Development and evaluation of an algorithm to identify users of Prolia(®) during the early postmarketing period using health insurance claims data

Capture of biologic and biosimilar dispensings in a consortium of U.S.-based claims databases: Utilization of national drug codes and Healthcare Common Procedure Coding System modifiers in medical claims

PURPOSE

To assess the capture of biologics (originator and biosimilar) in the Biologics and Biosimilars Collective Intelligence Consortium (BBCIC) Distributed Research Network (DRN), with a focus on medical claim National Drug Code (NDC), a new data field, and Healthcare Common Procedure Coding System (HCPCS) modifier.

METHODS

We conducted a repeated cross-sectional study among patients with medical and pharmacy benefits enrolled in insurance plans participating in the BBCIC DRN between 1 January 2013 and 30 September 2017. We calculated the proportion of medical claims with ≥1 NDC and identified select biologics using four different approaches: (a) specific HCPCS alone, (b) specific HCPCS and NDC, (c) non-specific HCPCS with NDC, and (d) HCPCS with modifiers (applicable to biosimilars). Numbers of dispensings were calculated for each biologic by approach and select patient and claim characteristics.

RESULTS

More than 1.5 million eligible participants contributed approximately 4 million person-years of data, including 1.2 billion medical claims. The proportion of medical claims with ≥1 NDC increased from 1.2% in 2013 to 3.0% in 2017. Medical claim NDCs identified 39% and 28% of vedolizumab dispensed in 2014 and 2015 and 30% of Epogen/Procrit dispensed overall. Out of 26,381 filgrastim biosimilar dispensings identified, 51% had a HCPCS modifier and 12% had a medical claim NDC for Zarxio. HCPCS modifiers and medical claim NDCs were present for 38% and 3% of all infliximab biosimilars dispensed (total n = 1,244).

CONCLUSIONS

Medical claim NDC and HCPCS modifier improves identification of select biologics without product-specific HCPCS code, thereby facilitating product-specific biologic research.

Medication persistence and risk of fracture among female Medicare beneficiaries diagnosed with osteoporosis

We examined the relationship between persistent osteoporosis medication use and fracture risk among female Medicare beneficiaries diagnosed with osteoporosis using Medicare claims, 2009-2012. Persistent use was associated with reduced risk of fracture and significantly lower total health care costs in the follow-up period. Results were consistent using different analytical methods.

INTRODUCTION

This study aimed to examine the relationship between medication persistence and fracture risk among female Medicare beneficiaries diagnosed with osteoporosis.

METHODS

Elderly female Medicare beneficiaries diagnosed with osteoporosis and initiated on osteoporosis medication January 1, 2009-June 30, 2011, were included. Persistent medication use was defined as continuous use (no gap ≥ 60 days) for 1 year or longer. The key outcome was fragility fracture. A difference-in-difference analysis was performed at the log scale of fracture rate using a Poisson regression model with months 1-6 before medication initiation as the pre-initiation period and up to 18 months after as the post-initiation period. Total health care costs were compared using a similar approach. Sensitivity analyses were conducted using different pre- and post-initiation periods.

RESULTS

The study included 294,369 patients; 32.9% were persistent osteoporosis medication users and 67.1% non-persistent (< 12 months continuous use). Fracture incidence rates were 16.2 per 100 patient-years pre-initiation and 4.1 post-initiation for persistent users; corresponding rates for non-persistent users were 19.0 and 7.3 per 100 patient-years. The adjusted post-/pre-initiation fracture rate ratios were 0.284 for persistent and 0.411 for non-persistent users. The ratio of the two rate ratios was 0.692 (persistent vs. non-persistent, p < 0.0001), suggesting a significantly greater fracture rate reduction for persistent users. Adjusted cost ratios were significantly lower for persistent users. Sensitivity analyses results were similar.

CONCLUSIONS

Persistent use of osteoporosis medications was associated with reduced risk of fracture and significantly lower total health care costs. Payers and patients would benefit from interventions aimed at improving medication persistence.

Real-world effectiveness of osteoporosis therapies for fracture reduction in post-menopausal women

Studies examining real-world effectiveness of osteoporosis therapies are beset by limitations due to confounding by indication. By evaluating longitudinal changes in fracture incidence, we demonstrated that osteoporosis therapies are effective in reducing fracture risk in real-world practice settings.

INTRODUCTION

Osteoporosis therapies have been shown to reduce incidence of vertebral and non-vertebral fractures in placebo-controlled randomized clinical trials. However, information on the real-world effectiveness of these therapies is limited.

METHODS

We examined fracture risk reduction in older, post-menopausal women treated with osteoporosis therapies. Using Medicare claims, we identified 1,278,296 women age ≥ 65 years treated with zoledronic acid, oral bisphosphonates, denosumab, teriparatide, or raloxifene. Fracture incidence rates before and after treatment initiation were described to understand patients' fracture risk profile, and fracture reduction effectiveness of each therapy was evaluated as a longitudinal change in incidence rates.

RESULTS

Fracture incidence rates increased during the period leading up to treatment initiation and were highest in the 3-month period most proximal to treatment initiation. Fracture incidence rates following treatment initiation were significantly lower than before treatment initiation. Compared with the 12-month pre-index period, there were reductions in clinical vertebral fractures for denosumab (45%; 95% confidence interval [CI] 39-51%), zoledronic acid (50%; 95% CI 47-52%), oral bisphosphonates (24%; 95% CI 22-26%), and teriparatide (72%; 95% CI 69-75%) during the subsequent 12 months. Relative to the first 3 months after initiation, clinical vertebral fractures were reduced for denosumab (51%; 95% CI 42-59%), zoledronic acid (25%; 95% CI 17-32%), oral bisphosphonates (23%; 95% CI 20-26%), and teriparatide (64%; 95% CI 58-69%) during the subsequent 12 months.

CONCLUSION

In summary, reductions in fracture incidence over time were observed in cohorts of patients treated with osteoporosis therapies.

Assessment of off-label use of denosumab 60 mg injection during the early postmarketing period using health insurance claims data

Off-label use of denosumab 60 milligram (mg) injection was assessed within an administrative claims database. The completeness of claims to assess off-label use was investigated with medical record review. Potential denosumab 60 mg off-label use was observed based on claims, but many had evidence of on-label indications based on medical record review.

INTRODUCTION

Denosumab 60 mg injection is approved in the USA to treat patients at high fracture risk due to postmenopausal osteoporosis, male osteoporosis, and hormone therapy for the treatment of prostate and breast cancers. Its RANK ligand-inhibiting effect makes it a candidate for the off-label treatment of other conditions mediated by the rate of bone resorption by osteoclasts. To better understand its utilization patterns, we assessed off-label use of denosumab 60 mg within an administrative claims database.

METHODS

Definite, probable, and possible denosumab 60 mg users were identified during the early postmarketing period within a claims database of a US healthcare insurer. Medical record review confirmed a sample of these users. Off-label use among definite and probable users and all chart-confirmed users was classified using claims-derived age, dose interval, and diagnosis and treatment received relative to the administration date. Among chart-confirmed users classified as off-label, patient characteristics related to treatment indication were abstracted from medical records to investigate the completeness of claims to study off-label medication use.

RESULTS

Off-label use was identified based on claims in approximately 25 % of definite and probable denosumab 60 mg users and 35 % of chart-confirmed users. Medical record review identified evidence of on-label indications in 81 % of chart-confirmed users classified as off-label in claims.

CONCLUSIONS

Many of the off-label denosumab 60 mg users had diagnoses or treatment consistent with on-label indications based on medical record review, suggesting these are under-recorded in claims data. It is warranted to be cautious when using administrative databases to assess off-label medication use.