In 2012, the Center for Medicare and Medicaid Services (CMS) began the Hospital Readmission Reduction Program (HRRP), a program to reduce reimbursement to inpatient prospective payment system hospitals with excess readmissions for specified diseases. The diseases that are currently included in this program are acute myocardial infarction, heart failure, and pneumonia. The HRRP defines readmission as an unplanned admission for any reason within 30 days of discharge.1
There are several known patient variables that can increase the risk of 30-day readmission after an index admission for pneumonia. Patients with more underlying comorbidities, often determined by a Charlson comorbidity score greater than or equal to 2, are at an increased risk of readmission.2,3 In addition, specific diseases such as chronic lung disease, cancer, anxiety, and depression have been shown to increase a patient's risk of readmission.2,4 Age greater than or equal to 65, income less than $43,000, and having 3 or more previous admissions may also increase a patient's risk of readmission.2,4
In addition to patient variables, there are several institutional variables that can influence readmissions. Compliance with CMS process of care measures has shown to help decrease 30-day readmissions.5 Institutional specific information, such as time of year and the floor from which patients are discharged, has been associated with a difference in readmission rates for patients with an index admission of pneumonia, chronic obstructive pulmonary disease (COPD), or heart failure.6 And discharging patients with either clinical instability or uncompensated comorbidities has been shown to increase rates of readmission.2,7
There are limited studies that specifically address the impact of antibiotics on 30-day readmissions. A study by Metersky and colleagues8 showed that antibiotic choice in bacteremic pneumonia, specifically the use of macrolides, decreased mortality, and risk of 30-day readmission. In addition, patients are often treated unnecessarily or with overly broad antibiotics that could cause harm, such as infections due to Clostridium difficile.9,10 A study of antibiotic use in COPD exacerbations showed that patients experienced reduced readmissions due to COPD but increased readmissions due to C. difficile when they received antibiotics.11 Given the lack of available data evaluating antibiotics in a typical hospital setting, we performed a study to determine the impact of antibiotic choice on 30-day readmissions after an index admission for pneumonia.
MATERIALS AND METHODS
A retrospective cohort study was conducted between May 1, 2013, and March 31, 2014. It was approved by the university and hospital institutional review boards on April 11, 2014, and May 7, 2014, respectively. Patients with an admission diagnosis of pneumonia based on ICD-9 codes of 480.xx, 481, 482.xx, 483.x, 485, 486, and 487 were identified through our electronic medical record system and included in the study. Patients were excluded if they died during the index admission, were transferred to another hospital, or were discharged with hospice. Variables collected included sex, age greater than or equal to 65, Charlson comorbidity score, Hematocrit less than 30, greater than or equal to 3 readmissions, instability at discharge, mismatch between organism sensitivity and antibiotic, current cancer, a history of anxiety or depression, and chronic lung disease.3,12 Both the index admission diagnosis code and the readmission diagnosis code, if readmitted within 30 days, were documented. Use of any antibiotic, classified as either empiric or treatment, was recorded.
Empiric antibiotics were defined as starting within 24 hours of admission and lasting for less than or equal to 3 days. Treatment antibiotics were defined by use lasting longer than 3 days. A readmission for a recurrent pneumonia was determined by ICD-9 codes of 480.xx, 481, 482.xx, 483.x, 485, 486, and 487 or primary of 038.0–038.99, 518.81, and 518.84 with a secondary diagnosis of pneumonia.4 Instability at discharge was defined as temperature greater than 37.8°C, heart rate greater than 100 per minute, respiratory rate greater than 24 per minute, systolic blood pressure less than 90 mm Hg, oxygen saturation less than 90%, unable to tolerate oral intake, or abnormal mental status.7
Each antibiotic used was assessed as being used empirically or for treatment. Our sample was split into 2 groups: the first group—our “analysis” group—consists of the patients who were administered antibiotics commonly used at our facility for pneumonia: piperacillin/tazobactam, ceftriaxone, azithromycin, vancomycin, tobramycin, aztreonam, or amoxicillin/clavulanate; and the second group—our comparison group—consists of the patients who were administered one of the 12 excluded antibiotics (see full list in Table 1 notes). Although frequently used for pneumonia at our facility, for statistical purposes, levofloxacin was included in the comparison group. Our analysis compared the 30-day readmissions for the patients in the analysis group to those in our comparison group.
Descriptive statistics were used for patient demographics, antibiotics used, and mean Charlson comorbidity score. Multivariate logistic regressions were performed. The reported P values, which are calculated with heteroskedasticity-robust standard errors, are based on 2-sided tests with significance at the 0.05 and 0.01 levels reported accordingly. Analyses were completed using the Stata statistical software package (Stata/SE version 12.0; StataCorp, College Station, TX).
A total of 271 patients were included in the study; details are provided in Table 1. The percentage of patients readmitted for any reason was 18.8%, and the percentage due to recurrent pneumonia was 4.06%. Table 1 additionally shows how the analysis and comparison group means differ. In general, the analysis and comparison groups were very similar, but there are some notable differences. First, there were a higher percentage of males in the analysis group. Also, the analysis group tended to have more instability on discharge than the comparison group but was less likely to have cancer. Finally, the analysis group was more likely to have a Hematocrit less than 30.
A total of 598 individual empiric antibiotic courses, spread among 20 different antibiotics, were used. Individual treatment antibiotic regimens included 474 courses spread among 29 different antibiotics. The most common antibiotics used both empirically and for the treatment of pneumonia were levofloxacin (35.8, 41.8), piperacillin/tazobactam (18.1, 15.6), ceftriaxone (15.2, 7.8), azithromycin (13.9, 8.6), vancomycin (6.4, 4.9), and amoxicillin/clavulanate (0, 5.5), percentages respectively.
Probability of readmission
Empiric tobramycin was associated with a 31.2% increase in the probability of 30-day readmission for any reason compared with the comparison group antibiotics (P < 0.01) (see Table, Supplemental Digital Content 1, http://links.lww.com/AJT/A27, which shows the regression analysis for the empiric antibiotic courses). A total of 7 readmissions occurred in patients who received empiric tobramycin and one of those 7 (14.3%) was for a diagnosis of acute renal failure.
Additionally, empiric aztreonam was associated with a 13.7% increase in the probability of 30-day readmission for a recurrent pneumonia compared with the comparison group (P < 0.05). A similar relationship shows for treatment with aztreonam, with an associated increase in probability of 30-day readmission for pneumonia of 13.5% over the comparison group (P < 0.05) (see Table, Supplemental Digital Content 2, http://links.lww.com/AJT/A28, which shows the regression analysis for treatment antibiotic courses).
Our hospital readmission rate is consistent with the literature, although due to data availability, we were only able to determine readmissions to our own facility. The readmission risk factors that our hospital can control, such as instability at discharge and a mismatch between organisms and antibiotic sensitivity, were low. This could account for our rate of recurrent pneumonia, which is lower than in the literature.4,6
Other studies have assessed antibiotic impact on either pneumonia readmissions or clinical response with varying results. Our study did not corroborate any of their findings, specifically, the favorable impact of macrolides, amoxicillin/clavulanate, amoxicillin, or linezolid.8,13,14 However, we chose to use ICD-9 codes to define a diagnosis of pneumonia, although this is likely an overly broad definition, to capture a “real-world” use of antibiotics. In addition, we included all patients admitted to our hospital with a diagnosis of pneumonia, including patients from nursing facilities and those with recent hospitalizations or antibiotic use. Therefore, our definition of pneumonia is not strictly community acquired and includes a larger cross section of patients. Also, it is possible that up to 50% of patients may have received unnecessary or inappropriate treatment.7 Including all patients likely improved our ability to determine the influence of adverse effects of antibiotic use on readmissions.
Our groups were fairly well balanced, with slightly more risk factors for readmission found in the analysis group. We attempted to control for as many risk factors as possible that contribute to general and pneumonia readmissions. However, new data on what influences readmissions is ever evolving. We included age greater than 65, but another study did not find that age affected their pneumonia readmission rates, so it is unknown how this could have affected our results.15 In addition, due to lack of available data, we were unable to assess income, which could have affected our outcomes. Adherence to pneumonia processes of care has been shown to affect 30-day readmission rates; however, since our institutional compliance to these measures is high, our study did not collect and incorporate this information into our analysis.5
The risk of readmission for any reason seen with tobramycin, including one subject readmitted with acute renal failure, suggests that adverse reactions due to tobramycin could be influencing this risk. Because of this information, our hospital reevaluated dosing and therapeutic drug monitoring for aminoglycosides. Since our study uses a pneumonia definition broader than community acquired, the use of aminoglycosides in these patients may have been warranted. However, given our results, further exploration into appropriateness could help decrease readmissions.
The association between aztreonam, in both empiric and treatment periods, and an increase in probability of 30-day readmission for recurrent pneumonia is not surprising given its narrow spectrum, and less reliable coverage against Pseudomonas aeruginosa and Enterobacteriaceae, which can be dependent on dosing.16,17 These findings give a future target for review of appropriate use and dosing. The results of our small-scale study will hopefully encourage further research on how antibiotic choice could be used to help reduce readmissions and could be a new evaluative tool for antimicrobial management programs.
1. Centers for Medicare and Medicaid Services. Readmissions Reduction Program. 2014. Available at: http://www.cms.gov/Medicare/Medicare-Fee-for-Service Payment/AcuteInpatientPPS/Readmissions-Reduction-Program.html
. Accessed September 20, 2014.
2. Caplastegui A, Espana Yandiola PP, Quintana JM, et al. Predictors of short-term rehospitalization following discharge of patients hospitalized with community acquired pneumonia. Chest. 2009;136:1079–1085.
3. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–383.
4. Mather JF, Fortunato GJ, Ash JL, et al. Prediction of pneumonia 30-day readmissions: a single-center attempt to increase model performance. Respir Care. 2014;59:199–208.
5. Lee JS, Nsa W, Hausmann LRM, et al. Quality of care for elderly patients hospitalized for pneumonia in the United States, 2006 to 2010. JAMA Intern Med. 2014;174:1806–1814.
6. Dharmarajan K, Hsieh AF, Lin Z, et al. Diagnoses and timing of 30-day readmissions after hospitalization for heart failure, acute myocardial infarction or pneumonia. JAMA. 2013;309:355–363.
7. Halm EA, Fine MJ, Kapoor WN, et al. Instability on hospital discharge and the risk of adverse outcomes in patients with pneumonia. Arch Intern Med. 2002;162:1278–1284.
8. Metersky ML, Ma A, Houck PM, et al. Antibiotics for bacteremic pneumonia: improved outcomes with macrolides but not fluoroquinolones. Chest. 2007;131:466–473.
9. Dellit TH, Owens RC, McGowan JE, et al. Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America guidelines for developing an institutional program to enhance antimicrobial stewardship. Clin Infect Dis. 2007;44:159–177.
10. Slimings C, Riley TV. Antibiotics and hospital-acquired clostridium difficile infection: update of systematic review and meta-analysis. J Antimicrob Chemother. 2014;69:881–891.
11. Rothberg MB, Pekow PS, Lahti M, et al. Antibiotic therapy and treatment failure in patients hospitalized for acute exacerbations of chronic obstructive pulmonary disease. JAMA. 2010;303:2035–2042.
12. Micek ST, Lang A, Fuller BM, et al. Clinical implications for patients treated inappropriately for community-acquired pneumonia in the emergency department. BMC Infect Dis. 2014;14:61.
13. Blasi F, Ostermann H, Racketa J, et al. Early versus later response to treatment in patients with community-acquired pneumonia: analysis of the REACH study. Respir Res. 2014;15:6.
14. Caffrey AR, Morrill HJ, Puzniak LA, et al. Predictors of clinical success among a national veterans affairs cohort with methicillin-resistant staphylococcus aureus pneumonia. Clin Ther. 2014;36:552–559.
15. Ranasinghe I, Wang Y, Dharmarajan K, et al. Readmissions after hospitalization for heart failure, acute myocardial infarction, or pneumonia among young and middle-aged adults: a retrospective observational cohort study. PLoS Med. 2014;11:e1001737.
16. Mandell LA, Wunderink RG, Anzueto A, et al. Infectious Diseases Society of America/American Thoracic Society consensus guidelines on the management of community-acquired pneumonia in adults. Clin Infect Dis. 2007;44:S27–S72.
17. Burgess DS, Frei CR. Comparison of b-lactam regimens for the treatment of gram-negative pulmonary infections in the intensive care unit based on pharmacokinetics/pharmacodynamics. J Antimicrob Chemother. 2005;56:893–898.