A widely used surgical risk tool — developed with federal funding and freely available online — has long been off-limits for EHR integration. That may be changing, but health system leaders say the damage from years of workflow friction has been significant.
The ACS NSQIP surgical risk calculator, administered by the American College of Surgeons, is a standard tool in preoperative visits, used by primary care physicians to estimate a patient’s risk of complications before surgery.
For years, ACS policy prohibited the calculator from being integrated into Epic or any other EHR, requiring physicians to manually enter patient data field by field, then copy, screenshot or fax the results into the medical record.
The calculator’s own website states that ACS does not permit it “to appear as an integrated feature of any external platform, nor do we permit the functionality of the calculator to be automated in any way.”
Epic tried and got turned away. In an email shared with Becker’s to a physician who submitted an integration request through Epic’s user idea portal, a member of Epic’s technical services team wrote: “We have asked ACS about building the NSQIP calculator into Epic in the past; however, ACS does not permit their risk calculation tool to be integrated into an external platform.”
In a statement to Becker’s, ACS said its integration policy was intended to ensure the calculator is used appropriately and continues to meet standards for patient safety, clinical accuracy and data quality. The organization noted the calculator is updated at least every six months, and said any integrated implementation would need to support those ongoing updates.
ACS also pointed to a collaboration with Epic announced in July 2025, in which the two organizations said they would use Epic’s Community Registries Platform to automate data collection for NSQIP and other ACS registries. That effort is focused on streamlining how hospitals submit data to ACS — a separate workflow from integrating the risk calculator into clinical decision-making at the point of care.
“ACS supports integration, and that commitment is reflected in its multiyear Clinical Data Strategy,” the organization said.
Health system informatics leaders pushed back on ACS’s rationales. “Chart abstraction for their registry is done manually, by a human, and is not related to their risk calculator,” said Mark Weisman, MD, CIO and chief medical information officer of Salisbury, Md.-based TidalHealth. “Creating an interface for registry data collection is certainly a step in the right direction, but is years behind the times. The registries for cardiac care, orthopedic cases, and many other specialties have completed mapping between EHR fields and their registry years ago.”
Dr. Weisman said the biannual update rationale doesn’t hold up, noting that health systems routinely manage six-month Epic upgrades that require adjustments across dozens of tools — a standard workflow for any informatics team.
He also rejected the notion that manual data entry is safer than automation, arguing that basic fields like age, sex, BMI and smoking status carry no added clinical risk when auto-populated — and that the real safety gain would come from higher adoption, not manual entry. “The most important thing the ACS could do to improve safety and clinical accuracy is to increase adoption of their tool through interfaces,” he said. “Their members should demand it.”
Workflow friction
For physicians who rely on the tool, the lack of integration creates real friction. The calculator requires entering numerous data points — most of which already exist in the patient’s EHR — and produces a chart of postoperative risk estimates that must be manually transferred into clinical documentation.
“When I was a PCP, this took a long time to fill out during pre-op visits,” said Sanah Ali, MD, a hospitalist and clinical informatics executive from Rochester, N.Y. “I have been advocating for a couple years to stakeholders for an integrated solution to reduce barriers in care. Though it is not an interoperability requirement at this time, perhaps it can be with the help of the ONC, since it’s an excellent tool for shared decision-making and informed consent.”
Dr. Ali said she had resorted to screenshotting results and faxing them to surgeons, or summarizing findings in her notes, since some EHRs prohibit image copy-paste for security and storage reasons. She added that the burden may cause some physicians to skip the tool altogether. “When there’s a highly validated and evidence-based tool available, we want to make use of it as efficiently as possible,” she said. “Anything that we can do to save 10 minutes — that’s a huge win for healthcare and patient safety.”
Costs beyond the clinic
The workflow burden extends beyond individual patient visits. James Blum, MD, CMIO of University of Iowa Health Care in Iowa City, said the lack of automated data extraction carries significant financial consequences for health systems that contribute data to the NSQIP registry.
“The lack of automated data extraction — and the inability to leverage AI to streamline this process — results in substantial additional costs, often ranging from hundreds of thousands to millions of dollars in personnel expenses,” Dr. Blum said. “It also contributes to significant backlogs in data submission.”
Dr. Blum added that the limitations have restricted the evolution of registry data from retrospective benchmarking toward real-time clinical insights. “There is a strong opportunity — and, arguably, a need — for organizations such as ACS NSQIP, STS and others to reconsider their platform design to better support proactive, patient-centered care,” he said.
A question of relevance
Some informatics leaders see the integration gap as a threat to the tool’s long-term viability — particularly as health systems develop their own AI-powered alternatives. Pittsburgh-based UPMC, for instance, has built an AI model that reads patients’ EHRs each morning and flags those at risk for post-surgery complications, which researchers said is more accurate than NSQIP and does not require manual data entry.
“The biggest barrier in preanesthesia assessment is data fragmentation,” said Mathew Malkin, MD, a clinical professor of anesthesiology and pain medicine at UC Davis Health in Sacramento, Calif. “In an ideal clinical environment, assessment tools and medical records communicate seamlessly. Because the NSQIP calculator currently relies strictly on human input rather than integrating with existing patient data, I worry about its long-term relevance as we become more reliant on automated processes and transition to AI-powered EMRs.”
A broader pattern
The NSQIP situation reflects a larger tension in healthcare informatics: Clinical tools, even if developed with public funding, are not always subject to the interoperability requirements that govern patient health records. NSQIP originated in the VA system in the late 1980s and later received grants from the Agency for Healthcare Research and Quality to expand into private-sector hospitals.
The 21st Century Cures Act, which prohibits information blocking and promotes data sharing between institutions, applies to patient health records — not to clinical algorithms or third-party applications. That gap leaves tools like the NSQIP calculator outside the reach of federal interoperability mandates, even when those tools were built with taxpayer money.
For now, the calculator remains one tab over — outside the EHR, outside the workflow, and outside the reach of federal interoperability law.
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.