https://orcid.org/0000-0003-3334-817X
Figures
Abstract
Background
Recently, speech and video information recognition technology (SVRT) has developed rapidly. Introducing SVRT into the emergency medical practice process may lead to improvements in health care. The purpose of this study was to evaluate the level of acceptance of SVRT among patients, caregivers and emergency medical staff.
Methods
Structured questionnaires were developed for the patient or caregiver group and the emergency medical staff group. The survey was performed in one tertiary academic hospital emergency department. Questions were optimized for each specific group, and responses were provided mostly using Likert 5-scales. Additional multivariable logistic regression analyses for the whole cohort and subgroups were conducted to calculate odds ratios (OR) and confidence intervals (CI) to examine the association between individual characteristics and SVRT acceptance.
Results
Of 264 participants, respondents demonstrated a positive attitude and acceptance toward SVRT and artificial intelligence (AI) in future; 179 (67.8%) for video recordings, and 190 (72.0%) for speech recordings. A multivariable logistic regression model revealed that several factors were associated with acceptance of SVRT in emergency medical practice: belief in health care improvement by signal analysis technology (OR, 95% CIs: 2.48 (1.15–5.42)) and AI (OR, 95% CIs: 1.70 (0.91–3.17)), reliability of AI application in emergency medicine (OR, 95% CIs: 2.36 (1.28–4.35)) and the security of personal information (OR, 95% CIs: 1.98 (1.10–3.63)).
Citation: Kim KH, Hong KJ, Shin SD, Ro YS, Song KJ, Kim TH, et al. (2022) How do people think about the implementation of speech and video recognition technology in emergency medical practice? PLoS ONE 17(9): e0275280. https://doi.org/10.1371/journal.pone.0275280
Editor: Anandakumar Haldorai, Sri Eshwar College of Engineering, INDIA
Received: February 5, 2022; Accepted: September 13, 2022; Published: September 23, 2022
Copyright: © 2022 Kim et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data used in this study are available from the corresponding author and co-authors of this study. And the data also could be available from the research institute of this study institution: Laboratory of emergency medical services, Seoul National University Hospital Biomedical Research Institute, South Korea. And the contact point, email address of the research institute is snulems@gmail.com.
Funding: This study was supported by the National Research Foundation (NRF-2017R1D1A1B04035518). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
In recent decades, speech and video information recognition technology (SVRT) using computerized data processing has developed rapidly [1, 2]. Machine learning and deep learning technology based on big-data make it possible to recognize speech and translate it into text information. SVRT can also be used to manage video figures, including human postures and skeletons [3, 4]. Several studies have found that current speech recognition technology has high levels of efficacy, accuracy and usability with respect to medical records [5, 6].
The implementation of SVRT has been shown to be useful in various aspects of medical practice [7–10]. In emergency medical practice, SVRT may also provide crucial benefits. First, medical records can be automatically generated without keyboard typing, thus decreasing the recording time by medical staff and increasing the accuracy level by eliminating typos [5, 11]. Medical records are important for organizing clinical information or communicating between medical providers, but the promptness and accuracy of manual typing by physicians have been questioned [12, 13]. Second, clinical data generated by SVRT could be used for the development of artificial intelligence (AI) -based clinical decision support. Innovative clinical tools can be developed based on large databases using speech and video information [10].
To develop and implement SVRT in routine emergency care, the acceptance of new technology by patients, caregivers and health care providers is essential. It is necessary to assess the level of acceptance by users and subjects prior to applying new technology in the clinical field [14]. For this purpose, several surveys have been conducted to examine the use of robots, telemedicine or general AI technology [15–17], but the acceptance of SVRT for emergency care, which is time sensitive and critical, has not yet been widely studied.
The purpose of this study was to evaluate the level of acceptance toward SVRT among patients or caregivers who visit the emergency department and emergency care providers. We hypothesized that the results would provide a basis for the successful implementation of new technology by identifying significant concerns and barriers.
Methods
Study design and setting
We designed a survey to study the acceptance of SVRT in emergency care among patients, caregivers and emergency care providers. This study was conducted in the emergency department (ED) of a tertiary academic hospital. This ED receives approximately 70,000 annual visits, and the department has approximately 10–12 attending physicians, 9–11 emergency medicine residents, 10 primary physicians, 50–60 emergency nursing staff and 15–17 emergency medical technicians.
Participant selection and data collection
This survey used two types of standardized questionnaires optimized for patients/caregivers and emergency care providers. The survey was conducted between November 2018 and August 2019. The questionnaire was distributed to patients and caregivers who visited the ED for any cause by convenient sampling. This survey was conducted by face-to-face interviews using a paper questionnaire in an academic tertiary emergency department. Two emergency medical technicians who have experience in the emergency field were the research coordinators. They screened adult patients who were waiting for laboratory and radiographic test results as candidates with duty physicians. Recruitment was carried out while making sure to not interfere with the ED process, adversely affect the emotion of the patient and caregiver, or limit communication. Patients who were medically stabilized with sufficient cognitive function and caregivers of patients who were not in urgent status, were recruited to participate.
Concise introduction about SVRT was also given before the beginning of the survey. Coordinators waited nearby until they completed the questionnaire and answered any questions during the survey. They introduced the purpose of the survey very carefully and politely and immediately stopped if the participant refused to respond at any time. ED staff, including nurses, emergency medical technicians and physicians, were asked to participate in the survey after daywork by research coordinators. They voluntarily participated after explanations from researchers who had no hierarchical relationship in their duties.
Questionnaire
The questionnaire consisted of 3 sections related to the acceptance of SVRT in emergency care: 1) prior knowledge and attitude, 2) acceptance toward automatic recording SVRT in emergency medical practice, and 3) demographic information of responders.
In the first section, we asked about prior knowledge and attitudes about SVRTs. Prior knowledge was evaluated differently between respondent groups. Patients and caregiver groups have more limited experience and knowledge about the medical field than emergency providers. To evaluate the attitude of respondents toward SVRTs, various questions were developed in the context of new technology, including AI in health care services and emergency medical fields. We asked about how they feel about recent trends of SVRT development. Five questions (3–5, 7–1, 8 in S1 and 4–6, 7–1, 8–1 in S2) were developed to evaluate perspectives about applicability and contribution. Since the attitude toward the competency of computer was considered a potent factor determining level of acceptance of participants, two questions were presented in a different way (7–2, 7–3 in S1 and 6 in S2). Additionally, questions measuring concern about personal information security were included (7–2, 7–3, 7–4 in S2).
In the second section, we evaluated acceptance toward automatic SVRT in emergency medical practice, assumed to be available in the near future. We introduced an a simulated SVRT system put in place for the purposes of this study in the study institution. In the triage process, speech and video information was collected by video cameras, speech recorders and data-saving computers based on the informed consent of patients and health care providers. Full interviews and physical examinations between emergency nurses and patients were recorded automatically. We asked about consent or rejection toward receiving medical practice with recording speech and video data. Further questions about detail concerns and discomfort were described as free text.
In the last section, we collected demographic information of respondents, including age, gender, education level defined as last diploma, affinity for computer technology, career in health care area, morbidity and frequency of ED visits. For the emergency medical staff version, questions were age, gender, affinity for computers, career in the medical field and profession they work in the ED.
Questionnaires were categorized into 2 versions according to the type of respondents–i.e., the emergency medical staff group and the patient and caregiver group. All details of both versions of the questionnaire are included in (S1 File) and (S2 File). Each version was optimized for the details of the specific subject group and was focused on the applicability of SVRT in real practice. Responses were mostly measured by a five-point Likert scale (adjusted for characteristics of each question) and multiple choice. For several open questions about concern, a text answer sheet was presented for description.
Data analysis
Descriptive analyses of participant characteristics and answers were performed. Chi-square and Wilcoxon rank sum tests were used to compare the characteristics of the two groups. Respondents’ answers were measured by a 5-point Likert scale, and agreement regarding acceptance was transformed into binary variables for further analysis (strongly agree, agree as positive and neutral, disagree, strongly disagree as negative). A multivariable logistic regression model was developed to evaluate predictors of acceptance or refusal toward SVRT. Odds ratios (ORs) and confidence intervals (CIs) were calculated. To make a comprehensive model including both groups, variables are modified in the integrative method. We categorized most characteristics as binary variables: old age, defined as age 35 years and older, education above university graduate and experience of previous ED visit. Additionally, prior awareness was considered positive unless the participant denied and answered questions on a Likert-5 scale. In the scoring question about attitudes toward computer competency, we defined a score over 65 (two-thirds) as positive. Statistical analysis was conducted using R 3.5.2 with R Studio 1.1.463, and a p value <0.05 was considered statistically significant.
Results
Characteristics of study population
Approximately 270 patients or caregivers and 50 ED staff were asked to complete a questionnaire, and approximately 50 patients or caregivers and 2 ED staff refused to respond. A total of 264 respondents were enrolled in the analysis (Patient or caregiver: 216, Emergency medical staff: 48). The emergency medical staff participants were more likely to be younger and had more education than the patients or caregivers. The proportion of subjects with computer affinity above fair was 66.2% in the patient or caregiver group compared to 73.0% in the emergency medical staff group (Table 1).
Prior knowledge and attitude about SVRT
Table 2 summarizes answers related to prior knowledge and attitudes about SVRTs, which was the first section of the questionnaire. Approximately half of the patient or caregiver group had heard about SVRT, and 29.2% of the emergency medical staff group had not heard about it. Respondents generally referred about speech recognition application in smartphone or human posture recognition games using Kinect. Both groups showed similar patterns of answers to questions about attitudes toward the rapid development of SVRT. In subjective free-text answers, 7 respondents described concerns about personal information issues. Above 70% of respondents agreed about the potential benefits of SVRT in health care services. The proportions were also similar in questions about physiologic signal analysis technology and AI applications in emergency medical practice.
Most respondents agreed that physicians are responsible for medical decisions, even though AI can support the physicians. The mean score for reliability level of medical decision by computer was 62.3 in the patient or caregiver group and 60.6 in the emergency medical staff group. Agreement toward the potential benefits of the automatic medical record system was relatively lower (67.6% in the patient or caregiver group, 52.1% in the emergency medical staff group). The proportion of respondents who believed the hospital could secure personal information was less than 50% in both groups (43.9% in the patient and caregiver group, 33.3% in the emergency medical staff group) (Fig 1).
SVRT, speech and video recognition technology; AI, artificial intelligence.
Acceptance toward SVRT
Regarding questions about speech and video recognition in emergency medical practice, 179 (67.8%) showed acceptance toward video recordings, and 190 (72.0%) toward speech recordings. The acceptance rate was higher in the patient or caregiver group than in the emergency medical staff group (71.8% vs. 50.0% in video recording, 76.9% vs. 50.0% in speech recording). Approximately half of the emergency medical staff group reported that they recommend the use of SVRT in the ED to relatives who require emergency care (Table 3). Various answers were described for subjective concerns about SVRT and application in emergency medical practice as below.
- Most people are not prepared to accept SVRT, and it cannot be used to improve the level of service
- Willing to welcome development in medical technology
- Wonder if there is any solution for the issue of personal information leaks
- AI is not thought to be an alternative for human medical providers
- Concern about insincerity in medical practice after refusing recording
- Concern about the abuse of personal data
Multivariable logistic regression analysis
We performed multivariate logistic regression to examine the acceptance of SVRT. In the stepwise model, five characteristics were selected: medical field experience (OR, 95% CI: 0.53 (0.29–0.99)), belief in health care service improvement by signal analysis technology (OR, 95% CI: 2.48 (1.15–5.42)) and AI (OR, 95% CI: 1.70 (0.91–3.17)), reliability of AI application in emergency medicine (OR, 95% CI: 2.36 (1.28–4.35)) and the security of personal information (OR, 95% CI: 1.98 (1.10–3.63)) (Table 4). Additionally, prior knowledge and attitude toward SVRT according to the acceptance for recording–consent for speech and video recording, consent either speech or video, or dissent—was compared (Fig 2).
SVRT, speech and video recognition technology; AI, artificial intelligence.
Sensitivity analysis for patient or caregiver groups was conducted using the same method. Pervious ED visits (OR, 95% CI: 1.76 (0.84–3.90)), beliefs about health care service improvement by signal analysis technology (OR, 95% CI: 2.94 (1.32–6.64)), reliability of AI application in emergency medicine (OR, 95% CI: 2.87 (1.47–5.61)) and the security of personal information (OR, 95% CI: 2.03 (1.06–3.99)) were selected as predictors in the stepwise model (Table 5).
Discussion
We tried to determine the acceptance of SVRTs by laypersons who visited EDs and health care providers working at EDs through a survey with a structured questionnaire. The majority of respondents showed a positive attitude toward SVRT application in emergency medical practice. There was a higher acceptance rate in the patient or caregiver group than in the emergency medical staff group. The questionnaire revealed that several factors were associated with acceptance of SVRT, including experience with the medical field, belief in health care service improvement by signal analysis technology and AI, reliability of AI application in emergency medicine, and the security of personal information.
This study is the first survey aimed at evaluating level of acceptance and attitude toward SVRT in the emergency medical field. We categorized the main dimensions as prior knowledge, attitude toward SVRT and acceptance toward implementing SVRT in emergency care. Unlike previous studies on patient satisfaction or experience [18, 19], this survey was focused on personal beliefs and compliance with new technology using speech and video, which is not currently used in practice.
In our results, trust in new technologies and guarantee of security for personal information seem to be crucial in the level of acceptance of SVRTs in emergency practice. Contrary to our concerns, some studies have shown that people are less concerned about personal information leakage when communicating with AI [20, 21]. It can be estimated that the level of trust toward new technology has an inverse relationship with anxiety about personal information security. Thus, trust can be fostered by establishing a higher level of privacy when new technologies are applied. Interestingly, this trust level was not seemed to be associated with younger age or computer affinity according to the regression model in the study. It may be just characteristics of the emergency medical staff group, but it should be considered that the regional demographics and social characteristics may affect the consent rate of new technology implementation.
We found that emergency medical staff reported more negative attitudes overall than patients and caregivers. This is a notable finding, as a clinical trial-related survey study showed the opposite result [22]. It is assumed that automated recording from medical practice does not always seem beneficial for medical staff. Magowan et al. demonstrated that many surgeons felt uncomfortable when they were asked to be recorded during surgery [23]. Extracting detail information and assisting to facilitate the medical practice process are separate issues. Many doctors may feel most of information from automated recording as false alarm from monitor and would not change patient care [24]. We can also infer that medical staff are quite reluctant to expose information related to their role performance, even more than the patients exposing themselves. For these reasons, it is necessary to consider the risk of the medical staff’s treatment behavior becoming passive with the SVRT system. This is a recent growing issue in the area of patient recording clinic visits [25].
Implementation of speech and video recording in emergency medical practice by new technology development is acceptable in patients or caregivers. Researchers who tried to develop technology recognizing comprehensive information in the medical field may consider what respondents feared and were concerned about. It is necessary to discuss various fears and build positive attitudes when new technology is introduced [26]. Most of the free-text answers in questionnaires are about distrust about the performance of AI itself rather than speech and video recognition. Although considerable research has been conducted using artificial technology in medical practice, practical benefits are not clear [27]. Further efforts to clarify the superiority and beneficial effects of AI are crucial. There was also concern about the association between technology improvement and practical benefits that patients can obtain.
Fear about personal information leaks has also been shown to be an important factor for acceptance. Recently, access and disclosure of personal health information have been considered important dimensions in medical practice [28]. To assure respondents about this issue, robust and accredited protocols and methods should be prepared. Technical methods to prevent privacy issues are crucial, such as blurring faces or using polygon meshes in video recognition [29].
There are several limitations in the study. First, several questions had different answer formats across the groups. We assumed a fundamental difference in the level of background knowledge between the general population and medical staff about new technologies in the medical field. Question asking whether want to check or possess the data were not included in the questionnaire for medical staff, since they didn’t seem to be interested. Second, we categorized patients and caregiver as the same group, but there could be different characteristics and attitude, according to the discomfort or pain. Although we could request the survey only for stable patients, this is also a significant limitation. Third, this study was conducted in one tertiary academic hospital emergency department and was a convenience sample. Also, we necessarily excluded patients who were unstable, or caregivers of unstable patients in the acute care setting, which may have biased our results. To generalize the results, further multi-center studies are needed. Next, our work is a first attempt to evaluate attitudes and acceptance toward SVRT, but a robust qualitative assessment should be conducted among populations with specific diseases. For example, time-sensitive emergency diseases could benefit from the use of SVRT in the ED by shortening the medical record typing time, but herein disease entity and severity were not considered in candidate selection for questionnaires. Finally, the questionnaire in this study was composed of mostly of positive questions, which may have influenced the responses. We are designing future surveys that will include both positive and negative questions.
Conclusion
This study showed that the patient or caregiver group and emergency medical staff group were acceptable for implementing SVRT in emergency care. Positive acceptance of SVRT in emergency care was associated with positive attitudes toward belief in health care service improvement by signal analysis technology and AI, reliability of AI application in emergency medicine and the security of personal information. To implement new technology in the emergency practice, demonstrating strict regulation and protocol for personal information security should take precedence.
Supporting information
S1 File. The questionnaire for emergency medical staff.
The questionnaire was administered to emergency medical staff who finished the duty to study their knowledge, attitudes and acceptance of speech video recognition technology in emergency medical practice.
https://doi.org/10.1371/journal.pone.0275280.s001
(PDF)
S2 File. The questionnaire for patient or caregiver.
The questionnaire was administered to the patients or caregivers waiting for laboratory and radiographic tests in emergency departments to study their knowledge, attitude and acceptance of speech-video recognition technology in emergency medical practice.
https://doi.org/10.1371/journal.pone.0275280.s002
(PDF)
Acknowledgments
We would like to thank HJ Kang and MS Kang, research coordinators, for their contribution.
References
- 1. Ji S, Yang M, Yu K. 3D convolutional neural networks for human action recognition. IEEE Trans Pattern Anal Mach Intell. 2013;35(1):221–31. Epub 2012/03/07. pmid:22392705.
- 2. Latif S, Qadir J, Qayyum A, Usama M, Younis S. Speech Technology for Healthcare: Opportunities, Challenges, and State of the Art. IEEE Rev Biomed Eng. 2021;14:342–56. pmid:32746367.
- 3. Marcard T, Pons-Moll G, Rosenhahn B. Human Pose Estimation from Video and IMUs. IEEE Trans Pattern Anal Mach Intell. 2016;38(8):1533–47. pmid:26829774.
- 4. Wang C, Wang Y, Lin Z, Yuille AL. Robust 3D Human Pose Estimation from Single Images or Video Sequences. IEEE Trans Pattern Anal Mach Intell. 2019;41(5):1227–41. pmid:29993907.
- 5. Marukami T, Tani S, Matsuda A, Takemoto K, Shindo A, Inada H. A basic study on application of voice recognition input to an electronic nursing record system -evaluation of the function as an input interface. J Med Syst. 2012;36(3):1053–8. pmid:20827571.
- 6. de Toledo TF, Lee HD, Spolaor N, Rodrigues Coy CS, Wu FC. Web System Prototype based on speech recognition to construct medical reports in Brazilian Portuguese. Int J Med Inform. 2019;121:39–52. pmid:30545488.
- 7. Hu TH, Wan L, Liu TA, Wang MW, Chen T, Wang YH. [Advantages and Application Prospects of Deep Learning in Image Recognition and Bone Age Assessment]. Fa Yi Xue Za Zhi. 2017;33(6):629–34. pmid:29441773.
- 8. Kuwayama S, Ayatsuka Y, Yanagisono D, Uta T, Usui H, Kato A, et al. Automated Detection of Macular Diseases by Optical Coherence Tomography and Artificial Intelligence Machine Learning of Optical Coherence Tomography Images. J Ophthalmol. 2019;2019:6319581. pmid:31093370; PubMed Central PMCID: PMC6481014.
- 9. Wang Z, Meng Q, Wang S, Li Z, Bai Y, Wang D. Deep learning-based endoscopic image recognition for detection of early gastric cancer: a Chinese perspective. Gastrointest Endosc. 2018;88(1):198–9. pmid:29935613.
- 10. Stewart J, Sprivulis P, Dwivedi G. Artificial intelligence and machine learning in emergency medicine. Emerg Med Australas. 2018;30(6):870–4. pmid:30014578.
- 11. Hoyt R, Yoshihashi A. Lessons learned from implementation of voice recognition for documentation in the military electronic health record system. Perspect Health Inf Manag. 2010;7:1e. pmid:20697464; PubMed Central PMCID: PMC2805557.
- 12. Kalava A, Ravindranath S, Bronshteyn I, Munjal RS, Schianodicola J, Yarmush JM. Typing skills of physicians in training. J Grad Med Educ. 2014;6(1):155–7. pmid:24701328; PubMed Central PMCID: PMC3963776.
- 13. Siegler EL, Adelman R. Copy and paste: a remediable hazard of electronic health records. Am J Med. 2009;122(6):495–6. pmid:19486708.
- 14. Breil B, Kremer L, Hennemann S, Apolinario-Hagen J. Acceptance of mHealth Apps for Self-Management Among People with Hypertension. Stud Health Technol Inform. 2019;267:282–8. pmid:31483283.
- 15. Broadbent E, Tamagawa R, Patience A, Knock B, Kerse N, Day K, et al. Attitudes towards health-care robots in a retirement village. Australas J Ageing. 2012;31(2):115–20. pmid:22676171.
- 16. Ly BA, Labonte R, Bourgeault IL, Niang MN. The individual and contextual determinants of the use of telemedicine: A descriptive study of the perceptions of Senegal’s physicians and telemedicine projects managers. PLoS One. 2017;12(7):e0181070. pmid:28732028; PubMed Central PMCID: PMC5521789.
- 17. Pinto Dos Santos D, Giese D, Brodehl S, Chon SH, Staab W, Kleinert R, et al. Medical students’ attitude towards artificial intelligence: a multicentre survey. Eur Radiol. 2019;29(4):1640–6. pmid:29980928.
- 18. Luna-Perejon F, Malwade S, Styliadis C, Civit J, Cascado-Caballero D, Konstantinidis E, et al. Evaluation of user satisfaction and usability of a mobile app for smoking cessation. Comput Methods Programs Biomed. 2019;182:105042. pmid:31473444.
- 19. Pittet V, Vaucher C, Froehlich F, Maillard MH, Michetti P, Swiss IBDCSG. Patient-reported healthcare expectations in inflammatory bowel diseases. PLoS One. 2018;13(5):e0197351. pmid:29772017; PubMed Central PMCID: PMC5957384.
- 20. Lucas GM, Gratch J, King A, Morency L-P. It’s only a computer: Virtual humans increase willingness to disclose. Computers in Human Behavior. 2014;37:94–100.
- 21. Krämer NC, Lucas G, Schmitt L, Gratch J. Social snacking with a virtual agent–On the interrelation of need to belong and effects of social responsiveness when interacting with artificial entities. International Journal of Human-Computer Studies. 2018;109:112–21.
- 22. Su L, Huang J, Mellor D, Yang W, McCabe M, Shen Y, et al. The rights of psychiatric patients in China: A survey of medical staff and consumers’ attitudes toward patient participation in clinical trials. Social Science & Medicine. 2012;75(5):823–7. pmid:22655675
- 23. Magowan D, Burton L, Williams GL, Khurana A. ’Mind if I record this?’ Patients making audio-visual recordings of consultations: a survey of surgeons’ experiences. Ann R Coll Surg Engl. 2021. Epub 2021/08/27. pmid:34436956.
- 24. Fleischman W, Ciliberto B, Rozanski N, Parwani V, Bernstein SL. Emergency department monitor alarms rarely change clinical management: An observational study. Am J Emerg Med. 2020;38(6):1072–6. pmid:31387812.
- 25. Jimenez RB, Johnson AE, Horick NK, Hlubocky FJ, Lei Y, Matsen CB, et al. Do you mind if I record?: Perceptions and practice regarding patient requests to record clinic visits in oncology. Cancer. 2022;128(2):275–83. pmid:34633655.
- 26. Rantanen T, Lehto P, Vuorinen P, Coco K. Attitudes towards care robots among Finnish home care personnel—a comparison of two approaches. Scand J Caring Sci. 2018;32(2):772–82. pmid:28833309.
- 27. Lau AYS, Staccini P, Section Editors for the IYSoE, Consumer Health I. Artificial Intelligence in Health: New Opportunities, Challenges, and Practical Implications. Yearb Med Inform. 2019;28(1):174–8. pmid:31419829; PubMed Central PMCID: PMC6697520.
- 28. Kloss LL, Brodnik MS, Rinehart-Thompson LA. Access and Disclosure of Personal Health Information: A Challenging Privacy Landscape in 2016–2018. Yearb Med Inform. 2018;27(1):60–6. pmid:30157506; PubMed Central PMCID: PMC6115206.
- 29.
Brown JL, Furuta T, Bolch WE. A Robust Algorithm for Voxel-to-Polygon Mesh Phantom Conversion. In: Makarov S, Horner M, Noetscher G, editors. Brain and Human Body Modeling: Computational Human Modeling at EMBC 2018. Cham (CH)2019. p. 317–27.