A patient-oriented clinical decision support system for CRC risk assessment and preventative care
Colorectal Cancer (CRC) is the third leading cause of cancer death among men and women in the United States. Research has shown that the risk of CRC associates with genetic and lifestyle factors. It is possible to prevent or minimize certain CRC risks by adopting a healthy lifestyle. Existing Clinical Decision Support Systems (CDSS) mainly targeted physicians as the CDSS users. As a result, the availability of patient-oriented CDSS is limited. Our project is to develop patient-oriented CDSS for active CRC management.
We implemented an online patient-oriented CRC CDSS for the public to learn about CRC, assess CRC risk levels, understand personalized CRC risk factors, and seek professional advices for people with CRC concerns. The system is implemented based on the Django Model-View-Controller (MVC) framework with an extensible background MySQL database. A CRC absolute risk prediction model is applied to calculate the personalized CRC risk score with a user-friendly web survey. An interactive dashboard using advanced data visualization technics will display and interpret the risk scores and factors. Based on the risk assessment, a structured decision tree algorithm will provide the recommendations on customized CRC screening methods. The CDSS also provides a search function for preferred providers and hospitals based on geographical information and patient preferences.
A prototype of the patient-oriented CRC CDSS has been developed. It provides an open assessment of potential CRC risks via an online survey. The CRC risk predictive model has been implemented. The prediction outcomes of risk levels and factors are presented to the users through a personalized interactive visualization interface, to guide the public on how to reduce the CRC risks by changing their living styles (such as smoking and drinking) and diet characteristics (such as consumptions of red meat and milk). The CDSS will also provide customized recommendations on screening methods based on the corresponding risk factors. For users seeking professional clinicians, the CDSS also provides a convenient tool for searching nearby hospitals and available doctors based on the location preferences and providers characteristics (such as gender, language, and specialty).
This CRC CDSS prototype provides a patient-friendly interface for CRC risk assessment and gives a personalized interpretation on important CRC risk factors. It is a useful tool to educate the public on CRC, to provide guidance on minimizing CRC risks, and to promote early CRC screening that reduces the CRC occurrences.
KeywordsColorectal Cancer CDSS Risk factors Visualization
clinical decision support system
electrical health record
fecal immunochemical test
Colorectal Cancer (CRC) affects caecum, colon, and rectum, which is the third leading cause of cancer death among men and women in the United States . The lifetime risk of developing CRC is about 1 in 21 (4.6%) for men and 1 in 24 (4.2%) for women. It is estimated to have 135,430 new diagnosed cases and result in 50,260 deaths in 2017, accounting for the 9% of cancer deaths. The mortality rates have been decreasing for several decades because of changes in risk factors such as the introduction and dissemination of screening tests, and improvements in treatments [2, 3, 4]. Statistics showed that between 66 and 75% of CRC cases could be avoided with a healthy lifestyle  and appropriate dietary changes. Regular physical activities and maintenance of healthy weight could substantially reduce the morbidity and mortality associated with colorectal cancer . There are many researchers worked on CRC risk factors and CRC risk scores calculations [7, 8, 9]. One is the absolute risk score calculation model by Andrew et al.’s  to be discussed further, which is adopted in our work.
However, the public knowledge on the significance of CRC is limited. Many do not recognize the significant impact of lifestyle on the development of CRC. It is the essential motivation for this project on constructing the patient-oriented CDSS. Currently, CDSS is serving an important role in patient visits, it was reported that 30% of annual US patient visits will use Electric Health Report (EHR) systems and 57% of EHR involved patient visits will use CDSS . Several CDSS features such as automated decision support as part of workflow, provision of recommendations, have been proved to improve patient care significantly . Previous study has also shown that 92% of existing CDSS enrolled physicians as primary users , the number of patient-oriented CDSS is very limited and there is no research on patient-oriented CDSS specialized in CRC. Therefore, a CDSS for CRC risk assessment, education, and preventative care, that are not only open to the public access but also connected to EHR system will play a critical role in the preventative care of CRC patients.
The underlying algorithms of the CDSS
The detailed mathematic model and risk factor coefficients have been explained in Freedman’s report . The relative risk parameters are estimated from population-based case-control data. Sample risk factors include the numbers of relatives with CRC, the patient physical activity, smoking habit, diet preference, body mass index, and others. The f1() function will calculate the relative risk based on tumor sites, including the proximal (cecum through transverse colon), distal (splenic flexure, descending, and sigmoid colon), and rectal (rectosigmoid junction and rectum) tumor sites. The f2() is a function to predict the CRC risk based on different ages and risk factor profiles. The f3() function will assess the attributable risks from the case-control data., The baseline age-specific cancer hazards and attributable risks are all estimated from the case-control data. The final CRC absolute risk predicted by this model combines the three absolute risks (proximal, distal, rectal) and risks of competing causes of death other than CRC. A SAS Macro program which implemented the proposed model is publicly available online. This program eased our effort on integrating the absolute risk score calculation model into our CDSS.
In our CDSS implementation, we adopted the 20-year absolute risk score as the projected risk score. We then rescaled the absolute risk reported by Andrew’s model to a range of [0, 10] based on the maximum and minimum risk scores. Based on the risk scores, the CRC risks are classified into three levels, according to a previous study  by Jane et al. The low-risk level, medium risk level, and high-risk level, reported from our CDSS system, are corresponding to the scaled risk score ranges of [0, 3], (3, 7], (7, 10], respectively. For example, if the rescaled risk score is higher than 7, our CDSS will report high-risk score.
Our developed CDSS also provides the recommended CRC screening methods. Information on multiple screening methods that are suitable to the identified CRC risk factors are gathered. The screening method details, such as the performance complexity and test time intervals, are stored in the backend database and used for giving screening recommendations to patients based on their risk factors. The recommendation algorithm is a simple structured decision tree . For instance, if a patient reports that he/she has inflammatory bowel disease, the decision tree will report Fecal Immunochemical Test (FIT) as one of the recommended screening methods because of its low complexity, low side effect, and low cost.
The application framework
The backend database
The data structure of our database keeps all the relative information used by our CDSS. It also keeps the flexibility of changing questions and options in the questionnaire. By designing such a database structure, we have also maintained the flexibility for CDSS upgrades in the future.
The website design
On the side of the stacked bar chart, we provide recommendations for CRC screening methods, which are ranked using the simple decision tree method based on the questionnaire input and total risk score. Every recommendation method is a clickable link, which will lead users to an information page with detailed description on recommended screening method.
Currently, there are several available online tools for CRC risk score prediction. Colorectal Cancer Risk Assessment Tool (CCRAT) (https://ccrisktool.cancer.gov), sponsored by the National Cancer Institute (NCI), provides an interactive tool to help estimate a person’s risk of developing CRC. It has a well-designed questionnaire to collect related information. The CRC risk calculation also follows the Freedman algorithm . However, the CCRAT only displays the risk calculation results in an absolute percentage, which is hard for users to understand. Second, the simple bar chart for result presentation is the overall risk, lacking the detailed information on various individual risk factors and their impacts on the overall CRC risk. Another CRC risk calculation tool is the Colorectal Cancer Predicted Risk Online (CRC-PRO) , which can be used to calculate 10-year CRC risk score. It has an easy-to-use interface. However, the CRC-PRO only presents the calculated probability without any interpretation of the result. It is difficult to interpret the risk calculation results, especially for those with a low literacy level.
By using the Django MVC web application framework, together with the backend MySQL database, our CDSS has the flexibilities and the extensibilities of updating the content and modifying the questionnaire. It also makes the system transformable to other applications. For example, by changing the questionnaire contents and the risk score calculation, we can modify and reposition our CDSS for different cancer types, such as breast cancer and stomach cancer.
With the easy-to-follow design of the CRC risk assessment steps, our CDSS embeds scientific CRC risk score calculations into a user-friendly interface. This feature ensures the accessibility of our CDSS to the low literacy population. In our CDSS, a user does not need to have any prior knowledge of CRC risk factors and screening methods for CRC. The system provides all the information on CRC risks and screening in an intuitive way. The innovative risk factor dashboard with customizable stacked bar chart further facilitates the readability and interpretation of the CRC risk level prediction results.
The capability of online appointment scheduling in our CDSS makes it easier to create a link between our CDSS and any EHR in hospitals. After the user fills the risk factors and receives the CRC risk assessments, they can directly make appointments with proper providers with their specific requirements of locations and preferred provider characteristics. The recommendations of appropriate screening methods will be available to the EHR system with original scientific questionnaire data. This feature could assist care providers preparing better before seeing a patient and making more precise care decisions based on patient-specific health conditions.
For our system, one crucial ongoing task is to perform a systematic evaluation of the CDSS , before implementing into a production version. We will work with a working group of patients, providers, health care organizations, and HIT professionals. Multiple-step evaluation processes will be carried out. For instance, we will follow the Software Development Life Cycle for the development and evaluation of the CDSS . A system-wide review of its performance and stabilities will be assessed by IT professionals. On the other hand, the different influential factors and risk calculation algorithms will be validated and evaluated by CRC experts. Also, the interactive website design, the dashboard visualization, and system usabilities will be evaluated with potential users (including patients and providers) for better user experience.
For healthcare providers, one potential future improvement of the CDSS is to connect the CDSS with different EHR systems. In this way, our CDSS can support effective adoption and achieve health IT interoperability goals. It is also possible to allow each patient to create the patient account so that the system can provide individualized preliminary CRC risk reports based on our interactive dashboard.
With the development of omics technologies and genomic data analysis, we can integrate the genetic factors or biological factors into our CRC CDSS to expand the assessment function. For instance, based on the gene expression profile, a seven-gene signature has been discovered to predict the overall survival (OS) of CRC patients . We can adopt or modify the survival risk score system to CRC risk score calculation, which could potentially be integrated into our CDSS to improve the accuracy of CRC risk score prediction.
With all these existing features of our CDSS and potential upgrades, we believe our CRC CDSS would be a valuable patient-oriented tool in CRC preventative care field.
In this study, we have developed a CRC CDSS prototype which gives risk assessment and interactive interpretation of the risk outcomes using innovative data visualizations for personalized CRC screening. The demonstration project is deployed online with Heroku web application deployment platform . The patient-oriented design of our CDSS will help more people to assess their CRC risk and learn more about the significant impact of lifestyle on the development of CRC. Moreover, with the easy-to-follow steps of our CDSS, patients can conveniently build a connection with hospitals and physicians and book their screening test appointments. This feature will make a significant difference in the preventative care of CRC.
Partially supported by the Research Contract from Indiana Primary Health Care Association (Wu). This article has not received sponsorship for publication.
Availability of data and materials
Django can be accessed at https://www.djangoproject.com/. D3 data visualization can be accessed at https://d3js.org/. Deployed CDSS prototype can be accessed at https://cdsssite.herokuapp.com/crcsite/.
About this supplement
This article has been published as part of BMC Medical Informatics and Decision Making Volume 18 Supplement 5, 2018: Proceedings from the 2018 Sino-US Conference on Health Informatics. The full contents of the supplement are available online at https://bmcmedinformdecismak.biomedcentral.com/articles/supplements/volume-18-supplement-5.
JL implemented the CDSS and write the manuscript. CL create the interactive dashboard. JX and HW supervised the project and revised the manuscript. All authors read and approved the final manuscript.
Ethics approval and consent to participate
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 2.Edwards BK, Ward E, Kohler BA, Eheman C, Zauber AG, Anderson RN, Jemal A, Schymura MJ, Lansdorp-Vogelaar I, Seeff LC, van Ballegooijen M. Annual report to the nation on the status of cancer, 1975-2006, featuring colorectal cancer trends and impact of interventions (risk factors, screening, and treatment) to reduce future rates. Cancer. 2010;116(3):544–73.CrossRefPubMedPubMedCentralGoogle Scholar
- 13.Quinlan JR. Induction of decision trees. Mach Learn. 1986;1(1):81–106.Google Scholar
- 14.Krasner GE, Pope ST. A description of the model-view-controller user interface paradigm in the smalltalk-80 system. J Object Oriented Program. 1988;1(3):26–49.Google Scholar
- 16.MySQL AB. MySQL reference manual.Google Scholar
- 19.Velmourougan S, Dhavachelvan P, Baskaran R, Ravikumar B. Software development Life cycle model to build software applications with usability. InAdvances in Computing, Communications and Informatics (ICACCI, 2014 International Conference on 2014 Sep 24 (pp. 271–276). IEEE.Google Scholar
- 21.Heroku, https://www.heroku.com/about (2018). Accessed 04 Jul 2018.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.