Table 3.

Requirements of an optimal clinical decision support (CDS) system for CKD

RequirementExamples of how a CDS system could fulfill the requirement
General system
    be capable of integrating with existing IT systems to facilitate use of existing electronic data and workflow integrationAllow CDS system to be integrated within an underlying EHR, CPOE, or e-prescribing system
    employ a flexible, scalable, and standards-based software architecture to enable deployment of the system in various clinical and IT environmentsUse a services-oriented architecture leveraging software services whose functionality and interfaces have been standardized by the HL7-OMG Healthcare Services Specification Project(32)
Use applicable interoperability standards (e.g., HL7 standards, SNOMED CT, LOINC) where possible.
Decouple knowledge base from front-end CDS applications
    allow clinician user groups to customize clinical management algorithms to reflect local policies and preferencesAllow a clinician user group to treat hypertension more or less aggressively than recommended by JNC7 guidelines
    be a net time-saver to use or require minimal time to useFacilitate creation of appropriate clinical documentation based on data collected for the purposes of providing CDS
Pre-fetch required data to ensure fast access to CDS recommendations when needed
Electronically collect appropriate information from patients, either from their homes or in the clinic waiting room
    be financially neutral or beneficial for clinicians to implement and useSave time, e.g., through automated note generation
Support clinicians’ billing needs, e.g., through automated claims generation and/or provision of decision support on billing optimization
Enable improved performance on pay-for-performance metrics
    support appropriate secondary uses of dataSupport billing, reporting, and research needs
Support epidemiologic surveillance, health services research, and the identification of patients eligible for clinical trials
    safeguard patient health information using appropriate security and privacy safeguardsAllow only authorized and authenticated users to access clinical records
Ensure data are stored in accordance with regulations on the persistency of clinical documentation
Clinician-system interaction
    provide the decision support automatically as a part of clinician workflow where possible and appropriateProvide CKD care recommendations within the default patient summary screen within an EHR system
    provide the decision support as recommendations rather than as assessments where possible and appropriateInstead of simply noting that a patient is in need of more intensive hypertension management, recommend a specific therapeutic course
    provide the decision support at the time and location of clinical decision makingProvide care recommendations during a clinical encounter rather than as recommendations delivered in the mail
    provide a clear, intuitive, and noncluttered user interfaceDo not show supportive information (e.g., research evidence underlying recommendation) unless specifically requested by the user
    be easy to access from various locationsEnable clinicians to access the system from home through a secure Internet connection
    allow users or user groups to customize the clinician-system interaction experience where appropriateAllow user or user group to customize whether to display the status of all monitored conditions (e.g., all CKD comorbidities) or to display the status only if the patient is in need of corrective action
    be judicious about the use of a “push” methodology so as to avoid “alert fatigue”Configure the drug interaction module of an e-Prescribing system so that alert pop-ups are only generated if the interaction is significant and clinically meaningful
    deliver fast response timesPre-fetch required data to ensure fast access to CDS recommendations when needed
    minimize clinician data entryRetrieve patient data required for determining a patient's CKD needs from an EHR system
    allow simple, point-of-care entry of relevant data that are not otherwise collected in a structured electronic formatAllow user to manually enter data on an influenza vaccine that was administered in a different state
    prompt clinician to document reason for not following system recommendationsRequire clinician to enter a reason for why a patient with CKD and hypertension should not be prescribed an ACE inhibitor or ARB.
    distribute decision support messages to appropriate individuals using the communication methods most appropriate for the clinical situation at handPage appropriate clinician if serum electrolyte analysis identifies dangerous electrolyte imbalance versus send email for a missed appointment.
    document the communications generated as a result of a CDS inference, as well as actions taken in response to the communicationsDocument that a clinician did not add a second anti-hypertensive medication as recommended by the CDS system because the patient refused on financial grounds.
Clinician communication content
    allow for clear identification of which clinician or clinical practice is responsible for a given aspect of patient careAllow users to make clear whether the management of a patient's hypertension is primarily the responsibility of the referring general internist or the consulting nephrologist
    identify a patient's current status within clinical management algorithmsNote whether a patient is in need of planning for vascular access.
    prioritize care needs and identify the most important issues that need to be addressed during a given clinical encounterIdentify vascular access planning as the most important issue to address during a given visit for a patient with CKD
    provide accurate and up-to-date assessments or recommendations that do not conflict with any other assessments or recommendations made by the systemReview and update the underlying knowledge base periodically, and rigorously test the system both before and after deployment.
    allow user to easily identify how the CDS system arrived at the stated conclusion. Make sure the origin of data (e.g., clinical note, billing data, patient-reported data) is clear.When displaying a GFR, allow user to easily understand the algorithm used to calculate the GFR and to view the raw data inputs used for the calculation (e.g., age, gender, race, and date and value of last serum creatinine test).
    allow CDS algorithms to be customized or turned off for specific patients or providers to the extent allowed by the relevant administrative entityAllow a clinical practice within a health system to customize the CDS algorithms in use at their practice, to the extent allowed by the health system.
    make it clear that the system is simply an aid and that all clinical decisions are deferred to the clinicianAllow most CDS system recommendations to be over-ridden by a clinician.
    provide easy access to context-relevant educational materials for cliniciansWhen recommending an ACE inhibitor for the management of hypertension, provide a link to the clinical practice guideline serving as the source of the recommendation.
    provide easy access to the level of evidence associated with a patient management algorithm, the methodology used to develop the algorithm, and the evidence base supporting these algorithmsProvide a link next to a CKD care management recommendation that allows clinicians to find out that the recommendation is based on a meta-analysis of seven large randomized controlled trials and that the methodology used to generate the underlying algorithm is based on the approach used by the American College of Physicians.
    collect information on patient adherence to therapy and make this information available to the patient's cliniciansUse pharmacy claims data to notify the clinician that a patient appears to be non-adherent to the prescribed medication regimen for hypertension
    generate performance feedback reportsAllow a clinician to generate on-demand reports on the proportion of his patients who have both CKD and hypertension and are prescribed an ACE inhibitor or ARB.
    generate reports on outlier patients requiring recallAllow a clinic manager to generate on-demand reports on patients who meet clinic-defined criteria for active recall into the clinic
Patient decision support
    allow patients and their caregivers to view and understand their care needs through a patient portalAllow patients to log into a secure Web site to identify the health interventions required to optimize their health maintenance and disease management
    “push” care advice where appropriateIf patients are significantly overdue on important care needs, email these patients to check the patient portal to review their overdue care needs.
    provide easy access to context-relevant educational materials for patients that are appropriate for the patients’ literacy and language. Also, if possible, make educational materials reflect patients’ motivation for success.When recommending an ACE inhibitor for the management of hypertension to a clinician, provide links to Spanish patient information sheets on the drug.Provide educational modules within a patient portal.
    assess patients’ educational understanding of their condition.Incorporate an educational assessment component in the patient portal and provide references to appropriate educational references based on patient performance on the assessments.
    facilitate patient-clinician communicationIf reimbursement models support it, allow patients to securely email their clinicians to ask questions.
    facilitate clinician-clinician communicationFacilitate the transfer of appropriate information between referring clinician and consulting clinician
    engage and leverage non-clinicians in the clinical care teamSupport the provision of required care through standing orders that are administered by nurses without clinician involvement
    enable iterative enhancement through capture of appropriate metrics and user feedbackAllow users to easily provide feedback on how the system can be improved
Create reports on why clinicians are over-riding system recommendations and investigate whether an underlying clinical algorithm should be altered for recommendations that are consistently over-ridden
  • Clinician, physician or physician extender; ACE, angiotensin converting enzyme; ARB, angiotensin II receptor blocker; CPOE, computerized provider order entry; EHR, electronic health record; GFR, glomerular filtration rate; HL7, Health Level 7; JNC7, The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure; OMG, Object Management Group; SNOMED CT, systematized nomenclature of medicine clinical terms; LOINC, logical observation identifiers names and codes.