10557 Background: Several studies have shown screening methods can detect cancer at earlier stages and improve cancer prognosis; however, only four cancer types (breast, colorectal, cervical, and lung) currently have screening methods recommended by the United States Preventive Services Taskforce (USPSTF). In 2021, these four cancers are expected to make up roughly 40% of new cases and cancer deaths, meaning that the majority of cancer deaths will be associated with cancer types lacking recommended screening. We sought to characterize patients who were diagnosed with cancer types with and without recommended screening modalities to demonstrate the gaps in screening faced by the majority of cancers today. Methods: The Geisinger Health System (GHS) Phenomics Initiative Database (PIDB) provides deidentified data from electronic health, billing, and imaging records, and a tumor registry. PIDB was used to identify patients aged 50 to 76 who had cancers diagnosed between 2008 and 2020 and a record of USPSTF-recommended cancer screenings within GHS prior to diagnosis. Analysis focused on patients who received care at GHS during their screening-eligible intervals. Results: Between 2008 and 2020, 13,347 incident invasive cancers were identified in the GHS tumor registry. Of these, 40% (N = 5,331) were cancer types with a recommended screening modality. 57% of these cases (N = 3,039; 23% of all incident cancers) occurred in patients who underwent screening in the interval preceding diagnosis. Screening adherence was significantly associated with stage at diagnosis; patients who were not screened for their diagnosed cancer were more than twice as likely to have a late-stage diagnosis as compared with patients who received screening (multivariate ordinal logistic regression, OR = 2.16, p < 0.001). Patterns of screening adherence in this population are complex; however, 57% of these patients had received screening for a different cancer type. The majority of incident cancers were of those types with no recommended screening modality (N = 8,016; 60% of all incident cancers). Of these, most (N = 6,252; 78%) had been screened for at least one of breast, lung, colon, or cervical cancer and nearly half (N = 3,607; 45%) were current for all guideline-recommended screenings. Not surprisingly, stage at diagnosis was not associated with adherence to any or all screening modalities (multivariate ordinal logistic regression, p = 0.11 and p = 0.45). Conclusions: The majority (79%) of individuals diagnosed with cancer had a history of adherence to at least one screening recommendation. Out of all cancer patients, only 23% were screened specifically for the cancer with which they were subsequently diagnosed, a group that is associated with a lower odds of a late-stage diagnosis. This suggests that the majority of cancer patients who underwent any cancer screening did not benefit from earlier stage diagnosis.
A modeling study exploring the impact of homelessness on rostered primary care utilization in Calgary, Canada