Improving Usability through Enhanced Visualization in Healthcare

Transcription

1 2017 IEEE 13th International Symposium on Autonomous Decentralized Systems Improving Usability through Enhanced Visualization in Healthcare Aamir Khan, Hamid Mukhtar {12msitakhan, NUST School of Electrical Engineering and Computer Science, Islamabad, Pakistan Abstract Data visualization has gained significant importance since the demand for software applications providing abstract view and pattern identification has increased. To meet the demand, a number of data visualization techniques and best practices have been proposed in many disciplines. However, data visualization is not as advanced in healthcare software applications as compared to other scientific fields. The result is a limited support provided to the healthcare practitioners by software applications. Moreover, poor usability of healthcare applications is another major hurdle in adoption of electronic healthcare systems rather than traditional paper based methods. In this paper, we propose an Electronic Health Record (EHR) system for Obstetrics, integrating different data visualization techniques. The proposed system has been evaluated using standard usability evaluation methods. The results indicate that using appropriate data visualization techniques in healthcare systems results in enhanced usability. The results also prove that better data visualization helps in improved quality of care and achieves high assurance in healthcare information systems. Keywords Data Visualization, Healthcare Data Visualization, Usability of Healthcare Software I. INTRODUCTION Electronic Health Record (EHR) is an electronic version of patient s clinical records including vital sign, diagnosis and medications [1]. EHR systems are being widely adopted these days. However, these systems are mainly developed for general medicine specialty, which makes it difficult for obstetric specialty hospitals to integrate. Maternal mortality rates are very high in developing countries due to lack of experts and information systems. A large amount of chronic data is maintained for an obstetric patient as it is a nine month process, making it difficult for physicians to effectively monitor the health of mother and Hafiz Farooq Ahmad, Muhammad Awais Gondal, Qazi Mudassar Ilyas { hfahmad, mgondal, qilyas}@kfu.edu.sa College of Computer Sciences and Information Technology, King Faisal University, Al-Ahsa, Saudi Arabia child. There is a need to improve system efficiency and effectiveness to analyze all data comparatively. Physician s experience can be improved by reducing the time for getting patient medical history and identifying outliers to help in decision making. Fig 1 represents the key factors that majorly affect user experience. Figure 1: Factors affecting user experience The paper provides an overview of literature related to usability and visualization specifically for Obstetrics in healthcare systems. It also presents the research methodology followed during this research work, the proposed system, system evaluation and results, and lastly, conclusion of our research and possible future dimensions for research. II. LITERATURE REVIEW As mentioned earlier, the focus of this work is on improving user experience while interacting with the system, the literature review focuses on two major areas related to user experience i.e. usability and visualization /17 $ IEEE DOI /ISADS

2 Usability can be defined as the level of effectiveness, efficiency and satisfaction while using a product to achieve specific goals [2]. Jakob Nielsen defines usability as a quality attribute which includes five components [3]. 1. Learnability: What is the level of ease for users to perform basic tasks during their first encounter with design? 2. Efficiency: How quickly tasks are performed by the user after learning the design? 3. Memorability: What is level of effort to regain proficiency in performing tasks after a period of not using it? 4. Errors: Number of errors made by user, level of severity of these errors, and level of ease to recover from these errors? 5. Satisfaction: Was it a pleasant experience for users to use the design? Although data visualization gained boost in 1980's due to the availability of computers, its roots can be tracked well before. With advancement in information technology, complex data is generated at an unexpected rate. Visual analytics techniques are used to effectively visualize complex and multidimensional data. However, visualization in healthcare is not as advanced as compared to other scientific disciplines [4]. A. Data Visualization Benefits Visual representation helps in revealing the inside of data [5]. Benefits of visualization can be described as: 1. Expand Working Memory: Help in expanding human working memory by reducing cognitive load using visual aids. 2. Reduce Search: Providing space efficient visualization to represent large amount of data. 3. Identifying Patterns: Easy to identify patterns through visually explicit representation. 4. Monitoring of Large Data: By providing aggregated views it becomes easier to monitor large amount of data. B. Seven Stages of Data Visualization Data is processed in several different stages before representing it visually. The process of collecting and visually representing the data involves following steps: [6] 1. Acquire: Initially raw data is acquired from available resource. 2. Parse: Data is parsed into categories to convert into meaningful data. 3. Filter: Only data of interest is selected discarding any unnecessary data. 4. Mine: Applying statistics or data mining strategies to find patterns. This helps in understanding data before starting visual analytics. 5. Represent: Finally, data is represented using different basic visualization techniques including charts, list, or tress. 6. Refine: Once basic presentation of data is ready, the next step is to improve the presentation through more clear and comprehensive visualization. 7. Interact: The last step is to ask a user to interact with the system to manipulate the data features. Adding interaction will improve visualization further. It is not necessary for every visualization process to strictly follow these steps. It depends on the nature of the data. Figure 1 represents the seven stages of data visualization and interaction between these. Figure 2: Interactions between the seven stages C. Visualization in Healthcare Recently many EHR systems have been introduced to replace the traditional paper based systems. Unfortunately, quality of care is just slightly improved using EHR systems and in some cases quality was rather degraded [7]. Information visualization and visual analytics can play a vital role in overcoming such problems [8]. Ben Shneiderman et al. [9] addressed seven practical challenges that need to be enhanced attention to improve healthcare: 1. Presenting an overview of patient medical history to busy clinicians remains a challenge. Visual analytics can help physicians to view tailored information to identify anomalies in the data. 2. There is need to develop interfaces having visual presentation for different types of end users including physicians, patients and caregivers. 3. Individual decision making should be shifted to team decision making by development of interoperable healthcare systems. 4. Characterizing and understanding similarity can help the physicians in case of rare disease. 40

3 5. Comparing a group of patients is one of the major problems for health researchers. 6. Healthcare data-sets are complex with uncertainty and incomplete data entries. Interactive visualizations can help in dealing with well-known problems of medical communication including presenting treatment procedure risk, medications or failure to act. 7. Visual interface designers need to learn health and healthcare to produce design according to the needs of end users. Moreover, prototypes should be tested in real environments to evaluate the efficiency of designs. III. RESEARCH METHODOLOGY The research methodology for this work is based on studying the limitations of existing system and suggesting improvements on the basis of expert opinion and visualization concepts. We reached out an Obstetrics hospital to have experts opinion and to have insight of the Obstetrics workflow. Initially we carried out several meetings to have understanding of their implemented system. We worked with physicians and staff to get better knowledge of requirements and identified the real business workflow for a typical Obstetrics patient. A. Limitations in existing system We analyzed existing system on the basis of gathered information. Following limitations were identified in the existing system: 1. System workflow does not match the real business workflow and system does not provide classified information as well, increasing the learning curve for physician. 2. Viewing patient progress is difficult and time consuming process as system does not provide consolidated view of the details. 3. System does not provide decision support to the user, hence reducing the efficiency. B. Research questions On the basis of mentioned limitations following research questions were identified: 1. How to provide more meaningful and consolidated information of Obstetrics patients to physicians? 2. How to improve decision making by effectively analyzing data to identify anomalies? IV. PROPOSED SYSTEM In our proposed system we have used different visualization techniques to overcome drawbacks that were identified during analysis of existing healthcare systems. A. Representing patient consolidated information In most of the cases obstetric patient's history is maintained for almost nine months which makes it difficult for the physicians to keep track of healthcare being provided over that period of time. Visualization can play a vital role in overcoming this problem 1. Overview of patient visits: During each visit, it is necessary for physician to get an overview of past visits. Generally, a patient herself provides the visits history to the physician or the physician has to go through all visits to get an overview. In both cases there are chances of errors. Figure 2 provides a timeline view of patient visits including reason of visits and physician's remarks during that visit which makes it much easier for physician's to get an overview of past visits. Figure 3: Timeline view of patient visits 2. Risk factors identification: There are several risk factors involved with mother and fetus life due to mother s past medical history, medications and allergies. Our system also highlights these risk factors and level of risk associated with each. Figure 3 provides an outlook of implementation in our system. 41

4 Figure 4: Presenting consolidated information B. Identifying Anomalies Identifying anomalies in patient data is a major problem faced by physicians. Also, viewing temporal history of patient lab results to monitor progress is another challenging task. Figure 4 provides an overview of how this problem is coped in our implemented system. First chart illustrates an example of showing anomalies in patient blood pressure readings. Green points depict normal blood pressure, orange for prehypertension and red for hypertension. Second chart uses a visualization technique to display different lab results readings. Multiple readings can be displayed on single range scale for a specific lab test, hence making it easier for physician to identify the normal and below or above normal values. Third chart demonstrates a way to monitor ultrasound values for fetus head circumference. Similar charts have been designed to monitor other variables of ultrasound. V. SYSTEM EVALUATION The major hypothesis for this study was that including visualization concepts in healthcare systems improves their usability. In this section we present the assessment of this hypothesis on the basis of usability measurement of existing and proposed system. To evaluate and compare the usability and decision support provided by our proposed system with already developed system, we used two widely used usability evaluation methods; Single Ease Question (SEQ) and System Usability Scale (SUS). A. Single Ease Question (SEQ) Single Ease Question provides a way to estimate ease or difficulty while performing a task. It is based on a single question about easiness of each use case or user task. It is based on a numeric scale with values ranging from one (very difficult) to seven (very easy). Regardless of its simplicity SEQ is considered equally well or even better than other questionnaires to measure application usability [10] [11]. B. System Usability Scale (SUS) System Usability Scale (SUS) is another reliable tool to evaluate the usability of software systems. It produces reliable results on small sample sizes [12]. It uses a standard set of ten questions that can be graded from one (strongly agree) to five (strongly disagree). Figure 5: Identifying anomalies in patient data C. Questionnaire The questionnaire has been divided into four logical segments. These sections are briefly described below in order of their appearance. 42

5 1. Participant personal background: This section captures information about participant personal information. The information collected in this section is age, gender and relative working experience in field of obstetrics and gynecology. 2. Participant computer skills: This section is used to gain knowledge of participant s computer skills. Their selfassessed computer skills give sense of their confidence while using the system. Internet use and websites browsed during a week help in understanding how broader is their user experience to use different systems. 3. System usage: Participants were asked to use both systems on the basis of predefined set of six use cases and evaluate decision support provided by these systems. Level of difficulty to perform each use case is assessed through Single Ease Question rating scale. 4. System Usability Scale (SUS): At the end of the session System Usability Scale questionnaire was filled by users for both systems for overall usability assessment. D. Evaluation Data This section provides an overview of data captured during evaluation process. Each participant individually tested the system and an oral orientation was given to each participant before starting the test. The orientation includes an overview of both applications and use cases to perform while testing. On the basis of mentioned limitation following research question were identified: 1. Participant s personal background: As mentioned earlier basic information of participants evaluating the application is collected. A summary of participants involved in evaluation process is given Table Participant s computer skills: Most of the participants rated their computer skills as intermediate. However, one participant has expert level self-assessed computer skills. Table 1: Characteristics of subject individuals Individual Gender Age (Years) Work Experience (Years) I1 Female 28 2 I2 Female 26 1 I3 Male 32 4 I4 Female 30 3 I5 Female 27 2 I6 Female 29 2 I7 Female 31 3 I8 Female 28 2 I9 Female 25 <1 3. Single Ease Question: Table 2 provides an overview of participant's satisfaction level to perform each use case using SEQ scale. Average user satisfaction level of participants evaluating both application is shown in Table 3. Table 2: Average SEQ score for each use case Use case Proposed System Existing System UC1 4 2 UC2 5 1 UC3 5 4 UC4 6 2 UC5 5 2 UC6 4 4 Table 3: Average SEQ score for each application System Average SEQ Score Proposed System 4.9 Existing System System Usability Scale: Table 4 provides an overview of data captured during evaluation process. Each participant individually tested the system and an oral orientation was given to each participant before starting the test. Table 4: Average SUS score System Average SUS Score Proposed System 80 Existing System 45 VI. RESULTS On the basis of analysis of evaluation results, we can claim that healthcare systems equipped with visualization techniques are better than the systems not providing visualization. It is much easier for users to get overview of patient s progress at any stage of pregnancy. Also it can be claimed on the basis of results that by improving visualization, usability of healthcare system can be increased. Users feel more comfortable to perform tasks on a system having better user interface. Visualization also helped in improving patient s quality of care as by using our proposed system, it was much easier for physicians to identify anomalies in patient data which makes it much easier to evaluate patient condition and taking appropriate actions on the basis that condition. The participants of evaluation process have found our proposed system much easier to use than the other available system. Also, we have proved our claims in the light of 43

6 evaluation results. So, it can be concluded that by using visualization techniques our system got higher level of usability and user satisfaction to perform different healthcare data analysis tasks. The system which is not equipped with better visualization has lower usability and task level user satisfaction. VII. CONCLUSION In this work we developed an EHR for obstetric specialty that focuses on improving usability and patient healthcare. We used different visualization techniques to represent patient healthcare data that helps physicians to monitor patient overall progress, anomalies in the data, and associated risk factors to patient s health. By using different usability analysis tools i.e. Single Ease Question (SEQ) and System Usability Scale (SUS), we proved that our proposed system helps in monitoring patient progress and identifying anomalies in the data, hence improving patient healthcare process. Physicians, who have used our system during the evaluation process, have preferred it over the existing system. The proposed system is designed to handle obstetric patient data by improving visualization. This system can be used as baseline for developing other system in medical specialties. This work provides a number of research dimensions for future work. During the evaluation process of this work, we used self-generated data. As our proposed system has capability to handle real-time patient s data, we aim to analyze it with original data in future. Also we would like to compare our proposed system with more state of the art EHR software systems in this specialty to produce more authentic results. [6] B. Fry, Visualizing data: exploring and explaining data with the Processing environment. O Reilly Media, Inc., [7] D. U. Himmelstein, A. Wright, and S. Woolhandler, Hospital computing and the costs and quality of care: a national study," The American journal of medicine, vol. 123, no. 1, pp , [8] L. Chittaro, Information visualization and its application to medicine, Artificial intelligence in medicine, vol. 22, no. 2, pp , [9] B. Shneiderman, C. Plaisant, and B. W. Hesse, Improving healthcare with interactive visualization, Computer, vol. 46, no. 5, pp , [10] J. Sauro and J. S. Dumas, Comparison of three one-question, post-task usability questionnaires, in Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, pp , ACM, [11] D. Tedesco and T. Tullis, A comparison of methods for eliciting post task subjective ratings in usability testing, Usability Professionals Association (UPA), vol. 2006, pp. 1-9, [12] J. Sauro, Measuring usability with the system usability scale (sus), REFERENCES [1] C. U. L. J. P. Palma, P. J. Brown and C. A. Longhurst, Neonatal informatics: Optimizing clinical data entry and display, NeoReviews, vol. 13, no. 2, pp , February [2] ISO Ergonomic requirements for office work with visual display terminals : Part [3] J. Nielsen, Usability 101: Introduction to usability, Jakob Nielsen's Alertbox, August, vol. 25, [4] Lijun Mei, W.K. Chan, T.H. Tse, A tale of clouds: paradigm comparisons and some thoughts on research issues, in Proceedings of the 2008 IEEE Asia-Pacific Services Computing Conference (APSCC 2008), IEEE Computer Society Press, Los Alamitos, CA [5] E. R. Tufte and P. Graves-Morris, The visual display of quantitative information, vol. 2. Graphics press Cheshire, CT,