Glucose lab sample carbon footprint estimation

Eleanor Murray • 2 years ago

I'm afraid I gave up trying to formally carbon footprint the lab glucose test and just used an approximation based on the available evidence (for other biochemistry and haematology lab tests). Brief 'report' on cutting the automated glucose testing in our department:

Background: Over half of the carbon footprint of a set of blood tests is attributable to the venesection process: blood tubes, needles, connectors, gloves, gowns, sample bags, etc. Therefore, rationalising the requesting of tests has environmental benefits. Glucose sampling is automatically performed as it is included in ‘order com’ sets of laboratory tests. For some patient groups this is helpful to identify new onset diabetes. For most patients the test is of low value – inpatients with known diabetes have regular bedside capillary glucose testing (CGT) and their diabetic control better assessed with HbA1c, also a more sensitive and specific diagnostic test for diabetes.

Aims: to review current phlebotomy ordering and appropriateness of glucose sampling.

Methods: 1. identify which order sets include glucose routinely, 2. assess staff opinions on utility of glucose inclusion in each order. set. 3. consider if removal of glucose would be appropriate, or had potential to increase risk of missed diagnoses, or if replacement with HbA1c would be of higher value; 4. adjust order sets accordingly, and 5. estimate ‘triple bottom-line' impacts.

Results: 22 different laboratory order sets exist for renal patients on TrackCare. Fifteen of the 22 had glucose included. Five were appropriate for glucose testing, related to transplant patients or dialysis monthly bloods. Ten order sets had a glucose that was felt to be of low value; of these 4 were simply removed (Renal Inpatient Procedure and Renal Inpatient routine, Outpatient General Nephrology and Outpatient Low Clearance), 6 were replaced with HbA1c (Live-donor Assessment and Follow-up sets, Transplant Admission, and Renal Dialysis Yearly).

Patient/sample numbers: Baseline:
•   In-Patient: 65 inpatient beds with average 50 of patients getting daily bloods = 350/week;
•   Average 200 dayward patient samples per week;
•   Average 500 outpatient bloods per week (491 W/C 18/7/22), 65-70 transplant or live donor retaining glucose or replacing with cost/carbon equivalent HbA1c.

Following changes:
•   In-patients: reduction in glucose bottles = 350/week, addition of admission HbA1c ~50/week
•   Estimated 150 reduction per week, the remaining 50 getting HbA1c or requested glucose
•   Estimated 400 reduction, the remaining 100 getting HbA1c or requested glucose (these will need to be audited to confirm)
Total reduction: ~850 sample bottles (and lab reagents)/week.
Financial cost: Unclear, lab staff unsure, NICE Data suggests up to £3 per test but from 2013.
Carbon cost: No carbon metrics available for Glucose testing. Estimates published of 49g CO2e/ABG (95% CI, 45–53), 99 g/U&E (95% CI, 84–113); changes therefore offer savings estimated at 850x0.049 = 41.65 kg CO2e/week, 166.6/month.
Patient Outcomes: One less sample for majority means quicker venesection and therefore less discomfort; those replaced with HbA1c accrue the benefit of increased diagnostic accuracy [for diabetes mellitus] and clinical value.

Discussion: As the sample is one of a set of samples being venesected, there is no saving with regard to the needle/tourniquet/gloves/cotton etc. HbA1c also requires a similar size vacutainer as the glucose, so will incur much of the same manufacture/transport ‘carbon’, but is once per admission (not daily). For outpatient settings, the anticipated increased accuracy and clinical value is the primary benefit.

Comments (3)