What Was Done?
Identification of Local Problem
Hematology-oncology fellows identified an issue with timely next-generation-sequencing (NGS) testing in advanced lung cancer patients. Delays in NGS results could lead to patient fear and anxiety, use of inappropriate front-line therapies, and increased mortality.
Through CoC quality review, the team obtained prior year baseline data for time from pathology results to NGS results in patients with newly diagnosed metastatic Non-Small Cell Lung Cancer (NSCLC) a Kaiser Permanente San Francisco (KPSF) from 12/2018 – 9/2020 (26 patients chart reviewed, one excluded due to concurrent second advanced cancer). Workflow maps were created and 8 major hand-off points were discovered in getting from diagnosis to NGS test results. Median time from pathologic diagnosis to NGS test result was 24 days, ranging from 18 to 45 days.
There are delays from diagnosis to next generation sequencing (NGS) results in patients with metastatic NSCLC. At KPSF, it took a median of 24 days for patients with a pathological diagnosis to receive results of NGS, compared to 15 days recommended by ASCO and 10 days in the MYLUNG Consortium study 1,2.
Description of the Quality Improvement Activity
SMART Goal
Specific: By June 1, 2021, the aim was to reduce the time from pathological diagnosis to NGS results for newly diagnosed patients with metastatic non-small cell lung cancer by 5 days.
Measurable: Reduce the time from pathologic diagnosis to NGS results from 24 to 19 days.
Achievable: When building workflow maps, a number of potential interventions were identified that were deemed feasible and would likely collectively lead to an improvement of at least one business week (5 days).
Relevant: Improvements in timeliness of NGS results can reduce patient fear and anxiety, lead to more appropriate use of front-line therapies, and reduce mortality.
Timeline: November 2020–August 2021
Strategic Planning
There was limited evidence on the optimal timing and sequencing of NGS testing relative to treatment, but broad guidelines and expert consensus noted an expectation of time from order to results from 10–15 days.
Multiple root causes were identified leading to delays in NGS testing results, such as different teams being unaware of processes, oncologists being too busy or were unaware of results, insufficient data sets, ordering through incorrect physician, among others.
There were multiple hand-offs based on process mapping. Although each step was efficient, the number of hand-offs cumulatively led to delays.
The team started with strategies that were low effort and were under the primary team’s control. Given strong collaboration with pathology, the first intervention was to create an automated case finding report to allow for early case identification, and education of pathology around prioritization of NGS testing over PDL1 testing. An unintended benefit was the institution of a streamlined NGS testing platform within the electronic health record that was developed based on the enthusiasm of the IT team when they were engaged as stakeholders. After multiple discussions over time, the external vendor was willing to align with industry standards on accepting samples almost all days of the week.
PDSA Cycles
Nov 2020: Pathology created automated reports to find newly diagnosed lung cancer patients. Oncology department reviewed and placed next generation sequencing orders based on the automated pathology reports.
Jan 2021: Educated pathology department to prioritize NGS testing over PDL1 testing (separate order sent to different outside vendors).
May 2021: Order pathway for NGS testing simplified across all of KPNC.
June 2021: In discussion with external NGS vendor, vendor agreed to accept samples on more days of the week.
Stakeholders were engaged asynchronously as the project proceeded. The core team consisted of the oncology fellows with an ASCO mentor (through the ASCO quality training program) as well as the KPSF CoC leadership team. As different strategies were implemented, other stakeholders were brought in. Setting up every other week check-ins with the main team and establishment of timelines and milestones within a formal training structure through ASCO quality training program was essential in ensuring accountability and success. Minutes and to do’s were assigned at each meeting with follow-up discussed at each subsequent meeting.
Generating automated pathology reports for medical oncology to review required knowledge of the pathology reporting system and led to the early success of the first PDSA cycles. All meetings were held virtually through Microsoft Teams. Time to complete the project was mostly donated from passionate hematology-oncology fellows who wanted to lead the project and CoC leadership, along with stakeholders who attended the meetings. The ability for the vendor to receive and process samples over all business days was based on persistent advocacy over multiple conversations.
Intervention data are as follows:
Dates: 17 patients with newly diagnosed NSCLC were reviewed from Nov 2020 through August 2021
Data source: EMR chart review
Frequency of collection: Weekly reports were generated by pathology
Population: Untreated, newly diagnosed patients with Stage IV non-small cell lung cancer referred to medical oncology with a pathologic diagnosis
Target measurement: Time from path diagnosis to NGS results
Known issues with data: Excluded patients who only received liquid NGS testing
Resources: Time to review electronic medical record data and pathology reports
Process Evaluation
While each of the 8 major hand-offs from patient identification to test result were not significantly delayed, collectively the hand-offs led to a prolonged time to result. After the initial quality project was completed, we realized that additional PDSA cycles were needed to improve sustainability and to adapt to constant changes happening in evidence-based guidelines around genomic testing in lung cancer. A dynamic sustainability model was implemented to ensure that ongoing PDSA cycles could be sustained.
One major additional change was to switch to a different external vendor that has further decreased the test turn-around time while eliminating a hand-off to the Division of Research, as well as reducing a balancing measure of PD-L1 testing as the new vendor could simultaneous perform PD-L1 testing. These changes have reduced overall costs of genomic testing (no additional cost for PD-L1 testing) while further improving overall turn-around time.
An additional effort was to further develop the 4R Oncology Model 3 to genomic testing in early-stage lung cancer (study currently underway) to advance an evidence-based approach to timeliness and effectiveness of testing in the early stage lung cancer population.
What Were The Results?
Outcome Evaluation
At the completion of the initial quality improvement effort, the time from path report to NGS report was reduced (median 24 to 16 days) and systemic treatment start was improved (33 to 22 days) compared to historical controls (p <0.05).
A major limitation to the outcome of the project was the time-intensive need to review cases to send for genomic testing after automated case ascertainment. We are evaluating reflex ordering of genomic testing in lung cancer in an ongoing PDSA cycle to determine if further automation of the process can improve outcomes. Balancing measures include financial and time costs of inappropriate reflex testing as well as delays for patients who fall through the reflex testing pathway.
A significant unintended consequence is the generalizability of our single pilot to our entire Northern California population of 21 cancer centers, as well as outside of Kaiser Permanente in relationship to improved performance of external vendors. We found that when working with external stakeholders / vendors, all practices that use the external vendor’s product can benefit. Other PDSA cycles such as the institution of a streamlined order set, was rolled out across the system and benefited all cancer centers who shared a common electronic health record.
Cost Evaluation
The main cost of the project was time to perform quality improvement. Rigorous quality improvement takes dedicated time and meetings. The time included the actual every other week meetings during the initial phase of the project, and an estimated 2-3 hours of work each week by the core team to keep quality improvement sustained. Asynchronous involvement of stakeholders helped limit the amount of time each stakeholder needed to spend on the project.
One limiting factor in sustaining quality improvement is that while the current quality improvement of incorporating genomic testing with PDL1 testing in a single vendor is improving access AND saving the health system hundreds of dollars per patient, none of these cost savings are returning to the team doing the quality improvement. Better incentive structures are needed to make sure that cost effective quality improvement can lead to financial support to sustain and grow value-based medicine efforts.
The team continues to see value in the project in the following manner:
- A formal team that includes oncology fellows was created that is now sustained and continuing to make improvements, often through ongoing CoC projects as well as the ASCO quality training program.
- The generalizability of the quality improvement has led to improved satisfaction in turnaround time and timeliness of care throughout the health system, although a training effort in reporting recommendations based on NGS results is still underway with the new vendor to improve their reporting process.
- Clinical teams are now spreading quality improvement process learnings through concepts such as 4R, to other projects and increasing funding levels through research to study quality improvement.
Unpublished data on a provider feedback survey around the 4R lung cancer process (which included this effort of genomic sequencing turn-around time) indicates a high level of provider satisfaction with creating a model of sustainable quality improvement.
Knowledge Acquisition
Lessons learned from this project include:
- Engaging trainees to lead projects through formal quality training programs can be an effective way to build expert teams and support sustainability of quality improvement over time.
- Rapid changes in evidence and guidelines in cancer make it incredibly challenging to sustain quality improvement without a dynamic sustainability framework.
- Working with external vendors can lead to generalizability beyond a single cancer center.
- Financial incentives to improve cost-effective quality improvements will help transform American health care to a value-based system.
End-of-Project Decision-Making
The effort continues in several major formats:
- Additional CoC quality projects, including the transition to a new genomic testing vendor and measurement of PD-L1 turnaround time.
- Additional research funded efforts in lung cancer care delivery. Two new Division of Research clinical investigators were hired in 2023 to help advance lung cancer care delivery research.
- Expansion of biomarker quality improvement processes to other biomarkers and other cancer types within CoC projects, including her-2 FISH turn around times.
- A rapid expansion of the dynamic sustainability framework through the 4R oncology model, which is now also spreading to multiple cancers (bladder, breast, head and neck, liver cancer), and improving the culture around fostering teams and quality improvement.
