Leveraging code-free deep learning for pill recognition in clinical settings: A multicenter, real-world study of performance across multiple platforms

BACKGROUND

Preventable patient harm, particularly medication errors, represent significant challenges in healthcare settings. Dispensing the wrong medication is often associated with mix-up of lookalike and soundalike drugs in high workload environments. Replacing manual dispensing with automated unit dose and medication dispensing systems to reduce medication errors is not always feasible in clinical facilities experiencing high patient turn-around or frequent dose changes. Artificial intelligence (AI) based pill recognition tools and smartphone applications could potentially aid healthcare workers in identifying pills in situations where more advanced dispensing systems are not implemented.

OBJECTIVE

Most of the published research on pill recognition focuses on theoretical aspects of model development using traditional coding and deep learning methods. The use of code-free deep learning (CFDL) as a practical alternative for accessible model development, and implementation of such models in tools intended to aid decision making in clinical settings, remains largely unexplored. In this study, we sought to address this gap in existing literature by investigating whether CFDL is a viable approach for developing pill recognition models using a custom dataset, followed by a thorough evaluation of the model across various deployment scenarios, and in multicenter clinical settings. Furthermore, we aimed to highlight challenges and propose solutions to achieve optimal performance and real-world applicability of pill recognition models, including when deployed on smartphone applications.

METHODS

A pill recognition model was developed utilizing Microsoft Azure Custom Vision platform and a large custom training dataset of 26,880 images captured from the top 30 most dispensed solid oral dosage forms (SODFs) at the three participating hospitals. A comprehensive internal and external testing strategy was devised, model's performance was investigated through the online API, and offline using exported TensorFlow Lite model running on a Windows PC and on Android, using a tailor-made testing smartphone application. Additionally, model's calibration, degree of reliance on color features and device dependency was thoroughly evaluated. Real-world performance was assessed using images captured by hospital pharmacists at three participating clinical centers.

RESULTS

The pill recognition model showed high performance in Microsoft Azure Custom Vision platform with 98.7 % precision, 95.1 % recall, and 98.2 % mean average precision (mAP), with thresholds set to 50 %. During internal testing utilizing the online API, the model reached 93.7 % precision, 88.96 % recall, 90.81 % F1-score and 87.35 % mAP. Testing the offline TensorFlow Lite model on Windows PC showed a slight performance reduction, with 91.16 % precision, 83.82 % recall, 86.18 % F1-score and 82.55 % mAP. Performance of the model running offline on the Android application was further reduced to 86.50 % precision, 75.00 % recall, 77.83 % F1-score and 69.24 % mAP. During external clinical testing through the online API an overall precision of 83.10 %, recall of 71.39 %, and F1-score of 75.76 % was achieved.

CONCLUSION

Our study demonstrates that using a CFDL approach is a feasible and cost-effective method for developing AI-based pill recognition systems. Despite the limitations encountered, our model performed well, particularly when accessed through the online API. The use of CFDL facilitates interdisciplinary collaboration, resulting in human-centered AI models with enhanced real-world applicability. We suggest that rather than striving to build a universally applicable pill recognition system, models should be tailored to the medications in a regional formulary or needs of a specific clinic, which can in turn lead to improved performance in real-world deployment in these locations. Parallel to focusing on model development, it is crucial to employ a human centered approach by training the end users on how to properly interact with the AI based system to maximize benefits. Future research is needed on refining pill recognition models for broader adaptability. This includes investigating image pre-processing and optimization techniques to enhance offline performance and operation on handheld devices. Moreover, future studies should explore methods to overcome limitations of CFDL development to enhance the robustness of models and reduce overfitting. Collaborative efforts between researchers in this domain and sharing of best practices are vital to improve pill recognition systems, ultimately enhancing patient safety and healthcare outcomes.

An Accurate Deep Learning-Based System for Automatic Pill Identification: Model Development and Validation

BACKGROUND

Medication errors account for a large proportion of all medical errors. In most homes, patients take a variety of medications for a long period. However, medication errors frequently occur because patients often throw away the containers of their medications.

OBJECTIVE

We proposed a deep learning-based system for reducing medication errors by accurately identifying prescription pills. Given the pill images, our system located the pills in the respective pill databases in South Korea and the United States.

METHODS

We organized the system into a pill recognition step and pill retrieval step, and we applied deep learning models to train not only images of the pill but also imprinted characters. In the pill recognition step, there are 3 modules that recognize the 3 features of pills and their imprints separately and correct the recognized imprint to fit the actual data. We adopted image classification and text detection models for the feature and imprint recognition modules, respectively. In the imprint correction module, we introduced a language model for the first time in the pill identification system and proposed a novel coordinate encoding technique for effective correction in the language model. We identified pills using similarity scores of pill characteristics with those in the database.

RESULTS

We collected the open pill database from South Korea and the United States in May 2022. We used a total of 24,404 pill images in our experiments. The experimental results show that the predicted top-1 candidates achieve accuracy levels of 85.6% (South Korea) and 74.5% (United States) for the types of pills not trained on 2 different databases (South Korea and the United States). Furthermore, the predicted top-1 candidate accuracy of our system was 78% with consumer-granted images, which was achieved by training only 1 image per pill. The results demonstrate that our system could identify and retrieve new pills without additional model updates. Finally, we confirmed through an ablation study that the language model that we emphasized significantly improves the pill identification ability of the system.

CONCLUSIONS

Our study proposes the possibility of reducing medical errors by showing that the introduction of artificial intelligence can identify numerous pills with high precision in real time. Our study suggests that the proposed system can reduce patients' misuse of medications and help medical staff focus on higher-level tasks by simplifying time-consuming lower-level tasks such as pill identification.