If we cannot measure it, we cannot improve it: Understanding measurement problems in routine oral/dental assessments in Canadian nursing homes—Part I

TitleIf we cannot measure it, we cannot improve it: Understanding measurement problems in routine oral/dental assessments in Canadian nursing homes—Part I
Publication TypeJournal Article
Year of PublicationSubmitted
AuthorsHoben M, Yoon MN, Lu L, Estabrooks CA
JournalGerodontology
Volumen/a
Issuen/a
ISBN Number0734-0664
Abstract

Abstract Objective To compare Resident Assessment Instrument—Minimum Data Set 2.0 (RAI) oral/dental items collected by nursing home (NH) care staff to (a) assessments collected by trained research assistants (RAs) and (b) “gold standard” clinical assessments by dental hygienists (DHs). Background Routine collection of RAI oral/dental items is mandatory in most Canadian NHs. However, the performance of these items is less than optimal and oral/dental problems are severely under-reported. Accurate assessment is a prerequisite for preventing, detecting and treating oral health problems. Not knowing the reasons for performance problems is a barrier to improving performance of the RAI oral/dental items. Materials and Methods We included 103 NH residents from 4 NHs in Edmonton, Alberta, Canada. Using Kappa statistics, we compared the agreement of residents’ last (no older than 90 days) RAI assessment with RAI assessments completed by trained RAs and “gold standard” clinical assessments by DHs. We also assessed the inter-rater reliability (IRR) of RA and DH assessments. Results Care staff assessments had poor agreement with RA and DH assessments (Kappa < 0.2 for most items). RAs and DHs identified more oral/dental problems than care staff. However, IRR of RA assessments was low (Kappa < 0.7 for 7/9 items). IRR of DH assessments was acceptable (Kappa > 0.7) for most items. Conclusions The quality of RAI oral/dental assessments can be improved by better training care staff and ensuring appropriate time to do the assessments. However, remaining problems—even with trained RAs—suggest that rewording some of the items or supplementing them by more robust tools may be required.

URLhttps://onlinelibrary.wiley.com/doi/abs/10.1111/ger.12449
DOI10.1111/ger.12449