Using Nursing Home Antibiograms to Improve Antibiotic Prescribing and Delivery AHRQ Annual Meeting September 10, 2012

Using Nursing Home Antibiograms to Improve Antibiotic Prescribing and Delivery AHRQ Annual Meeting September 10, 2012 Abt Team: Jeremiah Schuur, MD, MHS, Brigham and Women’s Hospital; Rosanna M. Bertrand, PhD, Donna Hurd, MSN Denver Health/University of Maryland Team: Jon Mark Hirshon, MD, MPH, PhD

Background Antibiotics are frequently prescribed in Nursing Homes (NHs) NH residents are frequently transferred to emergency departments (EDs) for acute conditions and given antibiotics Initial antibiotic decisions are empiric – based on: Patient factors (e.g. age, symptoms) Facility factors (type of NH, historical experience, formulary) Preference / knowledge, etc.

Background Antibiogram: A tool that presents local microbiologic sensitivity data to assist clinicians in making empiric prescribing decisions. Hospitals have used antibiograms to: Identify important local resistance patterns Increase antibiotic prescribing for acute infections Agency for Healthcare Research and Quality (AHRQ) contracted with two ACTION teams to investigate the feasibility of creating NH antibiograms: Can NH antibiograms be created? Can NH antibiogram programs be effectively implemented? Can antibiograms be transferred to local EDs Do antibiograms affect prescribing?

Sample Antibiogram

Abt NH Antibiogram Project Funded by AHRQ ACTION Network: Contract # HHSA290200600011I TO13 Research Team Abt Associates Project Director: Rosanna Bertrand, PhD NH Clinical Expert: Donna Hurd, MSN; Jennifer DeAngelis, BA; Laura Goodman, BA Brigham and Women’s Hospital Principal Investigator: Jeremiah Schuur, MD, MHS Allen Gold, BA

Project Details Developed an antibiogram at three Boston-area NHs Used NH-specific micro data from the clinical lab Created a one page antibiogram with key findings Educated NH prescribers and staff about antibiogram Designed a process for antibiogram transfer to local ED Evaluated Provider knowledge and perceptions of antibiograms with survey and interviews Transfer rate of NH antibiogram to ED Prescribing patterns at NH and ED (pre/post)

Abt’s Key Findings Feasible to create NH antibiograms but technical limitations Simple to use NH micro data – difficult to integrate local hospital data Sample size issues limit the number of microorganism that can be described (CLSI standard is at least 30 results per organism) Most useful results for urine infections Most data came from urine cultures (75-90%) Significant resistance to common antibiotics among gram-negative organisms that frequently cause UTIs (E. coli, Klebsiella): Quinolone sensitivity for E. coli in urine was limited (45-60%) Prescribers report using antibiograms, but no evidence found When given an antibiogram with E.coli resistance to quinolones they said they would change their prescribing for urine infections. No significant change in use of quinolones or broad spectrum antibiotics

Key Challenges & Lessons Learned Changing NH prescriber behavior is challenging Diverse group of providers (MD/DO, NP, PA) Geographically separated Time pressed Antibiogram impact limited unless implemented as a QI program Does prescriber see antibiogram at time of prescribing? Is their prescribing behavior followed (QA/QI)? Ensuring ED prescribers receive NH antibiogram is challenging NH transition of care is complex -- another paper form transfer is difficult NH paper forms not always seen by ED prescribers (lost, not reviewed, etc.) Usefulness of an online antibiogram is limited unless available for all NHs Providers won’t look for a rare event

Denver Health/ University of Maryland- Nursing Home Antibiogram Project Funded by Agency for Healthcare Research and Quality ACTION Contract No. 290-2006-00-20, Task Order No. 9 Research Team Denver Health ACTION PI: Thomas D. MacKenzie, MD, MSPH Susan Moore, Josh Durfee University of Maryland Principal Investigators: Jon Mark Hirshon, MD, MPH, PhD and Jon Furuno, PhD.

Denver-Maryland NH Antibiogram Project Details Project implemented at 3 separate NHs in Maryland Urban, rural, suburban Chart reviews conducted At each NH and their affiliated acute care hospital Antibiogram created for each NH Pocket size, laminated index card Needs assessment performed through interviews with the infection control/QA nurse In-services conducted at each NH for nursing staff and physicians Shared findings from chart reviews Educated staff on use of the antibiogram Post implementation evaluation at one NH Prescribing patterns at one NH and affiliated ED (pre/post) Knowledge of ED physicians of antibiogram

Characteristics of the 3 NHs

Chart Reviews Findings Reviewed 623 NH and 216 ED (hospital) charts NHs differed in the number of residents transfer each month (9-20 transfers) UTI was most common infection and urine cultures were the most common type ordered Escherichia coli (E. coli) and Proteus mirabilis (P. mirabilis) were the two most prevalent organisms isolated in cultures NH Physicians prescribed mainly one antibiotic where ED physicians started patients on at least two. Oral antibiotics were prescribed in NH and IV in ED. NH physicians mainly prescribed fluoroquinolones (ciprofloxacin or levofloxacin) as an initial antibiotic (20-35% of initial prescribing instances). Empiric antibiotic prescribing was very common in all three NHs (84-87% of initial antibiotics prescribed). For the culture positive treated infections, empiric prescribing was correct only 25-33% of the time. Definitive antibiotic prescribing occurred in 32-52% of culture positive treated infections.

Evaluation of the Antibiogram in one NH/ED In NH, ciprofloxacin was again the most frequently prescribed initial antibiotic but was prescribed 15% of the time versus 22% . Cefuroxime was prescribed14% of the time, up from 3% during the initial chart review period. In NH, empiric prescribing was correct 45% of time, up from 30% No changes in prescribing patterns were found in the ED.

Key Challenges and Lessons Learned Challenges Availability of NH staff/physicians Only interviewed one fulltime ED physician who had not seen antibiogram Time required for chart review Small number of cultures Lessons Learned NH antibiograms can be created using chart review but this is time consuming NH relationships with their laboratories differ based on the NH NH antibiograms mainly represent UTIs as, 90 to 100 percent of suspected urinary tract infections are cultured, but rarely any other suspected infections are cultured. Culture information from ED transfers did not add to the NH antibiogram unless the NH had a high number of transfers

Using Nursing Home Antibiograms to Improve Antibiotic Prescribing and Delivery AHRQ Annual Meeting September 10, 2012