Antibiotic prescribing practice in residential aged care facilities - health care providers' perspectives

Widespread and inappropriate antibiotic use in residential aged care facilities (RACFs) has been widely reported.1,2,3,4 This is especially concerning given emerging evidence of antibiotic resistance in RACFs.5 Further, older people are particularly susceptible to the adverse consequences of antibiotic use, including Clostridium difficile infection.6 Thus, efforts to optimise antibiotic prescribing in this population are warranted.

Existing strategies to improve antibiotic use have largely focused on the acute care setting;7 however, different approaches are needed in the RACF setting due to differences in antibiotic prescribing behaviour and organisational resources. Several studies have proposed various factors leading to the widespread prescribing of antibiotics in RACFs, including difficulty in establishing clinical diagnosis of infection and lack of onsite diagnostic facilities.1,8,9 These, however, were based primarily on anecdotal presumption rather than the individual experience of relevant health care providers.

There may be unique challenges to improving antibiotic use in RACFs; however, limited data exist. Accordingly, this study reports on the organisational workflow and workplace culture influencing antibiotic prescribing behaviour and the perceived difficulties in optimising antibiotic use in RACFs.

Methods

This study involved high-level care RACFs affiliated with four major public health care services in metropolitan and regional Victoria, Australia. It forms part of a larger study exploring antibiotic prescribing practices in RACFs. Key health care providers, namely nurses, general practitioners and pharmacists servicing individual RACFs, were recruited using a combination of purposive and snowball sampling strategies.10 Institutional ethics approvals were obtained from all participating health care service networks and Monash University. Informed consent was sought from individual participants.

Senior executive nurses, nurse unit managers (NUMs) and registered nurses (RNs) were invited to participate in one-to-one interviews or focus groups. One-to-one interviews were conducted with GPs and pharmacists. Semistructured interview guides tailored to different health care providers' perspectives were used. All data collection and interviews were conducted by one or two interviewers (C J L and M K) between 8 January 2013 and 2 July 2013. Recruitment continued until data saturation was reached. All interviews were audio recorded and transcribed verbatim. Onsite observation of the working environment and documentation related to antibiotic prescribing was also undertaken. Field notes from onsite observations were compared with interview transcripts for discrepancies.

Data were analysed and coded for emergent themes using the framework approach.11 Data management was facilitated with NVivo version 9.0 (QSR). All transcripts were independently verified against audio recordings by C J L and M K. Data analyses were performed independently by C J L and M K for cross-validation purposes. Themes and codes were finalised at regular meetings involving all researchers.

Results

Characteristics of study sites and participants

Twelve public RACFs (with 30–100 beds per facility) within the four health care networks participated. Primary care was delivered by GPs from different practices (range, 1–19 GPs per RACF). Individual RACFs were serviced by external community pharmacies (for medication supply) and consultant pharmacists (for medication review). Sixty-one participants consented to interviews: 40 nurses (four executive nurses, 15 NUMs and 21 RNs), 15 GPs and six pharmacists (Appendix). Fifteen RNs participated in three focus groups (4–6 participants per group). Other participants were interviewed individually.

Emergent themes

These can be categorised into workflow- and culture-related factors.

Workflow-related factors

Logistical challenges with provision of medical care. An important concern cited by all informants was the lack of onsite doctors to provide immediate clinical assessment. Consequently, antibiotics were commonly prescribed by phone order, especially for minor or recurrent infections. Telephone prescription would not necessarily be followed by onsite review (Box 1, quote 1). Due to logistical barriers, GPs tended towards initiating antibiotics early rather than waiting and observing (Box 1, quotes 2 and 3). There was delay in reviewing antibiotic prescriptions, particularly among GPs without regular onsite visits. All stakeholder groups believed that reliance on locum doctors was associated with greater use of antibiotics (Box 1, quotes 4 and 5).

Pharmacy support. Half the RACFs (6/12) did not have access to onsite antibiotics for after-hours use, which sometimes hindered timely administration (Box 1, quote 6). Medication review for individual RACF residents was only performed annually by consultant pharmacists through a scheduled residential medication management review; as such, short-term courses of antibiotics were rarely reviewed. Most GPs and nurses felt that there was a limited role for pharmacists in influencing antibiotic prescribing (Box 1, quote 7). Pharmacists also perceived major challenges in guiding antibiotic use, including their offsite location, limited communication with GPs and lack of access to clinical notes (Box 1, quote 8).

Nurse-driven infection management. All participants acknowledged the significant role of nurses in driving infection management in RACFs, with mixed opinions about having such a nurse-led system. Some GPs felt confident with nursing assessment, relying primarily on nursing staff information to guide their decisions to prescribe antibiotics (Box 2, quote 1). Other GPs had negative views, commenting on rapid staff turnover, lack of experienced nurses and variability of assessment quality, especially from agency or casual nursing staff (Box 2, quote 2). Several GPs also raised concerns about overreporting and pressure to treat from nurses, leading to unnecessary antibiotic prescribing (Box 2, quote 3). From nurses' perspectives, some emphasised their influence on GPs in initiating or changing antibiotics (Box 2, quote 4). However, many considered their responsibility in infection management overwhelming, given existing staffing and workload issues. Some also indicated a lack of confidence and knowledge in advising about antibiotic use (Box 2, quotes 5 and 6).

Institutional policy and guidelines for antibiotic prescribing. None of the participating RACFs had an antimicrobial restriction policy. Prescribing was often based on residents' histories and antimicrobial susceptibility results, if available. In most instances, however, the type and dose of antibiotics were chosen without following guidelines or evidence, with few GPs citing use of the Australian Therapeutic guidelines: antibiotic.12 Indeed, several GPs had concerns that the guidelines are generally not applicable to the older RACF population. Pharmacists likewise claimed that choice of antibiotics used in this population did not normally follow the guidelines (Box 3, quotes 1 and 2).

There was no standardised method for infection surveillance across participating RACFs. Some (eight RACFs) used infection control practitioners and followed the McGeer definitions for infection surveillance,13 while others (four RACFs) had a self-initiated infection registry. Only one facility monitored long-term trends in antibiotic use and the benefits of this were highlighted (Box 3, quote 3).

It was routine practice at several RACFs to perform regular (monthly or bi-monthly) dipstick urinalysis for all residents, regardless of presence of symptoms. However, this practice was criticised as leading to overtreatment of asymptomatic bacteriuria, with consistent views across stakeholders supporting its abolishment (Box 3, quotes 4 and 5).

External expertise and diagnostic facilities. Most GPs rarely sought advice from infectious diseases specialists (Box 3, quote 6). External supports used included microbiologists at pathology services (regarding multidrug-resistant organisms) and mobile services from hospitals (eg, Mobile Assessment and Treatment Service and In-Reach service) for assistance in administering intravenous antibiotics.

Few of the RACFs (2/12) had onsite radiology or pathology services. Most GPs rarely ordered radiological investigations for chest infections, partly because of the difficulty in transferring debilitated residents to an external site (Box 3, quote 7). Additionally, delay in pathology sample collection often complicated the clinical decision (Box 3, quote 8). Interestingly, GPs had mixed views about the usefulness of urine cultures in guiding antibiotic treatment for urinary tract infections (UTIs) (Box 3, quotes 9 and 10).

Culture-related factors

Patient. Most GPs and nurses felt that resident frailty was an important factor in early initiation of antibiotic treatment, with many GPs also prescribing broader spectrum antibiotics (eg, amoxicillin–clavulanate as opposed to amoxicillin) for this reason (Box 4, quotes 1 and 2). Difficulties in assessing residents with behavioural problems or cognitive deficits also complicated the prescribing decisions. Among this population, correctly obtaining a urine sample for microbiological investigation was often impossible. Fever and typical urinary symptoms were often not observed in presumed UTIs, and therefore, the decision for antibiotic therapy frequently depended on less specific symptoms including changes in behavioural or functional status (Box 4, quotes 3 and 4).

Family. Pressure from family members was identified to influence antibiotic prescribing (Box 4, quotes 5 and 7). Often there were unrealistic expectations of antibiotics being prescribed for minor symptoms or to avert hospitalisation. Antibiotics were sometimes prescribed for residents in end-stage illness to fulfil family expectation.

Institutional. Several GPs felt institutional pressure to use antibiotics in order to avoid legal consequences and sometimes to prolong a resident's life inappropriately, with most nursing staff admitting to overreporting symptoms due to fear of litigation (Box 4, quotes 8 and 9). Both GPs and nurses emphasised the importance of advance care planning in guiding antibiotic prescribing decisions (Box 4, quote 10).

Discussion

To our knowledge, this is the first study that has explored the views of key health care providers about barriers and challenges to optimising antibiotic prescribing in RACFs. One of the major concerns raised was the logistical barrier associated with lack of onsite doctors. This places heavy responsibility on nurses for infection management, a role they are generally not trained to perform. Indeed, this study highlighted a perceived lack of knowledge and guidance regarding antibiotic use among nursing staff. Further guidance and support to the nursing staff is clearly needed. Additionally, the Therapeutic guidelines: antibiotic12 were deemed not relevant to the RACF population, highlighting an unmet need.

Extending the roles of pharmacists in antimicrobial stewardship (AMS) in the RACF setting has shown positive outcomes.14,15 There is potential for consultant pharmacists to provide additional support to nursing staff, particularly with regard to education about appropriate antibiotic use and facilitating surveillance of antibiotic use. Extensive antibiotic surveillance has been common practice in the United States and in European RACFs;16,17 however, such activities are relatively scarce in Australian RACFs. Monitoring of longitudinal trends of antibiotic use and benchmarking across RACFs will be a useful starting point to improve antibiotic use.

Another recurrent theme was the influence of routine dipstick urinalysis (regardless of symptoms) on overprescribing of antibiotics for asymptomatic bacteriuria. Despite studies showing that urine dipstick tests are unreliable for identifying older residents with laboratory evidence of UTI,18 half of the participating RACFs used routine full-ward tests. Anecdotally, positive dipstick urinalysis often led to initiation of antibiotics, especially among psychogeriatric residents. This is concerning, given that treatment for asymptomatic bacteriuria has been shown to contribute to the emergence of antibiotic resistance.19,20 Accordingly, nursing staff education highlighting evidence-based practice about diagnosis and treatment of UTI should be promoted.

Empiric antibiotic prescribing without pathological or radiological investigations was found to be common practice. Reassessing antibiotic therapy according to culture and susceptibility results is critical given the increasing occurrence of multidrug-resistant organisms in the RACF setting,21,22 and particularly helpful given the GPs' reliance on nursing staff to follow-up on the duration and outcomes of antibiotic treatment. Indeed, the recently revised McGeer criteria for infection surveillance has recommended mandatory urine culture for the diagnosis of UTIs.13 On the other hand, however, mandatory culture of urine samples regardless of obvious signs and symptoms could paradoxically lead to an increase in unnecessary antibiotic prescribing for asymptomatic bacteriuria. Thus, strict evidence-based guidelines for the indication of urine cultures and treatment of UTIs are warranted.

The pressure to prescribe antibiotics from nursing staff and family was reportedly a significant influence on antibiotic prescribing behaviour; notably, these factors are potentially modifiable. Antibiotic prescribing decisions in older patients are often difficult and controversial, particularly as part of end-of-life care.23 Studies have shown higher use of antibiotics and a greater risk of acquiring multidrug-resistant organisms among older people with advanced dementia or end-stage illness,24,25 highlighting the need to re-evaluate antibiotic prescribing in this group. Ideally, all new residents of RACFs should have an advance care plan, including decisions about future antibiotic therapy and palliation alternatives. This might prompt appropriate discussions with family and reduce pressure to prescribe in some situations.

In conclusion, significant issues with the existing organisational workflow and culture of RACFs have been identified that might contribute to poor antibiotic prescribing practices, underlying the need for targeted AMS initiatives in this setting. Further intervention should consider the limitations of institutional resources and health care professionals' working relationships within this environment. Importantly, this study has highlighted areas and modifiable factors that will assist in developing future AMS interventions.