Assessing the Impact of an Interdisciplinary Team Approach Using the ARMOR Protocol on the Rate of Psychotropic Medications and Other Quality Indicators in Long-Term Care Facilities

Polypharmacy involving the use of psychotropic medications are common^1,2 in the nursing home (NH) population, where frail older adults with multiple chronic conditions are more likely to receive such medications.^2,3 Antipsychotic medications are commonly used to manage the behavioral and psychological symptoms of dementia.⁴ According to the Centers for Medicare & Medicaid Services (CMS), 39.4% of NH residents who had cognitive impairments and behavioral problems, but no diagnosis of psychosis or related conditions, received antipsychotic drugs from July 2010 to September 2010.⁵

Several large studies have demonstrated an association between treatment with antipsychotics and increased morbidity and mortality in people with dementia⁶ and in older adults, in general.⁷ The economic impact of polypharmacy is also substantial, with annual medication-related issues costing $7.6 billion in nursing facilities alone.⁴ Therefore, achieving optimal medication management in frail NH residents, by reducing the burden of unnecessary medications, can potentiate a reduction in the cost of care and improve the quality of care for residents. 

Although multiple tools such as the screening tool to alert doctors to the right treatment (START), the Screening Tool of Older Persons’ Prescriptions (STOPP), the Good Palliative Geriatric Practice (GPGP) algorithm, and the Potentially Inappropriate Medications (PIMS) list appear promising, the efficacy of such interventions regarding their functional impact on the overall quality of care has not been adequately studied.^8-11 Traditional tools to monitor for problems related to drug prescribing in NHs include monthly visits from a pharmacist and the use of the Minimum Data Set (MDS), which requires monitoring difficult behaviors and drug use review by the CMS.^12,13 Outcomes of such efforts to improve drug prescribing in long-term care (LTC) have had limited results. Current literature on deprescribing for older patients in different care settings shows some promise with educational and multidisciplinary approaches.^14-16 However, most interventions do not highlight the use of any specific tool to achieve success, and they support a team-based intervention with education and training components.^16,17

The Assess, Review, Minimize, Optimize, Reassess (ARMOR) protocol, first published in 2009, was developed to approach polypharmacy in a systematic and organized fashion.¹⁸ The intent of the tool, which has been used in various settings, is to maintain a balance between evidence-based medicine and the physiology of aging; functional status, comfort, and mobility are upheld as the essential outcome measures for any medication change while using evidence-based strategies. The tool promotes the American Geriatrics Society multimorbidity management guidelines and was chosen for its comprehensive biopsychosocial approach, complete medication regimen review, and risk assessment feature.¹⁹ 

We chose to use the ARMOR tool for our intervention, as evidence suggests that a team-based approach is more promising in addressing inappropriate prescribing in older residents. In addition, the ARMOR tool has the potential to fulfill both components of training and education for the users. Additionally, a pilot study from the Wheaton Franciscan Healthcare system has reported some success in the reduction of multiple medications, falls, and pharmacy costs using the ARMOR tool.²⁰  

For our current study, we implemented the ARMOR protocol in an interdisciplinary team (IDT) setting to address the use of psychotropics and measured its impact on 7 CMS NH quality indicators.²¹ The CMS quality indicators include falls and activities of daily living (ADLs), and reflect the well-known adverse events associated with the use of psychotropic drugs. Moreover, an IDT-based setting was chosen to achieve effective communication and a direct interface with care providers, the lack of which are known barriers in addressing polypharmacy in NHs. The aim of this study was to examine the impact of a team-based intervention program on the following: (1) the rate of psychotropic prescribing pattern over a 1-year period and (2) the impact on other CMS quality indicators before and after the implementation of the intervention, using the ARMOR protocol, in 5 NHs. 

Methods

The project was a collaborative initiative between the Michigan Peer Review Organization (MPRO) and Nexcare Health Systems. It was approved by their institutional ethics committees, and informed consent was obtained from all residents and their families. 

Study Design

We conducted a 12-month, team-based intervention using a nonrandomized, single arm, pre- and post-baseline vs an intervention study design to investigate the impact of the ARMOR tool protocol with an IDT on the following 7 CMS NH quality indicators: antipsychotic use, moderate/severe pain, falls, antianxiety use, psychological symptoms reported as behaviors (ie, behaviors that are considered disruptive or potentially harmful to self and others, including caregivers), depression, and ADLs, as reported through the MDS.²¹ Our baseline measurements were taken from October 1, 2011 through September 30, 2012. We implemented our protocol in October 2012, and data were collected for 1 year, ending September 2013. 

Setting and Participants 

A total of 600 resident charts in 5 suburban LTC facilities of a for-profit organization in Michigan were evaluated during the intervention period. LTC residents with the following characteristics were included: 

• Were on psychotropic medications (ie, antidepressants, antianxiety, and antipsychotics; both as needed and routine use) 
• Experienced recent or frequent falls (as defined in the MDS manual as unintentional change in position coming to rest on the ground or unto the next lower surface in the last 30 days)
• Were reported to have behaviors causing distress to self or others 
• Had significant changes in their medical condition as noted on the MDS (during the last 30 days)

To reduce the reactive effects of the experimental procedure and to improve the external validity of the design, residents receiving short-term rehabilitation were not included. This is because the intervention was focused on LTC residents, and short-term residents represent a variable age cohort in NH populations. In addition, short-term residents were more likely to be discharged within a few weeks of admission (Figure 1). We also excluded residents receiving hospice and respite care.

Article continues below Figure 1

IDT Composition 

At each NH, our IDT consisted of several key care team members who were regularly interacting with the residents. Our core team consisted of the director of nursing, a nurse manager, a social worker, an activity director, and the medical director. Noncore members of the team included rehabilitation therapists, consultant pharmacists, dieticians, and certified nursing aides. Input from noncore members on the patient’s general health status (eg, any behaviors, food intake, participation in activities, or other notable concerns) and recommendations were requested periodically. The core team composition for all participating facilities remained unchanged. The consultant pharmacists were invited to all meetings but were only able to participate in a few of the meetings due to their scheduling limitations. 

Procedure

We streamlined the processes of the existing IDT by expanding the role of the medical director, identifying residents with psychotropic use and polypharmacy, and reviewing their medication regimens monthly using the ARMOR protocol. Charts were reviewed by the IDT during the intervention period using the previously mentioned inclusion criteria. The care plans of the residents were systematically reviewed for the indications of all of the prescribed medications using the ARMOR protocol. A comprehensive review of the residents’ functional status and biopsychosocial profile was completed with feedback and input from core IDT members. The core IDT received 2 hours of training before the implementation of resident review utilizing the ARMOR process. On average, each monthly intervention lasted for 1.5 hours for the duration of the pilot project. 

ARMOR Tool 

The ARMOR tool is a stepwise approach to assess a geriatric patient (Table 1). The IDT used the tool when reviewing charts of the residents who met the inclusion criteria. The 5 steps to the process are as follows: Assess, Review, Minimize, Optimize, Reassess.

Assess consists of first reviewing vitals and current health status as recorded in the chart by the physician and nursing staff and then assessing the individual resident charts for the prescribed psychotropic medications and certain groups of medications that have the potential to cause adverse outcomes. 

Review involves analyzing the prescribed medications for potential drug-drug and drug-disease interactions and any possible impact of each medication on the resident’s functioning (eg, balance, fall, appetite, bladder, or bowel function), mood, and behaviors. During this step, team members of each discipline share their observations and assessments with the entire IDT, discuss the need for psychotropic use, and identify any unmet need in cases involving disruptive behaviors as well as medications, pain, bowel issues, or bladder issues triggering such behaviors. The IDT weighs an individual medication’s benefits against primary body functions (ie, appetite, weight, pain, mood, vision, hearing, bladder, bowel, skin, swallowing, activity level). 

Minimize involves discontinuing all prescriptions that lack clear evidence of their usage in each patient and/or medications whose risks of use outweigh benefits of use and those that have a high potential for increasing changes of adverse events, such as falls or decline in function. 

Optimize seeks to improve medication regimens by addressing any duplication and redundancy. The goal is to try to follow the dictum of “one medication for one condition” wherever possible with team consensus. This step also involves adjusting renally cleared medications to creatinine clearance and dose adjustment for those that are metabolized in the liver for clearance. If the medication is found to be renally metabolized, then the dose is adjusted based on the creatinine clearance. If the medication is found to be hepatically metabolized, then the dose is adjusted based on liver function in patients with established hepatic disease. Additionally, in this step, consideration for gradual dose reduction for antidepressants is also undertaken to achieve optimal medication regimen. 

Reassess entails that the overall recommendations generated by the IDT throughout the process are communicated to the attending physician for the patient for final approval and implementation. The residents are reassessed in a follow-up team meeting.

Outcome Measures

NH quality indicators, used by the CMS, are metrics obtained from the MDS and are used to report quality of care in NH.²¹ We examined the impact of our intervention on 7 of these indicators from the monthly Certification And Survey Provider Enhanced Reports (CASPERs) generated from MDs data (Table 2): 

• Residents who are on psychotropic medications (antipsychotic use indicator)
• Residents who had falls (falls indicator)
• Residents who are exhibiting behaviors deemed disruptive or potentially harmful to self or others, including caregivers. Such behaviors range from simply refusing care to yelling, pacing, and other psychomotor activities (disruptive behaviors indicator)
• Residents who use antianxiety or hypnotic medications (antianxiety indicator) 
• Residents who reported depressive symptoms on screening (depression indicator)
• Residents who reported moderate to severe pain on screening during the reporting period (pain indicator)
• Residents who had an increase in the need for help to perform ADLs (increased need for assistance with ADL indicator) 

Data Analysis 

The chi-square test was used to compare differences between the baseline and the intervention period for the 7 outcome measurements. The rates were calculated by dividing the number of resident charts evaluated per ARMOR protocol every month by the number of residents in that facility for that month. The difference in rates from baseline to intervention was calculated by subtracting the baseline prescribing rate from the intervention group prescribing rate. A value of 0 indicated no change, an increase in rate is indicated by a negative value, and a positive value indicated a decrease in the rate of a measurement from baseline. For each quality indicator, a risk improvement rate, which represented the rate of change for an indicator, was calculated using the following formula: (baseline rate - intervention rate) / baseline rate x 100.

Descriptive statistics were used to describe outcome variables monthly for 12 months, for both baseline and intervention rates. We used Pearson product-moment correlation coefficient equations to measure the strength of the linear association between 2 outcome variables and an ordinary least squares method to fit a linear trend equation. Throughout the analyses, the .05 level on a 2-sided design-based test of significance represented the cutoff value for assessing statistical significance. All analyses were conducted with SAS 9.1.3 (SAS Institute Inc, Cary, NC). Our goals were to discern whether the rate of psychotropic prescribing changed compared with the intervention period and whether the reported rate of quality indicators increased or decreased compared with the baseline rates. 

Results

A total of 600 resident charts were evaluated monthly in the 5 facilities over the 12-month period. Two noncore team members were replaced in one facility during the intervention period due to administrative reasons. After 3 months, one facility lost a core member due to change of employment; however, this did not significantly affect the scheduled IDT process.  

The monthly rates of antipsychotic use were compared for the periods before and during intervention. The rates of antipsychotic use in the intervention period were observed to be consistently lower than the baseline rates (Table 3). The difference in the rate of antipsychotic use from baseline to intervention was consistently lower, ranging from 1.67% in the first quarter to 4.64% in the last quarter of team intervention. The average rate of reduction or antipsychotic risk improvement rate with intervention was 13.9%. 

The 7 measured CMS quality indicators showed significantly positive or negative linear trends during the intervention period (Table 4). A positive change from baseline to intervention means that the frequency of the measured indicator decreased, and a negative change from baseline to intervention means that the frequency of a given indicator increased. The rate of disruptive behaviors compared with the baseline rate declined significantly during the intervention period. The rate of disruptive behaviors decreased by 18.2%. The change from the baseline to intervention was positive since the frequency of the disruptive behaviors decreased. The rates of increased need for ADL and depression also showed a statistically significant downward linear trend, signifying that the need of assistance with ADLs and the frequency of reported depression decreased (Table 4 and Figure 2). 

During the intervention period, we noticed a negative trend in 3 CMS indicators; pain, falls, and antianxiety showed a statistically significant increase. In other words, the change from the baseline to the intervention period demonstrated an increase in the frequency of reported falls, an increase in reported pain, and an increase in the use of antianxiety medications. These results were unexpected. 

We also noticed an inverse relationship between 2 sets of CMS indicators. That is to say that a reduction in the use of psychotropic medications was associated with an improvement in ADLs and fewer reports of depression, and also was linked to an increase in the rate of falls and reports of pain. The lower use of antipsychotics also appears to unmask untreated anxiety, expressed in the results as the increased rate of antianxiety medications.

Discussion

This IDT-based intervention using ARMOR was associated with lower monthly rates of antipsychotic use compared with the baseline rates and a statistically significant decline in the rate of reported behaviors. Moreover, several CMS quality indicator rates showed statistically significant improvements between pre-intervention and intervention periods. The findings of this pilot study are encouraging and point to the importance of a team-based approach in addressing the systemic problem of polypharmacy, especially the use of psychotropics, in LTC facilities.

A key feature of our study was the use of an IDT-based approach with a focus on educating and training the care team. In our opinion, the education component is a lasting—albeit intangible—benefit of our efforts. The benefits of an IDT approach are supported by emerging evidence suggesting that it may be a promising factor in deprescribing strategies in older residents.^14-16

We found that the prescribing rate of antipsychotics (antipsychotic use indicator) in the intervention group over the 12-month period was significantly reduced compared with the baseline rate. A change was also observed in the rates of other quality indicators (Figure 2, Table 3, and Table 4) during the intervention period. We also observed an unexpected increase in the rates of falls and reported pain. Since the evidence suggests that the higher rate of psychotropic use is commonly associated with falls in older adults, one would expect a reduction in the rate of falls with a concomitant reduction in the rate of antipsychotic use.^22,23 The fact that this was not the case in our study points to the complex interplay of various system-based factors in caring for NH residents. A plausible explanation for this paradoxical increase in falls could be the reduction of so-called “chemical restraint” associated with psychotropics. 

Moreover, the reduction of psychotropics increases the activity level of residents, as indicated by the simultaneous reduction in the need for assistance with ADLs with intervention. Arguably, it could also enable residents to be able to better express their previously “inexpressible needs,” such as unexpressed pain.^24,25 However, the traditional poor staffing ratio of nursing aides to nurse per patient—a familiar reality in NH care—could make it difficult to meet the needs of increasingly verbal and mobile residents.^26,27 We postulate that this may account for an increase in the rates of the quality indicators of reported pain and falls. Given the limited duration of intervention and the scope of our pilot, however, it is not sufficient to conclude this with certainty. In the future, this potentially unique relationship of different indicators in NH settings should be further explored.

Another surprising finding in our intervention is the reduction in reported depression on MDS screening and an unexpected increase in the rate of antianxiety use with the intervention. These findings raise interesting questions about the interplay between antipsychotic use, reported pain, falls, and the use of antianxiety agents. This raises an important question regarding the use of antianxiety medications in the context of previously “inexpressible needs” and the possibility of pain and pain-related behaviors being perceived by caregivers as anxiety. Unfortunately, our study did not measure the prescribing pattern of pain medications, so this remains conjecture at this point.

A hard copy of the ARMOR protocol was used during the intervention. We anticipate that an electronic interface of the ARMOR standardized processes and the use of the electronic medical record would further reduce the time necessary for reviewing medications. We posit that the ARMOR pilot intervention is reproducible with a few changes tailored to the uniqueness of team dynamics and needs.  

It was not the primary intention or scope of the study to assess the relationships between the rates of measured indicators. Thus, definite conclusions cannot be elucidated from our analysis. However, our findings point toward a need to further explore the complexities inherent in NH systems of care to strive for a meaningful interpretation of quality indicators in the future.  

The limitations of our study include a relatively short duration of intervention and a lack of randomization. Other limitations of the pilot study lie within the methodology and complexity of existing care models in NHs. Nonrandomization of the sample in our baseline vs intervention design could have effects on the internal validity of the findings. Future studies should utilize control groups to increase the rigor and validity of these findings. 

Conclusion

This pilot study is an attempt to explore the impact of an IDT-based intervention on psychotropic prescribing rate in the LTC setting. These preliminary findings are encouraging and highlight the importance of a team-based approach in addressing polypharmacy in LTC facilities. A randomized controlled intervention study is needed to further explore the impact of this intervention. 

References

1. van der Spek K, Gerritsen DL, Smalbrugge M, et al. PROPER I: frequency and appropriateness of psychotropic drugs use in nursing home patients and its associations: a study protocol. BMC Psychiatry. 2013;13:307. doi:10.1186/1471-244X-13-307

2. Kamble P, Chen H, Sherer J, Aparasu RR. Antipsychotic drug use among elderly nursing home residents in the United States. Am J Geriatr Pharmacother. 2008;6(4):187-197. 

3. Moga DC, Taipale H, Tolppanen AM, et al. A comparison of sex differences in psychotropic medication use in older people with Alzheimer’s disease in the US and Finland. Drugs Aging. 2017;34(1):55-65. doi:10.1007/s40266-016-0419-5

4. United States General Accounting Office. Adverse Drug Events, The Magnitude of Health Risk Is Uncertain Because of Limited Incidence Data. https://www.gao.gov/new.items/he00021.pdf. Published January 2000. Accessed March 10, 2019.

5. Center for Medicare Advocacy. Reducing antipsychotic drug use in nursing homes: save residents’ lives, save Medicare billions of dollars. Center for Medicare Advocacy website. https://www.medicareadvocacy.org/reducing-antipsychotic-drug-use-in-nursing-homes-save-residents-lives-save-medicare-billions-of-dollars/#_edn2. Accessed March 1, 2019.

6. Schneider LS, Dagerman KS, Insel P. Risk of death with atypical antipsychotic drug treatment for dementia: meta-analysis of randomized placebo-controlled trials. JAMA. 2005;294(15):1934-1943.

7. Maher RL, Hanlon JT, Hajjar ER. Clinical consequences of polypharmacy in elderly. Expert Opin Drug Saf. 2014;13(1):57-65. 

8. Barry PJ, Gallagher P, Ryan C, O’Mahony D. START (screening tool to alert doctors to the right treatment)--an evidence-based screening tool to detect prescribing omissions in elderly patients. Age Ageing. 2007;36(6):632-638. 

9. Gallagher P, Mahoney D. STOPP (Screening Tool of Older Persons’ potentially inappropriate Prescriptions): application to acutely ill elderly patients and comparison with Beers’ criteria. Age Ageing. 2008:37(6):673-679. doi:10.1093/ageing/afn197

10. Garfinkel D, Mangin D. Feasibility study of a systematic approach for discontinuation of multiple medications in older adults: addressing polypharmacy. Arch Intern Med. 2010;170(18):1648-1654. doi:10.1001/archinternmed.2010.355

11. Hamilton H, Gallagher P, Ryan C, Byrne S, O’Mahony D. Potentially inappropriate medications defined by STOPP criteria and the risk of adverse drug events in older hospitalized patients. Arch Intern Med. 2011;171(11):1013-1019.

12. Cooper JA, Cadogan CA, Patterson SM, et al. Interventions to improve the appropriate use of polypharmacy in older people: a Cochrane systematic review. BMJ Open. 2015;5(12):e009235. 

13. Dahl LJ, Wright R, Xiao A, Keeven A, Carr DB. Quality improvement in long term care: the psychotropic assessment tool (PAT). J Am Med Dir Assoc. 2008;9(9):676-683.

14. Pasina L, Marengoni A, Ghibelli S, et al. A multicomponent intervention to optimize psychotropic drug prescription in elderly nursing home residents: an Italian multicenter, prospective, pilot study. Drugs Aging. 2016;33(2):143-149. 

15. Milos V, Rekman E, Bondesson A, et al. Improving the quality of pharmacotherapy in elderly primary care patients through medication reviews: a randomised controlled study. Drugs Aging. 2013;30(4):235-246. 

16. Martin P, Tamblyn R, Ahmed S, Benedetti A, Tannenbaum C. A consumer-targeted, pharmacist-led, educational intervention to reduce inappropriate medication use in community older adults (D-PRESCRIBE trial): study protocol for a cluster randomized controlled trial. Trials. 2015;16:266. 

17. CGS 37th Annual Scientific Meeting: Integrating Care, Making an Impact. Canadian Geriatr J. 2017;20(3):198-239. doi:10.5770/cgj.20.285

18. Haque R. ARMOR: a tool to evaluate polypharmacy in elderly persons. Annals of Long-Term Care: Clinical Care and Aging. 2009;17(6):26-30. 

19. American Geriatrics Society Expert Panel on the Care of Older Adults with Multimorbidity. Guiding principles for the care of older adults with multimorbidity: an approach for clinicians. J Am Geriatr Soc. 2012;60(10):e1-e25. 

20. Denison Bub L. Medication reduction policy in two long term care facilities. Poster presented at: 15th Annual 2011 NICHE Conference.  https://archive.constantcontact.com/fs028/1105747634773/archive/1107255169257.html. Accessed March 1, 2019.

21. Castle NG, Ferguson JC. What is nursing home quality and how is it measured? Gerontol. 2010;50(4):426-442. doi:10.1093/geront/gnq052

22. Woolcott JC, Richardson KJ, Wiens MO, et al. Meta-analysis of the impact of 9 medication classes on falls in elderly persons [published correction appears in Arch Intern Med. 2010;170(5):477]. Arch Intern Med. 2009;169(21):1952-1960.

23. de Jong MR, Van der Elst M, Hartholt KA. Drug-related falls in older patients: implicated drugs, consequences, and possible prevention strategies. Ther Adv Drug Saf. 2013;4(4):147-154.  

24. Parás-Bravo P, Paz-Zulueta M, Alonso-Blanco MC, Salvadores-Fuentes P, Alconero-Camarero AR, Santibañez M. Association among presence of cancer pain, inadequate pain control, and psychotropic drug use. PLoS One. 2017;12(6):e0178742. 

25. Burfield AH, Wan TT, Sole ML, Cooper JW. Behavioral cues to expand a pain model of the cognitively impaired elderly in long-term care. Clin Interv Aging. 2012;7:207-223. 

26. Bowblis JR. Staffing ratios and quality: an analysis of minimum direct care staffing requirements for nursing homes. Health Serv Res. 2011;46(5):1495-1516. 

27. Zhang X, Grabowski DC. Nursing home staffing and quality under the Nursing Home Reform Act. Gerontologist. 2004;44(1):13-23.