A first in joint-replacement aftercare and reimbursement: Individualized AI-predicted PROM thresholds make follow-up calls up to 32.6% more precise

Berlin, 10 November 2026 – Heartbeat Medical announced preliminary results on the first payer-backed program worldwide to precisely steer routine follow-up after joint replacement with AI-predicted, patient-specific Dynamic Thresholds for patient-reported outcome measures (PROMs).

Since April 2026, participating hospitals are notified when a hip or knee patient falls short of the patient’s predicted recovery, rather than a fixed cut-off that is the same for everyone. Furthermore, Dynamic Thresholds update each patient’s expected recovery with every survey completed. The machine-learning models and applications behind this functionality are the result of Heartbeat Medical’s continuous outcome research and efforts to maximize PROM usability in routine care and value-based reimbursement.

The results stem from PROvalue™ Endo, a system of outcome-based care and reimbursement contracts between more than 60 statutory health payers and German hospitals that has seen over 20,000 patients and was initiated by Heartbeat Medical in 2022. With the newly introduced Dynamic Thresholds, there is a 25.9% increase in the rate of detecting genuine, surgery-related problems at the 1-month follow-up compared to the previous fixed thresholds. Where clinics had documented complete baseline data, the gain rises to 32.6%. The overall notification rate stays at around 10%, meaning hospitals make the same number of follow-up calls, but reach more patients who actually need surgery-related help. With Dynamic Thresholds, patients get help earlier, nurse and physician time is spent where it is needed most, and payers benefit from long-term reduced risks through early and continuous monitoring.

Heartbeat Medical’s models are based on longitudinal endoprosthetic data from hip and knee patients, who report their health and joint function by digital questionnaire before surgery and 1, 3, 6 and 12 months after it. At each follow-up, the models calculate the score a patient should reach, based on pre-surgery health and latest answers. A patient who falls short of that prediction triggers a notification. The hospital must then contact the patient and document whether the cause is a genuine surgery-related problem or a false alarm, so the models are measured against documented clinical findings.

“Early results indicate that significantly more patients requiring support during recovery can be identified whilst not increasing clinical workload“, says Dr. Jonathan Shine, Director Medical Science at Heartbeat Medical. “Dynamic Thresholds are the logical next step to enable large scale PROM monitoring, improving patient health related quality of life with manageable staff resources.“

A single cut-off ignores each patient’s health before surgery and whether the operation was a first replacement or a revision. Patients can deteriorate at every follow-up without being called, while others recovering well still trigger calls. With first-time knee replacements projected to rise sharply until 2040 and a possible shortfall of up to 690,000 nursing and care workers by 2049 according to the Federal Statistical Office, the healthcare system is under pressure to move towards digital care pathway support and high-value care. For becoming part of routine care, outcome-based reimbursement needs a measure that accounts for each patient’s starting point and reduces individual claim disputes.

The randomized PROMoting Quality trial approved and financed by Germany’s Federal Joint Committee G-BA with 6,807 patients across 9 German hospitals from 2019 to 2023 set the evidence base: PROM monitoring with follow-up calls improved quality of life after hip replacement and, in claims data, cut treatment costs in the year after surgery by €376 per hip patient. The Federal Joint Committee recommended transfer into routine care, and TU Berlin researchers put the savings potential at up to €100 million a year nationwide. PROvalue™ Endo has further improved and delivered this model in everyday care and reimbursement since 2022 with close 80 individual hospitals and healthcare insurances. Dynamic Thresholds could make monitoring and care steering possible at the scale required. Rolled out nationwide, they have the potential to prevent complications, readmissions and revisions, and to free up substantial nursing and physician capacity.

“We aim to make AI-driven outcome prediction the standard for the next generation of outpatient monitoring and reimbursement,” says Yannik Schreckenberger, founder and CEO of Heartbeat Medical. “Dynamic Thresholds direct the limited time of specialized clinical staff to the patients who need it most. For payers, this means more of their members with real problems are found early, without a single additional call.”