2015 Beers criteria medication review in assisted living facilities

Jessica C. Chun, DNP, NP-C (Adult Geriatric Nurse Practitioner), Susan J. Appel, PhD, ACNP-BC, FNP-BC, CCRN, FAHA, FAANP (Professor), & Steven Simmons, MD, ABIM-BC (Internal Medicine Physician)

Journal of the American Association of Nurse Practitioners


Abstract

Background and purpose: The elderly population is expected to double by 2050 with falls and hospitalizations due to adverse drug events having a major effect on health and quality of life. With the release of the revised 2015 American Geriatrics Society (AGS) Beers criteria, usage of potentially inappropriate medications (PIMs) should be studied to determine their effect on falls and hospitalizations in frail populations such as those inassisted living facilities. 

Methods: This quality improvement project used a retrospective chart review on residents from a purposive sample of two assisted living facilities in Northern Virginia. Residents were aged ≥65 and lived at the facility for at least 6 months and were not enrolled in hospice and/or palliative care or living in the dementia unit. The 2015 AGS Beers criteria were used to evaluate the effect of PIMs on falls and hospitalization rates.

Conclusions: This project did not find statistical significance between PIMs and falls (p = .276). A favorable, but not statistically significant trend, was noted between PIMs and hospitalizations (p = .079).

Implications for practice: Understanding the effect of PIMs on falls and hospitalizations could help providers improve prescribing practices for the elderly population who are at the greatest risk for potential adverse effects from polypharmacy.


Background and purpose

By 2050, the elderly population within the United States is expected to nearly double from 43.1 million in 2012 to 83.7 million (Ortman, Velkoff, & Hogan, 2014). The population aged ≥65 are more sensitive to the side effects of medications and are at a greater risk for adverse drug reactions(ADRs). Impaired renal function affects drugs cleared by filtration, but acute and chronic kidney disease has been shown to have an effect on hepatic and gastrointestinal enzymes and transporters as well (Miners, Yang, Knights, & Zhang, 2017). Two out of three older Americans have multiple chronic conditions that increase the risk for polypharmacy and adverse drug events (Centers for Disease Control, 2016). These changes associated with aging, multiple comorbidities, and polypharmacy place the elderly at an increased risk for falls and hospitalizations. Medicationmetabolism and absorption are affected by decreased drug clearance, low albumin levels, and metabolic changes associated with aging (Dagli & Sharma, 2014). The Beers criteria were created as a way for providers to assess and closely monitor drug use among the older adults with the goal of decreasing ADRs (American Geriatrics Society 2015 Beers Criteria Update Expert Panel, 2015).

The Beers criteria guidelines were first established in 1991 in a report whose primary author was Dr. D. H. Beers (AGS, 2015). The criteria are an evidence-based list of medications that are known to cause side effects in the elderly because of physiologic changes associated with aging (AGS, 2015). The most recent iteration in 2015 was completed by an expert panel consisting of geriatric specialists in medicine, nursing, pharmacy, research, and quality measures. The panel also included ex-officio members from the Center for Medicare and Medicaid Services, National Committee for Quality Assurance, and the Pharmacy Quality Alliance. These guidelines were intended for all prescribing providers caring for patients aged ≥65 in the United States with the exception of those on hospice and/or palliative care. The Beers criteria are relevant as a number of hospitalizations in elderly are due to an adverse drug event, with a percentage of those being a result of potentially inappropriate medications(PIMs). These guidelines are easily implementable and are provided as pocket-sized references or applications for smartphones for providers to carry and refer to. The guidelines are organized by disease or syndrome to facilitate locating a class of medication that should be avoided inthe elderly population (AGS, 2015). The guidelines are designed to be one aspect of medicationmanagement and should be used in conjunction with other tools such as the screening tool of older people's prescriptions and screening tool to alert to right treatment (START) criteria (AGS,2015).

Falls due to potentially inappropriate medications

It is estimated that one-fourth of Americans aged ≥65 fall each year (National Council on Aging, 2015). Falls are the leading cause of fatal injury and one of the top 10 leading causes of death inpeople aged ≥65 (Centers for Disease Control [CDC], 2017). Falls lead to over 1,200,000 hospitalizations (usually because of a hip fracture or fall) for nonfatal injuries in 2015 (CDC, 2017). Multiple studies have assessed the effect of PIMs on falls in the elderly population. Medications included in the Beers criteria—hypnotics, tricyclic antidepressants, serotonin reuptake inhibitors, and serotonin–norepinephrine reuptake inhibitors—were all found to be associated with increased falls (Park, Satoh, Miki, Urushihara, & Sawada, 2015).

In a study of 1,016 patients aged above 70, researchers in Ireland found that 44% had PIMs prescribed before fall (McMahon, Cahir, Kenny, & Bennett, 2014). They further noted that after fall, there was a reduction in alpha blockers prescribed (p = .04) and an increase in the amount of short- and intermediate-acting benzodiazepines (p = .0002). Likewise, in a retrospective chart reviewby Slaney, MacAulay, Irvine-Meek and Murray (2015), there was an 8% increase noted in the probability of a fall (est. = 3.26; SE, 0.55; p < .001) among nonacute inpatients within a hospital in Canada related to PIMs. A retrospective case-control study of 801 patients in Singapore found no difference in PIM usage in patients admitted for falls between the control and sample cases. The investigators, however, noted an increase in the prescription of PIMs upon discharge after a fall (Chen et al., 2016). Although studies to date have varied results on the effect of PIMs on fall rates, they seem to indicate that regular review of medications and reduction of polypharmacy in the elderly is likely to reduce adverse outcomes in this population.

In short, the majority of studies regarding PIMs and falls seem to be observational. Furthermore, these studies have shown varied results based on PIM usage and their correlation to falls. Therefore, there is a need for continued research regarding the effect of PIMs on falls. Based on the current literature, reducing PIMs associated with falls could be important to the safety of patients whenever appropriate.

Hospitalizations due to potentially inappropriate medications

Hospitalization due to adverse drug events accounts for up to 6–12% of hospital admissions inthe elderly with PIMs being one of the main risk factors for or predictors of hospital admission (Parameswaran Nair et al., 2016). A review of Medicare beneficiaries found that up to one third of beneficiaries each month were prescribed a PIM with the number rising to one half over a 12-month span (Jirón et al., 2016). The discrepancies in methodologies used in studies and varying results to date have made the existing body of evidence inconclusive (Patterson, 2014). A systematic review of the literature by Patterson (2014) suggested that future research should focus on clinical end points such as hospital admissions associated with inappropriate prescribing. A more robust and clinically based series of studies are needed to draw more confident conclusions regarding the effect of PIMs on hospitalization rates. This will ultimately drive better adoption of practice change.

Prescribing deficits

In a study of six vignettes assessing provider's prescribing knowledge (N = 89) regarding PIMs, 14% of providers who scored ≤4 used the Beers criteria, whereas 31% scoring of those provider's scoring ≥5 out of six used the Beers criteria (Ramaswamy et al., 2011). Improving the knowledge base of prescribers regarding the Beers criteria could decrease the use of PIMs inthe population of patients aged ≥65. Corbi et al. (2015) found that they were able to reduce the number of PIMs prescribed but were using a decision support tool on a personal digital device and providing education to providers (p = .02). Another important finding of this study was that the increase in PIMs was associated with age, polypharmacy, and lack of decision support tool. A multicomponent quality improvement initiative at the Veterans Affairs Medical Center that included a geriatric pharmacology lecture, review of the 2012 Beers criteria at quarterly journal club, and reminder cards identifying the top five PIMs resulted in a decrease from 9.4 ± 1.5% PIMs prescribed to 4.6 ± 1% (p < .001) (Stevens et al., 2015). Increasing provider knowledge regarding PIMs using Beers can improve prescribing practices and ultimately reduce negative outcomes.

Medication oversight in assisted living facilities

There is currently no standardized regulation regarding pharmacists' oversight for review of PIMs in assisted living facilities. Thus, there is wide variability in both provider and pharmacist oversight of medications in the assisted living population (Stefanacci & Haimowitz, 2014). According to a study by the 2010 Centers for Disease Control National Survey of Residential Care Facilities, only 68% of respondents from assisted living facilities had doctor or pharmacist on staff who regularly review medications of residents (Centers for Disease Control [CDC], 2011). The inconsistent regulation and oversight between facilities create an opportunity for implementation of evidence-based tools to improve outcomes.

Given the variability of medication oversight within assisted living facilities and long-term care facilities, creating an environment of regular medication oversight is one large area for quality improvement. Because the assisted living population are largely aged above 65, providers require a specialized understanding of the potential adverse effects of medications being prescribed to the elderly. The 2015 Beers criteria can be used as an evidence-based tool that facilitates a standardized method for reviewing PIMs and customizing care specifically for this population. In short, quantifying how medication review will affect outcomes is a critical first step.

Theoretical framework

A literature search revealed minimal use of theory or conceptual frameworks to guide change of practice related to PIM. The mid-range theory used to guide this project was Lewin's change theory. There are three phases in Lewin's change theory: unfreezing, movement, and refreezing (Wojciechowski, Pearsall, Murphy, & French, 2016). The project focused on the unfreezing stage. Data were collected and disseminated regarding the use of PIMs within the facility and the effect (or lack thereof) on falls and hospitalization rates. The next phase will be the period of change. Based on the findings, protocols related to the use of PIMs and identification of PIMs can be identified within each facility. Education regarding their use can then be disseminated to providers and administrators. A tool to assist with easy identification of PIM should be developed to assist providers with clinical decision making. This process should be evaluated for ongoing effectiveness.

Plan-do-study-act cycle

The following model for improvement was completed for this quality improvement process. The effect of PIMs on falls and hospitalization was assessed to improve outcomes at two assistedliving facilities in Northern Virginia. A plan-do-study-act (PDSA) tool was used. The following is a summation.

  1. Plan

Determine the effect of PIMS on falls and initial hospitalization.

Step: Medication review

Steps to execute are the following: Obtain approval from the institutional review board. Obtain consent from participants as needed. Identify patients ≥65 who have lived at facility ≥6 months. Perform chart audit using 2015 Beers criteria for PIMs, number of hospitalizations, number of falls, applicable diagnoses and laboratory values. Enter de-identified data into SPSS software. Run statistical analysis and analyze data to determine the correlation between PIMs and hospitalization and fall rates. Report findings to administrators, pharmacists, medical directors, and providers.

2. Do:

Statistics were run to evaluate the effect of PIMs on falls and hospitalizations rate. No statistical significance was noted between PIMs and falls and hospitalizations as discussed later.

3. Study:

This review demonstrated that PIMs did not have a direct effect on falls but did demonstrate some insights toward hospitalizations. Discussion related to findings with administrators determined that a longer period would benefit findings. The study also found that PIMs are frequently started during hospitalization.

4. Act:

Reassess the data for a longer period. Review hospital discharge lists for PIMs and assess ongoing need. Train staff in chart review using American Geriatrics Society (AGS) 2015 Beerscriteria.

The following is a discussion of the methods and analysis performed during this PDSA cycle.

Methods

After institutional review board approval was granted, this quality improvement project used a retrospective chart review to determine whether having fewer or no PIMs prescribed resulted inless falls and hospitalizations among elderly patients residing in assisted living facilities. Residents were classified into two groups and analyzed base on whether or not they have been prescribed PIMs as defined by the Beers criteria.

Sample and setting

Residents were recruited from a purposive sample of two assisted living facilities in the Northern Virginia area. Residents were aged ≥65 and lived at the facility for at least 6 months and were not enrolled in hospice and/or palliative care and not living in the dementia unit.

Data collection

PIMs were identified based on the American Geriatrics 2015 Beers criteria. Permission to use the tool was obtained from the AGS. Other data collected were age, weight, creatinine, number of falls, and number of hospitalizations. Each PIM as identified by the Beers criteria was assessed for their effects on falls and hospitalizations.

Statistical methods

The independent variable of PIM (no and yes) is nominal, and the three dependent variables of falls (no and yes), hospitalization (no and yes), and 30-day rehospitalization rate (no and yes) are nominal. Two binary logistic regressions were chosen to examine the effect of PIMS on falls and initial hospitalization (Laerd Statistics, 2015). The binary dependent variables (yes, 01) would be falls and initial hospitalization and the one nominal independent variables (x) would be PIMS. Data were collected regarding the number of patients with rehospitalizations in less than 30 days.

Results

A retrospective chart review was performed on 132 charts of residents currently residing in two assisted living facilities in Northern Virginia from January 2016 to October 2017. Of the 132 participants, 95 met all inclusion criteria of the project. A total of 37 charts were excluded for the following reasons: nine participants were excluded for being under the age of 65, five participants were excluded for being enrolled in a hospice, 20 participants were excluded for living at the facility less than the required time to participate, and three participants were excluded secondary to incomplete data for evaluation.

There were 95 participants in the project who were selected from a purposive sample of two assisted living facilities in Northern Virginia. There were 65 (68.4%) participants who identified as female. Of the 95 participants, 13 (13.7%) had the primary assisted living facility admission diagnosis classified as Alzheimer disease, 8 (8.4%) had depression, 14 (14.7) had diabetes, 35 (36.8%) had heart disease, 6 (6.3%) had respiratory disease, and 19 (20%) were classified as other based on admitting diagnosis to the assisted living facility. Physiologic characteristics are illustrated in Table 1; the mean age of the participants was 83.89, with a range of 65–99 (SD ± 7.915). The mean weight was 153.12 pounds. The average creatinine level was 1.25, and the average CrCL level was 48.703.

A total of 95 PIMs were prescribed to this population, 21 (22.1%) participants were on one PIM, 21 (22.1%) participants were on two PIMs, 9 (9.5%) participants were on three PIMs, 5 (5.3%) participants were on four PIMs, and 1 (1.1%) participant was on five PIMs. Medications were classified into organ system based on the 2015 Beers classification. The most commonly prescribed classification of PIMs was central nervous system agents, which were prescribed 29 times. Of participants being prescribed multiple PIMs, 6 (6.3%) were prescribed >1 central nervous system agent. Other frequently prescribed PIMs included gastrointestinal (17), pain medications (16) and anticholinergics (15), endocrine (11), and cardiovascular agents (7), respectively (Table 2).

 

Analysis

To work toward reducing hospitalization rates and falls, the following question was devised: Are adults aged ≥65 who live in assisted living facilities, prescribed PIMs, at an increased risk for hospitalizations and/or falls, compared with adults, who have no prescribed PIMs, during the previous 6 months? The variables of numbers of PIMS, frequency of falls, initial hospitalization, and numbers of 30-day rehospitalization were treated as discrete count variables.

Two binary logistic regressions were chosen to be used to examine the effect of PIMS on falls and initial hospitalization (Laerd Statistics, 2015). The binary dependent variables (yes, 01) would be falls and initial hospitalization, and the one nominal independent variables (x) would be PIMS. The binomial logit estimated models for these two cases is logit(Y) = β0 + β1 X 1 + ε.

Number of falls

To understand the effects of PIM usage on the likelihood that participants would fall, a binomial logistic regression was performed. A test of a fall model against a constant model was not statistically significant with PIMs as a predictor of falls (p = .276). The model explained 1.7% (Nagelkerke R 2) of the variance in falls and correctly classified 51.6% of cases as no or yes for falls. The predictor variable, of PIM usage, was not statistically significant. There was no association with PIM usage and the number of falls (Table 3–5).

Numbers of initial hospitalizations

A binomial logistic regression was performed to understand the effects of PIM usage on the likelihood that participants will be hospitalized. The logistic regression model was not statistically significant but did show a favorable statistical trend (p = .079). The model explained 4.4% (Nagelkerke R 2) of the variance in hospitalizations and correctly classified 63.2% of cases as no or yes for hospitalizations. The predictor variable of PIM usage was not statistically significant. There was no association with PIM usage and initial hospitalization (Tables 6–8).

Number of 30-day rehospitalizations

Of the 35 patients who had a hospitalization, only four had a rehospitalization within a 30-day period. Given the small sample size of n = 4, a logistic regression was not performed given the inability to determine statistical significance. All 4 (100%) of those with a rehospitalization within 30 days were prescribed PIMs; 2 (50%) of the patient were on two PIMs and 2 (50%) of the patients were on three PIMs. Of the patients discharged on PIMs, 2 (50%) had a new PIM added at discharge.

Limitations

Limitations of this study included the small sample size and short follow-up duration. A longer follow-up period would have allowed for a larger sample size and a longer length of time to track hospitalizations and rehospitalizations. Also, other potential risks factors for hospitalization were not assessed. Regarding falls, administration logs related to administration time of medications before timing of falls was not recorded to determine whether timing of medication had an effect of falls. Also, this study looked at the 2015 Beers criteria as a whole and did not look specifically at medications designated as a high risk for causing falls. Because this was a retrospective chart review, it was beyond the scope of this study to determine whether other medications had been tried and failed before continuing a PIM therefore justifying an ongoing need for a PIM despite potential other alternatives.

Conclusion

Elderly patients with multiple comorbidities are at risk for falls and hospitalization. Reducing factors that may lead to hospitalization such has prescribing PIMs whenever possible is crucial. Most patients (60%) in this assisted living population were taking at least one PIMs as classified by the 2015 Beers criteria. Although we found no statistical significance between PIM usage and fall rates in this population, there was a favorable statistical trend between PIM usage and hospitalizations. All patients who had a rehospitalization were on at least two PIMs. The elderly may benefit from careful monitoring of their medication regimen and for close consideration of medications on the 2015 Beers criteria with the goals of trying to reduce PIMs whenever appropriate. Using a consistent method to review and reduce PIMs may have a positive outcome on hospital admissions and falls among the elderly. Reviewing medications profiles for ongoing PIM usage after hospitalization to prevent future hospitalizations should be considered. Providers should become familiar with the Beers criteria as one tool for clinical decision making as appropriate. Further quality improvement projects should be done to determine whether reduction in PIM usage prospectively has a positive effect on hospitalization and 30-day rehospitalization rate given the small sample size of this study.

Implications for practice

Although this project demonstrated a favorable, but not statistically significant, relationship between PIMs and hospitalizations, the sample size was small and other potential risks factors for hospitalization were not assessed. Efforts should be made in the future to follow patients prospectively regarding PIMs, falls, hospitalizations, and justification for ongoing usage to gain an ongoing understanding of usage where applicable. Patient complexity, medicationadherence, and other factors that may have resulted in patient hospitalization were not accounted for within the project. Other issues that affected decisions related to prescribing that were not captured in the chart review were failed medication regimens, social factors, and worsening disease. Despite the limitations, benefit exists and providers should still continue to use the 2015 Beers criteria among the elderly population as a guide for clinical decision making when considering medication regimens for the elderly. Providers should be educated on the proper use of the 2015 Beers criteria. Likewise, providers should document rationales for medication selection and the need for continued used of PIMs to ensure optimal care and communication within the team.


References

American Geriatrics Society 2015 Beers Criteria Update Expert Panel. (2015). American Geriatrics Society 2015 updated beers criteria for potentially inappropriate medication use in older adults. Journal of the American Geriatrics Society. Advance online publication. doi: 10.1111/jgs.13702.

Centers for Disease Control. (2016). The state of aging and health in America. Retrieved from https://www.cdc.gov/aging/agingdata/ data-portal/state-aging-health.html.

Centers for Disease Control [CDC]. (2011). National survey of residential care facilities. Retrieved from ftp://ftp.cdc.gov/pub/ Health_Statistics/NCHS/Dataset_Documentation/NSRCF/2010/ 2010NSRCF_FacilityPublicUseFileDataDictionary.pdf.

Centers for Disease Control. Injury prevention & control. (2017). Webbased injury statistics query and reporting system (WISQARS). Retrieved from https://www.cdc.gov/injury/wisqars/.

Chen, L. L., Ng, H. T., Siow, B., Ong, K. Y., Tay, J. Y., Cheen, H. H., … Mamun, K. (2016). PHP58: The 2012 updated beers criteria in falls related hospitalization in older adults. Value in Health: the Journal of the International Society for Pharmacoeconomics and Outcomes Research, 19, A824.

Corbi, G., Gambassi, G., Pagano, G., Russomanno, G., Conti, V., Rengo, G., … Ferrara, N. (2015). Impact of an innovative educational strategy on medication appropriate use and length of stay in elderly patients. Medicine, 94, e918–e918. doi: 10.1097/MD.0000000000000918.

Dagli, R. J., & Sharma, A. (2014). Polypharmacy: A global risk factor for elderly people. Journal of International Oral Health, 6, i.

Jiron, M., Pate, V., Hanson, L. C., Lund, J. L., Jonsson Funk, M., & St ´ urmer, ¨ T. (2016). Trends in prevalence and determinants of potentially inappropriate prescribing in the United States: 2007 to 2012. Journal of the American Geriatrics Society, 64, 788–797.

Laerd Statistics (2015). Binomial logistic regression using SPSS Statistics. Statistical tutorials and software guides. Retrieved from https://statistics.laerd.com/

McMahon, C. G., Cahir, C. A., Kenny, R. A., & Bennett, K. (2014). Inappropriate prescribing in older fallers presenting to an Irish emergency department. Age and Ageing, 43, 44–50.

Miners, J. O., Yang, X., Knights, K. M., & Zhang, L. (2017). The role of the kidney in drug elimination: Transport, metabolism, and the impact of kidney disease on drug clearance. Clinical Pharmacology and Therapeutics, 102, 436–449.

National Council on Aging. (2015). Falls prevention facts. Retrieved from https://www.ncoa.org/news/resources-for-reporters/getthe-facts/falls-prevention-facts/.

Ortman, J., Velkoff, V. & Hogan, H. (2014). An aging nation: The older population in the United States. Retrieved from https://www. census.gov/prod/2014pubs/p25-1140.pdf.

Parameswaran Nair, N., Chalmers, L., Peterson, G. M., Bereznicki, B. J., Castelino, R. L., & Bereznicki, L. R. (2016). Hospitalization in older patients due to adverse drug reactions: The need for a prediction tool. Clinical Interventions in Aging, 11, 497–505.

Park, H., Satoh, H., Miki, A., Urushihara, H., & Sawada, Y. (2015). Medications associated with falls in older people: Systematic review of publications from a recent 5-year period. European Journal of Clinical Pharmacology, 71, 1429–1440.

Patterson, S. M. (2014). Interventions to improve the appropriate use of polypharmacy for older people. Cochrane Database of Systematic Reviews. Retrieved from http://search.ebscohost. com/login.aspx?direct=true&db=chh&AN=CD008165&site=ehost-live.

Ramaswamy, R., Maio, V., Diamond, J. J., Talati, A. R., Hartmann, C. W., Arenson, C., & Roehl, B. (2011). Potentially inappropriate prescribing in elderly: Assessing doctor knowledge, confidence and barriers. Journal of Evaluation in Clinical Practice, 17, 1153–1159.

Slaney, H., MacAulay, S., Irvine-Meek, J., & Murray, J. (2015). Application of the beers criteria to alternate level of care patients in hospital inpatient units. Canadian Journal of Hospital Pharmacy, 68, 218–225.

Stefanacci, R. C., & Haimowitz, D. (2014). The perfect INTERACTions to reduce hospitalizations. Geriatric Nursing, 35, 466–470.

Stevens, M. B., Hastings, S. N., Powers, J., Vandenberg, A. E., Echt, K. V., Bryan, W. E., … Vaughan, C. P. (2015). Enhancing the quality of prescribing practices for older veterans discharged from the emergency department (EQUiPPED): Preliminary results from enhancing quality of prescribing practices for older veterans discharged from the emergency department, a Novel Multicomponent Interdisciplinary Quality Improvement Initiative. Journal of the American Geriatrics Society, 63, 1025–1029.

Wojciechowski, E., Pearsall, T., Murphy, P., & French, E. (2016). A case review: Integrating Lewin’s theory with lean’s system approach for change. Online Journal of Issues in Nursing, 21, 1.

Loading