(1074) Insulin Pen Needle Fill Rates Among Insured Adults with Type 1 and Type 2 Diabetes in the US, 2019-2021: A Retrospective Claims Analysis

Objectives: Insulin pen needles (PNs) are intended for a single subcutaneous (SQ) injection; however, patients with diabetes often reuse PNs, a practice that can cause PN tip deformity, breakage, and contamination, and that has been significantly associated with lipohypertrophy and injection pain. The study objective was to compare the actual annual fill rate of insulin PNs with the expected fill rate (with single use) based on insulin regimen.

Methods: This retrospective study drew on US claims data (1-Oct-2018 to 31-Dec-2022) from the Aetna fully-insured Commercial (CFI) and Medicare Advantage (MA) databases. Eligible patients were ≥18 years old with type 1 or type 2 diabetes mellitus (T1D or T2D); had at least 1 claim for SQ insulin via PNs from 1-Jan-2019 to 31-Dec-2021 (index claim) with continuous medical/pharmacy eligibility for 3 months before index and 1 year after (follow-up), and, during follow-up, had ≥2 claims for their insulin regimen. We excluded patients receiving hospice or palliative care or using mail order prescriptions. The “expected” injections/day by insulin regimen were 4 for T1D basal+prandial insulin and, for T2D: 1 for basal long-acting insulin, 2–4 for basal+prandial, and 2 for premixed. Whether the annual median actual-to-expected ratio (A:E) for number of PNs equaled 1 was evaluated using t-tests with 2-sided p-values.

Results: In the CFI database (N=10,854), mean age was 53 years; 55% men; 10% T1D; mean social vulnerability index (SVI) was 0.46 (possible range 0 to 1=most vulnerable). The A:E medians ranged from 0.41 (T1D basal+prandial cohort) to 0.82 (T2D basal cohort; all p< 0.001). For T2D basal+prandial, A:E medians were 0.68 and 0.34 for 2 and 4 expected injections/day, respectively. Mean numbers of 100-count boxes/patient in 1-yr follow-up were 4.8 overall, and 5.5 vs. 4.0 with SVI of < 0.25 vs. ≥0.75, respectively. In the MA database (N=32,495; mean age of 70; 47% men; 2% T1D; mean SVI 0.49), the A:E medians ranged from 0.55 (T1D basal+prandial) to 1.10 (T2D basal and basal+prandial), with A:E median of 0.55 for 4 expected T2D basal+prandial injections/day (all p< 0.001). Mean boxes/patient in 1-yr follow-up were 5.5 overall, and 5.7 vs. 5.2 with SVI of < 0.25 vs. ≥0.75, respectively.

Conclusions: The actual number of PNs dispensed was fewer than the expected number by insulin regimen for most cohorts, suggesting the need for educational support to ensure an adequate supply of PNs with prescribed insulin regimens, particularly for more vulnerable communities, where fewer PNs per patient were dispensed.