Abstract
OBJECTIVE
To determine the incidence and variability of antibiotic use in facilities which provide chronic care and to determine how often clinical criteria for infection are met when antibiotics are prescribed in these facilities.
DESIGN
A prospective, 12-month, observational cohort study.
SETTING
Twenty-two facilities which provide chronic care in southwestern Ontario.
PARTICIPANTS
Patients who were treated with systemic antibiotics over the study period.
MEASUREMENTS
Characteristics of antibiotic prescriptions (name, dose, duration, and indication) and clinical features of randomly selected patients who were treated with antibiotics.
RESULTS
A total of 9,373 courses of antibiotics were prescribed for 2,408 patients (66% of all patients in study facilities). The incidence of antibiotic prescriptions in the facilities ranged from 2.9 to 13.9 antibiotic courses per 1,000 patient-days. Thirty-six percent of antibiotics were prescribed for respiratory tract infections, 33% for urinary infections, and 13% for skin and soft tissue infections. Standardized surveillance definitions of infection were met in 49% of the 1,602 randomly selected patients who were prescribed antibiotics. Diagnostic criteria for respiratory, urinary, and skin infection were met in 58%, 28%, and 65% of prescriptions, respectively. One third of antibiotic prescriptions for a urinary indication were for asymptomatic bacteriuria. Adverse reactions were noted in 6% of prescriptions for respiratory and urinary infections and 4% of prescriptions for skin infection.
CONCLUSIONS
Antibiotic use is frequent and highly variable amongst patients who receive chronic care. Reducing antibiotic prescriptions for asymptomatic bacteriuria represents an important way to optimize antibiotic use in this population.
Keywords: antibiotic use, chronic care facilities, infection
Antibiotics are frequently prescribed to older adults who reside in long-term care facilities. The median rate of systemic antibiotic use reported in this population is 7 courses per 1,000 resident-days, which is similar to the incidence of antibiotic use in acute care hospitals.1–5 Although the spread of multiresistant bacteria poses the greatest challenge to acute care hospitals,6 the detection of antibiotic-resistant bacteria in long-term care facilities, including multiresistant Enterobacteriaceae, enterococci resistant to aminoglycosides or vancomycin, and methicillin-resistant Staphylococcus aureus, has led to increasing concern about the use of antibiotics in this population.7–15 The potential for bacterial resistance, adverse drug reactions, and the financial costs associated with antibiotic use warrant that they be prescribed judiciously to older persons residing in long-term care facilities.16 Although bacterial infections in this population are common,17,18 between 22% and 89% of antibiotic prescriptions in residents of long-term care facilities have been described as inappropriate.2,3,19,20
For clinicians, the decision whether or not to prescribe an antibiotic to a long-term care facility resident can be difficult. Often, clinical signs are nonspecific, and the resident may not be capable of expressing symptoms due to cognitive, hearing, or speech impairment. Frequently, important microbiological diagnostic sources, such as sputum cultures, cannot be obtained, and diagnostic tools, such as x–rays, are not easily available. It is not surprising that the prescription of antibiotics for residents of long-term care facilities is most often empiric, i.e., without the benefit of Gram stain or culture results.16 Relatively few studies have evaluated the appropriateness of the initiation of empiric antibiotics in residents of long-term care facilities.2,3,19
We conducted a prospective study to determine the incidence and variability of antibiotic use in facilities providing chronic care in southwestern Ontario. Using standardized consensus definitions of infection for long-term care facilities,21 we also determined how often clinical criteria for infection were met when antibiotics were prescribed in these facilities.
METHODS
Study Facilities
The majority of patients who receive chronic care in Ontario are older than 65 years and require more than 3 hours of nursing care per day.22 This population is similar to residents of skilled nursing homes in the United States. In Ontario, chronic care is provided both by chronic care hospitals and by acute care hospitals with chronic care beds. Therefore, both types of facilities were enrolled in this study. Specialized chronic care hospitals (such as respiratory, spinal cord, or psychiatric facilities) and units providing primarily rehabilitation or palliative care were excluded. All 33 eligible facilities providing chronic care in south-central and southwestern Ontario were invited to participate in the study.
Each of the participating facilities and the institutional review board of the University of Toronto approved the study.
Data Collection
Patients who were being treated with systemic antibiotics were identified prospectively over the 12-month study period (November 1996 to October 1997) by each facility's infection control practitioner through daily review of pharmacy records. Data collected included the name of the antibiotic prescribed, route, start and stop date of administration, and site of infection for which the antibiotic was prescribed. Number of beds, staffing characteristics, and antibiotic-related policies were also recorded for each study facility.
To document the clinical features present in patients for whom antibiotics were prescribed, a random sample of 10 antibiotic courses, stratified by the presumed site of infection, was selected at 4-month intervals. Since smaller facilities generate fewer prescriptions, all antibiotic courses were reviewed for facilities with less than 75 beds. Data abstracted from patient charts included patient demographic, clinical, and laboratory information. Outcomes of antibiotic treatment, including adverse reactions and death, were also recorded. Since lower respiratory, urinary, and skin and soft tissue infections comprise the majority of bacterial infections in long-term care facilities,17,18 prescriptions for these indications were assessed to see if they fulfilled diagnostic criteria based on definitions of infection for long-term care facilities21(Table 1)In the absence of a validated reference standard for assessing antibiotic use, these clinical definitions, although intended for surveillance purposes, provide a conservative standard for assessing antibiotic prescribing. To ensure accurate data abstraction, an audit of the chart abstraction at each facility for at least 5 of the first 10 prescriptions was performed. Data abstraction was considered acceptable if 80% or more of the items matched.
Table 1.
Definitions21
| Infection Type | Symptoms and Signs |
|---|---|
| Lower respiratory tract infection | At least 3 of the following: |
| a) New or increased cough | |
| b) New or increased sputum production | |
| c) Fever (≥38°C) | |
| d) Pleuritic chest pain | |
| e) New or increased physical findings on chest exam | |
| f) New/increased shortness of breath or respiratory rate more than 25 per minute or worsening mental or functional status | |
| Urinary tract infection | 1. If the resident does not have a chronic indwelling catheter and has at least 3 of the following signs and symptoms: |
| a) Fever (≥38°C) or chills | |
| b) New or increased burning pain on urination, frequency, or urgency | |
| c) New flank or suprapubic pain or tenderness | |
| d) Change in character of urine | |
| e) Worsening of mental or functional status | |
| 2. If the resident has an indwelling catheter and has at least two of the following signs or symptoms: | |
| a) Fever (≥38°C) or chills | |
| b) New flank or suprapubic pain or tenderness | |
| c) Change in character of urine | |
| d) Worsening of mental or functional status | |
| Skin and soft tissue infection | One of the following criteria must be met: |
| 1. Pus present at a wound, skin, or soft tissue site | |
| 2. Four or more of the following signs or symptoms: | |
| a) Fever (≥38°C) or worsening mental/functional status; and/or, at the affected site, the presence of new or increasing | |
| b) Heat | |
| c) Redness | |
| d) Swelling | |
| e) Tenderness or pain | |
| f) Serous drainage |
Statistical Analysis
Data entry was performed using EPI-INFO (Version 6.02, CDC, Atlanta, Ga). Logistic regression using binomial random effects (Egret for Windows, Cytel Software Corporation, Cambridge, Mass) was performed to assess factors potentially associated with fulfillment of standardized criteria for infection. Values for variables at the level of the facility were clustered; all individuals prescribed antibiotics in the same facility shared the same value for these variables. A random effects model was used so as to capture the systematic differences between facilities that may not have been associated with other measured variables. Models for lower respiratory tract infection, urinary infection, and skin and soft tissue infection were constructed.17,18 Potential risk factors for the prescription of antibiotics for asymptomatic bacteriuria, defined as a positive urine culture in the absence of urinary symptoms, were also assessed using a random effects model. For all models, variables with P <.2 were selected using a backwards stepwise approach. Variables with P≥.05 were removed from the model. To estimate the extent to which the variation in the rate of antibiotic use was explained by the infection rate, an analysis of variance was performed using PROC REG (SAS, Version 7.0, Cary, NC). Since infection rates were not directly assessed in this study, a rate for each type of infection (lower respiratory, urinary, skin and soft tissue) was estimated by multiplying the rate of antibiotic use (number of courses of antibiotics per 1,000 patient-days) by the proportion of episodes for which antibiotics were prescribed that met the surveillance definition for each type of infection. This calculation is based on the assumption that episodes that met surveillance definitions were true infections.
RESULTS
Twenty-two of 33 eligible facilities (10 chronic care and 12 acute care hospitals) agreed to participate in the study. Facilities that participated were more likely to have more chronic beds (mean, 182 beds) than those that refused (mean, 93 beds) (P <.001). Ten study facilities were situated in metropolitan areas, 6 in semirural areas, and 6 in rural areas (Table 2)Nine of the facilities had an academic affiliation. Facilities with such an affiliation had more beds on average than facilities without an academic affiliation (mean number of beds, 443 vs 232, respectively; P =.03). All facilities had a pharmacy and therapeutics committee and most (18 of 22) had antibiotic restriction policies.
Table 2.
Characteristics of the Study Facilities
| Characteristic | n(%) |
|---|---|
| Location | |
| Metropolitan | 10 (45) |
| Semirural | 6 (27) |
| Rural | 6 (27) |
| Pharmacy and therapeutics committee | 22 (100) |
| Antibiotic subcommittee | 18 (82) |
During the 12 months of the study, there were 9,373 courses of antibiotics prescribed in 1,276,053 patient-days of surveillance, for a mean of 7.3 courses of antibiotics per 1,000 patient-days (Table 3)The incidence of antibiotic prescriptions in the 22 facilities ranged from 2.9 to 13.9 antibiotic courses (or 138 to 211 defined daily dosages) per 1,000 patient-days. Antibiotics were prescribed on 6% of all patient-days. Over the 12-month study period, 2,408 (66%) of the 3,656 patients in study hospitals received at least 1 course of antibiotics. Of these, 36% received only 1 course of antibiotics, 21% received 2 courses, 13% received 3 courses, and 28% received 4 courses or more. The mean duration of a course of antibiotics was 9 days (range, 1 to 365 days).
Table 3.
Description of Antibiotic Prescriptions
| Characteristic | |
|---|---|
| Antibiotics prescribed, n | 9,373 courses |
| Incidence of antibiotic prescriptions, mean per 1,000 patient-days | 7.3 courses |
| Residents exposed to at least 1 course antibiotics, % | 66 |
| Mean, duration of antibiotics, d (range) | 9 (1 to 365) |
| Indications, % | |
| Lower respiratory tract infection | 36 |
| Urinary infection | 33 |
| Skin and soft tissue infection | 13 |
| Miscellaneous other infections | 9 |
| Prophylaxis | 5 |
| Fever with no known focus | 2 |
Thirty-six percent of antibiotics were prescribed for lower respiratory tract infections, 33% for urinary tract infections, 13% for skin and soft tissue infections, 5% percent for prophylaxis, and 2% for fever where the focus of infection was unknown. Nine percent of prescriptions were for a variety of other presumed infections, including otitis media, superficial eye infections, and gastroenteritis. The most commonly prescribed antibiotics were trimethoprim-sulphamethoxazole, ciprofloxacin, amoxicillin, and cephalexin (17%, 17%, 13%, and 9% of antibiotics prescribed, respectively). Ciprofloxacin and amoxicillin were the most commonly prescribed antibiotics for lower respiratory infections (20% and 17%, respectively) (Table 4)For urinary infections, trimethoprim-sulphamethoxazole, norfloxacin, and ciprofloxacin were most commonly prescribed (33%, 15%, and 15%, respectively). Cloxacillin and cephalexin (29% and 24% of prescriptions, respectively) were the most commonly prescribed antibiotics for skin infections. Seventy-six percent of prescriptions were administered orally, 13% were administered through a gastrostomy or jejunostomy tube, 8% were administered intravenously, and 3% were administered intramuscularly. Sixty-five percent of prescriptions were orders written after the patient was assessed by a physician, 30% were verbal orders; the origin of the order for 5% of the prescriptions was not clear.
Table 4.
The 10 Most Commonly Prescribed Antibiotics for Respiratory, Urinary, and Skin or Soft Tissue Indications
| Respiratory Tract Infections | Urinary Tract Infections | Skin and Soft Tissue Infections | |||
|---|---|---|---|---|---|
| Antibiotic | Prescriptions, n(%) | Antibiotic | Prescriptions, n(%) | Antibiotic | Prescriptions, n(%) |
| Ciprofloxacin | 690 (20) | Trimethoprim-sulfa | 1,047 (33) | Cloxacillin | 367 (29) |
| Amoxicillin | 593 (17) | Norfloxacin | 477 (15) | Cephalexin | 305 (24) |
| Cefuroxime | 411 (12) | Ciprofloxacin | 469 (15) | Ciprofloxacin | 211 (17) |
| Trimethoprim-sulfa | 335 (10) | Amoxicillin | 375 (12) | Clindamycin | 44 (4) |
| Cephalexin | 247 (7) | Nitrofurantoin | 231 (7) | Amoxicillin | 41 (3) |
| Clindamycin | 169 (5) | Cephalexin | 126 (4) | Trimethoprim-sulfa | 41 (3) |
| Erythromycin | 169 (5) | Ofloxacin | 85 (3) | Erythromycin | 38 (3) |
| Cefaclor | 138 (4) | Gentamicin | 65 (2) | Metronidazole | 12 (1) |
| Clarithromycin | 127 (4) | Ampicillin | 54(2) | Tetracycline | 17 (1) |
| Ceftriaxone | 96 (3) | Cefuroxime | 38 (1) | Amoxicillin-clavunate | 8 (1) |
Clinical and Laboratory Features
The mean age of the 1,602 patients whose charts were reviewed was 75 years; 51% of patients were male, and 61% were totally dependent in activities of daily living. Cough and fever were the most common clinical features of the 646 patients with presumed respiratory infection (Table 5)An elevated white blood cell count (>10 × 109/ L) was noted in 54% of the 215 patients in whom the test was obtained. Fifty percent of patients receiving antibiotics for a respiratory indication had a chest x-ray, and, of these, 50% had an infiltrate compatible with pneumonia. Five-hundred eighty-two patients, 40% with indwelling urinary catheters, were treated with antibiotics for a urinary indication. New or increased urinary incontinence and fever were the most common symptoms (Table 5). Thirty percent of patients treated with antibiotics for a urinary indication had asymptomatic bacteriuria. Of the 474 patients treated for skin and soft tissue infections, 64% percent were treated for cellulitis, 23% for infected decubitus ulcers, and 5% for an infection at either a gastrostomy or jejeunostomy site. The most common clinical findings at the site of infection were erythema and tenderness (Table 5). The white blood cell count was elevated in 25% of the 88 patients in whom it was ordered.
Table 5.
Clinical Features of Patients Who Received Antibiotics for Respiratory, Urinary, or Skin and Soft Tissue Indications
| Indication | Clinical Features | Patients, n(%) |
|---|---|---|
| Respiratory | Cough | 75 |
| Fever | 39 | |
| Tachypnea | 31 | |
| Change in mental status | 29 | |
| Rhinorrhea | 12 | |
| Urinary | New or increased incontinence | 27 |
| Fever | 23 | |
| Change in mental status | 21 | |
| Dysuria | 15 | |
| Frequency | 13 | |
| Skin and soft tissue | Erythema | 73 |
| Tenderness | 57 | |
| Swelling | 51 | |
| Purulent drainage | 42 | |
| Warmth | 38 | |
| Change in mental status | 12 | |
| Fever | 11 |
Antibiotic Prescriptions and Standardized Long-term Care Definitions of Infection
Overall, 49% of antibiotic prescriptions for infection met standardized surveillance definitions. Fifty-eight percent of prescriptions reviewed met diagnostic criteria for lower respiratory infection, 28% met standardized criteria for urinary infection, and 65% met standardized definitions for skin infection.
For respiratory prescriptions, univariate analysis showed that diagnostic criteria for lower respiratory tract infection were less likely to have been met when the facility where the antibiotics were prescribed had an antibiotic subcommittee (odds ratio [OR], 0.51; 95% confidence intervals [95% CI], 0.29 to 0.90, P =.02) (Table 6)The following variables were entered in the multivariable model: total beds, presence of physician on duty, restricted reporting, availability of an infectious diseases consultant, restricted antibiotic formulary. Logistic regression analysis using binomial random effects revealed that the likelihood of indications for a prescription meeting standard definitions of lower respiratory tract infection increased with increasing number of total beds in the facility (OR, 1.15 per 100 beds; 95% CI, 1.04 to 1.27; P =.01) and decreased with an antibiotic subcommittee at the facility (OR, 0.45; 95% CI, 0.27 to 0.71; P =.01).
Table 6.
Univariate and Logistic Regression Analysis for Lower Respiratory Tract Infection*
| Variable | Unadjusted OR (95% CI) | P Value | Adjusted OR (95% CI) | P Value |
|---|---|---|---|---|
| Total beds, per 100 beds | 1.10 (1.04 to 1.21) | .06 | 1.15 (1.04 to 1.27) | .01 |
| Chronic beds, per 100 beds | 1.02 (0.80 to 1.29) | .90 | — | — |
| Presence of physician on duty | 1.52 (0.89 to 2.60) | .13 | — | — |
| Restricted reporting | 1.66 (0.83 to 3.2) | .15 | — | — |
| Availability of an infectious diseases consultant | 1.44(0.84 to 2.47) | .18 | — | — |
| Infection control practitioner full-time equivalents | 0.91 (0.74 to 1.12) | .38 | — | — |
| Restricted antibiotic formulary | 1.60 (0.95 to 2.7) | .08 | — | — |
| Presence of an antibiotic subcommittee | 0.51(0.29 to 0.90) | .51 | 0.45 (0.27 to 0.71) | .01 |
| Seven-day (vs 10-day) automatic stop date for antibiotic orders | 0.92 (0.49 to 1.72) | .79 | — | — |
Binomial random effects used to assess factors associated with prescriptions meeting definitons for lower respiratory tract infection. Adjusted odds ratios (OR) of variables kept in the final model are presented. CI indicates confidence interval.
When factors associated with meeting urinary diagnostic criteria were examined by univariate analysis, diagnostic criteria were more likely to have been met for prescriptions from larger facilities (more total beds, OR, 1.28 per 100 beds; 95% CI, 1.05 to 1.56, P =.01) and from facilities with more chronic beds (OR, 1.71 per 100 beds; 95% CI, 1.22 to 2.39; P =.01) (Table 7). These variables, along with restricted reporting were retained for the logistic regression random effects model. The final model showed that prescriptions from larger facilities (OR, 1.21 per 100 beds; 95% CI, 1.01 to 1.44; P =.04) and from facilities with more chronic beds (OR, 1.5 per 100 beds; 95% CI, 1.04 to 2.19; P =.03) were associated with meeting diagnostic criteria, while prescriptions from facilities where antibiotic reporting was restricted were less likely to meet definitions (OR, 0.34; 95%CI, 0.14 to 0.86; P =.02). There were no significant predictor variables for prescriptions for asymptomatic bacteriuria in either univariate or multivariable analysis.
Table 7.
Univariate and Logistic Regression Analysis for Urinary Tract Infection*
| Variable | Unadjusted OR (95% CI) | P Value | Adjusted OR (95% CI) | P Value |
|---|---|---|---|---|
| Total beds, per 100 beds | 1.28 (1.05 to 1.56) | .01 | 1.21 (1.01 to 1.44) | .01 |
| Chronic beds, per 100 beds | 1.71 (1.22 to 2.39) | .01 | 1.5 (1.04 to 2.19) | .03 |
| Presence of physician on duty | 0.97 (0.27 to 3.44) | .97 | — | — |
| Restricted reporting | 0.52 (0.19 to 1.36) | .18 | 0.34 (0.14 to 0.86) | .02 |
| Availability of an infectious diseases consultant | 0.97 (0.31 to 3.0) | .95 | — | — |
| Infection control practitioner full time equivalents | 1.49 (0.87 to 2.55) | .14 | — | — |
| Restricted antibiotic formulary | 1.21 (0.49 to 3.01) | .67 | — | — |
| Presence of an antibiotic subcommittee | 1.56 (0.61 to 4.1) | .35 | ||
| Seven day (vs 10-day) automatic stop date for antibiotic orders | 0.57 (0.17 to 1.94) | .37 |
Binomial random effects used to assess factors associated with prescriptions meeting definitons for urinary tract infection. Adjusted odds ratios (OR) of variables kept in the final model are presented. CI indicates confidence interval.
For skin and soft tissue infection, the availability of an infectious diseases consultant was the only variable associated with meeting the definition of infection in univariate analysis (OR, 2.31; 95% CI, 1.07 to 4.97; P =.03) (Table 8)This was the only variable retained in the multivariable model.
Table 8.
Univariate and Logistic Regression Analysis for Skin and Soft Tissue Infections*
| Variable | Unadjusted OR (95% CI) | P Value | Adjusted OR (95% CI) | P Value |
|---|---|---|---|---|
| Total beds, per 100 beds | 1.13 (0.95 to 1.35) | .21 | — | — |
| Chronic beds, per 100 beds | 0.92 (0.69 to 1.24) | .61 | — | — |
| Presence of physician on duty | 0.70 (0.32 to 1.54) | .38 | — | — |
| Restricted reporting | 0.74 (0.25 to 2.2) | .59 | — | — |
| Availability of an infectious diseases consultant | 2.31 (1.07 to 5.0) | .03 | 2.31 (1.07 to 5.0) | .03 |
| Infection control practitioner full-time equivalents | 1.06 (0.58 to 1.95) | .85 | ||
| Restricted antibiotic formulary | 1.17 (0.50 to 2.72) | .72 | ||
| Presence of an antibiotic subcommittee | 0.95 (0.37 to 2.46) | .92 | ||
| Seven-day (vs 10-day) automatic stop date for antibiotic orders | 0.70 (0.24 to 2.02) | .51 |
Binomial random effects used to assess factors associated with prescriptions meeting definitons for skin and soft tissue infections. Adjusted odds ratios (OR) of variables kept in the final model are presented. CI indicates confidence interval.
The analysis of variance revealed that the surrogate infection rate (number of episodes treated with antibiotics multiplied by the proportion of episodes meeting surveillance definitions for infection) accounted for 62%, 87%, and 89% of the variance in the antibiotic prescription rate for lower respiratory, urinary, and skin and soft tissue infections, respectively (adjusted R2, 0.62, 0.87, and 0.89, respectively; P <.001 for each model).
Adverse Reactions and Deaths
For both urinary tract and respiratory infections, 6% of prescriptions resulted in an adverse reaction, while 4% of prescriptions for skin infection were complicated by an adverse reaction. Nausea accounted for 34% of adverse reactions, rash for 19%, diarrhea for 11%, and yeast infection for 8%. Adverse reactions occurred in 30% of prescriptions for amoxicillin-clavulinic acid, 8% of prescriptions for trimethoprim-sulphamethoxazole, 8% of prescriptions for clindamycin, 4% of prescriptions for ciprofloxacin, 4% of prescriptions for cefuroxime, 4% of prescriptions for cephalexin, and 3% of prescriptions for amoxicillin. Death within 3 weeks of symptom onset occurred in 10% of antibiotic prescriptions for a respiratory indication and in 2% of episodes where antibiotics were prescribed for urinary or skin indications.
DISCUSSION
Our findings indicate that antibiotics are frequently prescribed for patients who receive chronic care in Ontario. The mean incidence of antibiotic use in the 22 study facilities (7.3 courses per 1000 patient-days) falls within the reported range of antibiotic use in other types of long-term care facilities (2 to 12 courses per 1,000 resident-days).1–4 Exposure to at least 1 course of antibiotics occurred in 66% of patients in our study, similar to the report by Montgomery et al. where 76% of nursing home residents received at least 1 course of antibiotics over a 12-month period.2 The majority (82%) of prescriptions were for respiratory, urinary, and skin infections, which is consistent with previous reports.3,4,19,20
Variability in the incidence of antibiotic prescriptions between facilities (2.9 to 13.9 antibiotic courses per 1,000 patient-days) was considerable. Most of the variance in the antibiotic prescription rate, at least for urinary and skin and soft tissue infections (adjusted R2, 0.87 and 0. 89, respectively), was explained by differences in our estimates of infection rates. This data suggests that baseline risks for urinary and skin and soft tissue infections between the study facilities differed considerably. These findings are in contrast to the report by Mylotte, where 50% of the variance in overall antibiotic prescriptions in a cohort of nursing home patients was unexplained by rates of infection.1 The lower variance explained by the estimated lower respiratory infection rate (62%) in our study suggests that factors other than the infection rate were important in driving antibiotic use for lower respiratory tract infections. A limitation of this analysis of variance is that the infection rate was calculated assuming that existing surveillance definitions of infection reflect true infections; however, no better definition of infection currently exists.
Although most of the variability in rates of antibiotic prescription for urinary tract infections appears to be explained by underlying infection rates, 30% of prescriptions for urinary indications were for asymptomatic bacteriuria. Asymptomatic bacteriuria, the presence of significant bacteriuria in the urine without urinary symptoms, is a common condition in elderly, institutionalized individuals. The prevalence of asymptomatic bacteriuria increases with age, occurring in up to 50% of elderly women and 35% of men who reside in long-term care facilities.23 Asymptomatic bacteriuria should be treated in populations at high risk of developing subsequent infection, such as children and pregnant women. However, there is compelling evidence to support not treating asymptomatic bacteriuria in older adults who are institutionalized. Randomized controlled trials of antibiotic treatment for asymptomatic bacteriuria in elderly, institutionalized individuals have demonstrated harm rather than treatment benefit,24–27 and the risk of selecting for antimicrobial resistance is of increasing concern.13–17 Despite this, our findings are in keeping with previous studies: older institutionalized adults with asymptomatic bacteriuria frequently receive antibiotic therapy.2,3,19 These data highlight the need to understand why the evidence of clinical trials has not been adopted into practice, and for an interventional strategy to reduce antibiotic prescribing for asymptomatic bacteriuria in older persons who reside in long-term care facilities.
Relatively few studies have evaluated the appropriateness of antibiotic prescribing in long-term care facilities. Zimmer et al. developed minimum criteria based on Centers for Disease Control surveillance definitions of infection.19,28 The authors conducted a survey of patients on systemic antibiotics in 42 skilled nursing homes and found that 62% of prescriptions met minimum criteria for initiating antibiotics.19 Eighty-eight percent of prescriptions for respiratory tract infections and 65% of prescriptions for urinary infection met criteria. Using criteria adapted from Zimmer et al., Montgomery and colleagues reported that 57% of prescriptions in their study met the criteria.2 Eighty percent of lower respiratory tract infections, 30% of urinary infections, and 70% of skin infections met minimum criteria. In our study, standardized definitions of infection were met in about half of antibiotic prescriptions. Our results are similar to those of Montgomery et al., who found that 28% of prescriptions for urinary infection and 70% of those for skin infection met their criteria.2 However, a lower proportion of prescriptions for lower respiratory tract infections met definitions in our study (58% compared with the 87% reported by Montgomery et al.2). This is probably a reflection of the more stringent criteria for infection used in our study.
Size of the facility was independently associated with fulfillment of definitions for respiratory and urinary infections (OR, 1.15 per 100 beds and OR, 1.21 per 100 beds, respectively); however, the magnitude of these effects are minimal and therefore of little clinical importance. Greater opportunity for educational exposure in larger compared with smaller facilities is a potential explanation for these results, since facilities with a university affiliation were more likely to be large. For skin and soft tissue infections, prescriptions from facilities with the availability of a microbiologist or infectious disease physician were twice as likely to meet definitions compared with those without these specialists. This provides some evidence to support specialist consultation. In contrast, presence of an antibiotic subcommittee (OR, 0.45; 95% CI, 0.27 to 0.74) was associated with failure for meeting criteria. Of the 6 facilities with an antibiotic subcommittee, 5 had a subcommittee on the basis of being affiliated with an acute care facility, 1 as part of a comprehensive drug evaluation program. None could identify specific difficulty with antibiotic use in their facility as a reason for implementing the subcommittee. Similarly, for urinary infections, restricted reporting (selective antibiotic reporting by the microbiology laboratory) in the facilities was instituted because of microbiology laboratory practice standards, not because of difficulty with antibiotic prescribing. It is likely that these associations represent confounding by one or more unmeasured variables. For example, it is possible that characteristics pertaining to individual patients (which were not collected) may have played a role.
Trimethoprim-sulphamethoxazole and ciprofloxacin were the most commonly prescribed antibiotics, each accounting for 17% of prescriptions. The fact that ciprofloxacin was the most commonly prescribed antibiotic for respiratory tract infection, accounting for 20% of prescriptions, is disconcerting. The activity of ciprofloxacin against Streptococcus pneumoniae, an important cause of pneumonia in this population,29 is only moderate.30 The frequent use of ciprofloxacin for respiratory infections is particularly concerning given the increasing rates of quinolone resistance amongst isolates of Streptococcus pneumoniae obtained from older adults.31 Although our aim was not to systematically evaluate the use of specific antimicrobial agents, this observation is in keeping with a report where half of specific antimicrobials were considered inappropriate.20 Strategies to help improve physicians' choice of antimicrobials for patients receiving chronic care are needed.
Strengths of this study include the prospective collection of data and the use of multivariable modeling to determine factors associated with meeting criteria for infection. This study evaluated antibiotic prescribing in view of clinical features present at the time of the initiation of therapy. A general limitation of assessing empiric antibiotic use is that since microbiological diagnosis is not assessed it is possible that prescriptions for lower respiratory tract infections caused by viral agents, and therefore not requiring antibiotics, are classified as appropriate. A specific limitation of this study is that the standardized definitions of infection used may have been too stringent. These definitions, although clinical, were intended for surveillance purposes and not to assess appropriateness of antibiotic use. However, in the absence of a validated reference standard for appropriateness, we believe that using these standards represents a reasonable, albeit conservative, guide for assessing antibiotic use. A major challenge to researchers is to develop better diagnostic assessments for infection of older adults residing in long-term care facilities.
This study demonstrates that antibiotic use is extensive and highly variable amongst patients who receive chronic care. Reducing antibiotic prescriptions for asymptomatic bacteriuria represents an important way to optimize antibiotic use in this population.
Acknowledgments
This study was supported by a grant from the Physicians Services Incorporated Foundation.
We would like to thank the infection control practitioners, administrators, and other staff at the participating study facilities for their help in data collection. We also thank Lorraine Moss for help in data management.
REFERENCES
- 1.Mylotte JM. Measuring antibiotic use in a long-term care facility. Am J Infect Control. 1996;24:174–9. doi: 10.1016/s0196-6553(96)90009-7. [DOI] [PubMed] [Google Scholar]
- 2.Montgomery P, Semenchuk M, Nicolle LE. Antimicrobial use in nursing homes in Manitoba. J Geriatr Drug Ther. 1995;9:55–74. [Google Scholar]
- 3.Warren JW, Palumbo FB, Fitterman L, Speedie SM. Incidence and characteristics of antibiotic use in aged nursing home patients. J Am Geriatr Soc. 1991;39:963–72. doi: 10.1111/j.1532-5415.1991.tb04042.x. [DOI] [PubMed] [Google Scholar]
- 4.Katz PR, Beam TR, Brand F, Boyce K. Antibiotic use in the nursing home. Physician practice patterns. Arch Intern Med. 1990;150:1465–8. [PubMed] [Google Scholar]
- 5.Smith PW. Nursing home infection control: a status report. Infect Control Hosp Epidemiol. 1998;19:366–9. doi: 10.1086/647831. [DOI] [PubMed] [Google Scholar]
- 6.Goldmann DA, Weinstein RA, Wenzel RP, et al. Strategies to prevent and control the emergence and spread of antimicrobial-resistant microorganisms in hospitals. A challenge to hospital leadership. JAMA. 1996;275:234–40. [PubMed] [Google Scholar]
- 7.Armstrong-Evans M, Litt M, McArthur MA, et al. Control of transmission of vancomycin-resistant Enterococcus faecium in a long-term-care facility. Infect Control Hosp Epidemiol. 1999;20:312–7. doi: 10.1086/501623. [DOI] [PubMed] [Google Scholar]
- 8.Greenaway CA, Miller MA. Lack of transmission of vancomycin-resistant enterococci in three long-term-care facilities. Infect Control Hosp Epidemiol. 1999;20:341–3. doi: 10.1086/501628. [DOI] [PubMed] [Google Scholar]
- 9.Wiener J, Quinn JP, Bradford PA, et al. Multiple antibiotic-resistant Klebsiella and Escherichia coli in nursing homes. JAMA. 1999;281:517–23. doi: 10.1001/jama.281.6.517. [DOI] [PubMed] [Google Scholar]
- 10.Terpenning MS, Bradley SF, Yan JY, Chenoweth CE, Jorgensen KA, Kauffman CA. Colonization and infection with antibiotic-resistance bacteria in a long-term care facility. J Am Geriatr Soc. 1994;42:1062–9. doi: 10.1111/j.1532-5415.1994.tb06210.x. [DOI] [PubMed] [Google Scholar]
- 11.Wingard E, Shlaes JH, Mortimer EA, Shlaes DM. Colonization and cross-colonization of nursing home patients with trimethoprim-resistant gram-negative bacilli. Clin Infect Dis. 1993;16:75–81. doi: 10.1093/clinids/16.1.75. [DOI] [PubMed] [Google Scholar]
- 12.Chenoweth CE, Bradley SF, Terpenning MS, et al. Colonization and transmission of high-level gentamicin-resistant enterococci in a long-term care facility. Infect Control Hosp Epidemiol. 1994;15:703–9. doi: 10.1086/646841. [DOI] [PubMed] [Google Scholar]
- 13.Gaynes RP, Weinstein RA, Chamberlin W, Kabins SA. Antibiotic-resistant flora in nursing home patients admitted to the hospital. Arch Intern Med. 1985;145:804–7. [PubMed] [Google Scholar]
- 14.Rice LB, Willey SH, Papanicolaou GA, et al. Outbreak of ceftazidime resistance caused by extended-spectrum β-lactamases at a Massachusetts chronic-care facility. Antimicrob Agents Chemother. 1990;34:2193–9. doi: 10.1128/aac.34.11.2193. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Muder RR, Brennen C, Goetz AM, Wagener MM, Rihs JD. Association with prior fluoroquinolone therapy of widespread ciprofloxacin resistance among gram-negative isolates in a Veterans Affairs medical center. Antimicrob Agents Chemother. 1991;35:256–8. doi: 10.1128/aac.35.2.256. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Nicolle LE, Bentley D, Garibaldi R, Neuhaus E, Smith P. Antimicrobial use in long-term care facilities. Infect Control Hosp Epidemiol. 1996;17:119–28. doi: 10.1086/647256. [DOI] [PubMed] [Google Scholar]
- 17.Nicolle LE, Strausbaugh LJ, Garibaldi RA. Infections and antibiotic resistance in nursing homes. Clin Microbiol Rev. 1996;9:1–17. doi: 10.1128/cmr.9.1.1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Strausbaugh LJ, Joseph C. In: Hospital Epidemiology and Infection Control. Mayhall CG, editor. Baltimore, Md: Williams and Wilkins; 1996. pp. 1151–70. [Google Scholar]
- 19.Zimmer JG, Bentley DW, Valenti WM, Watson NM. Systemic antibiotic use in nursing homes. A quality assessment. J Am Geriatr Soc. 1986;34:703–10. doi: 10.1111/j.1532-5415.1986.tb04301.x. [DOI] [PubMed] [Google Scholar]
- 20.Jones SR, Parker DF, Liebow ES, Kimbrough RC, Frear RS. Appropriateness of antibiotic therapy in long-term care facilities. Am J Med. 1987;83:499–502. doi: 10.1016/0002-9343(87)90761-3. [DOI] [PubMed] [Google Scholar]
- 21.McGeer A, Campbell B, Emori TG, et al. Definitions of infection for surveillance in long-term care facilities. Am J Infect Control. 1991;19:1–7. doi: 10.1016/0196-6553(91)90154-5. [DOI] [PubMed] [Google Scholar]
- 22.Chambers LW, Labelle R, Gafni A, Goeree R. Ontario Report. McMaster University Centre for Health Economics and Policy Analysis; 1992. The Organization and Financing of Public and Private Sector Long Term Care Facilities for the Elderly in Canada; pp. 92–8. Working Paper. [Google Scholar]
- 23.Nicolle LE. Asymptomatic bacteriuria in the elderly. Infect Dis Clin North Am. 1997;11:647–62. doi: 10.1016/s0891-5520(05)70378-0. [DOI] [PubMed] [Google Scholar]
- 24.Nicolle LE, Bjornson J, Harding GK, MacDonnel JA. Bacteriuria in elderly institutionalized men. N Engl J Med. 1983;309:1420–5. doi: 10.1056/NEJM198312083092304. [DOI] [PubMed] [Google Scholar]
- 25.Nicolle LE, Mayhew JW, Bryan L. Prospective randomized comparison of therapy and no therapy for asymptomatic bacteriuria in institutionalized elderly women. Am J Med. 1987;83:27–33. doi: 10.1016/0002-9343(87)90493-1. [DOI] [PubMed] [Google Scholar]
- 26.Ouslander JG, Schapira JF, Schnelle JF, et al. Does eradicating bacteriuria affect the severity of chronic urinary incontinence in nursing home residents? Ann Intern Med. 1995;122:749–54. doi: 10.7326/0003-4819-122-10-199505150-00003. [DOI] [PubMed] [Google Scholar]
- 27.Abrutyn E, Mossey J, Berlin JA, et al. Does asymptomatic bacteriuria predict mortality and does antimicrobial treatment reduce mortality in elderly ambulatory women. Ann Intern Med. 1994;120:827–33. doi: 10.7326/0003-4819-120-10-199405150-00003. [DOI] [PubMed] [Google Scholar]
- 28.Garner JS, Bennett JV, Schechler WE, et al. Brachman PS, Eichhoff TC, et al. Proceedings of the International Conference on Nosocomial Infections. Chicago: American Hospital Association; 1971. Surveillance of nosocomial infections. [Google Scholar]
- 29.Muder RR. Pneumonia in residents of long-term care facilities: epidemiology, etiology, management, and prevention. Am J Med. 1998;105:319–30. doi: 10.1016/s0002-9343(98)00262-9. [DOI] [PubMed] [Google Scholar]
- 30.Scully BE, Hooper DC, Wolfson JS. Quinolone Antimicrobial Therapy. 2nd ed. Washington, DC: American Society for Microbiology; 1993. Therapy of respiatory tract infections with quinolone antimicrobial agents. [Google Scholar]
- 31.Chen DK, McGeer A, de Azavedo JC, Low DE. Decreased susceptibility of Streptococcus pneumoniae to fluoroquinolones in Canada. N Engl J Med. 1999;341:233–9. doi: 10.1056/NEJM199907223410403. [DOI] [PubMed] [Google Scholar]