Effect of Physician Orders for Life-Sustaining Treatment Documents on Antimicrobial Use and Antimicrobial-Resistant Organism Infections in Nursing Home Residents

To the Editor—Widespread emergence of antimicrobial-resistant organism (ARO) infections represents a growing public health crisis [1, 2]. ARO infections are associated with increased morbidity, mortality, and cost [3]. Risk factors include antimicrobial exposure and nursing home residence [4–6]. Physician Orders for Life-Sustaining Treatment (POLST) forms are widely used in our region to direct healthcare compatible with the preferences of patients who may face death in the ensuing year [7–10]. Antimicrobial use is one topic about which patients or their decision makers indicate treatment preferences. They may request (1) no antibiotics except if needed for comfort; (2) no invasive (intramuscular/intravenous) antibiotics; or (3) aggressive treatment. After institutional review board approval, we performed a retrospective pilot study to determine whether patients who ordered comfort-only antimicrobial use received fewer days of therapy than those who ordered invasive or aggressive therapy, and whether fewer days of therapy correlated with fewer ARO infections. The study population consisted of 44 residents from a single nursing home in La Crosse, Wisconsin, all of whom had specified preferences for antimicrobial use in a POLST document (22 comfort-focused and 22 cure-focused) and who died between 1 September 2007 and 31 March 2008. AROs were defined as methicillin-resistant Staphylococcus aureus, vancomycin-resistant enterococci (VRE), and gram-negative rod (GNR) bacteria resistant to 1 or more of the following: third-generation cephalosporins (ceftazidime only for Pseudomonas aeruginosa), fluoroquinolones, or carbapenem antimicrobials. The cohorts were demographically and clinically similar (Table 1). Urinary tract infection was the most common indication for antimicrobial therapy. Two patients in the cure-focused cohort had documented new ARO infections (1 VRE and 1 GNR), compared with none in the comfort-focused cohort. Despite their disparate preferences regarding antimicrobial use, the cohorts' mean days of antimicrobial therapy per 1000 patient-days were not significantly different (92.7 vs 98.2 days, respectively; P = .498). Clinical Data of 44 Nursing Home Residents Near the End of Life With POLST Forms Indicating Either a Comfort or Cure Focusa of Antimicrobial Use Abbreviations: CI, confidence interval; IQR, interquartile range; POLST, Physician Orders for Life-Sustaining Treatment; UTI, urinary tract infection. a Comfort-focused group, patients with POLST orders for no antibiotics except if needed for comfort; cure-focused group, patients with POLST orders for either aggressive treatment or no invasive (intramuscular/intravenous) antibiotics. b As reported on the patient's Minimum Data Set administrative data. A lower score suggests a higher functional state. cP = .498 using Wilcoxon rank-sum test. Clinical Data of 44 Nursing Home Residents Near the End of Life With POLST Forms Indicating Either a Comfort or Cure Focusa of Antimicrobial Use Abbreviations: CI, confidence interval; IQR, interquartile range; POLST, Physician Orders for Life-Sustaining Treatment; UTI, urinary tract infection. a Comfort-focused group, patients with POLST orders for no antibiotics except if needed for comfort; cure-focused group, patients with POLST orders for either aggressive treatment or no invasive (intramuscular/intravenous) antibiotics. b As reported on the patient's Minimum Data Set administrative data. A lower score suggests a higher functional state. cP = .498 using Wilcoxon rank-sum test. These findings surprised us because POLST was previously found to be highly effective in directing patient preferences regarding cardiopulmonary resuscitation (CPR) [7–10]. Although it is unclear why POLST forms did not affect antimicrobial use, on the basis of record review and discussions with nursing home providers we speculate that medical events that necessitate consideration of CPR may be more likely to trigger review of the POLST form for direction. Furthermore, the perception that antimicrobial use is innocuous may bias staff from communicating POLST antimicrobial preferences to providers. Although it is not clear that antimicrobials provide comfort superior to that of high-quality palliative treatments, when decision makers are presented with an option to use antimicrobials, they nearly always choose it, regardless of POLST orders. This is in keeping with the findings of Hammes et al [10], who found that 20 of 115 (17%) patients received antimicrobials on the authority of their decision makers, despite their POLST preferences for comfort-only use. In conclusion, we did not find POLST forms to be effective in decreasing antimicrobial use in residents near the end of life. Additional research to understand the reasons that antimicrobials are prescribed in comfort-focused patients and whether antimicrobials offer benefit over aggressive palliative treatments is required. Financial support. The work was supported by the Gundersen Medical Foundation (Summer Research Fellowship to E. D.). Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.