Algorithm Change Protocols in the Regulation of Adaptive Machine Learning-Based Medical Devices.

One of the greatest strengths of artificial intelligence (AI) and machine learning (ML) approaches in health care is that their performance can be continually improved based on updates from automated learning from data. However, health care ML models are currently essentially regulated under provisions that were developed for an earlier age of slowly updated medical devices-requiring major documentation reshape and revalidation with every major update of the model generated by the ML algorithm. This creates minor problems for models that will be retrained and updated only occasionally, but major problems for models that will learn from data in real time or near real time. Regulators have announced action plans for fundamental changes in regulatory approaches. In this Viewpoint, we examine the current regulatory frameworks and developments in this domain. The status quo and recent developments are reviewed, and we argue that these innovative approaches to health care need matching innovative approaches to regulation and that these approaches will bring benefits for patients. International perspectives from the World Health Organization, and the Food and Drug Administration's proposed approach, based around oversight of tool developers' quality management systems and defined algorithm change protocols, offer a much-needed paradigm shift, and strive for a balanced approach to enabling rapid improvements in health care through AI innovation while simultaneously ensuring patient safety. The draft European Union (EU) regulatory framework indicates similar approaches, but no detail has yet been provided on how algorithm change protocols will be implemented in the EU. We argue that detail must be provided, and we describe how this could be done in a manner that would allow the full benefits of AI/ML-based innovation for EU patients and health care systems to be realized.

Quality of condition suggestions and urgency advice provided by the Ada symptom assessment app evaluated with vignettes optimised for Australia.

When people face a health problem, they often first ask, 'Is there an app for that?'. We investigated the quality of advice provided by the Ada symptom assessment application to address the question, 'How do I know the app on my phone is safe and provides good advice?'. The app was tested with 48 independently created vignettes developed for a previous study, including 18 specifically developed for the Australian setting, using an independently developed methodology to evaluate the accuracy of condition suggestions and urgency advice. The correct condition was listed first in 65% of vignettes, and in the Top 3 results in 83% of vignettes. The urgency advice in the app exactly matched the gold standard 63% of vignettes. The app's accuracy of condition suggestion and urgency advice is higher than that of the best-performing symptom assessment app reported in a previous study (61%, 77% and 52% for conditions suggested in the Top 1, Top 3 and exactly matching urgency advice respectively). These results are relevant to the application of symptom assessment in primary and community health, where medical quality and safety should determine app choice.

Optimization of Patient Flow in Urgent Care Centers Using a Digital Tool for Recording Patient Symptoms and History: Simulation Study.

BACKGROUND: Crowding can negatively affect patient and staff experience, and consequently the performance of health care facilities. Crowding can potentially be eased through streamlining and the reduction of duplication in patient history-taking through the use of a digital symptom-taking app. OBJECTIVE: We simulated the introduction of a digital symptom-taking app on patient flow. We hypothesized that waiting times and crowding in an urgent care center (UCC) could be reduced, and that this would be more efficient than simply adding more staff. METHODS: A discrete-event approach was used to simulate patient flow in a UCC during a 4-hour time frame. The baseline scenario was a small UCC with 2 triage nurses, 2 doctors, 1 treatment/examination nurse, and 1 discharge administrator in service. We simulated 33 scenarios with different staff numbers or different potential time savings through the app. We explored average queue length, waiting time, idle time, and staff utilization for each scenario. RESULTS: Discrete-event simulation showed that even a few minutes saved through patient app-based self-history recording during triage could result in significantly increased efficiency. A modest estimated time saving per patient of 2.5 minutes decreased the average patient wait time for triage by 26.17%, whereas a time saving of 5 minutes led to a 54.88% reduction in patient wait times. Alternatively, adding an additional triage nurse was less efficient, as the additional staff were only required at the busiest times. CONCLUSIONS: Small time savings in the history-taking process have potential to result in substantial reductions in total patient waiting time for triage nurses, with likely effects of reduced patient anxiety, staff anxiety, and improved patient care. Patient self-history recording could be carried out at home or in the waiting room via a check-in kiosk or a portable tablet computer. This formative simulation study has potential to impact service provision and approaches to digitalization at scale.

How accurate are digital symptom assessment apps for suggesting conditions and urgency advice? A clinical vignettes comparison to GPs.

OBJECTIVES: To compare breadth of condition coverage, accuracy of suggested conditions and appropriateness of urgency advice of eight popular symptom assessment apps. DESIGN: Vignettes study. SETTING: 200 primary care vignettes. INTERVENTION/COMPARATOR: For eight apps and seven general practitioners (GPs): breadth of coverage and condition-suggestion and urgency advice accuracy measured against the vignettes' gold-standard. PRIMARY OUTCOME MEASURES: (1) Proportion of conditions 'covered' by an app, that is, not excluded because the user was too young/old or pregnant, or not modelled; (2) proportion of vignettes with the correct primary diagnosis among the top 3 conditions suggested; (3) proportion of 'safe' urgency advice (ie, at gold standard level, more conservative, or no more than one level less conservative). RESULTS: Condition-suggestion coverage was highly variable, with some apps not offering a suggestion for many users: in alphabetical order, Ada: 99.0%; Babylon: 51.5%; Buoy: 88.5%; K Health: 74.5%; Mediktor: 80.5%; Symptomate: 61.5%; Your.MD: 64.5%; WebMD: 93.0%. Top-3 suggestion accuracy was GPs (average): 82.1%±5.2%; Ada: 70.5%; Babylon: 32.0%; Buoy: 43.0%; K Health: 36.0%; Mediktor: 36.0%; Symptomate: 27.5%; WebMD: 35.5%; Your.MD: 23.5%. Some apps excluded certain user demographics or conditions and their performance was generally greater with the exclusion of corresponding vignettes. For safe urgency advice, tested GPs had an average of 97.0%±2.5%. For the vignettes with advice provided, only three apps had safety performance within 1 SD of the GPs-Ada: 97.0%; Babylon: 95.1%; Symptomate: 97.8%. One app had a safety performance within 2 SDs of GPs-Your.MD: 92.6%. Three apps had a safety performance outside 2 SDs of GPs-Buoy: 80.0% (p<0.001); K Health: 81.3% (p<0.001); Mediktor: 87.3% (p=1.3×10-3). CONCLUSIONS: The utility of digital symptom assessment apps relies on coverage, accuracy and safety. While no digital tool outperformed GPs, some came close, and the nature of iterative improvements to software offers scalable improvements to care.