Are There Differences in Treatment Patterns, Cost, and Quality Indicators By Site of Care for First Line Treatment for Non-Hodgkin's Lymphoma and Chronic Lymphocytic Leukemia (NHL/CLL)?

Background: Hospital acquisitions of community clinics in the United States have led to a shift in oncology infusion therapy from physician office (PO) to hospital outpatient (HO) settings. Studies in commercially insured populations suggest that inherent differences between sites of care (SOC) can impact cancer treatment delivery and overall health care costs. This study utilizes a predominantly Medicare population to examine differences in treatment patterns, cost, and quality of care among patients with NHL/CLL receiving infusion chemotherapy and/or rituximab in a HO versus PO setting.

Methods: Patients ≥ 18 years initiating infusion therapy in 2008-2012 with at least 2 claims with a diagnosis of NHL or CLL occurring 30 or more days apart were identified from Humana medical claims data. The index date was the date of the first NCCN recommended infusion therapy. SOC attribution (HO vs PO) was based on where patients received ≥ 90% of their infusions. Differences by SOC in duration of first line treatment, number of infusions, and quality of care indicators, such as the use of infusions or hospitalizations within 30 days of death, were evaluated using Χ² and Wilcoxon Rank Sum tests. Median and interquartile range for duration of treatment and number of infusions by SOC were reported. Health care costs were determined by the sum of plan and patient costs for medical and pharmacy claims in the 6-months following the index date. Oncology costs, including supportive care, were computed from claims with cancer specific ICD-9 diagnostic codes. To control for the impact of case mix differences by SOC on costs, generalized linear models adjusting for age, sex, comorbidity, total health care cost in the pre-index period and geographic region were conducted.

Results: A total of 1,859 patients with a diagnosis for NHL or CLL were identified and 68% (1,262) received infusion therapy in the PO setting. Medicare beneficiaries represented 85% (1,587) of the study sample. Mean comorbidity index was higher among HO [3.7 ± 2.4 (SD)] compared to PO patients [3.3 ± 2.2 (SD)], p=0.0001. The proportion of patients receiving certain treatment regimens differed by SOC. Rituximab monotherapy was received by 24.5% (146) of HO and 14.1% (178) of PO patients, p<0.0001. Rituximab with chemotherapy was received by 63.3% (378) of HO and 72.7% (917) of PO patients, p<0.0001. Treatment regimens consisting of chemotherapy only did not differ by SOC, 12.1% (72) of HO and 13% (164) of PO patients, p<0.57. Treatment duration did not differ by SOC among those receiving any Rituximab therapy (p>0.05). Among those receiving chemotherapy only, treatment duration was shorter in the HO setting with 78.5 days (22.5-111) versus the PO setting with 112 days (59-162.5) p=0.0002. In the HO setting, there were fewer infusions for patients receiving chemotherapy and Rituximab, HO 6 (5-10), PO 7 (5-12), p=0.012 and chemotherapy only HO 6 (3-10), PO 7.5 (5-12), p=0.006.

In multivariate analyses, total healthcare costs were 22% higher among patients in the HO ($60,536) compared to the PO ($49,800) setting, p< 0.0001. Total oncology-related health care costs were also 24% higher in HO compared to PO, $58,033 versus $46,652 respectively, p<0.0001.

There were no statistically significant differences in the quality of care indicators by SOC, including the use of infusions or hospitalizations within 30 days of death among Medicare patients. Among 427 Medicare patients who died, use of infusions within 30 days of death was 25.7% (36) for HO and 23.7% (68) for PO,p=0.6479. Hospitalizations within 30 days of death occurred among 67.1% (94) of HO and 66.9% (192) of PO patients, p=0.9599.

Conclusion: This study, among the first to utilize a mostly Medicare NHL/CLL population, found differences by SOC in treatment patterns and cost, but not in the area of quality, defined as infusion and hospitalization within 30 days of death. Patients receiving care in the HO setting had a shorter duration of therapy and fewer infusions, but had higher total healthcare costs than those in the PO setting. As care shifts from the PO to HO setting, future studies should assess the impact of SOC on the delivery of care and health care costs associated with current therapeutic options for NHL/CLL.