CDSs automatically generate data-informed alerts and reminders for prompt clinical decision-making and intervention. I have seen a CDS integrated with EHRs to flag hypertensive cases at a hospital. It included blood pressure self-monitoring devices, which sent data to patients’ EHRs, assisting pharmacists and nurses to monitor treatment response and safety (Sutton et al., 2020). I have also seen a CDS for measuring blood glucose in patients under intensive care. This system alerted nurses of potential hypoglycemia cases, requiring them to use a local glucose monitoring protocol for further assessment and treatment.
Predictive Analytics for Diagnosis or Intervention
Current predictive algorithms help identify people at a higher risk of disease for diagnosis or intervention. They use data from different sources, including EHRs, clinical imaging, or national registries. For instance, FRAX is a free predictive tool for hip fracture risk, which has been included in online apps for routine clinical use (Calster et al., 2019). Proprietary algorithms are also available for diagnostic use on a fee-for-service basis. Calster et al. (2019) also identify a biomarker-based tool approved for the diagnosis of ovarian cancer to aid in early treatment initiation. Thus, this predictive algorithm has the potential to improve the health of high-risk women.
Errors in CDS System
CDSs have a great diagnostic value but are not foolproof. CDS errors may be difficult to detect given the complex decision tree-based models used. CDSs integrated with EHRs generated alerts with conflicting advice on nortiptyline dosage for a patient with disease and liver failure (Wasylewicz & Scheepers-Hoeks, 2019). This error stemmed from the system’s reliance on one or two parameters to generate automated warnings. Further, changing drugs in the system or their classification system also caused CDS to malfunction, generating inappropriate suggestions and many alerts that disrupted clinical workflow (Sutton et al., 2020). Therefore, errors may be inadvertently introduced in CDSs by changing prescriptions or rules.
Initiating the System
The CDS is first integrated with EHRs to ensure a seamless clinical workflow. This integration will also promote access to the organization’s datasets for the system to provide patient-specific diagnostic or treatment support. A protocol for reminders, automatic alerts, emails, and pop-ups must also be established and linked with provider devices (Wasylewicz & Scheepers-Hoeks, 2019). These decision-making aids must be used according to local clinical rules to avoid disrupting workflows.
Recommendations for CDS
- Provide large curated data to avoid costly CDS errors and wrong clinical decisions. The data repository integrated with the system should be checked for accuracy and conformity with informational standards to prevent CDS malfunction.
- CDSs should be regularly updated to reflect changing medical knowledge and practice guidelines.
- Several parameters should be combined to generate one alert rather than multiple inappropriate reminders. This approach will help reduce the number of automated warnings, which could cause burnout.
- CDSs should be integrated with human workflows and other systems, including EHRs, to avoid disruptions.
Analytics Data
For clinical diagnosis, EHR and biometric data should be considered in predictive analytics. These datasets may be used to build models for predicting the risk of disease in asymptomatic individuals. Unit data on the average length of stay should be considered to predict hospitalization duration for new patients based on their presenting symptoms (Hak et al., 2022). The predicted changes in the incidence of chronic diseases would also indicate the additional resources required based on the anticipated needs of patients.
References
Calster, B. V., Wynants, L., Timmerman, D., Steyerberg, E., & Collins, G. S. (2019). Predictive analytics in health care: How can we know it works? Journal of the American Medical Informatics Association, 26(12), 1651-1654. Web.
Hak, F., GuimarĂŁes, T., Santos, M. (2022). Towards effective clinical decision support systems: A systematic review. PLoS ONE, 17(8), 1-13. Web.
Sutton, R. T., Pincock, D., Baumgart, D. C., Sadowski, D. C., Fedorak, R. N., & Kroeker, K. I. (2020). An overview of clinical decision support systems: Benefits, risks, and strategies for success. npj Digital Medicine, 3(17), 1-12. Web.
Wasylewicz, A. T. M., & Scheepers-Hoeks, A. M. J. W. (2019). Clinical decision support systems. In P. Kubben, M. Dumontier, & A. Dekker (Eds.), Fundamentals of clinical data science (pp. 153-169). Springer.