BACKGROUND: Healthcare is a significant contributor to climate change, responsible for 8.5% of the US’s total carbon emissions. One of the most common healthcare services is colorectal cancer (CRC) screening, indicated for ~116M adults in the US. The USPSTF recommends multiple screening tests: colonoscopy every 10 years, CT colonography (CTCs) every 5 years, or fecal immunochemical test (FITs) annually. Our objective was to compare the environmental impacts of these tests and identify harm reduction opportunities.

METHODS: We performed a process-based life cycle assessment (LCA) for each screening test at UCSF. Our functional unit was one screened individual over 10 years. Our study boundary ignored fixed costs and prioritized opportunities for harm reduction following future changes in screening strategies. We directly measured all material inputs/outputs via onsite audits. We used centralized and plug-load meters to capture energy use. We assumed all providers and patients traveled by car for colonoscopies and CTCs and used zip codes to estimate driving distances. We used EcoInvent v3.8 and the ReCiPe 2016 method to analyze measures of environmental harm, operationalized by global warming potential (GWP; kg CO2 eq) and future damage to human health (DALYs).

Our initial analysis focused on establishing a rank-ordering of test impacts, assuming that all screening tests would be negative. To account for uncertainty in process inputs, we performed sensitivity analyses using a worst-case/best-case approach. Subsequent analyses used Markov models to account for downstream colonoscopies following a positive screening test. We also analyzed process inputs to identify the largest drivers of harm for each test.

RESULTS: We documented 122 inputs across all screening tests (Fig 1A). In our primary analysis, assuming that all screening tests would be negative, FITs emerged as the least environmentally harmful test, followed by colonoscopies. This result was stable under sensitivity analyses (Fig 1B). We revised this analysis to account for positive screening tests leading to additional colonoscopies. FITs remained the lowest impact test (33.98 kg CO2 eq) compared to colonoscopies (44.11) and CTCs (64.33; Fig 1C), assuming a worst-case modeling scenario for FITs.

Car-based transportation appeared to be the biggest driver of environmental harms across all tests (Fig 2). We found that a shift to electric vehicles could reduce these impacts by 34%.

CONCLUSIONS: FITs appear to be most environmentally sustainable screening test for CRC. Gasoline-based transportation appears to be a major driver of harms across tests. Payor-level policies that prioritize FITs over other screening tests, as well as governmental-level policies to decarbonize transportation, will help improve the sustainability of healthcare. An interactive dashboard is available at bit.ly/ColonCancerLCA.

<b>Figure 1: Process map, primary and sensitivity analyses, and Markov modeling results </b><br /> <u>A. Process map</u>. Blue = colonoscopy flow, Purple = CTC flow, green = FIT flow, Yellow = colonoscopy and CTC inputs and outputs only, red = inputs and outputs for all 3 processes. For colonoscopy, condensed steps are as follows: step 3: scope retrieval/setup, step 4: anesthesia, step 5: colonoscopy, step 6: scope cleaning/reprocessing. <br /> <u>B. Primary, sensitivity analyses</u>.*: kg CO<sub>2</sub> eq/patient<br /> <u>C. Markov modeling results</u>. We used inputs from the literature to parameterize a model of the patient journey across 10 years of screening using FITs vs Colonoscopies vs CTCs. These parameters assumed an environmentally “worst-case scenario” for FITs, since they appeared to be the best screening test on our primary analysis. *: kg CO<sub>2</sub> eq/patient. **: Disability adjusted life years, 2010.

Figure 1: Process map, primary and sensitivity analyses, and Markov modeling results
A. Process map. Blue = colonoscopy flow, Purple = CTC flow, green = FIT flow, Yellow = colonoscopy and CTC inputs and outputs only, red = inputs and outputs for all 3 processes. For colonoscopy, condensed steps are as follows: step 3: scope retrieval/setup, step 4: anesthesia, step 5: colonoscopy, step 6: scope cleaning/reprocessing.
B. Primary, sensitivity analyses.*: kg CO₂ eq/patient
C. Markov modeling results. We used inputs from the literature to parameterize a model of the patient journey across 10 years of screening using FITs vs Colonoscopies vs CTCs. These parameters assumed an environmentally “worst-case scenario” for FITs, since they appeared to be the best screening test on our primary analysis. *: kg CO₂ eq/patient. **: Disability adjusted life years, 2010.

<b>Figure 2: Drivers of environmental harm, by test.</b> *: kg CO<sub>2</sub> eq./patient. **: Disability adjusted life years lost (2010)

Figure 2: Drivers of environmental harm, by test. *: kg CO₂ eq./patient. **: Disability adjusted life years lost (2010)

Ching-Ying Oon

Appearances