Research in Gerontological Nursing

Research Brief 

Predicting Off-Label Antipsychotic Medication Use in a Randomly Selected Nursing Home Sample Based on Resident and Facility Characteristics

William E. Mansbach, PhD; Ryan A. Mace, MS; Kristen M. Clark, BS; Isabella M. Firth, MA; Jacqueline K. Breeden, MBA

Abstract

Reducing off-label antipsychotic medication use for behavioral and psychological symptoms of dementia (BPSD) in nursing home residents has been a centerpiece of government regulation, but without insight into utilization differences based on resident and facility characteristics. To examine whether resident and facility characteristics can predict off-label antipsychotic medication treatment for BPSD, residents prescribed antipsychotic medication (N = 216) from 17 Maryland nursing facilities were randomly selected. Based on physician diagnoses, 59.7% of participants were taking off-label antipsychotic medications for BPSD. Hierarchical logistic regression results suggest that dementia level (indicated by Brief Cognitive Assessment Tool scores) and age, but not facility characteristics, significantly predict greater likelihood of using off-label antipsychotic medications for BPSD. Having moderate-severe dementia was associated with more than a four-fold increase in off-label antipsychotic medication use for BPSD. Off-label use of antipsychotic medications for BPSD remains high, especially for older nursing home residents with more severe dementia, indicating that more targeted reduction approaches are needed.

[Res Gerontol Nurs. 2016; 9(6):257–266.]

Abstract

Reducing off-label antipsychotic medication use for behavioral and psychological symptoms of dementia (BPSD) in nursing home residents has been a centerpiece of government regulation, but without insight into utilization differences based on resident and facility characteristics. To examine whether resident and facility characteristics can predict off-label antipsychotic medication treatment for BPSD, residents prescribed antipsychotic medication (N = 216) from 17 Maryland nursing facilities were randomly selected. Based on physician diagnoses, 59.7% of participants were taking off-label antipsychotic medications for BPSD. Hierarchical logistic regression results suggest that dementia level (indicated by Brief Cognitive Assessment Tool scores) and age, but not facility characteristics, significantly predict greater likelihood of using off-label antipsychotic medications for BPSD. Having moderate-severe dementia was associated with more than a four-fold increase in off-label antipsychotic medication use for BPSD. Off-label use of antipsychotic medications for BPSD remains high, especially for older nursing home residents with more severe dementia, indicating that more targeted reduction approaches are needed.

[Res Gerontol Nurs. 2016; 9(6):257–266.]

In 2005, the U.S. Food and Drug Administration (FDA; 2005) determined that the use of antipsychotic medications for behavioral disturbances in older adults with dementia is associated with increased mortality. In response, the Centers for Medicare & Medicaid Services (CMS; 2014b) have set specific targets for reducing antipsychotic medication use in American nursing homes and established the National Partnership to Improve Dementia Care. The Partnership issued guidelines based on the evidence that (a) dementia prevalence is high in nursing homes, and (b) the frequent use of antipsychotic medications for treating behavioral disturbances in patients with dementia is often inappropriate. Approximately 70% to 80% of U.S. nursing home residents likely have some type of cognitive impairment (CMS, 2013; Mansbach, Mace, & Clark, 2014), and many studies estimate that 50% to 60% have dementia (Magaziner et al., 2000; Mansbach, MacDougall, Clark, & Mace, 2014). Furthermore, behavioral and psychological symptoms of dementia (BPSD), such as physical and verbal aggression, delusions and other perceptual distortions, and anxiety and depression, are also common in patients with dementia (Kales, Gitlin, & Lyketsos, 2015).

Antipsychotic medications are commonly used to treat BPSD (Ballard, Waite, & Birks, 2006), despite not being FDA approved for such purposes, modest efficacy, and multiple adverse effects (FDA, 2005; Kamble, Sherer, Chen, & Aparasu, 2010; Lee et al., 2004). The prescription of antipsychotic medication is a primary intervention strategy for treating dementia-related behavioral problems, such as combativeness and verbal aggression, arguably the most common type of BPSD among nursing home residents (Salzman et al., 2008). Fortunately, the stakeholder base for rightsizing antipsychotic medication use for BPSD is growing. In fact, the American Psychiatric Association recently released new evidence-based recommendations on the use of antipsychotic medications for the treatment of BPSD that emphasize lower, judicious use in combination with nonpharmacological interventions (Reus et al., 2016).

The current study is the product of a public–private collaboration focused on improving care for nursing home residents with dementia in the state of Maryland. Stakeholders involved in this project include the Maryland Office of Health Care Quality, a long-term care trade organization, nursing home medical directors, and the study investigators. An important goal of the group is to develop “best practice” recommendations that will improve treatment of BPSD and help nursing homes meet federal guidelines for reducing antipsychotic medication use. Although nursing home operators and prescribers are aware of the need to reduce antipsychotic medication use, especially off-label use, there are few evidence-based guidelines instructing them how best to accomplish this goal. Understanding which resident and facility variables predict off-label use of antipsychotic medications for BPSD could help identify residents most likely to be prescribed these drugs and help nursing homes focus their drug reduction initiatives in an evidence-based manner.

To address this issue, resident and facility variables predictive of antipsychotic medication use among nursing home residents identified by previous research were examined. Several studies have confirmed a significant association for male gender (U.S. Government Accountability Office [GAO], 2015) and the diagnosis of dementia (Chen et al., 2010; Kamble, Chen, Sherer, & Aparasu, 2009; Lucas et al., 2014; Monette et al., 2012) with antipsychotic medication use. Research investigating associations with other resident characteristics is less conclusive (e.g., race, age) and understudied (e.g., education, marital status). Lucas et al. (2014) found that African American individuals were more likely to be prescribed antipsychotic medications than Caucasian individuals, whereas Chen et al. (2010) found that Caucasian individuals were more likely than African American individuals to be prescribed these drugs. Similarly, younger (Chen et al., 2010; Lucas et al., 2014) and older residents (Kamble et al., 2009) have both demonstrated a significant association with antipsychotic medication use. A number of facility characteristics have been associated with antipsychotic medication use, including high bed capacity, high facility-level antipsychotic medication use rates, low staffing, and facility for-profit ownership (Chen et al., 2010; GAO, 2015; Kamble et al., 2009; Lucas et al., 2014). However, these studies typically conflate general and off-label antipsychotic medication use. Hence, despite advances in identifying resident and facility characteristics associated with general antipsychotic medication prescribing, further investigation is needed to determine which factors are significantly predictive of off-label use for BPSD.

Two primary questions were investigated pertaining to off-label use of antipsychotic medications in nursing home residents. First, among residents prescribed antipsychotic medications, how common is off-label use of these drugs for BPSD? Although antipsychotic medication use in nursing homes in general has been well-documented, far less is known about use of antipsychotic medications specific to BPSD, especially behavioral disturbances. Second, are there specific resident and facility characteristics that predict off-label use of antipsychotic medications for BPSD? As a potentially important resident characteristic, levels of dementia are included in the current prediction model based on the idea that antipsychotic medication treatment for BPSD may better be understood by stage of dementia, rather than by dementia as a comprehensive or unity diagnosis. In a review of the literature, no previous studies that took this more granular approach of predicting off-label antipsychotic medication use for BPSD were found.

Method

Participants

Long-term care residents in 17 skilled nursing facilities in Maryland were randomly selected for the current prospective, cross-sectional, population-based study. Study eligibility required the use of antipsychotic medications prescribed by their attending nursing home providers as of February 1, 2014. This study investigated off-label prescribing patterns pertinent to behavioral problems (as opposed to all BPSD), as they are often targets of antipsychotic medication interventions. The rationale for this was three-fold: (a) behavioral issues, particularly symptoms of aggression, are among the most common in residents with dementia (Salzman et al., 2008); (b) behavioral problems are associated with higher risk management challenges, potential harm to residents, and facility survey deficiencies; and (c) behavioral symptoms are among the most challenging problems for facility staff and are associated with caregiver burnout (Cocco, Gatti, de Mendonça Lima, & Camus, 2003).

Of the 565 eligible residents across the 17 facilities, 231 were randomly selected for the current study. Inclusion in the final sample required a completed Brief Cognitive Assessment Tool (BCAT; Mansbach, MacDougall, & Rosenzweig, 2012) to assess levels of cognitive functioning, proficiency in English, and recent history of at least one behavioral disturbance. Residents were excluded (9.4% of those randomly selected) if they had medical or psychiatric impairments too severe to complete the BCAT. Residents with severe dementia were not excluded so long as they could complete the BCAT. Table 1 presents key demographic features for the final sample of the 216 randomly selected residents who met the above criteria and were included in the data analysis.


Select Demographics and Clinical Characteristics of Participants (N = 216)

Table 1:

Select Demographics and Clinical Characteristics of Participants (N = 216)

Skilled Nursing Facilities

The 17 skilled nursing facilities were recruited through a three-step process. First, letters containing the scope and goals of the research were sent to all nursing home administrators whose facilities were members of the long-term care trade organization (78 skilled nursing facilities). Second, all facilities who expressed interest in participating were contacted, and more detailed information about the study was provided. Third, administrators who indicated interest in participating were invited to attend a 2-hour orientation, after which consent to participate was obtained. All facilities were long-term care settings judged to be representative of U.S. nursing homes in general, in the sense that they had relatively equivalent proportions of residents with dementia (to one another and U.S. facilities). None were designated as “psychiatric” specialty care facilities, of which there are relatively few in the United States.

Procedure

Data for the current study were collected and procedures approved as part of the public–private collaboration to improve care for nursing home residents with dementia in the state of Maryland. The Medical Ethics Committee from each nursing home reviewed and approved all study procedures. Data from facilities and residents were de-identified in two ways. First, information pertaining to facility characteristics and most resident characteristics were from public use datasets collected by others (i.e., CMS). The data analysis was a secondary use of this publicly available data. Second, specific resident characteristics that were not public use datasets (i.e., BCAT scores) were stripped of identifying information by the participating facilities, who maintained the coding. The researchers had no access and no capability to identify residents at any time.

The goals of this nursing-based project were to (a) better understand current prescribing practices as they pertain to off-label use of antipsychotic medications for residents with behavioral disturbances, and (b) determine if specific resident characteristics, including dementia levels, and facility features predicted off-label antipsychotic medication use for BPSD. In this effort, nursing and clinical staff from participating facilities were trained and demonstrated proficiency in using the BCAT to assess cognitive functioning. The BCAT was selected as the cognitive measure because it has been found to account for an additional 47% of the variance in dementia diagnoses over and above the Brief Interview for Mental Status (BIMS; Chodosh et al., 2008; Mansbach, Mace, et al., 2014), which is the Minimum Data Set (MDS) 3.0 mandated cognitive instrument in U.S. nursing homes.

Each facility identified residents eligible for study participation (i.e., all residents prescribed antipsychotic medications for behavioral problems) based on individual institutional pharmacy lists. Using a computerized pseudorandom number generator, a study investigator randomly selected up to 15 eligible residents from each facility. This process allowed each eligible resident to have an equal likelihood of being selected for this population-based study. The decision to select up to 15 residents was based on balancing a sufficient sample size for the study objectives and analyses, with the additional time burden this project demanded from facility staff. To prevent potential bias, facility staff were masked from the participant selection process, and the researchers responsible for random selection were blind from all identifying information of the residents. All randomly selected participants subsequently completed informed consent and the BCAT with trained facility staff over a 1-month period (February to March 2014). Participants who were unable to complete the BCAT, for medical or physical reasons, were excluded and replaced by additional randomly selected participants. Individual facility social work (57%), nursing (35%), and recreation (8%) staff, who were trained and demonstrated proficiency in administering the BCAT, completed all of the testing for this study. BCAT tests completed by facility staff were audited to ensure the accuracy of cognitive assessment administration and scoring. Scoring errors on individual BCAT items, none of which changed the cognitive category of participants, were corrected and additional BCAT training was provided to facility staff. No significant differences were found between facility testers or the staff disciplines on BCAT scores (p > 0.05).

The most recent monthly physician order sheets (February 2014) were used to obtain dementia and psychiatric diagnoses for the current study. Psychiatric and dementia diagnoses were determined by the attending physician and were based on one or more of the following: (a) medical history provided on admission (e.g., via transfer summary, community physician, patient, or family); (b) informal assessment and interview; (c) transcription of diagnoses from consultant behavioral health providers; (d) other informants inclusive of facility staff, family, and community providers; and (e) MDS information. The BCAT was used to identify levels of cognitive functioning according to score ranges reported by Mansbach et al. (2012). The attending physicians who prescribed antipsychotic medication for patients with dementia did so without knowledge of BCAT scores.

The off-label use of antipsychotic medication was operationalized as prescribing these drugs for behavioral problems (e.g., physical aggression, verbal aggression) in nursing home residents with identified dementia, but without either schizophrenia or bipolar disorder. Participants with schizophrenia and bipolar disorder were identified as having FDA-approved conditions for these medications. The broader bipolar diagnosis was included rather than the specific diagnosis of bipolar mania. Participants with bipolar disorder either had a current or previous episode of mania.

Selection of Resident and Facility Characteristics

A two-step process was used to select resident and facility characteristics for the current model predicting off-label antipsychotic medication use for BPSD. First, characteristics previously reported in the literature as potentially predictive of antipsychotic medication use, as well as characteristics commonly associated with BPSD, were identified (Chen et al., 2010; GAO, 2015; Kamble et al., 2009; Lucas et al., 2014; Monette et al., 2012). Second, two medical directors from facilities not participating in the study were asked to review the list of resident and facility for comprehensiveness. Both suggested adding the Five-Star Quality Rating System used in U.S. nursing homes as important facility characteristics (CMS, 2014c). Despite concerns about the Nursing Home Compare data (Rahman & Applebaum, 2009), CMS quality ratings are widely used by patients and providers and considered acceptable for use in research studies (Sangl, Saliba, Gifford, & Hittle, 2005). In the most current version of the Five-Star system, antipsychotic medication use is included in the calculation of star ratings, although this was not the case at the time the medical directors made their suggestions. Resident characteristic data were obtained through MDS information and administration of the BCAT. Facility characteristic data were found in the publically available CMS (n.d.) Nursing Home Compare database. Table 1 and Table 2 present descriptive characteristics of the resident and facility variables, respectively.


Select Characteristics of Participating Facilities (N = 17)

Table 2:

Select Characteristics of Participating Facilities (N = 17)

Materials

Brief Cognitive Assessment Tool. The BCAT (Mansbach et al., 2012) was selected and approved as the objective assessment tool for this Maryland nursing-based project. This 21-item instrument was designed to assess cognitive functioning, and it can be individually administered (by licensed provider or paraprofessional) in approximately 15 minutes. It has demonstrated strong psychometric properties in nursing home facilities (Mansbach, Mace, et al., 2014). The BCAT produces a total score suggesting specific cognitive levels, and it yields a Contextual Memory Factor score and Executive Control Functions Factor score. In the original development study, psychometric analyses confirmed strong evidence for internal reliability (Cronbach's alpha = 0.92), test–retest reliability (r = 0.99), and evidence of the construct validity of BCAT score inferences through convergent, discriminant, and predictive validity analyses (Mansbach et al., 2012). Mansbach et al. (2012) did not find evidence that BCAT scores were biased with respect to several demographic characteristics (e.g., education, race, marital status) or type of residency (e.g., assisted living, independent living). The BCAT is sensitive to the full spectrum of cognitive functioning, including normal cognition, mild cognitive impairment (MCI), mild dementia, and moderate-severe dementia (Mansbach et al., 2012).

Statistical Analysis

Statistical analyses were performed using SPSS version 20. Descriptive statistics were used to report resident and facility characteristics. Hierarchical logistic regression was used to determine the extent to which the resident (Step 1) and facility (Step 2) characteristics significantly predicted off-label use of antipsychotic medications for BPSD. This multivariate approach has been commonly used to examine predictors of antipsychotic medication use among nursing home residents (Chen et al., 2010; GAO, 2015; Kamble et al., 2009; Lucas et al., 2014; Monette et al., 2012). Although the literature does not provide specific sample size requirements for logistic regression, the current study sufficiently meets a recommended observation-to-predictor ratio (Peng, Lee, & Ingersoll, 2002).

Results

Of the 231 randomly selected nursing home residents, who were all prescribed antipsychotic medications to treat behavioral problems, 216 met eligibility criteria for data analysis. Participants ranged in age from 41 to 100, with a mean age of 78.4 (SD = 12.1 years). As shown in Table 1, 68.1% of participants were female, 63.4% were Caucasian, 43.5% were widowed, 47.2% had 12 years of education or less, and 42.1% had at least some post-secondary education. Based on physician diagnoses, 59.7% of participants in the entire sample were identified as taking off-label antipsychotic medications for BPSD. The remaining 40.3% of the sample was not classified as taking off-label antipsychotic medications for BPSD due to a diagnosis of schizophrenia or bipolar disorder (23.6%), or the absence of dementia (16.7%). Based on BCAT scores, 81.8% (N = 129) of participants prescribed off-label antipsychotic medications for BPSD had moderate-severe dementia, whereas only 18.2% had mild dementia.

As shown in Table 3, a hierarchical logistic regression model was developed to determine the extent to which the resident and facility characteristics predicted off-label use of antipsychotic medications for BPSD. First, resident characteristics were entered into the model at Step 1. As depicted in Step 1 of Table 3, the total variance explained by the model as a whole was between 43.7% (Cox and Snell R2) and 59.1% (Nagelkerke R2), with 83.7% of participants correctly classified as using off-label antipsychotic medications for BPSD (χ2 [6, 216] = 109.23, p < 0.001). Of the resident characteristics, only BCAT scores (Wald χ2 = 25.04, p < 0.001, odds ratio [OR] = 0.92), age (Wald χ2 = 20.06, p < 0.001, OR = 1.12), and race (Wald χ2 = 5.36, p < 0.05, OR = 0.48) were significant. After entry of the facility characteristics at Step 2, the total variance explained by the model as a whole was between 47.1% (Cox and Snell R2) and 63.6% (Nagelkerke R2) (χ2 [2, 357] = 293.64, p < 0.001). The addition of the facility characteristics in Step 2 did not improve the classification of participants as using off-label antipsychotic medications for BPSD. In the final model, only BCAT scores (Wald χ2 = 22.59, p < 0.001, OR = 0.91) and age (Wald χ2 = 17.35, p < 0.001, OR = 1.12) remained significant. For every additional point on the BCAT, participants were 8.7% less likely to be prescribed off-label antipsychotic medications for BPSD. With every year increase in age, participants were 12.3% more likely to use off-label antipsychotic medications for BPSD. In the final model, none of the facility characteristics were significant (p > 0.05). BCAT scores and age had a significant and small correlation (r = −0.33, p < 0.001).


Summary of a Logistic Regression Analysis for Predicting Off-Label Antipsychotic Medication Use for Behavioral and Psychological Symptoms of Dementia

Table 3:

Summary of a Logistic Regression Analysis for Predicting Off-Label Antipsychotic Medication Use for Behavioral and Psychological Symptoms of Dementia

Discussion

Two primary questions were investigated in the current study. First, nursing home residents prescribed antipsychotic medications were randomly selected to determine the rate of off-label use for behavioral problems. Based on physician diagnoses, approximately 60% of the sample was identified as taking off-label antipsychotic medications for these BPSD. The current rate appears to be much lower than the 86% reported by Kamble et al. (2010) and 83% reported by the U.S. Department of Health and Human Services (2011). However, these differences may simply reflect recent successful CMS initiates to lower general antipsychotic medication use in nursing homes. In fact, CMS (2016) reports a 15% reduction in antipsychotic medication use in U.S. nursing homes in 2013 and a cumulative reduction of 27% through the third quarter of 2015. One of the challenges in comparing the current study findings to CMS data is that the sample was based on a cohort of residents in Maryland nursing homes prescribed antipsychotic medications, whereas the CMS national data are derived from a general cohort of nursing home residents (not specific to those prescribed antipsychotic medications). Although lowering general antipsychotic medication use may indeed decrease off-label use for BPSD, it does little to improve the understanding of resident or facility characteristics predictive of treatment with these medications. Moreover, it does not allow targeting off-label antipsychotic medication reductions in those residents most at risk for adverse reactions to these drugs.

The second research question aimed to address these knowledge gaps. Whether specific resident and facility characteristics significantly predicted off-label use of antipsychotic medications for BPSD were examined. Only two characteristics were significant in the final hierarchical logistic regression model, with level of cognitive functioning (indicated by BCAT scores) as the strongest predictor followed by age. As expected, BCAT scores and age were significantly associated; however, only 10.6% of their variance was shared, suggesting that each independently predicted off-label use. Specifically, increased severity of cognitive impairment and age were associated with greater risk of taking off-label antipsychotic medications for BPSD. Although advancing age and severity of dementia have been associated with a higher concentration of BPSD (Kales et al., 2015), the current study confirms that off-label antipsychotic medication use also becomes more prevalent.

Race was also significant at Step 1, such that Caucasian participants were associated with a higher risk of taking off-label antipsychotic medications for BPSD. However, race became non-significant with the entry of facility characteristics at Step 2. The literature on race-based variation in antipsychotic medication prescribing is mixed with respect to whether this variable is predictive of general antipsychotic use (GAO, 2015). The current finding pertains to off-label use only, and does not support race as a significant predictor when considering both resident and facility characteristics. Another variable previous research has identified as a significant resident characteristic is gender, with males more likely to be prescribed antipsychotic medications than females (GAO, 2015; Kamble et al., 2009; Lucas et al., 2014). However, support for this finding was not found in the current study. The differences in results could be explained by the fact that these studies examined antipsychotic medication use more generally, whereas the current study investigated off-label use for BPSD specifically.

No facility characteristics were found to predict off-label use in the final model. This finding was surprising and contrasts with some previous studies. For example, nurse staffing (Hughes, Lapane, & Mor, 2000; Kim & Whall, 2006) and facility size (Hughes et al., 2000) have been found to be significantly associated with antipsychotic medication use. Methodological differences might account for the current discrepant results. Kim and Whall (2006) used a dementia cohort sample for their study, whereas in the current study, residents prescribed antipsychotic medications were used as the primary cohort. A study by Hughes et al. (2000) was based on a retrospective, online survey of resident characteristics, whereas the current study used a prospective, cross-sectional design. The current findings are also interesting in light of the goals of the Five Star Quality Rating System, which is a tool created by CMS in 2008 to help consumers select and compare skilled nursing care centers (CMS, 2014c). The rating system has become a proxy measure for resident care in U.S. nursing homes, but in the current study, staffing and quality measures did not predict off-label antipsychotic medication treatment for BPSD. Since data were collected for the current study, the Five Star Quality Rating System added antipsychotic medication measures as key metrics. Future studies should explore whether inclusion of this new metric predicts off-label use for BPSD.

The current findings suggest that a useful first step toward lowering off-label antipsychotic medication use for BPSD would be to assess residents with instruments that have demonstrated ability to identify specific dementia levels. The BIMS, which is currently used in U.S. nursing homes as part of the MDS, is suitable for paraprofessional use and has evidence for its utility in identifying residents with any cognitive impairment and severe cognitive impairment (Saliba et al., 2012). However, the BIMS has demonstrated poor sensitivity for identifying dementia and differentiating among dementia stages (Mansbach, Mace, et al., 2014); therefore, it may not be an ideal tool for identifying the higher risk category of residents with more severe dementia when administered alone. In addition to the BIMS, providers may elect to use cognitive measures with greater sensitivity to specific dementia levels. The administration of both the BIMS and BCAT has been found to account for 88% of the variance in cognitive diagnoses among nursing home residents (Mansbach, Mace, et al., 2014; Mace, Mansbach, & Clark, 2016).

Accurately identifying residents with more advanced dementia and who are relatively older than their peers could alert prescribers to be particularly careful about the risks/rewards of off-label antipsychotic medications for treating BPSD. After all, the syndrome of dementia is not monolithic, but encompasses different cognitive and functional levels, and BPSD do not appear to be evenly distributed across dementia levels (Hersch & Falzgrat, 2007). In the current study, having moderate-severe dementia was associated with more than a four-fold increase in off-label antipsychotic medication use for BPSD relative to milder dementia. Providers who use nonpharmacological interventions for BPSD may do well to target these residents, as they may be at risk for higher morbidity and mortality based on their age and cognitive status. This approach would be consistent with CMS initiatives to change the antipsychotic medication prescribing culture in nursing homes by slowing the “medication first” automatic response (CMS, 2014a).

Limitations

There are limitations to the current study that should be considered. Residents too medically or psychiatrically impaired to complete the BCAT (i.e., 9.4% of those randomly selected) were excluded. Although the sample included residents with moderate-severe dementia (60.6%), the current findings may not generalize to those with medical or psychiatric impairments too severe to tolerate cognitive assessment. Second, whereas the residents in the current study were randomly selected from facility pharmacy lists, the participating facilities were not randomly selected. Seventy-eight Maryland nursing homes were contacted, but only 17 agreed to participate. Based on feedback from non-participating facilities, a barrier for participation was the time investment required of participating staff. Perhaps those facilities who chose to participate had the staff resources to do so, or perhaps they were more motivated to reduce off-label antipsychotic medication use in their institutions. Third, interrater reliability data are not available for the BCAT due to a lack of facility staff resources (i.e., time for cognitive screening was limited). Although previous psychometric evaluations of the BCAT have shown strong reliability, the omission of interrater reliability in the current study is noteworthy because it is a major predictor of off-label antipsychotic medication use. Fourth, off-label prescribing patterns pertinent to all BPSD were not investigated; behavioral problems (e.g., combativeness, verbal aggression) were the focus. For a more complete understanding of off-label prescribing patterns, future studies should examine the association between other BPSD and antipsychotic medication treatment, as well as possible differences among specific behavioral problems (e.g., physical versus verbal aggression) in specific contexts (e.g., during personal care, during meals).

Conclusion

There are many challenges in managing BPSD in nursing home residents, and antipsychotic medications remain a mainstay of intervention. The consensus that residents with dementia who are prescribed antipsychotic medications are at increased risk for medical complications and increased lethality highlights the importance of understanding off-label prescribing patterns in a more detailed and nuanced manner. The current findings suggest that considering resident and facility characteristics may help facilities concentrate their efforts in reducing off-label antipsychotic medications for those who are most vulnerable. Studies that investigate other potential predictors of off-label use are urgently needed to create best practices for dementia care.

References

  • Ballard, C.G., Waite, J. & Birks, J. (2006). Atypical antipsychotics for aggression and psychosis in Alzheimer's disease. The Cochrane Collaboration, (5). Retrieved from http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD003476.pub2/abstract;jsessionid=1C73E383959B8FF190B15AAE05A9B6B8.f04t02?deniedAccessCustomisedMessage=&userIsAuthenticated=false
  • Chen, Y., Briesacher, B.A., Field, T.S., Tjia, J., Lau, D.T. & Gurwitz, J.H. (2010). Unexplained variation across US nursing homes in antipsychotic prescribing rates. Archives of Internal Medicine, 170, 89–95. doi:10.1001/archinternmed.2009.469 [CrossRef]
  • Chodosh, J., Edelen, M.O., Buchanan, J.L., Yosef, J.A., Ouslander, J.G., Berlowitz, D.R. & Saliba, D. (2008). Nursing home assessment of cognitive impairment: Development and testing of a brief instrument of mental status. Journal of the American Geriatrics Society, 56, 2069–2075. doi:10.1111/j.1532-5415.2008.01944.x [CrossRef]
  • Centers for Medicare & Medicaid Services. (n.d.). Nursing home results: 228 nursing homes in Maryland. Retrieved from http://www.medicare.gov/nursinghomecompare/search.html
  • Centers for Medicare & Medicaid Services. (2013). Nursing home data compendium 2013 edition. Retrieved from http://www.cms.gov/Medicare/Provider-Enrollment-and-Certification/CertificationandComplianc/downloads/nursinghomedatacompendium_508.pdf
  • Centers for Medicare & Medicaid Services. (2014a). Interim report on the CMS National Partnership to Improve Dementia Care in Nursing Homes: Q4 2011 – Q1 2014. Retrieved from http://www.cms.gov/Medicare/Provider-Enrollment-and-Certification/SurveyCertificationGenInfo/Downloads/Survey-and-Cert-Letter-14-19.pdf
  • Centers for Medicare & Medicaid Services. (2014b). National Partnership to Improve Dementia Care exceeds goal to reduce use of antipsychotic medications in nursing homes: CMS announces new goal. Retrieved from http://www.cms.gov/Newsroom/MediaReleaseDatabase/Press-releases/2014-Press-releases-items/2014-09-19.html
  • Centers for Medicare & Medicaid Services. (2014c). Nursing home compare five-star quality rating system: Year five report. Retrieved from https://www.cms.gov/Medicare/Provider-Enrollment-and-Certification/CertificationandComplianc/Downloads/NHC-Year-Five-Report.pdf
  • Centers for Medicare & Medicaid Services. (2016). Update report on the National Partnership to Improve Dementia Care in Nursing Homes. Retrieved from https://www.cms.gov/Medicare/Provider-Enrollment-and-Certification/SurveyCertificationGenInfo/Downloads/Survey-and-Cert-Letter-16-28.pdf
  • Cocco, E., Gatti, M., de Mendonça Lima, C.A. & Camus, V. (2003). A comparative study of stress and burnout among staff caregivers in nursing homes and acute geriatric wards. International Journal of Geriatric Psychiatry, 18, 78–85. doi:10.1002/gps.800 [CrossRef]
  • Hersch, E.C. & Falzgrat, S. (2007). Management of the behavioral and psychological symptoms of dementia. Clinical Interventions in Aging, 2, 611–621. doi:10.2147/CIA.S1698 [CrossRef]
  • Hughes, C.M., Lapane, K.L. & Mor, V. (2000). Influence of facility characteristics on use of antipsychotic medications in nursing homes. Medical Care, 38, 1164–1173. doi:10.1097/00005650-200012000-00003 [CrossRef]
  • Kales, H.C., Gitlin, L.M. & Lyketsos, C.G. (2015). Assessment and management of behavioral and psychological symptoms of dementia. BMJ, 350, h369. doi:10.1136/bmj.h369 [CrossRef]
  • Kamble, P., Chen, H., Sherer, J.T. & Aparasu, R.R. (2009). Use of antipsychotics among elderly nursing home residents with dementia in the US: An analysis of national survey data. Drugs & Aging, 26, 483–492. doi:10.2165/00002512-200926060-00005 [CrossRef]
  • Kamble, P., Sherer, J., Chen, H. & Aparasu, R. (2010). Off-label use of second-generation antipsychotic agents among elderly nursing home residents. Psychiatric Services, 61, 130–136. doi:10.1176/ps.2010.61.2.130 [CrossRef]
  • Kim, H. & Whall, A.L. (2006). Factors associated with psychotropic drug usage among nursing home residents with dementia. Nursing Research, 55, 252–258. doi:10.1097/00006199-200607000-00005 [CrossRef]
  • Lee, P.E., Gill, S.S., Freedman, M., Bronskill, S.E., Hillmer, M.P. & Rochon, P.A. (2004). Atypical antipsychotic drugs in the treatment of behavioral and psychological symptoms of dementia: Systematic review. BMJ, 329, 75. doi:10.1136/bmj.38125.465579.55 [CrossRef]
  • Lucas, J.A., Chakravarty, S., Bowblis, J.R., Gerhard, T., Kalay, E., Paek, E.K. & Crystal, S. (2014). Antipsychotic medication use in nursing homes: A proposed measure of quality. International Journal of Geriatric Psychiatry, 29, 1049–1061. doi:10.1002/gps.4098 [CrossRef]
  • Mace, R.A., Mansbach, W.E. & Clark, K.M. (2016). Rapid cognitive assessment of nursing home residents: A psychometric comparison of two brief instruments. Research in Gerontological Nursing, 9, 35–44. doi:10.3928/19404921-20150522-05 [CrossRef]
  • Magaziner, J., German, P., Zimmerman, S.I., Hebel, J.R., Burton, L., Gruber-Baldini, A.L. & Kittner, S. (2000). The prevalence of dementia in a statewide sample of new nursing home admissions aged 65 and older: Diagnosis by expert panel. The Gerontologist, 40, 663–672. doi:10.1093/geront/40.6.663 [CrossRef]
  • Mansbach, W.E., MacDougall, E.E., Clark, K.M. & Mace, R.A. (2014). Preliminary investigation of the Kitchen Picture Test (KPT): A new screening test of practical judgment for older adults. Aging, Neuropsychology, and Cognition, 21, 674–692. doi:10.1080/13825585.2013.865698 [CrossRef]
  • Mansbach, W.E., MacDougall, E.E. & Rosenzweig, A.S. (2012). The Brief Cognitive Assessment Tool (BCAT): A new test emphasizing contextual memory, executive functions, attentional capacity, and the prediction of instrumental activities of daily living. Journal of Clinical and Experimental Neuropsychology, 34, 183–194. doi:10.1080/13803395.2011.630649 [CrossRef]
  • Mansbach, W.E., Mace, R.A. & Clark, K.M. (2014). Differentiating levels of cognitive functioning: A comparison of the Brief Interview for Mental Status (BIMS) and the Brief Cognitive Assessment Tool (BCAT) in a nursing home sample. Aging and Mental Health, 18, 921–928. doi:10.1080/13607863.2014.899971 [CrossRef]
  • Monette, J., Alessa, W., McCusker, J., Cole, M., Voyer, P., Champoux, N & Belzile, E. (2012). Association of resident and room characteristics with antipsychotic use in long-term care facilities (LTCF). Archives of Gerontology and Geriatrics, 55, 66–69. doi:10.1016/j.archger.2011.06.011 [CrossRef]
  • Peng, C.J., Lee, K.L. & Ingersoll, G.M. (2002). An introduction to logistic regression: Analysis and reporting. Journal of Educational Research, 96, 3–14. doi:10.1080/00220670209598786 [CrossRef]
  • Rahman, A.N. & Applebaum, R.A. (2009). The nursing home Minimum Data Set assessment instrument: Manifest functions and unintended consequences–past, present, and future. The Gerontologist, 49, 727–735. doi:10.1093/geront/gnp066 [CrossRef]
  • Reus, V.I., Fochtmann, L.J., Eyler, A.E., Hilty, D.M., Horvitz-Lennon, M., Jibson, M.D. & Yager, J. (2016). The American Psychiatric Association practice guideline on the use of antipsychotics to treat agitation or psychosis in patients with dementia. American Journal of Psychiatry, 173, 543–546. doi:10.1176/appi.ajp.2015.173501 [CrossRef]
  • Saliba, D., Buchanan, J., Edelen, M.O., Streim, J., Ouslander, J., Berlowitz, D. & Chodosh, J. (2012). MDS 3.0: Brief interview for mental status. Journal of the American Medical Directors Association, 13, 611–617. doi:10.1016/j.jamda.2012.06.004 [CrossRef]
  • Salzman, C., Jeste, D.V., Meyer, R.E., Cohen-Mansfield, J., Cummings, J., Grossberg, G.T. & Zubenko, G.S. (2008). Elderly patients with dementia-related symptoms of severe agitation and aggression: Consensus statement on treatment options, clinical trials methodology, and policy. Journal of Clinical Psychiatry, 69, 889–898. doi:10.4088/JCP.v69n0602 [CrossRef]
  • Sangl, J., Saliba, D., Gifford, D.R. & Hittle, D.F. (2005). Challenges in measuring nursing home and home health quality: Lessons from the First National Healthcare Quality Report. Medical Care, 43(3 Suppl.), I24–I32. doi:10.1097/00005650-200503001-00005 [CrossRef]
  • U.S. Department of Health and Human Services. (2011). Medicare atypical antipsychotic drug claims for elderly nursing home residents. Retrieved from https://oig.hhs.gov/oei/reports/oei-07-08-00150.pdf
  • U.S. Food and Drug Administration. (2005). Public health advisory: Deaths with antipsychotics in elderly patients with behavioral disturbances. Retrieved from http://www.fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/ucm053171.htm
  • U.S. Government Accountability Office. (2015). Antipsychotic drug use: HHS has initiatives to reduce use among older adults in nursing homes, but should expand efforts to other settings. Retrieved from http://www.gao.gov/assets/670/668221.pdf

Select Demographics and Clinical Characteristics of Participants (N = 216)

Characteristicn (%)
Age (mean, SD)78.4 (12.1)
Gender
  Female147 (68.1)
  Male69 (31.9)
Race
  Caucasian137 (63.4)
  African American74 (34.3)
  Other5 (2.4)
Marital status
  Widowed94 (43.5)
  Single, never married55 (25.5)
  Married37 (17.1)
  Divorced30 (13.9)
Years of education
  <1219 (8.8)
  1283 (38.4)
  13 to 1529 (13.4)
  1634 (15.7)
  >1623 (10.6)
  Trade school5 (2.3)
  Missing23 (10.6)
BCAT identified cognitive levels
  No dementia11 (5.1)
  Mild cognitive impairment36 (16.7)
  Mild dementia38 (17.6)
  Moderate-severe dementia131 (60.6)

Select Characteristics of Participating Facilities (N = 17)

Characteristicn (%)
Geographic location
  Suburban14 (82.4)
  Rural2 (11.8)
  Urban1 (5.9)
Ownership status
  Profit9 (52.9)
  Non-profit8 (47.1)
Licensing
  Stand-alone13 (76.5)
  Continuing care4 (23.5)
Bed size
  <992 (11.8)
  100 to 19910 (58.8)
  >2005 (29.4)
Antipsychotic medication usea
  Low (<10%)2 (11.8)
  Moderate (10% to 19%)10 (58.8)
  High (≥20%)4 (23.5)
CMS staffing rating
  5 stars1 (5.9)
  4 stars11 (64.7)
  3 stars1 (5.9)
  2 stars4 (23.5)
CMS quality measure rating
  5 stars13 (76.5)
  4 stars4 (23.5)

Summary of a Logistic Regression Analysis for Predicting Off-Label Antipsychotic Medication Use for Behavioral and Psychological Symptoms of Dementia

Step 1: Participant LevelStep 2: Facility Level
PredictorsWald χ2BSEORWald χ2BSEOR
Participant
  Age20.06**0.110.021.1217.35**0.120.280.12
  Sex1.66−0.650.510.521.04−0.550.540.58
  Education1.030.170.161.180.440.130.191.13
  Race5.36*−0.750.320.481.23−0.430.390.65
  Marital status1.410.220.191.252.460.320.201.37
  BCAT25.04**−6.410.170.9222.59**−0.900.020.91
Facility
  Location1.080.610.591.85
  Ownership0.15−0.250.660.78
  Licensing2.581.230.773.44
  Antipsychotic medication rate3.350.930.512.52
  Bed size0.12−0.080.600.92
  CMS staffing0.450.210.321.24
  CMS quality0.560.460.611.58
  Model χ2109.23120.82
  Classification83.7%82.6%
   R20.44 to 0.590.47 to 0.63
Authors

Dr. Mansbach is Founder and CEO, Mr. Mace is Research Assistant, and Ms. Clark is Senior Vice President, Mansbach Health Tools, LLC, Simpsonville; and Ms. Firth is President and Ms. Breeden is Grant Coordinator, LifeSpan Network, Columbia, Maryland.

Dr. Mansbach has ownership rights of the Brief Cognitive Assessment Tool, the cognitive instrument used in this study. The remaining authors have no potential conflicts of interest, financial or otherwise. The authors thank The Beacon Institute and Maryland Office of Health Care Quality for their support of this research.

Address correspondence to William E. Mansbach, PhD, Founder and CEO, Mansbach Health Tools, LLC, P.O. Box 307, Simpsonville, MD 21150; e-mail: wmansbach@thebcat.com.

Received: June 27, 2016
Accepted: September 08, 2016
Posted Online: September 27, 2016

10.3928/19404921-20160920-03

Sign up to receive

Journal E-contents