Related Experiment Videos
The P-A-I-N MMPI classification system: a critical review
Michael E Robinson1, Glenn I Swimmer, Dean Rallof
1Department of Clinical and Health Psychology, University of Florida, Gainesville, FL 32610 U.S.A. St. Charles Hospital Pain Management Center, Toledo, OH 43605 U.S.A. Department of Psychology, Bowling Green State University, Bowling Green, OH 43403 U.S.A.
Pain
|May 1, 1989
Summary
The Costello algorithm for clustering chronic pain patient profiles using the Minnesota Multiphasic Personality Inventory (MMPI) was too restrictive. Further validation of empirically derived clusters is needed before adopting literature-based methods.
Area of Science:
- Psychological assessment
- Pain management
- Data clustering algorithms
Background:
- The Minnesota Multiphasic Personality Inventory (MMPI) is a widely used tool for psychological assessment.
- Clustering algorithms are employed to group patients with similar MMPI profiles for research and clinical purposes.
- Literature-based clustering methods, like Costello et al., offer a standardized approach but require validation.
Purpose of the Study:
- To compare the Costello et al. literature-based MMPI clustering algorithm with a standard clustering procedure for chronic pain patients.
- To evaluate the restrictiveness and applicability of the Costello algorithm on a local patient sample.
- To assess the utility of MMPI clusters in predicting treatment response.
Main Methods:
- A literature-based clustering algorithm (Costello et al.) was applied to MMPI profiles of chronic pain patients.
- A standard clustering procedure was also used for comparison.
- The performance of the Costello algorithm was assessed based on its ability to classify MMPI profiles.
Main Results:
- The Costello et al. algorithm was found to be too restrictive, failing to classify 69% of the MMPI profiles in the local sample.
- This indicates significant limitations in its direct applicability to diverse patient populations.
- The study highlights the need for caution in adopting literature-based methods without local validation.
Conclusions:
- It may be premature to adopt literature-based MMPI clustering methods like Costello et al. without further validation.
- Empirically derived, locally relevant clusters are crucial for accurate patient classification.
- The predictive validity of local clusters in forecasting treatment response should be established before considering broader adoption of standardized algorithms.