Nursing Home Antibiotic Stewardship Policy and Antibiotics Use: 2013 – 2017

Abstract

Objectives:

Reducing inappropriate nursing home (NH) antibiotic usage by implementing stewardship programs is a national priority. Our aim is to evaluate the influence of antibiotic stewardship programs on antibiotic use rates in NHs over time.

Design:

Retrospective, repeated cross-sectional analysis.

Setting and Participants:

Long-term residents not receiving hospice care in freestanding NHs that participated in one or both surveys in 2013 and 2017.

Methods:

Survey data were merged with the Minimum Data Set and the Certification and Survey Provider Enhanced Reporting data. Our outcome was a binary indicator for antibiotic use. The main predictor was the NH antibiotic stewardship policy intensity. Using multivariate linear regression models adjusting for resident and facility characteristics that differed between the two years, we calculated antibiotic use rates in 2013 and 2017 for all residents, those with Alzheimer’s disease, and those with any infection including urinary tract infections (UTIs).

Results:

Our sample included 317,003 resident assessments from 2013 and 267,537 assessments from 2017, residing in 953 and 872 NHs, respectively. NH antibiotic stewardship policy intensity increased from 2013 to 2017 (p < 0.01) and among all NH residents, including those with Alzheimer’s disease, antibiotic use rate decreased (p < 0.05), with 45% of the decline attributable to strengthening stewardship programs. For most residents, policy intensity was associated with decreased usage in residents with UTI. However, among Alzheimer’s disease residents with a UTI, this association did not persist.

Conclusions and Implications:

While there was a decrease in antibiotic use in 2017, more time is needed to see the full impact of antibiotic stewardship policy into practice. Adjustments to programs that directly address barriers to implementation and appropriate UTI antibiotic use for residents with Alzheimer’s disease are necessary to continue strengthening NH antibiotic stewardship and improve care.

Brief summary:

Implementing nursing home antibiotic stewardship programs is a national priority and adjustments may be needed due to COVID-19, barriers to implementation, and for care of vulnerable populations.

Introduction:

In the approximately 15,600 nursing homes (NHs) in the United States (U.S.) there are an estimated 1.13 to 2.68 million infections yearly, with the most common being urinary tract infections (UTIs).¹ In NHs, residents are often prescribed antibiotics for suspected infections, without an infection diagnosis² and without meeting the minimal clinical criteria.³ Suspected infections are commonly diagnosed in Alzheimer’s disease and Alzheimer’s disease related dementias (AD/ADRD) residents who often have severe cognitive impairment and atypical clinical presentation of infections.^2,3 These oftentimes inappropriate and unnecessary antibiotic prescriptions⁴ lead to antibiotic overuse, antibiotic resistance, and colonization of multidrug resistant organisms for this already vulnerable NH population.³

Inappropriate antibiotic use is pervasive and has led to soaring rates of antibiotic resistance and obsolescence of common antibiotics have compelled health systems to implement methods to reduce antibiotic use. The Society for Healthcare Epidemiology, the Infectious Diseases Society of America, and the Pediatric Infectious Diseases Society, together have defined antibiotic stewardship as “coordinated interventions designed to improve and measure the appropriate use of antimicrobial agents by promoting the selection of the optimal antimicrobial drug regimen including dosing, duration of therapy, and route of administration” and is widely hailed as an effective methodology to reduce inappropriate antibiotic use across health care settings.⁵ Antibiotic stewardship programs promote appropriate use of antibiotics,⁶ have been effective in reducing inappropriate antibiotic usage in acute care settings, and may reduce inappropriate antibiotic use in NHs.

Antibiotic stewardship programs have been long implemented in acute care settings and only recently has been translated to the long-term care setting. In 2016, in one of the largest overhauls of nursing home policies and regulations, the Centers for Medicare and Medicaid Services (CMS) mandated an implementation plan that each long-term care facility must have 1) an infection prevention and control program in place by November 2016; 2) an antibiotic stewardship program in place by November 2017; and 3) a trained infection preventionist in place by November 2019.⁷ Following this CMS implementation, two national surveys were conducted to measure changes in NH antibiotic stewardship and UTI prevention from 2013 to 2017 and reported improvement in these outcomes.⁷

Our aim was to evaluate antibiotic use rates in NHs and intensity of antibiotic stewardship programs over time. Our hypothesis was that the strengthening of NH antibiotic stewardship programs would reduce overall antibiotic use.

Methods:

Sample

Two cross-sectional surveys, in 2013 and in 2017, were conducted with a national random sample of NHs across the U.S.^8,9 Freestanding NHs with at least 30 beds were identified from Certification and Survey Provider Enhanced Reporting (CASPER) data. Further details of survey sample and processes are published elsewhere.^8,9 In brief, we sampled 2500 NHs in the 2013 survey and received 988 completed responses (39% response rate). For the 2017 survey, we resurveyed the 988 NHs and added an additional random sample of 832 NHs, stratified by Quality Innovation Network-Quality Improvement Organizations (QIN-QIO)¹⁰ region and enrollment in CDC’s National Healthcare Safety Network Long-term Care Facility Component (NHSN).¹¹ QIN-QIO are CMS-funded regional collaboratives to improve NH infection control programs. We received 892 responses (49% response rate) for the 2017 survey and 456 NHs participated in both the 2013 and 2017 surveys.

Survey data were merged with concurrent CASPER and Minimum Data Set 3.0 (MDS) assessments. MDS assessments are performed for every Medicare beneficiary in NHs at admission, and every quarter thereafter as well as when there is a significant change in health status. The MDS has been evaluated as a valid and reliable measure for NH resident care.¹² We limited our analyses to long-term residents and included quarterly and annual assessments. We excluded assessments for residents under the Medicare hospice benefit (i.e., terminally ill residents no longer receiving curative treatments¹³) as these residents would receive different treatment as compared to the rest of the long-term population.

Measures

The outcome of interest was a binary antibiotic use indicator, measured by MDS item N0410F, which indicates the number of days that the resident received an antibiotic.¹⁴ This assessment-level variable counts the number of days the resident used antibiotics in a 7-day look back period. We developed a binary variable if the resident had four or more days of antibiotic use. The main independent variable was NH antibiotic stewardship policy intensity. Collected in both national surveys, the antibiotic stewardship measure included five policies, based on CDC’s “The Core Elements of Antibiotic Stewardship for Nursing Homes.”¹⁵ The policies were: 1) collection of data on antibiotic use; 2) having antibiotic-prescribing guidelines/therapeutic formularies; 3) having policies to restrict use of antibiotics; 4) reviewing cases for antibiotic appropriateness; and 5) providing feedback to clinicians on antibiotic prescribing. A standardized intensity score was calculated by summing the number of policies present in each NH, dividing by the maximum amount (5), and multiplying by 100 so that each index ranged from 0 (no policies) to 100 (all five policies).

We included resident and facility characteristics for risk adjustment based on established risk factors for increased antibiotic use.^2,3 Resident characteristics were identified from MDS assessments and included age, sex, race/ethnicity, marital status, active diagnoses of comorbidities and activities of daily living. Comorbidities included active diagnoses of AD/ADRD, heart failure, diabetes, asthma and chronic obstructive pulmonary disease, other respiratory disease, and urinary and bowel incontinence. We also examined recorded active diagnoses of UTI and created a variable indicating no recorded infection, i.e. no active diagnoses of pneumonia, wound, UTI or sepsis.

NH characteristics from CASPER included regional location, urban/rural location, part of chain organization, for profit ownership status, and CMS Five Star Quality Rating.¹⁶ Facility characteristics from the national surveys included participation in QIN-QIO infection prevention collaborative, high turnover of key staff (more than 3 in past 3 years), and nursing staff rates (hours per resident day).

Statistical analysis

Using a repeated cross-section design, we evaluated the change in antibiotic stewardship across 2013 and 2017 and examined association between policy intensity on antibiotic use. The unit of analysis was the MDS assessment. We examined differences in resident and facility characteristics between years with Pearson chi-square tests and t-tests as appropriate. We then calculated unadjusted and adjusted antibiotic use rates for each year, adjusting for all resident and facility characteristics.

To estimate the direct effect of increased antibiotic stewardship policy intensity on antibiotic use, the isolated contribution of policy intensity was calculated as the predicted antibiotic use when time was set to 2013 values of covariates and policy intensity to 2017, divided by the total adjusted change in antibiotic use. Multivariate ordinary least squares models were used to estimate the effect of policy intensity levels on antibiotic use. We stratified these models based on all assessments, assessments with a UTI diagnosis, and assessments with no recorded infections. We further stratified models based on active Alzheimer’s disease diagnosis. We also examined assessments with AD/ADRD but as the results largely matched the overall population, we do not show these results. All models were clustered at the NH level and controlled for year and all resident and NH characteristics. Sensitivity analyses were also done limiting the data to the NHs that participated at both time points; however, no significant differences were found between these and the full sample. Data were analyzed using Stata version 14 (Stata Corp., College Station, TX) and statistical significance was defined as p < 0.05.

Results:

Our sample consisted of 317,003 MDS assessments from 2013 representing 953 NHs and 267,537 MDS assessments from 2017 representing 872 NHs (Table 1). Resident and facility characteristics differed across the two time periods. There were lower percentages of assessments with active UTI (3% vs 5.3%) and greater percentage of those with no recorded infection (94.1% vs 91.8%) in 2017 compared to 2013, respectively. Although statistically significant, age, race/ethnicity, sex, marital status, comorbid conditions, and dependence on activities of daily living only differed slightly across years. There were also slight differences in NH characteristics, most importantly being that participation in infection control collaboratives increased (42% in 2017 vs 34% in 2013) and staff turnover decreased for all key staff in 2017 compared to 2013. From 2013 to 2017, antibiotic stewardship policies increased across all policies.

Table 1.

Resident and nursing home characteristics by survey year.

	2013	2017
N Nursing Homes	953	872
N MDS Assessments *	317,003	267,537
	% of assessments (unless otherwise noted)		p value
Resident characteristics
Infections
UTI	5.30	3.00	<0.01
No recorded infection*	91.78	94.08	<0.01
Age categories
65–75 years	14.50	16.17	<0.01
76–85 years	26.72	25.94	<0.01
86–95 years	36.59	35.47	<0.01
96 years or older	8.65	9.58	<0.01
Race/ethnicity
White	78.64	79.29	<0.01
Native American	0.61	0.90	<0.01
Asian	1.59	2.40	<0.01
African American	12.47	10.86	<0.01
Hispanic	5.24	4.66	<0.01
Hawaiian/Pacific Islander	0.41	0.45	0.03
Male	32.11	33.05	<0.01
Currently married	18.51	18.74	0.02
Comorbidities
Alzheimer’s disease	17.09	16.85	<0.01
Dementia other than Alzheimer’s	46.62	48.62	<0.01
Heart failure	20.05	20.94	<0.01
Diabetes	32.80	32.79	0.93
Asthma/COPD	20.21	22.31	<0.01
Respiratory disease	1.44	2.67	<0.01
Urinary incontinence	14.87	16.53	<0.01
Bowel incontinence	11.25	11.05	<0.01
Activities of daily living, mean+
Bed mobility	2.45	2.47	<0. 01
Self transfer	2.56	2.58	<0.01
Self locomotion	2.27	2.31	<0.01
Self dressing	2.74	2.71	<0.01
Self eating	1.45	1.39	<0.01
Self toileting	2.71	2.69	<0.01
Self hygiene	2.71	2.65	<0.01
Nursing home characteristics
Region
North	32.24	33.55	<0.01
Midwest	29.74	26.86	<0.01
South	29.12	27.71	<0.01
West	8.89	11.88	<0.01
Chain facility	51.02	51.06	0.80
Location
Urban	78.26	74.88	<0.01
Rural remote	14.39	15.46	<0.01
Rural urban adjacent	7.35	9.66	<0.01
For-profit facility	67.06	63.39	<0.01
CMS Quality Rating, mean	3.33	3.34	<0.01
Participation in infection control collaboratives	33.80	41.97	<0.01
Staff turnover (3 or more in past 3 years)
Infection preventionist	35.91	23.12	<0.01
Administrator	34.09	19.40	<0.01
Director of Nursing	34.70	27.33	<0.01
Occupancy rate, mean	85.13	83.60	<0.01
Nurse staffing rates (hours per resident day), mean
Certified nursing assistant	2.42	2.47	<0.01
Licensed practical nurse	0.80	0.82	<0.01
Registered nurse	0.70	0.74	<0.01
Antibiotic stewardship policies
Collection of data on antibiotic use	39.66	60.34	<0.01
Having antibiotic-prescribing guidelines/therapeutic formularies	34.14	65.86	<0.01
Having policies to restrict use of antibiotic	27.50	72.50	<0.01
Reviewing cases for antibiotic appropriateness	38.11	61.89	<0.01
Providing feedback to clinicians on antibiotic prescribing	35.24	64.76	<0.01

^*Abbreviations: CMS = Centers of Medicare & Medicaid; COPD = chronic obstructive pulmonary disease; UTI = urinary tract infection.

⁺Activities of Daily Living scores measure ability to perform tasks related to personal care. Each task (bed mobility, self transfer, self locomotion, self dressing, self eating, self toileting, self hygiene) contains a set of items scored on a scale of 1–5.

Antibiotic stewardship policy intensity increased in NHs drastically between 2013 and 2017 (p value < 0.01) (Figure 1). In 2013, 20% of NHs had no policies, 31% had one out of the five policies, 25% had two policies, 17% had three policies, 6% had four and only 1% had all five policies (policy intensity of 100). By 2017, the distribution flipped and 15% of NHs had all policies, 32% had four, 29% had three, 15% had two, 8% had one and 1% had no policies.

Figure 1:

Change in nursing home antibiotic stewardship policy intensities from 2013 to 2017.

Antibiotic use prevalence rates in 2017 decreased from 2013 (−0.28% difference, p < 0.05). After adjusting for resident and NH characteristics, the decrease was even greater (−0.45 difference) in the overall sample (Table 2). The change in antibiotic policy intensity resulted in 45% of this difference. Assessments with recorded Alzheimer’s disease had a −0.58 percent (6.92% to 6.34%) difference in antibiotic use from 2013 to 2017, respectively, and 81% of this decrease was attributable to improvements in antibiotic stewardship.

Table 2.

Antibiotic use rates and percent attributable to antibiotic stewardship policy changes.

	Unadjusted antibiotic use rate		Adjusted antibiotic use rate*		% change due to improvements in antibiotic stewardship policies*
	2013	2017	2013	2017
All assessments	8.28%	8.04%	8.54%	8.09%	45.15%
Assessments with AD	6.87%	6.33%	6.92%	6.34%	81.22%

Note: Unadjusted and adjusted antibiotic use rates in 2013 and 2017 were statistically significant at p<0.01.

^*Adjusted for all resident and nursing home characteristics as presented in Table 1.

Table 3 presents the cross-sectional, multivariate models with combined data from 2013 and 2017 for the overall population and for Alzheimer’s disease residents, predicting antibiotic use for each level of stewardship policy intensity, and adjusted for all resident and NH characteristics. First, we report on antibiotic stewardship policy intensity overall and for Alzheimer’s disease residents on all assessments. In the overall population, among the 555,400 assessments, antibiotic use rate (4 days or more) declined significantly by 0.0089 days only with an antibiotic stewardship intensity of 80 (4 policies in place) as compared to zero intensity (β = −0.0089, p = 0.03). Among assessments with recorded Alzheimer’s disease, antibiotic use decreased when NHs had policy intensities of 80 or 100 (β = −0.0176 and β = −0.0140, respectively, both p < 0.05).

Table 3.

Multivariable linear models estimating effects of antibiotic stewardship policy intensity on antibiotic use, overall and stratified by Alzheimer’s disease.

	Overall						Alzheimer’s disease
	All assessments		Recorded UTI		No recorded infection		All assessments		Recorded UTI		No recorded infection
	n = 555,400		n = 23,785		n = 515,304		n =92,066		n = 3,663		n = 86,601
Antibiotic stewardship policy intensity	β	P	β	P	β	P	β	P	β	P	β	P
0	Reference		Reference		Reference		Reference		Reference		Reference
20	−0.0036	0.32	−0.0347	0.01	−0.0019	0.54	−0.0106	0.05	−0.0503	0.10	−0.0093	0.06
40	−0.0039	0.30	−0.0188	0.18	−0.0041	0.20	−0.0096	0.08	−0.0178	0.54	−0.0120	0.01
60	−0.0038	0.34	−0.0312	0.04	−0.0023	0.50	−0.0099	0.09	−0.0347	0.27	−0.0097	0.06
80	−0.0089	0.03	−0.0361	0.02	−0.0063	0.07	−0.0176	<0.01	−0.0330	0.40	−0.0159	<0.01
100	−0.0032	0.53	−0.0490	0.01	−0.0002	0.97	−0.0140	0.05	0.0124	0.77	−0.0138	0.03

Note: All models included all resident and nursing home facility characteristics from Table 1.

Abbreviations: UTI = urinary tract disease.

Next, we discuss antibiotic stewardship policy intensity for the overall population and for Alzheimer’s disease residents with UTI and those with no recorded infections. Among the overall NH population, in assessments with a recorded UTI infection, antibiotic use decreased with increased policies, with the highest decrease associated with having all five policies (intensity = 100; β = −0.0490, p = 0.01). No differences in antibiotic use were found across policy intensities when there was no recorded infection in the overall population. There was no impact on antibiotic stewardship policy intensity on use in those with Alzheimer’s disease and UTI. In those with Alzheimer’s disease and no recorded infection, increased policies were associated with decreased antibiotic use (β = −0.0120 for 40 intensity, p = 0.01, β = −0.0159 for 80 intensity, p < 0.01, and β = −0.0138 for 100 intensity, p = 0.03).

Discussion:

From 2013 to 2017, antibiotic stewardship policy intensity increased in NHs while antibiotic use decreased, particularly for NHs with high antibiotic stewardship policy intensity (≥ 80%). Our results suggest that antibiotic stewardship programs did not have unintended consequences of restricting antibiotic use where clinically needed and may have reduced overall antibiotic use among the overall NH population, including those with Alzheimer’s disease. Antibiotic use decreased with increased antibiotic stewardship policy intensity, particularly for non-Alzheimer’s disease residents with UTIs. Our analyses show that 45% of antibiotic use decrease is attributable to improvements in antibiotic stewardship among all NH patients.

Our results of the impact of NH antibiotic stewardship programs on antibiotic use from 2013 to 2017 are promising given that overall, antibiotic use in NHs in the US has not decreased.² The increased intensity of NH antibiotic stewardship programs is potentially explained by growing attention to the global problem of inappropriate antibiotic usage. During the study time period, besides the CMS regulation change, there were significant publications promoting antibiotic stewardship in NHs including guidelines from the CDC (used as our antibiotic stewardship measure) in 2016,¹⁵ the Joint Commission in 2016,¹⁷ the Infectious Disease Society of American along with the Society for Healthcare Epidemiology of America in 2016,¹⁸ and AHRQ in 2017.^19,20 NH staff have expressed the need for this continued and consistent support and education from external and internal support systems to sustain antibiotic stewardship programs.²¹ Given the increased publications during the time period and the increased antibiotic stewardship intensity observed, high priority should be given to continue training NH staff.

Studies have found that having a trained infection preventionist strengthens antibiotic stewardship programs.^22,23 The CMS implementation plan mandated all NHs to have a trained infection preventionist in place by November 2019. The capacity of this implementation likely varied given that regulations require only part-time, unspecified time commitments, and vague training requirements.²² NH infection preventionists were most responsible for COVID-19 preparedness²⁴ and NHs with infection preventionists had lower COVID-19 incidence compared to NHs without an infection preventionist.²⁵ Regardless of incidence and that COVID-19 has viral origins, early in the pandemic there were increased prescriptions of antibiotics in NHs during the pandemic.²⁶ Time is needed to understand the impact of COVID-19 on NH antibiotic stewardship and presents an opportunity to refocus antibiotic stewardship programs to address specific NH needs and vulnerable populations.

Approximately half of all NH residents have AD/ADRD²⁷ and constitute a vulnerable population that is often overprescribed antibiotics for suspected infections without sufficient clinical criteria.³ NH residents with Alzheimer’s disease often times have multiple comorbidities, that coupled with aging can lead to atypical clinical presentations that potentially impact accuracy of UTI diagnosis and subsequent treatment.³ We found that antibiotic stewardship was associated with decreased antibiotic use for Alzheimer’s disease residents. Although some antibiotic stewardship guidelines exist for UTIs,²⁸ there are no specific guidelines for residents with Alzheimer’s disease. NH antibiotic stewardship guidelines should be updated to provide clear guidance for NH residents with Alzheimer’s disease, a large population in the NH that require specialized attention.

This study has limitations, most notably given the nature of the MDS data used. The MDS item N0410F used to determine the prevalence of antibiotic use has yet to be evaluated for validity and may underestimate the actual prevalence of antibiotic use. In addition, MDS does not provide prescriber information for medications, which may impact the frequency of antibiotic prescriptions.²⁹ However, MDS overall has acceptable validity and reliability for NH resident health status.¹² It is possible that there was misclassification bias of Alzheimer’s disease residents, misclassification of infections versus no infections, and classification of UTIs. Furthermore, we did not have access to resident health records and therefore cannot directly determine if the residents received inappropriate antibiotic use or if antibiotics were ever withheld in the presence of an infection diagnosis. Nevertheless, in sensitivity analyses we found that residents with AD/ADRD largely matched overall resident results. In addition, self-selection bias may have influenced our findings. For example, NHs that participated in the surveys may have stronger antibiotic stewardship programs in place. Finally, the timing of the second survey may have been too early to accurately determine the impact of an antibiotic stewardship program on antibiotic use since the deadline for establishing antibiotic stewardship programs was until the end of 2017.

Conclusion and Implications:

In conclusion, our results demonstrate a decrease in antibiotic usage but more time is needed to observe this translation of policy to practice, particularly with the COVID-19 pandemic. These findings highlight the potential for NH antibiotic stewardship policies to decrease inappropriate and unnecessary antibiotic use. Although NH antibiotic stewardship policy increased, the intensity varied across NHs where 15 percent had all 5 policies and approximately 30 percent had three or four, respectively. Furthermore, significant antibiotic use impacts were observed in NHs with relatively high antibiotic stewardship policy intensity (≥ 80%), these levels of intensity may be difficult to achieve for under resourced NHs. At the time of the study only 45% of NHs sampled had achieved this level, therefore future research should examine NH antibiotic stewardship policy intensity change over time, particularly if more NHs have achieved intensity greater than 80%. Our survey was based off of the CDC’s “The Core Elements of Antibiotic Stewardship for Nursing Homes”, and further work is needed to identify specific policies that are most effective or what are barriers that NHs experience in implementing certain policies, in order to best support them in implementing the policies. As always, preventing infections is the first step in reducing unnecessary antibiotic use. Future studies should evaluate how prevention efforts, particularly in residents with catheters and other medical devices moderate the effect of antibiotic stewardship programs on antibiotic use. Improving NH antibiotic stewardship programs is a national priority that will require time, reflection, and refocusing.