Constructing Instrument Mapping Using Markov Boundary Techniques: A Case Study with Two Quality of Life Instruments

In psychiatry, it is common to have multiple available instruments for a concept of interest. Different studies may choose to measure the same concept with different instruments for reasons such as time and resource constraints. This makes it difficult to compare findings between studies, since differences in findings may be due to the use of different instruments. In this paper, we described and demonstrated a two step process to identify a correspondence mapping between QLS and GAF, two quality of life instruments. Specifically, we applied a Markov Boundary discovery algorithm, HITON-MB, to identify the Markov Boundary of GAF using different QLS subscales and items as candidate Markov Boundary members. The mapping function is then derived by regressing the identified Markov Boundary members to GAF. We applied the two step procedure to two independ datasets to examine the QLS-GAF correspondence across different patient population and training conditions. We identified the intrapsychic foundation subscale of QLS to be a consistent Markov Boundary member of GAF. We also reported the quality of the derived mapping functions.