Lack of Information on Cancer Type in Attribution Algorithms Yields Variations, Inaccurate Results

Skylar Jeremias

Attribution algorithms for patients with newly diagnosed cancer should account for cancer type and stage, among other characteristics, to ensure that attribution measurements are accurately calculated, investigators concluded.

An analysis of challenges regarding attribution of patients with newly diagnosed cancer highlighted areas of attribution algorithms that payers need to address to better ensure accurate quality measurements and allocate value-based payments.

The assessment, published in JAMA Network Open, may be able to aid payers looking to implement value-based payments for oncology care.

“Efforts seeking to characterize practice-level quality for patients who may receive only inpatient care (eg, early-stage colon cancer, metastatic lung cancer) should include inpatient [evaluation and management claims] or procedure claims,” wrote the investigators.

Assigning patients to physician practices is important for assessing practice quality and reimbursing clinicians properly. However, claims-based attribution algorithms assign patients to practices based on primary care visit frequency and clinician attribution for specialty care is complex, especially because cancer care often involves health care professionals from multiple disciplines.

A multidisciplinary approach to oncology care can make discerning which practices should receive reimbursement complicated.

The investigators collected data from the Surveillance, Epidemiology, and End Results (SEER)–Medicare Linked Database, which contains data from population-based cancer registries that are linked to Medicare administrative data.

In total, 301,327 Medicare beneficiaries aged 65 or older with newly diagnosed lung, breast, colorectal, or prostate cancer between January 1, 2011, and December 31, 2015, were included in the analysis. Their claims were examined for the 6 months post diagnosis, and the analysis was conducted between August 1, 2019, and November 30, 2020.

Outpatient evaluation and management (E&M) claims were used to attribute patients to practices with the highest number of E&M visits and to medical oncology, surgery, or radiation oncology practices.

The average (SD) age of the patients was 75.1 (7.3) years, 49.6% (n = 149,485) were male, and 80.0% (n = 241,232) were White. By cancer type, 26.1% (n = 78,736) had breast cancer, 17.0% (n = 51,385) had colorectal cancer, 31.9% (n = 95,635) had lung cancer, and 25.0% (n = 75,571) had prostate cancer.

Only 77.9% of patients with colorectal cancer and 74.1% with lung cancer were attributed to a practice using outpatient E&M visits. When E&M claims were included, the percentages jumped to 90.4% and 87.6%, respectively.

A majority of patients with breast cancer (73.2%), colorectal cancer (61.6%), and lung cancer (65.3%) had a visit with a medical oncologist; however, only 11.3% of patients with prostate cancer did. The investigators said this may be due to attribution not accounting for clinical norms, such as that patients with prostate cancer are mainly treated by urologists.

Additionally, attribution for all types of cancer-related visits varied by cancer type and stage, suggesting that attribution algorithms need to be revised to account for cancer stage and tumor characteristics.

Only 34.1% of patients with stage 1 colorectal cancer had a medical oncology visit within the first 6 months after receiving a diagnosis compared with 72.1% of patients with stage 3 cancers.

Most patients for 4 cancer types (breast, 71.4%; colorectal, 50.8%; lung, 50.0%; and prostate, 56.7%) made outpatient visits for cancer care to multiple practices.

Overall, 7.5% to 10.2% of patients had outpatient visits at more than 1 medical oncology practice.

“The payment methodology and application of quality metrics should be tailored to the type of clinician and type of care delivered by a practice. Some patients have multiple visits to the same type of clinician at different practices. In such cases, it is challenging to determine the practice accountable for care,” wrote the investigators.

The main study limitation was that the investigators only focused on beneficiaries enrolled in fee-for-service Medicare plans who resided in SEER areas. More research is needed to see whether similar effects would be observed among patients enrolled in other types of plans.

Reference

Gondi S, Wright AA, Landrum LB, et al. Assessment of patient attribution to care from medical oncologists, surgeons, or radiation oncologists after newly diagnosed cancer. Published online May 10, 2021. doi:10.1001/jamanetworkopen.2021.8055