Validity and Reliability of HLD Treatment Need Indices Currently Used in the United States

Thesis (Master's)--University of Washington, 2020Introduction: State Medicaid agencies rely on treatment need indices to identify patients with orthodontic medical necessity. The Handicapping Labiolingual Deviations (HLD) index is the most common orthodontic index in the United States. States which have adopted the HLD index have modified it without supporting validation studies. The purpose of this study was to validate all state-modified HLD indices currently used in the United States and to investigate the implementation of the Washington state HLD modification, HLD(WA). Methods: To validate all current HLD indices, a panel of 17 orthodontists was recruited to assess the malocclusion of 204 consecutive patient records obtained from the University of Washington Department of Orthodontics archives. State HLD indices were subsequently applied to all 204 patients to predict state-specific orthodontic treatment need decisions. ROC curves were generated for each state-modified HLD indices to determine each index’ diagnostic value. We also computed the optimal point thresholds for all state-modified HLD indices. To evaluate the implementation of HLD(WA), the HLD prior-authorizations submitted on behalf of 418 Medicaid-eligible patients by five orthodontic clinics were gathered and compared to corresponding Washington state Medicaid program treatment decisions. We evaluated approval and denial rates among different clinics and compared applications approved and denied according to index guidelines versus reviewer override. Results: There were many qualitative differences among state-modified HLD indices. State-modified HLD indices had a mean between-state agreement of 79.1%, a mean sensitivity of 41.6%, and a mean specificity of 92.9%. The optimal point thresholds were lower than the current point thresholds for all state-modified HLD indices. When the implementation of HLD(WA) was evaluated, we found a mean approval rate of 35.4% and a mean denial rate of 52.4%. While total mean approval and denial rates were not significantly different among clinic types, there were significant differences among clinic types in the number of applications approved and denied according to index guideline versus reviewer override (p<.01). Conclusions: Our results suggest that state-modified HLD indices fail to identify a considerable percentage of handicapping malocclusions. There is also a lack of consensus among state Medicaid programs on what constitutes a handicapping malocclusion. In addition, our study indicates that there are differences in approval and denial rates among clinic types. These findings may exacerbate access to orthodontic care disparities among Medicaid-eligible populations