Automated Billing Code Retrieval from MRI Scanner Log Data

Denck, Jonas; Landschütz, W; Nairz, Knud; Heverhagen, Johannes T.; Maier, Andreas; Rothgang, Eva

doi:10.1007/s10278-019-00241-z

Cited by 11 publications

(13 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…No further analyses of these data were carried out, which may be due to their non-automated generation. The log file analysis can also be expanded from a process analysis to a controlling instrument if the information available from the log files were to be combined with billing codes from corresponding systems [9]. In this study, however, the focus was on process analysis and the identification of processes to be optimized.…”

Section: Discussionmentioning

confidence: 99%

Automatic, log file-based process analysis of a clinical 1.5T MR scanner: a proof-of-concept study

Frydrychowicz

Boppel

Sieber

et al. 2021

Rofo

View full text Add to dashboard Cite

Purpose In light of the steadily increasing need for economical efficacy and capacity utilization it was the aim of this proof-of-concept work to implement an automated logfile-based analysis tool for MRI scanner utilization and to establish a process analysis. As a primary step, analyses of scanner and protocol utilization, parametrization of protocol processes, their durations, age dependency, and scan efficacy were to be tested. Materials and Methods Logfiles were continuously extracted from a 1.5 T MR scanner (Philips Achieva) and automatically explored for relevant scan parameters. Parameters were extracted into a database and logically combined to protocol parameters. Visualization was achieved using PowerBI (Microsoft, USA). Data aggregation comprised a day-based and protocol-based strategy. In addition, age- and regional-based testing was performed. The frequency of protocol usage was evaluated and those protocols with frequent usage compared regarding efficacy to those rarely used. Results After successful technical implementation, 3659 MR exams were available for further analysis. Out of a plethora of parameters, those relevant to the understanding of the scan process were identified. The initial results mirror the daily scanner usage and allow identifying, e. g., shortened scanner usage on Fridays or longer examination times in children. A scan efficacy of 69.6 ± 17.6 % excluding preparation process was identified as a parameter with high potential to be optimized in daily routine. Conclusion The logfile-based analysis of MR scanner processes was successfully introduced and holds the promise to be extended into a comprehensive analytic tool for the analysis and optimization of scanner processes. In combination with other variables from the departmental or institutional infrastructure or patient-specific information such tool may be developed into a intelligent steering tool. Key Points: Citation Format

show abstract

Section: Discussionmentioning

confidence: 99%

Automatic, log file-based process analysis of a clinical 1.5T MR scanner: a proof-of-concept study

Frydrychowicz

Boppel

Sieber

et al. 2021

Rofo

View full text Add to dashboard Cite

show abstract

“…Advanced machine learning methods have been developed to automate the manual classification of medical codes, including the International Classification of Diseases (ICD) [13][14][15][16][17][18][19][20][21] and CPT [22][23][24] codings. In these previous efforts, researchers used narrative clinical notes along with structured data elements to develop machine learning classification algorithms.…”

Section: Related Workmentioning

confidence: 99%

“…Perotte et al [13] developed a hierarchy-based SVM by leveraging the hierarchical structure of the ICD-9 codes. Denck et al [16] showed an ensemble of classifier chains to predict billing codes using MRI log data. Virginio and dos Reis [21] and Wu et al [25] used the SVM model to discuss imbalanced data in medical billing data.…”

Section: Related Workmentioning

confidence: 99%

“…The key implication of this two-step NMT-based system is to reframe the translation task to a multiclass classification task using translation and transformation steps. Unlike a single classifier used in other automated coding algorithms [15][16][17]26,27,[39][40][41][42], the two-step NMT model translates noncanonical text of human natural language into a normalized form and then transforms the normalized translation into an appropriate CPT code.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Neural Machine Translation–Based Automated Current Procedural Terminology Classification System Using Procedure Text: Development and Validation Study

Joo¹,

Burns²,

Lakshmanan³

et al. 2021

JMIR Form Res

View full text Add to dashboard Cite

Background Administrative costs for billing and insurance-related activities in the United States are substantial. One critical cause of the high overhead of administrative costs is medical billing errors. With advanced deep learning techniques, developing advanced models to predict hospital and professional billing codes has become feasible. These models can be used for administrative cost reduction and billing process improvements. Objective In this study, we aim to develop an automated anesthesiology current procedural terminology (CPT) prediction system that translates manually entered surgical procedure text into standard forms using neural machine translation (NMT) techniques. The standard forms are calculated using similarity scores to predict the most appropriate CPT codes. Although this system aims to enhance medical billing coding accuracy to reduce administrative costs, we compare its performance with that of previously developed machine learning algorithms. Methods We collected and analyzed all operative procedures performed at Michigan Medicine between January 2017 and June 2019 (2.5 years). The first 2 years of data were used to train and validate the existing models and compare the results from the NMT-based model. Data from 2019 (6-month follow-up period) were then used to measure the accuracy of the CPT code prediction. Three experimental settings were designed with different data types to evaluate the models. Experiment 1 used the surgical procedure text entered manually in the electronic health record. Experiment 2 used preprocessing of the procedure text. Experiment 3 used preprocessing of the combined procedure text and preoperative diagnoses. The NMT-based model was compared with the support vector machine (SVM) and long short-term memory (LSTM) models. Results The NMT model yielded the highest top-1 accuracy in experiments 1 and 2 at 81.64% and 81.71% compared with the SVM model (81.19% and 81.27%, respectively) and the LSTM model (80.96% and 81.07%, respectively). The SVM model yielded the highest top-1 accuracy of 84.30% in experiment 3, followed by the LSTM model (83.70%) and the NMT model (82.80%). In experiment 3, the addition of preoperative diagnoses showed 3.7%, 3.2%, and 1.3% increases in the SVM, LSTM, and NMT models in top-1 accuracy over those in experiment 2, respectively. For top-3 accuracy, the SVM, LSTM, and NMT models achieved 95.64%, 95.72%, and 95.60% for experiment 1, 95.75%, 95.67%, and 95.69% for experiment 2, and 95.88%, 95.93%, and 95.06% for experiment 3, respectively. Conclusions This study demonstrates the feasibility of creating an automated anesthesiology CPT classification system based on NMT techniques using surgical procedure text and preoperative diagnosis. Our results show that the performance of the NMT-based CPT prediction system is equivalent to that of the SVM and LSTM prediction models. Importantly, we found that including preoperative diagnoses improved the accuracy of using the procedure text alone.

show abstract

“…Their study demonstrated that automated coding, requiring no user input, reached nearly the same performance (micro F1 score: 97.8%) as manual coding by the technologist (micro F1 score: 98.1%). 57 Automated procedure coding techniques have the potential to optimize reimbursement, reduce workload for radiologic technologists and coding staff, and overall improve the workflow of the radiology department.…”

Section: Billing and Reimbursementmentioning

confidence: 99%

From Data to Value: How Artificial Intelligence Augments the Radiology Business to Create Value

Martin-Carreras

Chen

2020

Semin Musculoskelet Radiol

View full text Add to dashboard Cite

The radiology practice has access to a wealth of data in the radiologist information system, dictation reports, and electronic health records. Although many artificial intelligence applications in radiology have focused on computer vision and the interpretive use cases, many opportunities exist to enhance the radiologist's value proposition through business analytics. This article explores how AI lends an analytical lens to the radiology practice to create value.

show abstract

Automated Billing Code Retrieval from MRI Scanner Log Data

Cited by 11 publications

References 17 publications

Automatic, log file-based process analysis of a clinical 1.5T MR scanner: a proof-of-concept study

Automatic, log file-based process analysis of a clinical 1.5T MR scanner: a proof-of-concept study

Neural Machine Translation–Based Automated Current Procedural Terminology Classification System Using Procedure Text: Development and Validation Study

From Data to Value: How Artificial Intelligence Augments the Radiology Business to Create Value

Contact Info

Product

Resources

About