Article Text

Download PDFPDF
  1. Lin Yan1,
  2. Lisa M Lix1,2,
  3. Verena Schneider-Lindner3,
  4. Gary F Teare1,2,
  5. David Blackburn3,
  6. Nianping Hu2,
  7. Yvonne Shevchuk3
  1. 1School of Public Health, University of Saskatchewan, Saskatoon, Canada
  2. 2Health Quality Council, Saskatoon, Canada
  3. 3College of Pharmacy and Nutrition, Saskatoon, Canada


Introduction The Resident Assessment Instrument Minimum Data Set (RAI-MDS) is a clinical tool intended to provide comparable and consistent information about the functional status and care needs of residents of long-term facilities (LTCFs). The RAI-MDS has been used in numerous studies about quality of care, including studies about the use of medications associated with potentially adverse health outcomes. However, the accuracy of many of the RAI-MDS data elements is unknown.

Objectives This study examined the accuracy of RAI-MDS data on antipsychotic medication use. The study objectives were to: (1) assess agreement between the RAI-MDS and provincial prescription drug administrative data and (2) identify patient and facility characteristics associated with disagreement between these data sources.

Methods This study adopted a retrospective cohort design. The observation period was fiscal years 2004/2005–2010/2011 (ie, 1 April 2004–31 March 2011). The study cohort consisted of all individuals who resided in one of Saskatchewan's 168 LTCFs and had continuous health insurance for at least one fiscal year during the observation period.

Data sources included the RAI-MDS, provincial prescription drug, and person registry databases. The RAI-MDS collects data on initial admission to a LTCF, at regular intervals throughout the residence period, and at the time of major changes in health status. Information on antipsychotic medications is collected at each assessment based on the previous seven-day period. The prescription drug database collects information on all prescription drugs for provincial residents eligible for drug benefits; antipsychotic drugs are identified using the American Hospital Formulary System of drug classification. The person registry database contains demographic and health insurance coverage information. All databases can be linked via a unique personal health identifier.

For each observation year, the number of LTCF residents for whom antipsychotic drug use was reported on at least one day in the RAI-MDS assessments and who had at least one dispensation for an antipsychotic drug in the prescription drug database was identified. Cohen's κ was used to estimate agreement between the clinical and administrative data. Multivariate logistic regression analyses were used to identify patient and facility characteristics associated with disagreement.

Results The study population consisted of 15 007 LTCF residents. The average age was 80.6 years and 67.5% of residents were women. In 2010/2011, 42.1% of residents had received antipsychotic medications according to the MDS-RAI data; this percentage was similar for all study years. A similar percentage of residents had received antipsychotic medications according to the prescription drug database. Over the entire observation period, κ=0.80 (95% CI 0.78 to 0.81). κ Statistics were slightly higher in older than in younger residents. Estimates of agreement showed little variation across study years, and ranged from 0.77 to 0.82. Sex, age and non-licensed facility residence was associated with disagreement.

Conclusions This comparison of MDS-RAI clinical data and prescription drug administrative data suggests that the MDS-RAI may provide accurate information about antipsychotic medication use in LTCF residents. MDS-RAI data can be used to develop quality indicators for LTCFs and to monitor trends and variation in prescribing practices.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.