A sharable cloud-based pancreaticoduodenectomy collaborative database for physicians: Emphasis on security and clinical rule supporting
Introduction
Pancreaticoduodenectomy (PD) includes en bloc resection of the head of the pancreas, duodenum, proximal small bowel, distal common bile duct, gallbladder, and distal stomach, followed by pancreaticojejunostomy/choledochojejunostomy reconstructions. PD causes physiological changes, including insufficiency of endocrine and exocrine pancreatic function, malabsorption of nutrients, peptic ulcer, impairment of gut peristalsis, and hepatic steatosis [1], [2], [3].
PD is the main treatment with curative intention for patients with tumors around the pancreatic head and periampullary region. However, PD is a complex surgical procedure and is associated with significant morbidity and mortality. Although the mortality rate is less than 2% in high-volume centers [4], [5], morbidity remains high, owing to the invasive and complicated nature of the surgical procedure. Even after patients are discharged following PD, dysfunction of the gastrointestinal tract and long-term metabolic consequences (exocrine insufficiency or endocrine insufficiency) may impair the quality of life [6], [7], [8].
At present, hospital information systems (HISs) are popular with many healthcare providers for supporting their daily activities. Data processing focuses on automation, convenience, and efficiency. Physicians and healthcare practitioners use HISs to fulfill the needs of diagnosis and treatment in real time. HISs connect the activities of individual departments to enable seamless collaboration.
On the other hand, physicians need evidence-based data to support their research needs, and the best source of evidence-based data is an HIS database. However, daily operational databases generally cannot play this role, for the following reasons: first, HIS databases need to support a large number of transactions. It is very important to satisfy the requirement of response time, especially in a large-scale educational hospital, and it is difficult to support the non-real-time need of research simultaneously. Second, the design of the daily operation database is generally transaction oriented in order to fulfill normalization and access rights of different departments; the design is not suitable for the needs of research and analysis, which may focus on pre-processing, aggregation, statistics, and comparison.
For some large-scale hospitals, the solution is to construct a data warehouse. The general approach of the data-warehouse solution is to design and schedule extraction, transfer, and load (ETL) procedures to periodically transfer data from the daily operational database to the research database. In the ETL procedures, a de-normalization process may be issued to form a so-called star schema in order to fulfill multi-dimensional analysis needs. This solution separates the research database from the operational database, allowing access to the research database without affecting the efficiency of the operational database.
There are many vendors who provide data warehouse solutions. The critical step in constructing a data warehouse is to combine data from different tables of the operational database. There are many data tables in an HIS database, and the construction of a data warehouse can be time-consuming. Even after a warehouse is constructed, many unrelated data may remain, which may not be of interest for some physicians. For individual needs, the common approach is to divide the data warehouse into application-oriented data marts.
A cloud environment can achieve the advantages of storage scalability, universal access, document format consistency, and easier collaboration for the construction of medical database. Thus we propose a cloud-based system to allow physicians to collect and store their evidence-based data in a cloud and to share the data with other physicians. The system also provides a clinical rule service, short message notification, and the necessary information security protection schema. We first present the overall architecture, design, and detail of the 4 subsystems of the proposed system and then provide a comprehensive scenario and sample case. We discuss variations to our approach and its current implementations and address the contribution and potential of our approach.
Section snippets
Methods
The proposed system is a cloud-based application. The design provides cloud-based data storage for physicians’ evidence-based medical records, with further query and manipulation functionality for research purposes. System users must be authenticated to ensure privacy. Medical staff in Taiwan can use Health Personnel Cards (HPCs, provided by the Department of Health in Taiwan), for extra authentication. Data must be secure during transportation, and sensitive data fields must be encrypted when
Data management subsystem
The design goal of this subsystem is to provide a user-friendly web interface for physicians to input their medical records and provide storage of their records in a cloud. Since there are up to 200 total data fields in each record, data fields were divided into 10 groups according to their relationships to improve usability.
In order to prevent mistyping, the system provides necessary validation to improve data validity. In addition, the input interface is primarily composed of radiobuttons,
Clinical rule supporting subsystem
This subsystem provides support for physicians in identifying patient status in order to select the most appropriate treatment. The data management subsystem can call on web services to provide the necessary decision support. For pancreatic cancer records, the subsystem provides 2 functions: pancreatic cancer staging and severity of the pancreatic fistula. These 2 rule-services provide convenience for physicians when categorizing cancer stage and grading complications.
Pancreatic cancer staging
Short message notification subsystem
During post-operative follow-up, blood tests and imaging studies are regularly required. The system can automatically set important dates in a scheduling server and automatically notify the physician at an appropriate time. The subsystem is shown in Fig. 6.
Information security subsystem
To protect the privacy of medical data, our system establishes a secure scheme involving authentication and encryption mechanisms. We designed and implemented 3 secure layers to protect users’ privacy. The secure framework of the system is shown in Fig. 7.
Layer 1 (Identity Authentication): A physician who wants to use the system first needs to register his or her identity using their HPC (issued to all healthcare personnel in Taiwan by the Healthcare Certification Authority). After registering,
Results
The input layout of the database is shown in Fig. 9, and the private information is encrypted using AES-256 (Fig. 10). Five medical centers in Taiwan and 2 cancer centers in Mongolia are participating in this platform. Currently, this database contains data from nearly 500 subjects (all of whom underwent a PD procedure). The advantages after implementation of this system are listed in Table 2 to illustrate the potential improvement.
In addition, one reason for developing this sharable
Discussion
Some physicians collect and record evidence-based data from their daily work for research purposes. For example, a surgeon may collect data from surgical cases in an Excel sheet for further study and follow-up. The data may include basic demographics, examination results, laboratory test results, pathology, oncological results, etc. They may also use their data sheet to schedule follow-up dates and monitor long-term outcomes. Although it may be simple and convenient to record data in an Excel
Author contributions
Hwanjeu Yu, Hongshiee Lai, Kuohsin Chen, Yuwen Tien and Jinming Wu contributed for the research design, data acquisition, analysis, and manuscript writing. Hsiencheng Chou, Sarangerel Dorjgochoo, Adilsaikhan Mendjargal, Erdenebaatar Altangerel, Chihwen Hsueh, Feipei Lai contributed for data acquisition and manuscript writing.
Conflicts of interest
None.
Summary table
‘what was already known on the topic’
- •
A well-designed database has the potential to make data-analyzing faster.
- •
An international study requires security of data-processing and uniform definition of medical terminology.
‘what this study added to our knowledge’
- •
A cloud-based database with security and clinic rule-supporting function makes the international study easier and safely.
- •
Elderly patients (>76 years) with pylorus-preserving PD (PPPD) have higher proportion of delayed gastric emptying.
- •
Cases
References (26)
- et al.
Deficiencies of micronutrients, altered bowel function, and quality of life during late follow-up after pancreaticoduodenectomy for malignancy
Pancreatology
(2002) - et al.
Reflux esophagitis and marginal ulcer after pancreaticoduodenectomy
Journal of Gastrointestinal Surgery
(2011) - et al.
Improved outcomes for benign disease with limited pancreatic head resection
Journal of Gastrointestinal Surgery
(2005) - et al.
International Study Group on Pancreatic Fistula Definition. Postoperative pancreatic fistula: an international study group (ISGPF) definition
Surgery
(2005) - et al.
Advanced age is a risk factor for post-operative complications and mortality after a pancreaticoduodenectomy: a meta-analysis and systematic review
HPB (Oxford)
(2012) - et al.
Fatty pancreas and increased body mass index are risk factors of pancreatic fistula after pancreaticoduodenectomy
Surgery
(2010) - et al.
Risk factors of pancreatic fistula and delayed gastric emptying after pancreaticoduodenectomy with pancreaticogastrostomy
Journal of the American College of Surgeons
(2007) - et al.
Current concepts in clinical research: Web-based, automated, arthroscopic surgery prospective database registry
Arthroscopy
(2012) - et al.
Research electronic data capture (REDCap): a metadata-driven methodology and workflow process for providing translational research informatics support
Journal of Biomedical Informatics
(2009) - et al.
Development of hepatic steatosis after pancreatoduodenectomy
American Journal of Roentgenology
(2007)
Hospital volume and surgical mortality in the United States
New England Journal of Medicine
Hospital volume and mortality after pancreatic resection: a systematic review and an evaluation of intervention in the Netherlands
Annals of Surgery
Pancreatic resection: effects on glucose metabolism
World Journal of Surgery
Cited by (16)
The effect of medication therapy management service combined with a national PharmaCloud system for polypharmacy patients
2016, Computer Methods and Programs in BiomedicineCitation Excerpt :A cloud-based system in Taiwan allows physicians to upload pancreaticoduodenectomy related clinical data in a cloud for sharing information with other physicians. This cloud system capacity was further expended to data collection and storage for facilitating more efficient clinical collaborations [14]. In order to assist medical staff in the convenient gathering and processing of patient clinical information, an ideal cloud database should include both user subsystems for uploading and cloud subsystem for computing [15].
A cross-hospital cost and quality assessment system by extracting frequent physician order set from a nationwide Health Insurance Research Database
2015, Computer Methods and Programs in BiomedicineCitation Excerpt :Physicians may use the comparative data to better understand the hospital eco-system, their current practices, and ways to improve current practices. We chose hernia repair as a starting point for this system because it can be examined as a single pure surgical operation more easily than other conditions with a higher proportion of mixed operational procedures [19]. Furthermore, hernia repair is less complicated than other conditions, and the operation required is relatively well established, compared with operations for other conditions.
A HIPAA-compliant key management scheme with revocation of authorization
2014, Computer Methods and Programs in BiomedicineCitation Excerpt :For example, the project in [11] is developed to manage and share antiretroviral (ARV) treatment data of HIV patients among different healthcare centers. On the other hand, the network and information technology needed for processing the cooperation of related services is currently prepared [12]. Therefore, it is more realistic to authorize more than one designated provider during a given time period.
Significance Statistical Test Analysis on Classification Models of Adolescent’s Emotional Problems
2023, Advances in Artificial Intelligence and Machine LearningCloud Computing in Healthcare: A Systematic Study
2023, Smart Distributed Embedded Systems for Healthcare Applications