CaP Calculator: An Online Decision Support Tool to Improve Evidence-based Doctor-Patient Communication for Clinically Localized Prostate Cancer

Reviewer: Christine Hill, MD
Abramson Cancer Center of the University of Pennsylvania
Last Modified: September 24, 2008

Presenter: M.S. Katz
Presenter's Affiliation: Saints Medical Center, Lowell, MA
Type of Session: Scientific


  • Prostate cancer is the most commonly diagnosed cancer in men living in the United States, and most patients have localized disease at the time of diagnosis.
  • At the time of prostate cancer diagnosis, patients are faced with a multitude of decisions regarding treatment, and are often offered surgical treatments, radiotherapy treatments, systemic hormonal modulation, close observation, and combinations of these as potential treatment options. Each modality is associated with certain risks and benefits, and patients must be provided with considerable amounts of information in order to make informed decisions.
  • Certainly, a plethora of information is available regarding each treatment option, although direct comparison of treatments to one another is not available.
  • In addition, several risk-stratification schemes exist to predict for various outcomes and to guide treatment decisions. 
  • Each prostate cancer treatment has been demonstrated to be associated with a distinct pattern of change in quality of life, which influences patient satisfaction (Sandra MG, NEJM, 2008).
  • Patient satisfaction after treatment for prostate cancer has been demonstrated to be improved when decision-making is shared between patients and physicians (Kim SP, Cancer Invest, 2001). Shared decision making has been demonstrated to decrease regret regarding treatment decisions (Gugliotta A, Actas Urol Esp, 2008).
  • The authors of this study present an on-line decision support tool, “CaP Calculator,” developed within their institution. The objective behind development of this tool was facilitation of evidence-based doctor/ patient communication regarding treatment decisions.

Materials and Methods

  • The authors performed a literature search using MEDLINE and identified peer-reviewed articles providing risk assessment tools for men with biopsy-proven prostate cancers.
    • Articles were chosen that predicted pathologic stage and treatment outcomes for men with clinical stage T1 – 3 prostate cancer.
    • Predictive models to provide estimates of extracapsular extension, seminal vesicle invasion, and lymph node involvement were included.  
    • Models predicting treatment outcomes including PSA failure after radical prostatectomy or radiotherapy and development of distant metastases were also included.
    • Models were excluded based on the following criteria:
      • Use of pathologic data from radical prostatectomy
      • Use of older staging methods or tumor grading methods that could not be converted into contemporary parameters.
      • Inclusion of subjects with radiation doses less than 65 Gray (Gy).
    • 23 studies were identified that met these criteria and were included in this study.
  • Predictive models were then entered into Microsoft Excel tables, and linked to data entered by users.
  • Information was then converted to a Web-based format utilizing LAMP (Linux, Apache, MySQL, Perl) software platform (


  • CaP Calculator was designed for use by healthcare providers, and offers information to assist both patients and clinicians in decision-making.
  • Users are required to log on to this system, and to enter information regarding the individual patient.
  • When users provide the patient’s AJCC clinical T stage, Gleason score, age, and pre-treatment PSA, they are provided with eight estimates of pathologic findings (three predicting risk of extracapsular extension, two predicting risk of seminal vesicle invasion, and three predicting lymph node invasion). These estimates are provided via known, recognized nomograms and methods for risk-prediction. CaP Calculator provides clinicians and patients with a summary of these findings that may be utilized in decision-making.
  • Entering further patient information, including biopsy information (number positive cores, number of cores taken, percent of positive cores, and number of cores with Gleason 4-5 disease seen) increases the number of estimates of pathologic stage to 16 total.
  • CaP Calculator provides users with 5-year estimates of PSA control following prostatectomy, external beam radiotherapy, and brachytherapy, as well as risk of distant metastases at 5 years.
  • CaP Calculator was made available beginning in May 2007. Since that time, use has increased from approximately 100 uses monthly to approximately 550 uses monthly.

Author's Conclusions

  • The authors conclude that CaP Calculator provides a method for aggregation of prostate cancer predictive models simultaneously in a user-friendly format that is available via the Internet.
  • They note that it may be a valuable online resource to improve doctor-patient communication, providing that usefulness and accuracy are validated.

Clinical/Scientific Implications

  • As the authors point out, various predictive tools have been validated for use in prostate cancer decision making, including methods of stratifying patients by risk of biochemical failure, as well as methods of predicting pathologic stage based on clinical features.
  • The multitude of treatment options facing prostate cancer survivors may be confusing and overwhelming to patients, and predictive tools may be helpful in guiding decision-making.
  • The authors have developed a tool allowing aggregation of these models into one document. Ideally, this document will provide patients with the ability to compare predictive findings, and to make educated decisions.
  • As the authors point out, shared decision making has been associated with improved patient satisfaction and decreased regret regarding treatment decisions in patients with prostate cancer.
  • Although the tool they have developed may be quite useful, it is associated with certain limitations:
    • Firstly, the aggregate of predictive models produced by this tool is only as strong as the models themselves. Although the models utilized have been validated in the clinical setting, each has limitations, and patients and clinicians should be aware of these.
    • Secondly, the tool has currently been designed for use by clinicians. A “plain language” version for patient use could add considerably to this tool by allowing patients to better access and understand the clinical data presented.
    • Lastly, patients who receive information produced by CaP Calculator must be provided with appropriate counseling and assistance with decision-making. Although this tool provides a comprehensive comparison of clinical research, it certainly is not a substitute for excellent physician-patient communication; rather, it may serve to assist with communication.
  • Despite these limitations, the authors have developed an interesting tool that may prove to be an important clinical resource. As the authors point out, its accuracy still requires validation; however, pending this, CaP Calculator may contribute to the resources available to patients and physicians considering decisions regarding a new prostate cancer diagnosis.