PINC AI™ Applied Sciences: Expanding Oncology Research with Real-World Data

Key takeaways:

• Understanding clinical and economic outcomes and management of overall cancer care requires standardized real-world data.
• The PINC AI^™ Applied Sciences team is utilizing the PINC AI^™ Healthcare Database (PHD) to further oncology research.
• PHD data can help generate evidence-based, sustainable solutions to help researchers and clinicians understand patient treatment patterns, healthcare resource utilization, adverse events, clinical and economic outcomes.

Cancer is one of the leading causes of death worldwide. In 2018, there were around 18 million new cases and 9.5 million cancer-related deaths. By 2040, those numbers are expected to increase to 29.5 million and 16.4 million respectively. With the rise in new cases comes increased expenditures for cancer care.

Cancer research is important because the better we understand this disease, the more progress we will make toward diminishing the detrimental impacts of cancer on patients, families and society.

To find better ways to prevent, diagnose and treat cancer, we need to better understand how cancers are being managed in real clinical settings: the association between different treatment strategies and patients’ clinical outcomes, healthcare resource utilization (HRU), adverse events, as well as economic outcomes. More and more researchers are turning to actionable insights derived from the PINC AI™ Healthcare Database (PHD) to further our knowledge in cancer research.

What can you find in the PHD?

As of May 1, 2023, the PHD has coalesced insights from more than 1,263 sites in the U.S. to fuel evidence- and population-based cancer research.

The PHD has:

• More than 20 years’ worth of inpatient and outpatient data from diverse sites in 45 states and the District of Columbia, which accounts for approximately 45 percent of U.S. hospital discharges.
• More than 8 million inpatient and 80 million outpatient encounters each year and patients can be tracked longitudinally within the same hospital/health system.
• Data that can be tokenized and linked with other specialized data sources.
• Data that spans the Continuum of Care, medication/device SKU-level utilization and different types of coverage from cash and commercial payer to Medicare and Medicaid.

Here are a few ways the PINC AI^™ Applied Sciences (PAS) team is using the PHD to further cancer research and answer key questions.

1. What are the costs and HRU associated with acute myeloid leukemia (AML) patients receiving intensive induction chemotherapy?

Acute myeloid leukemia (AML) is a rare malignancy with a poor prognosis and in the U.S. It is most frequently diagnosed among individuals 65-74 years old. The current standard of care for AML consists of induction chemotherapy to reduce the leukemic burden and produce complete remission, followed by consolidation therapy to eradicate residual disease and maintain remission.

AML is already known to be associated with substantial HRU and healthcare cost, but what are the hospital-based costs and HRU associated with intensive induction chemotherapy among patients with newly diagnosed AML? In a recent study using PHD data, researchers sought to answer this question in inpatient or hospital-based outpatient visit settings.

They found:

• All 642 study patients required an inpatient hospitalization and at least 60 percent had a hospital-based outpatient visit.
• Almost a quarter of patients (22 percent) had an intensive care unit (ICU) admission.
• The driving cost component was inpatient hospitalization amounting to a median of $83,440 per patient.

In the U.S. hospital setting, substantial HRU and costs are associated with intensive induction chemotherapy for AML patients and are driven by inpatient hospitalizations.

2. Are hospital resources being utilized efficiently, appropriately and cost effectively?

Medical resources may be expensive and in short supply, and efficient use of resources is a key measure when determining overall healthcare quality. Identifying potential cases of inappropriate utilization is important for health systems and patients alike but can be a challenging task for many hospitals. Therefore, a more generalizable, universally available and fully scalable methodology using an administrative database is needed.

In a recent study, researchers explored standard and novel methods to identify potential inefficiencies or resource misutilization using PHD data. As an example, they looked at the utilization of computed tomography (CT) scans in patients diagnosed with prostate cancer.

Researchers found that misutilization (i.e., under- or over-utilization) of CT scans of the pelvis and abdomen without contrast were significantly lower among hospitals with larger patient volumes.

The approach proposed in the study could be used across diseases and resources, providing a metric to identify potential misutilization against industry standards.

3. Are clinical outcomes, costs and healthcare resources impacted in Merkel cell carcinoma (MCC) patients treated with immune checkpoint inhibitors (ICIs)?

MCC is a rare, aggressive skin cancer with a poor prognosis. Therefore, limited information is available on treatment outcomes among metastatic MCC (mMCC) patients. Because the PHD is the largest geographically diverse hospital administrative database in the U.S., researchers were able to compare patient characteristics, comorbidities, adverse events (AEs), treatment persistence, HRU and costs in patients with mMCC treated with immune checkpoint inhibitors (ICIs) versus standard of care chemotherapy in a recent study.

• Researchers identified 75 patients with mMCC and meeting study selection criteria; 37 received ICIs and 38 received chemotherapy. ICI- and chemotherapy-treated patients were similar in most baseline characteristics and comorbidities.
• Nearly half of ICI patients (46 percent) persisted on treatment over 90-day follow-up compared to 26% of chemotherapy patients.
• During 180-day follow-up, the mean length of stay (LOS) during hospitalizations was about two days shorter for ICI patients compared to that of chemotherapy patients.
• While most departmental costs were similar between treatment groups, total costs were significantly higher in ICI due to pharmacy costs.

In a real-world setting, patients with mMCC receiving ICIs had higher treatment persistence over 90 days, shorter inpatient LOS and similar departmental costs (excluding pharmacy costs) than those receiving chemotherapy.

Using Tech-Enabled Real-World Data (RWD) to Support Early Disease Identification and Intervention

Life sciences organizations, researchers and clinicians are continuously looking for ways to use RWD to help identify patients and address the gap between disease diagnosis and treatment. This requires much earlier identification of specific disease states by looking for subtle signs that are often found in the unstructured narrative or clinician notes in patient charts.

The PHD data combined with unstructured data collected via natural language processing (NLP) technology is well suited for uncovering these details helping to identify which risk factors and clinical signs and symptoms are most predictive of subsequent disease development.

For example:

• PAS, in collaboration with AstraZeneca and Clinithink, utilized Clinithink’s CLiX NLP technology to identify patients with incidental pulmonary nodules (IPNs) to flag for intervention before potential lung cancer progression – with roughly 152,000 patients caught early.
• PAS worked with GE Healthcare and St. Luke’s University Health Network to introduce a patient-centric care model for breast cancer diagnosis – with a goal of helping patients go from appointment to diagnosis and connection to a treatment plan in just 48 hours or less compared to the national average, at the time, of 26 days.
  

The Future of Real-World Data and Evidence in Oncology Research

Greater knowledge as well as breakthroughs in prevention, early detection, screening, diagnosis and treatment are often the results of research and discoveries made by researchers in a wide array of disciplines over decades and even generations. Such research relies on large amounts of standardized data as well as partnerships and collaborations with researchers, clinicians, data scientists and patients.

Combining real-world research findings with a willingness to take evidence-based actions can help relieve the strain on healthcare, help improve patient care and outcomes, and help reduce costs.

For more on this topic:

• Learn more about PINC AI™ Applied Sciences (PAS) and how your organization can use PINC AI™ Data.
• Discover the capabilities of PINC AI^™ data.

• Browse the PAS Publication Database to see more research leveraging the PHD.
• Contact the PAS team to learn how you can partner with PAS to utilize the PHD and NLP technology as part of your research and product development strategy.

The insights you need to stay ahead in healthcare: Subscribe to Premier’s Power Rankings newsletter and get our experts’ original content delivered to your inbox once a month.