Sample size re-estimation for clinical trials with longitudinal negative binomial counts including time trends

In some diseases, such as multiple sclerosis, lesion counts obtained from magnetic resonance imaging (MRI) are used as markers of disease progression. This leads to longitudinal, and typically overdispersed, count data outcomes in clinical trials. Models for such data invariably include a number of nuisance parameters, which can be difficult to specify at the planning stage, leading to considerable uncertainty in sample size specification. Consequently, blinded sample size re-estimation procedures are used, allowing for an adjustment of the sample size within an ongoing trial by estimating relevant nuisance parameters at an interim point, without compromising trial integrity. To date, the methods available for re-estimation have required an assumption that the mean count is time-constant within patients. We propose a new modeling approach that maintains the advantages of established procedures but allows for general underlying and treatment-specific time trends in the mean response. A simulation study is conducted to assess the effectiveness of blinded sample size re-estimation methods over fixed designs. Sample sizes attained through blinded sample size re-estimation procedures are shown to maintain the desired study power without inflating the Type I error rate and the procedure is demonstrated on MRI data from a recent study in multiple sclerosis.     

Use of colony-stimulating factor primary prophylaxis and incidence of febrile neutropenia from 2010 to 2016: a longitudinal assessment

Background: Guidelines recommend primary prophylactic use of colony-stimulating factor (PP-CSF) when risk of febrile neutropenia (FN) – based on chemotherapy and patient risk factors – is high. Whether and how PP-CSF use may have changed over time (e.g. due to guideline revisions, increasing use of myelosuppressive regimens, controversy regarding inappropriate CSF use), and whether there has been a concomitant change in the incidence of FN, is unknown.

Methods: A retrospective cohort design and data from two US healthcare claims repositories were employed. The study population included patients who had non-metastatic cancer of the breast, colon/rectum, lung or ovaries, or non-Hodgkin’s lymphoma (NHL), and who received myelosuppressive chemotherapy regimens with an intermediate/high risk for FN. For each patient, the first cycle of the first course was characterized in terms of PP-CSF use and FN episodes. Crude incidence proportions for PP-CSF and FN during the first cycle were estimated by calendar quarter (2010–2016); multivariable logistic regression models were used to estimate quarter-specific adjusted mean probabilities of FN by PP-CSF use.

Results: The study population totaled 142,730 patients with breast cancer (61%), colorectal cancer (14%), NHL (11%), ovarian cancer (10%) or lung cancer (5%). PP-CSF use increased from 52% in 1Q2010 to 58% in 4Q2016; pegfilgrastim was the most commonly used agent (>96% across quarters). PP-CSF administration on the same day as chemotherapy ranged from 8 to 11% until 1Q2015, and increased to 64% by 4Q2016. Adjusted incidence proportions for FN in the first chemotherapy cycle ranged from 2.7% (95% CI: 2.3–3.0) to 3.7% (95% CI: 3.1–4.3) among those who did not receive PP-CSF, and was 2.6% (95% CI: 2.5–2.7) across quarters among those who received PP-CSF.

Conclusions: Although the use of PP-CSF is commonplace in current US clinical practice, underutilization in cancer patients receiving chemotherapy regimens with an intermediate/high risk for FN may still be an issue. Use of same-day PP-CSF increased markedly from the end of 2015, although this finding reflects (at least in part) increased uptake of pegfilgrastim delivered via an on-body injector as well as the recent change in clinical practice guidelines. Overall, patients receiving PP-CSF appear to have a lower risk of FN during the first cycle of chemotherapy.