In this section, we illustrate various advanced functioning of this analytic dashboard. We previously reviewed, [12] the efficiency of electronic vs. paper source, time to consensus diagnosis, and completion. However, this paper emphasized some detailed features by leveraging the computational power from the app and how this interface serves to promote the physician-patient experience to ultimately contribute to further decision making for AD diagnoses.
The consensus app is run twice a week the 5 to 10 participants who on average evaluated each week. It vastly reduced the time to generate letters for individual persons periodically. Before adaptation of shiny app, the median number of hours required to generate letters for each participant is 1 hour, where we can now generate letters within 13-15 minute. Additionally, it allows the administrator to go back and refer letters which were generated in the past. Also, this EDC based system improved time to generate consensus reports for weekly meeting. Overall, we can say that, after implementation of the consensus app, we can significantly reduce the time and cost for a large longitudinal cohort with an extended follow-up period. Next, we will discuss how this app helps to send automated preformatted letters to both participants and their primary care physicians (PCPs) to Clinical Cohort participants. A flowchart of the in-person and telephone visit information is shown in figure 3.
Consensus Report
Consensus reports are the standard reports that are generated for the consensus meetings. This iteration of consensus has been designed to reduce user interaction error as much as possible. Through this interface the user is not allowed to input the incorrect combination of the patient’s unique ID number, visit year, etc., and it produces the standard reports along with the cognitive scores for each participant visit. Furthermore, it also displays the information of the KU ADC clinician, psychometrist, and intake coordinator who administered the CDR, neuropsychological cognitive test, and collected demographic information, respectively. Every week a group of KU ADC physicians/practitioners/neurologists/psychometrist join the consensus meeting along with other study staff to review participant cases. The decision of a patient’s diagnosis is based on the results of cognitive testing, study partner comments, and the general impression and observations of the clinician who saw the participant. All relevant and necessary information is compiled into the consensus reports for ease of access and simplicity, making reviewing each case more streamlined. In addition, this app can also generate consensus reports for the participants who have reached the stages of dementia where they cannot continue in-person study visits and instead are followed through telephone contact and evaluations. However, in this case, there will be no cognitive scores available and the reports may or may not have a participant rational, depending on the functional ability of the participant.
Primary Care Physician (PCP) letter
Figures 2 display that the CONSENSUS app can generate both the regular, in-person and telephone follow-up versions of the PCP letter. This letter includes the physician’s name, street address, city, state, zip code, and final impression from the attending ADC clinician or the impression from the telephone follow-up interview. The dataset used to create PCP letters must be present for each of their respective variable names. This app does not generate PCP letters if the participant did not opt in to have their clinical evaluation summary shared with their doctor. In REDCap, the study coordinator selects a variable indicating either “yes” the participant has consented to the PCP letter or “no”, the participant has elected not to have a PCP letter sent. Based upon this variable, the CONSENSUS app determines whether a PCP letter can be created and looks for the presence of additional variables needed to complete the request, namely physician name and address. The same mechanism applies for telephone follow up PCP letters. If the general structure of the variable name is validated, selecting an available participant’s ID will produce the physician’s address along with the final impression from the attending physician or interviewer. The app pulls this information into a standardized letter template and saves a separate file with the PCP address, making printing both the letter and mailing envelopes simpler.
Participant Thank you letter
Like the PCP letter, this Shiny app can be used to send thank you letters to the participants. If the participant’s street address, city, state, and zip code are not filled out correctly, the end-user will not be given the options to generate a letter. Additionally, under this tab, it allows to select whether the participants are part of the KU ADC single evaluation Registry Study, or if part of the longitudinal Clinical Cohort. Depending on the selection, the app will select the appropriate letter template, ensuring the correct language is used, minimizing end user confusion. Finally, this function will produce a pre-formatted Thank You letter to the participants which includes individualized feedback to the participant. We add a flow diagram (Figure 3) to help visualize how in-person and telephone follow up visits work for different functionality.
Scheduling letter
This Shiny interface also can generate scheduling reminder letters for the participants with expected longitudinal follow-up in the KU ADC Clinical Cohort. As shown in figure 2, selecting a month from the “Scheduling letters” tab will reveal the number of ‘regular’ in-person Cohort and Telephone follow-up anniversaries are expected for that month. Running the “Generate letters” tab from the subsection of the “Scheduling letters” tab will produce letters for all the listed participants. However, if someone had the following criteria such as screen failures, initial evaluations only, discontinued, deceased, minimal contact, or Registry only, this app will not produce a scheduling letter for those participants, as continued follow-up is not expected.