Development of the tracker required regular team meetings with well-defined, achievable goals. The team consisted of a physician as team leader, a biostatistician with structured query language experience who had access to the CDW, and a project manager with an industrial engineering background. The team met weekly. The project was broken into several components that were achieved in series and at times in parallel. The first goal was determining whether an algorithm could be written to correctly identify patients with prostate cancer treated with radiotherapy at the Hunter Holmes McGuire VAMC who did not have metastatic disease.
By using various values available within the CDW, such as ICD 9 codes, CPT codes, PSA laboratory values, dates, and other information, the authors were able to create a successful algorithm. The ability to complete the algorithm in a short time frame wasfacilitated by several factors: a very small group, weekly meetings, good communication, easy to understand concepts across all disciplines, ability to quickly determine whether the results of the algorithm were accurate or not, and high perceived value of the end product that served to motivate the team members. Each meeting ended with clear action items and a scheduled time for the next meeting. Throughout the design and implementation process, the team discussed any problems, planned solutions, and reviewed the status of project deliverables.
Results
The tracker has already been useful for reengaging patients in care and ensuring PSA testing is occurring at appropriate intervals. Of the more than 50,000 veterans currently alive who have received care at the Hunter Holmes McGuire VAMC, 1,158 were treated with radiotherapy definitively for prostate cancer. A total of 455 (39%) prostate cancer survivors had not been seen in the clinic in the past 13 months. Of these patients, 294 were being followed appropriately elsewhere within the VA system. Meanwhile, 161 neither had a PSA level nor a prostate cancer follow-up appointment recorded in the past 13 months anywhere within the entire VA system. This yielded a loss-to-follow-up rate of 14% (161/1,158).
The authors found that 21 (13%) of patients had a PSA level > 2.0 ng/mL above the posttreatment nadir.9 The authors were able to review the charts of these 21 patients to assess whether or not they required or were suitable for salvage brachytherapy. Of these, 1 has been set up for salvage high-dose rate brachytherapy treatment. Out of 50,000 patients, the PSA tracker algorithm facilitated a focus on the 21 patients who were most likely to be in need, making it possible for a nurse practitioner and physician to spend just 3 hours looking at charts instead of 3,000 hours.
Sustained use of the tracker is critically important to the Hunter Holmes McGuire VAMC project team and for the care of its veterans. Funds to support sustaining the program have been approved for fiscal year 2016. Efforts are underway to try to scale up the program and test the feasibility of disseminating the program across the enterprise. The authors estimate that an experienced advanced care provider would spend about 8 hours a week reviewing charts, contacting patients in the program, sending letters, and reviewing nuanced cases. The program would still benefit from increased automation as well as identifying a method for obtaining appropriate workload credit for this unique program.