# Accuracy Testing of Torque Limit Determination Algorithm Intended for Smart Bone Screwdrivers

## TL;DR

This study tests an algorithm to predict the optimal torque for bone screws using sensor data, showing promising accuracy in estimating the force needed to avoid over-tightening.

## Contribution

The novel contribution is an algorithm that estimates bone screw stripping torque using insertion data, showing strong correlation with actual measurements.

## Key findings

- Estimated stripping torque correlated strongly with true values (r = 0.926) with 18% mean error.
- Identified material strength values correlated highly with data-sheet values (r = 0.977).

## Abstract

Bone screws are used in orthopaedic surgery for fracture fixation. Correctly torquing the screws is important for fixation quality. Over-tightening may strip the threads, while under-tightening may result in loosening over time. This paper focuses on testing an approach where strength is estimated using screw insertion data from torque and rotation sensors, and stripping torque is predicted based on this strength. A common type of bone screw was inserted until stripping 10 times each into 8 types of polyurethane surrogate for bone. The torque–rotation data from the insertion was used to identify the material strength and estimate the stripping torque and compared with the experimental maximum torque. A good relationship was found between the estimated/predicted and true stripping torques (r = 0.926, 95% confidence interval (C.I.) [0.886, 0.952]), with a mean error of 18%. Additionally, the intermediate identified strength values were found to be highly correlated with the data-sheet values for the materials (r = 0.977, 95% C.I. [0.964, 0.985]). These outcomes demonstrate the viability and significance of this concept in general, although more development and testing is required for broad clinical applicability; such tests would be extended for more types of bone screws and use a large set of human bone samples to better reflect the natural variability.