Inappropriate antipsychotic medication use among older adults with dementia is associated with increased risk [pdf] of hospitalization and death. In 2017, the rate of potentially inappropriate use was 16%, having fallen from 24% in 2011. While this decline has been substantial, further decreasing the rate is an important goal to protect the health and wellbeing of nursing home residents. The new target of the National Partnership to Improve Dementia Care in Nursing Homes is a 15% reduction in the rate by the end of 2019.
Over the years, there have been many potential efforts to address this problem. Human Rights Watch published a report describing the practice as one meant to control resident behavior. Increased staffing, training, and stricter regulatory enforcement are previously discussed solutions to reduce inappropriate antipsychotic medication use.
Busch, Cohen, and Konetzka recently published a paper in Medical Care assessing the relative contribution of resident versus facility characteristics associated with inappropriate antipsychotic medication use. The relative contributions are essential so that efforts to improve the quality of care in nursing homes can be targeted.
In their analysis, the authors used information from over 1 million residents in 14,699 nursing homes in the US in 2014. There was substantial variability in the unadjusted rates of potentially inappropriate antipsychotic medication use among nursing homes. The mean rate was 18%, and the interquartile range was 11.3%-23.7% (SD 11.1). In 7% of the facilities, the rate of inappropriate use was higher than 35%! While this may naively appear to indicate that facility-level characteristics would drive differences–they do not. Fully 93% of the explained variation in potentially inappropriate antipsychotic medication use was attributed to resident characteristics, and 7% was attributed to facility-level characteristics.
The authors used several methods to determine the proportion of variation explained by resident and facility-level characteristics. The use of fixed-effects for nursing homes accounts for both observed and unobserved facility characteristics. Using this model in comparison with a model that controlled explicitly for different characteristics such as ownership, location, staffing, showed that 90% of the variation attributed to facilities was from unobserved characteristics. Looking at the variation explained by resident characteristics, mental health diagnoses accounted for the majority (69%) with cognitive impairment, accounting for an additional 20%.
In addition to figuring out which characteristics accounted for variation in potentially inappropriate antipsychotic medication use, the authors took a step further and simulated what would happen at the population level. The most substantial reductions were from the effects of Alzheimer’s and other dementias (4.7%) and other psychotic disorders (5.1%). Even the most significant changes in facility-level characteristics were associated with population changes of less than 1%. Of course, it’s possible that some facility-level characteristics were addressed early with Medicare’s initiatives that started in 2012 and were primarily dealt with by 2014 (the year used in this study).
While this analysis isn’t causal–it does suggest that current efforts might have more substantial effects if they were refocused. This research implies that developing new clinical strategies to treat diagnoses commonly seen in nursing home residents, such as dementia, would likely have a more significant impact on population rates of potentially inappropriate antipsychotic medication use than strategies designed to target facilities based on their characteristics. These treatments could be pharmacological, behavioral, or some combination. Hopefully, these treatments would not only prevent harm to patients but also improve their wellbeing and functioning.
Jess Williams