n = 769) of SCD patients were eligible to receive HU, identified by three or more hospitalizations for crisis in a 12-month period, of whom 38% (292/769) received at least one HU prescription. Finally, for patients who received at least two subsequent HU prescriptions (n = 241), 15.4% (n = 37) had a medication possession ratio of ≥ 80% of their cumulative daily dose. The prevalence and the persistence of HU use in this Medicaid population were low. Our results suggest that a small subset of indicated SCD patients received HU prescriptions consistently. Early therapy drop out and low adherence rates were common in patients prescribed HU. Hydroxyurea (HU) is the only Food and Drug Administration approved oral agent to modify the disease course in sickle cell disease (SCD) [1]. Efficacy data from the Multicenter Study of Hydroxyurea (MSH) in sickle cell anemia supports HU use in symptomatic patients with sickle cell anemia [2]. Follow-up studies of MSH patients also showed cost effectiveness and improved survival for HU users [3-5]. Despite the efficacy and cost-effectiveness evidence obtained from the MSH, there is consensus that HU use is suboptimal in treatment of adults with SCD [6]. A study highlighting underuse of HU was published by Lanzkron et al. [7]. Examination of Maryland hospitalization data for sickle cell anemia between 1995 and 2003 revealed an increase in the number and costs of hospitalizations despite HU receiving an expanded indication for treating symptomatic patients with sickle cell anemia. Examination of pharmacy data from a Medicaid managed care organization in Maryland covering 2001–2005 found that 85.9% of patients with SCD never had a claim for a HU refill [8]. Poor adherence to medication in chronic disease patients may contribute to increased hospitalizations and medical costs [9, 10]. A recent publication addressing North Carolina Medicaid recipients with SCD found that only 35% of patients receiving HU were adherent based on a calculated medication possession ratio (MPR) [11]. By examining Medicaid claims high resource utilization has been documented for patients with SCD in Florida [12]. The objective of this study was to describe HU adoption and adherence in a Florida Medicaid population and to identify determinants of HU use in adults SCD patients. A total of 2301 Medicaid beneficiaries diagnosed with SCD were identified. The majority were female (63.5%), black (85.1%), with a mean age of 25.6 years (10.9 SD, 16.0–63.3 range). Approximately 73% (n = 1673) of the patients in the sample had at least one emergency department (ED) visit, and 88% (n = 2025) had at least one inpatient hospitalization. In addition, 36% (n = 827) received red blood cell transfusions, and only 4% (n = 101) received deferoxamine. Of all SCD patients, 33.4% (n = 769) had at least three inpatient hospitalizations for SCD with pain crisis in any 12-month period, and 17% (n = 393) received a prescription for a long-acting opioid medication. Finally, nearly 17% (n = 384) of the cohort had at least one prescription claim for HU during the study period. Of those 2301 SCD patients identified by ICD-9-CM codes representing hemoglobin SS and other genotypes, 33% (n = 769) were considered potential candidates for HU based on documentation of at least three inpatient hospitalizations for SCD-related pain crisis within a consecutive 12-month period. Table I summarizes patient demographics and medical resources utilization for potential candidates for HU therapy. Candidates for HU therapy were ∼25 years old (9.7 SD), with 87% (n = 688) blacks and 57% (n = 439) women. In addition, 86% (n = 662) had ED visits, while 64% (n = 495) received red blood cell transfusions, 39% (n = 302) received long-acting opioids, and 7% (n = 52) received deferoxamine. Prevalence of HU use was low; only 38% (n = 292) of patients had at least one prescription claim for HU. Hence, 62% (n = 477) of patients did not have any prescription for HU, indicating potential under use of this therapy. SCD patients were stratified as HU users and non-HU users. Overall, sufficient evidence existed that race (χ2 = 4.5, P = 0.034), red blood cell transfusions (χ2 = 16.3, P < 0.0001), iron chelation (χ2 = 11.1, P = 0.001), long-acting opioids (χ2 = 39.5, P < 0.0001), duration of Medicaid eligibility (t-test = −6.92, P < 0.0001) were significantly associated with HU use. A higher proportion of HU users also received red blood cell transfusions (73 vs. 59%), iron-chelation therapy (11 vs. 4%), or long-acting opioids (53 vs. 31%) when compared to non-HU users. Moreover, HU users had longer Medicaid enrollments than non-HU users (54 vs. 44 months). Conversely, the characteristics of HU users and non-HU users were comparably similar by patient age-groups, gender, and use of emergency department services. Of the 384 SCD patients with at least one HU prescription, 83% (n = 322) received one or more HU prescriptions with 28–30 days supply however, the number of prescriptions dispensed varied. Of these patients, 70% (n = 227) received at least three HU prescriptions during the study period; conversely, at least 30% (n = 95) discontinued HU therapy after two prescriptions. Of the 384 HU users, 63% (n = 241) received two or more HU prescriptions with a at least 28 days and dispensing lapses less than 90 days. Of these patients, 70% (n = 168) took less than 60% of their cumulative daily dose and had frequent lapses in their prescription refills. Conversely, ∼15% (n = 37) had a high MPR (MPR ≥ 80%), while an additional 15% (n = 36) had a MPR between 60 and 79%. See Figure 1 (Supporting Information). Adjusted odds ratios summarized in Table II indicated that SCD patients who received red blood cell transfusions (OR = 1.62, 95% CI = [1.15, 2.27]), iron chelation (OR = 2.25, 95% CI = [1.23, 4.12]), long-acting opioids (OR = 2.43, 95% CI = [1.96, 3.03]) had higher odds of using HU. Patient demographics (race, age, and gender), and use of emergency department medical resources did not have a significant effect on the odds of HU use. Our retrospective study of Florida Medicaid recipients found a low prevalence of HU use, which was consistent with previous studies by Lanzkron et al. [7, 8]. In the examination of the Maryland Medicaid data only 38% of patients who averaged 3 or more hospitalizations for every 12 month enrollment period had a HU refill claim [8]. A recent publication by Candrilli et al. using similar methodology as in our study examined 2000–2008 data for North Carolina Medicaid recipients. They reported that 35% of patients under the age of 65 had a HU MPR ≥ 0.80 which is significantly higher than the 15% for the Florida patient population. This difference could possibly be accounted for by better adherence of patients under age 16 who were included in the North Carolina study. In addition, there were two academic centers in North Carolina that participated in MSH which may have influenced the uptake of HU therapy in the state once the results of the study were available [2]. However, both studies indicate adults with SCD who would potentially benefit from HU were not receiving treatment. Patients who started on HU but discontinued therapy after two prescriptions as identified in this study would not accrue the potential benefits of treatment. Furthermore, only a small subset of the Florida patients received HU prescriptions consistently and had few lapses in their prescription refills. This observational study used Medicaid administrative claims data. Our results may not be generalizable to non-Medicaid SCD patients. Also, there are limitations in discerning disease severity from analysis of administrative claims data without verification of findings such as radiologic evidence of lung infiltrates in the diagnosis of acute chest syndrome. Thus, the criterion for HU therapy was hospitalizations for crisis. Finally, in the absence of actual information on patient compliance with their medications, using the medication possession rate may actually over-estimate patient compliance with HU. In conclusion, our observational study provides insights into HU use in adults diagnosed with SCD in a non-managed care Medicaid population 6–10 years after publication of the MSH trial. HU appeared to be underused in this population, as a substantial number of patients who had frequent painful episodes did not receive treatment during the study period. Furthermore, only a small subset of patients received prescriptions consistently, which agrees with the findings of Candrilli et al. [11]. It also emphasizes the importance of improving translation of evidence-based treatment guidelines into clinical practice and disseminating patient education in order to increase adoption and promote adherence to HU therapy in SCD patients. This retrospective cohort study used claims data from Florida Medicaid fee-for-service program from January 1, 2001 to December 31, 2005. The dataset includes information on monthly eligibility as well as any medical and pharmacy claims for medically-needy or low-income beneficiaries submitted to Medicaid for reimbursement. The study was approved by the University of Florida Institutional Review Board. Please see Supporting Information for methodology pertaining to subject identification and statistical analysis. Measures of medication persistency were defined as follows. Therapy adoption was defined as having at least one or more HU prescriptions dispensed with a drug supply of 28 days or longer. Therapy adherence was determined by calculating the MPR, defined as the cumulative daily dose dispensed (not including the last prescription), divided by the time period between first and last prescription refills dispensed for patients with at least two prescriptions for 28 days supply or more [13]. Because any treatment provided during inpatient hospital stays was not captured in the outpatient pharmacy claims data, length of hospitalization (days) was subtracted from the denominator (duration of treatment), MPR was categorized by quintile interval. A MPR of 100% indicated that the patient had the drug dispensed to complete the course. Additional details of methods are presented in Supporting Information. The authors would like to acknowledge the assistance of Dionne Mayhew PhD. No external funding or financial support was received for the study. Intangible support for the study was provided by the University of Florida Center for Medicaid and the Uninsured. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article. Jane Ritho*, Huazhi Liu*, Abraham G. Hartzema*, Richard Lottenberg , * Department of Pharmaceutical Outcomes and Policy, University of Florida, College of Pharmacy, Gainesville, Florida, Division of Hematology/Oncology, University of Florida, College of Medicine, Gainesville, Florida.