At baseline, practice sites complete a Baseline Drug Shortage Survey and Pharmacy Baseline Survey; Drug Shortage Incident Reports are completed in real time as cancer care delivery problems occur; and the Quarterly Follow-Up Survey Number Treated Report every 3 months for one year (4 total).

Ancillary studies

Each incident is defined by the drug in shortage and by each cancer care delivery decision (e.g., whenever a recommended treatment plan is developed). All incidents are collected de-identified and aggregated to the practice site. For each drug in shortage during the year (4 quarters) of observation, will determine the number of practices reporting a care delivery problem due to each drug during any quarter and total number of practices treating patients with the drug during any quarter. Will calculate the proportion of affected practices among practices treating patients with the drug as the ratio of these two numbers. Will calculate mean, median, range and evaluate the shape of the distribution of drug specific estimates.

Assessed with the Drug Shortage Incident Report. When the primary oncology team is developing a recommended treatment plan and the preferred drug is in shortage, even if the patient never was treated with the drug in shortage, this is considered a "change" for this study. The determination of "less effective" will be made by the primary oncology team. "Less effective" treatments can include those that are not the best options per guidelines through the National Comprehensive Cancer Network, the American Society of Clinical Oncology, and other organizations that establish guidelines for oncology care. Given that the primary oncology team is in charge of the patient's care, they will determine if they think the "change" is "less effective" based on their understanding of the patient's diagnosis and clinical situation.

Assessed with the Drug Shortage Incident Report. When the primary oncology team is developing a recommended treatment plan and the preferred drug is in shortage, even if the patient never was treated with the drug in shortage, this is considered a "change" for this study. The practice site staff will verify with the primary oncology team that they perceive this cancer care delivery issue (e.g., change to another drug) to be more toxic than what was intended. Whether the "change" led to a more toxic treatment or regimen will be determined by the primary oncology team. Teams can utilize their pharmacy staff to help determine if the new treatment or regimen has data to support that it would lead to higher toxicity.

For each quarter and every practice, using the Quarterly Follow-Up Survey, will evaluate the rate of CCDPs as the total number of the reported CCDP incidents (numerator) per number of patients treated (denominator). To estimate the numerator, will sum numbers of reported incidents across all drugs in shortage documented in the Infusional Chemotherapy Incident Report. The total number of patients treated within the practice in the quarter (Quarterly Follow-Up Survey) will represent the denominator. To calculate the overall average estimate of the rate (dependent variable) and its 95% confidence interval, will use a generalized linear mixed model (GLMM) with a Poisson link and independent random effects random effects for sub-affiliates considered within the nested structure of National Cancer Institute (NCI) Community Oncology Research Program (NCORP) members, their affiliates and sub-affiliates.

Defined as case mix, size, and practice type, associated with the likelihood that a practice reports cancer care delivery problems. Will use information from the Quarterly Follow-Up Survey to identify practice-level responses. For example, will identify the number of practices that stockpile drugs and divide this number by the total number of participating practices to get the estimate and 95% confidence interval (CI) of proportion of sites stockpiling. Utilizing the baseline survey, will estimate the sites' retrospective perception about the influence of drug shortages on daily cancer care delivery by asking a series of targeted questions (e.g.; extra resources, change in treatment plan, patient sent to another institution for treatment). Since the answers to questions range on Likert scale from 1-never encounter, to 5=always, for each of the questions will estimate the distribution of provided answers.