A study that was done in collaboration between Montefiore Medical Center and the Apos Israel research group is already influencing the way medical treatments are measured in clinical practice. The study aimed to develop a questionnaire to assess patients, based on gold standard questions, that would be short enough to be practical in a clinical atmosphere, while also being accurate enough to render viable results.  

The 6-question questionnaire that was developed is a proposed Patient-Reported Outcome Measure (PROM), which is a self-administrated questionnaire that assesses a patient’s health state, quality of life, and functional status associated with their health condition.  

PROMs are regularly used as part of clinical trials to assess a patient’s condition and to assess the efficacy of the treatment being measured. There are growing efforts to implement the use of PROMs as a standard of care in real-life clinical practice in order to monitor and improve treatment results in practice.  

One of the challenges of using PROMs in a clinical setting is the fundamental difference between a clinical trial setting and a busy clinical practice. The difference between a clinical trial setting vs. a clinical practice setting include the following main elements: 

  • Patients lack the motivation, time, attention, and patience to invest a long time in completing a series of questionnaires.  
  • Clinicians usually do not have the time, and, in some cases, are skeptical about the contribution of PROMs to their care-service. 

The purpose of this research project was to create a new tool that will better fit the busy work routine of a typical clinical practice.  

We used two gold-standard PROMs that are commonly used in clinical trials to assess MSK conditions including knee osteoarthritis, the WOMAC and the SF-36. A decision tree classifier supported by a linear mix model regression was applied to determine, identify, and categorize the most influential questions that determine the overall score in each of the questionnaires.  

Six questions (2 from WOMAC and 4 from SF-36) were identified that can predict over 84% of the overall score. The six-question subset from a total of 60 questions in the WOMAC and SF-36 QOL scales yield over 50% of the sensitivity of the full surveys at a fraction of the overall burden of time and effort. Future work should focus on the validation and extension of this tool in clinical practice.  

The original publication can be found here