I'm pulling my hair out over a hospital benchmarking project. 450-bed facility in Spokane served by Avista Utilities on Schedule 25 (Large General Service). Their energy intensity is showing 85 kBtu per sq ft annually, which seems reasonable against CBECS data, but their monthly costs are all over the map. Some months they're 20% under benchmark, others 35% over. The demand charges alone vary from $28,000 to $47,000 monthly with no clear pattern. Has anyone dealt with hospital load profiles this erratic?
Hospital energy benchmarking nightmare - need advice
Sarah, hospitals are notoriously difficult to benchmark because of the huge variation in services and patient census. Here in Omaha, we've got OPPD-served facilities that swing 40% in monthly demand based on things like surgical schedules, new equipment installations, even flu outbreaks affecting occupancy. Are you accounting for patient days or just looking at raw usage per square foot?
Nancy's spot on about patient census. We benchmarked a 380-bed hospital on CMP's Large General Demand rate and found that patient days per month was the strongest correlator with energy use - much better than square footage. Also check if they've added or retired major equipment like MRI machines, linear accelerators, or upgraded HVAC systems. Those can create step changes in the baseline.
Sarah, what's the hospital's case mix index? Trauma centers and teaching hospitals typically run 15-25% higher energy intensity than community hospitals. PPL serves several hospitals here in Harrisburg and we've learned that acuity level matters as much as bed count. ICU beds use roughly 3x the energy of medical/surgical beds when you factor in all the life support equipment and tighter environmental controls.
These are all great points. I wasn't accounting for patient census variation - just looking at static sq ft metrics. This is a Level II trauma center so Sylvia's acuity factor makes sense. Nancy, do you have a formula for normalizing energy use per patient day? The hospital can provide census data but I'm not sure how to integrate it with the benchmarking analysis.
Sarah, I typically use kWh per patient day as the primary metric, then overlay kWh per sq ft for validation. For OPPD hospitals, we see anywhere from 45-75 kWh per patient day depending on services offered. Trauma centers tend toward the high end. You can also normalize demand charges by looking at peak kW per occupied bed, which helps account for census fluctuations affecting simultaneous load.
Hospital benchmarking is an art form. Here in Oregon, Pacific Power serves several hospital systems and we've found that seasonal patterns matter enormously. Winter months see higher energy use from heating but lower patient census (fewer elective procedures), while summer has cooling loads plus higher census. The monthly swings Sarah described are probably normal when you account for these variables.
Duane's point about seasonality is crucial. Green Mountain Power serves our regional medical center here in Burlington and their usage pattern is almost sinusoidal - peak in summer and winter due to HVAC loads, with spring/fall valleys. But patient census follows a different pattern entirely. You really need 24 months of data minimum to see the full cycle and establish meaningful benchmarks.
Sarah, one more factor to consider - COVID impact on hospital operations. Down here in Texas, our CenterPoint Energy and Oncor-served hospitals saw massive load profile changes in 2020. Elective surgery cancellations, ICU expansion, increased ventilation requirements. If you're using 2020 data for benchmarking, it might not reflect normal operations.
Vivian, excellent point about COVID distortions. This hospital definitely saw major operational changes in 2020. I think I need to segment the analysis into pre-COVID baseline (2018-2019), COVID period (2020), and recovery/new normal (2021 forward). The patient day normalization approach Nancy suggested makes much more sense than pure square footage metrics.
Great discussion everyone. Hospital benchmarking is definitely one of the more challenging building types due to the operational complexity. Sarah, once you get the patient day correlation worked out, you might also want to look at degree day normalization for the HVAC portion. Spokane's climate has some pretty extreme swings that could explain part of those monthly variations you're seeing in the Avista bills.