Do Clinical Decision Support Systems Reduce Inappropriate Antibiotic Prescribing for Acute Bronchitis?

Citation

Terry, A. (Feb, 2017). Do Clinical Decision Support Systems Reduce Inappropriate Antibiotic Prescribing For Acute Bronchitis? Online Journal of Nursing Informatics (OJNI), 21(1), Available at http://www.himss.org/ojni

Abstract

ncomplicated acute bronchitis is a self-resolving, upper respiratory infection most often caused by a virus. Treatment with antibiotics is unnecessary and contributes to the development of antibiotic resistance. Despite this evidence, healthcare providers frequently succumb to significant pressure from patients and prescribe antibiotics. One strategy being implemented to address this problem is the use of clinical decision support (CDS) systems, either through electronic medical records (EMRs) or non-electronic methods. Nurse practitioners are increasingly utilizing CDS systems as part of the care team. This integrative review of the literature demonstrates that, when implemented correctly, CDS’s can help reduce inappropriate antibiotic prescribing by nurse practitioners for acute bronchitis.

Introduction

Acute bronchitis is a self-limiting upper respiratory infection, also known as the common cold, which leads to inflammation in the bronchi. Acute bronchitis is most often caused by a viral infection such as respiratory syncytial virus (RSV), adenovirus, and parainfluenza (Centers for Disease Control and Prevention (CDC), 2015).  The virus can cause flu-like symptoms for up to two weeks, but a cough can linger for up to eight weeks (National Institute of Health (NIH), 2015).  Acute bronchitis is estimated to be the 5th leading reason Americans seek healthcare from primary care providers, resulting in an average two to three day work absence with each illness (Sauver et al., 2013). Patients often seek prescriptions from their primary care providers after over-the-counter medications fail to alleviate their symptoms. There is a common public misconception that antibiotics are necessary for treating acute bronchitis, creating a large patient expectation that providers will routinely prescribe these potentially harmful medications.

Clinical decision support (CDS) systems are tools that utilize either electronic medical records (EMR) or paper methods to guide the evidence-basis for specific treatment during patient encounters. For example, a CDS might warn clinicians by providing computerized alerts in patient charts that are specific to the patient and diagnosis. In addition, a CDS can help prevent common mental errors, such as miscalculating medication dosages for children, as well as reinforcing national best practice guidelines.

Clinical Problem

Antimicrobial resistance is a growing problem worldwide. According to Ackerman, Gonzales, Stahl, and Metlay (2013), more than half of patients seen for acute bronchitis by their primary care providers are treated with antibiotics, contributing to increased antimicrobial resistance, increased adverse drug events, and increased health care costs.  One way CDS’s have been implemented is to improve the care of patients with acute bronchitis by reducing antibiotic use. According to McCullough, Zimmerman, and Rodriguez (2014), the American Academy of Pediatrics (AAP) and the American College of Physicians (ACP) have issued guidelines to reduce antibiotic use for acute bronchitis and upper respiratory tract infections. The Healthcare Effectiveness Data and Information Set (HEDIS) tracked antibiotic prescriptions for acute bronchitis through their campaign “Avoidance of antibiotic treatment in adults with acute bronchitis” (McCullough et al., 2014, p.1092). The aim of this integrated literature review is to answer the research question: In patients with acute bronchitis, do clinical decision support systems help reduce inappropriate treatment with antibiotics? In addition, this literature review will discuss strategies of how to address the practice gap between provider knowledge and provider practice regarding antibiotics in acute bronchitis.

Methodology

An integrative review of ten research studies using randomized control and cross sectional designs was completed to highlight methods that can be used to decrease antibiotic prescriptions for upper respiratory viral illnesses and acute bronchitis. Several sources were used to retrieve articles for this integrative review. A search was conducted by exploring sources on PubMed, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Google Scholar and Medline. Databases were accessed through the Massachusetts College of Pharmacy and Health Sciences and University of Massachusetts Medical School online libraries. Keywords such as “acute bronchitis”, “antibiotic resistance”, and “clinical decision support tools” were entered as search terms. An ancestry approach was used to obtain a comprehensive literature review and only primary resources were used. Inclusion criteria for the integrative review included adult and pediatric outpatient settings with intervention strategies with electronic and paper CDS’s. Randomized controlled studies and larger sample sizes were preferred to increase validity and quality of data. Qualitative and quantitative studies were analyzed and accepted for the final literature review. Exclusion criteria included studies older than five years, meta-analyses, and meta-syntheses. Complicated acute bronchitis studies were also excluded, such as studies on patients with a history of chronic bronchitis and chronic obstructive pulmonary disease.

Several studies were found that related to uncomplicated acute bronchitis, CDS’s, and antibiotic resistance.  To increase validity of the search a total of thirty articles were analyzed and ten of the most relevant articles were chosen based on strength of the study and common themes such as “antibiotic resistance” and “misuse of antibiotics”. In addition, resources such as the websites of the National Institute for Health (NIH) and Centers for Disease Control (CDC) were utilized for relevant statements and guidelines on the use of CDS’s for curbing antibiotic use for acute bronchitis.

Critical analysis of the data was achieved by using a literature review matrix tool. The literature review matrix tool facilitated the organization of data from selected studies that focused on CDS’s for treating acute bronchitis. After completing the matrix tool, similar study results and implications for nurse practitioner practice were evident. The data observed is more fully addressed in the discussion section. A diffusion-of-innovation theory and Logic model was used in McCullough et al. (2014).  The logic model or matrix was used to evaluate the effectiveness of CDS’s being put in place. The diffusion-of-innovation theory was developed by E.M. Rogers in 1962 and communicates how ideas are spread over time through a specific population, as in this case, use of CDS’s by health care providers to use evidence-based practice (Kaminski, 2011). Other researchers failed to mention a theoretical framework but the diffusion-of-innovation theory could also be applied to their studies.

Critique of Research

The following section presents a critique of the selected literature for the use of CDS’s and their impact on the treatment of acute bronchitis. Each study is examined individually and then contrasted with the rest of the reviewed literature.

A qualitative study performed by Gibbs, Wieber, Myers and Jenkins (2014), consisted of two surveys conducted two years apart, with a total of 337 participants during a weeklong continuing medical education (CME) review conducted by the University Of Colorado School Of Medicine. The surveys asked four questions regarding the importance of appropriate antibiotic use, the barriers that prevented it, and the strategies that would improve compliance. The results of the first survey, with a sample size of 176 respondents, demonstrated a 90% consensus that inappropriate antibiotic use is a “very big” or “big” problem in the United States, but only 44% thought this was a “very big” or “big” problem in their own practices.  Using a sample size of 161 respondents, the second survey identified the top strategy to prevent inappropriate antibiotic use as either providing direct patient education or providing patient education tools such as pamphlets or paper based CDS’s. Gibbs et al. (2014) found multiple barriers to proper prescribing, including high patient expectation, difficulty in breaking old patient and provider habits, and concern on the part of providers that patients would seek care elsewhere if patient expectations were not met. These obstacles were also identified in other studies, such as Ackerman et al. (2013).  Gibbs et al. (2014) further found that providers were more likely to prescribe antibiotics for patients who had a fever, patients that were older, and patients with a prolonged cough. Limitations to the study by Gibbs et al. (2014) included a qualitative design using results that were implied but may not be done in practice and the response rate was 56-65%. One of the strengths of this study is its two-year duration. Gibbs et al., (2014) suggested Continuing Medical Education (CME) providers experiment with multifaceted interactive campaigns to continue to improve practice. Nurse Practitioners, in particular, should consider utilizing CME courses to build on their education in order to provide appropriate evidence-based practice and affect practice change.

Ackerman et al. (2013) performed a cluster-randomized trial with primary care providers (physicians, nurse practitioners, and physician assistants) from 33 practices using print-based and EMR-based interventions in the form of treatment algorithms.  There were 55 clinicians recruited, however only 29 completed the year-end survey. Print-based interventions and EMR interventions were divided between sites. All practices received brochures that explained appropriate antibiotic prescribing for cough and cold illnesses that were given to patients with a chief complaint of cough at check-in.  Providers were also given lectures on how to properly use the intervention and further instruction on how to decrease antibiotic prescribing and patient expectation. The survey showed that 97% of providers agreed that antibiotic resistance was a major health concern and 93% agreed that overprescribing antibiotics is a major cause of resistance. Despite these high percentages, providers continued to prescribe antibiotics for acute bronchitis (Ackerman et al., 2013). The survey revealed major themes among providers including high patient antibiotic expectation, insufficient time of patient encounters, and the understanding that antibiotics were overprescribed. 72% of providers believed that patients would be dissatisfied if they did not receive antibiotics. Furthermore, providers often lacked confidence that chest x-ray could effectively rule-out pneumonia. Positive feedback about the EMR-based intervention included that it demonstrated providers followed “a more factual process” and it “encouraged good medical decision making”, whereas the paper-based intervention was thought to be helpful as a reference aide when discussing diagnosis and treatment plans with patients (Ackerman et al., 2013, p.6). Strengths to this study were its use of randomization and inclusion of practices that have high and low rates of antibiotic prescribing. Another strength was the user-friendly paper based intervention that was easily used during patient encounters. A limitation to this study was the difficulty experienced by clinicians in using the EMR-based decision aid tool. The providers felt the difficulty navigating the EMR paired with it being “too long and comprehensive” deterred them from actually using it (Ackerman et al., 2013, p.6). The study was also impacted by its limited geographic location and small sample size.

Gonzales et al. (2013) conducted a cluster-randomized trial among 33 primary care providers on antibiotic use in uncomplicated acute bronchitis.  Electronic and paper algorithms were used as interventions. Also, patient education pamphlets were provided through the CDC’s Get Smart program when the nurse was alerted through the EMR at the paper-based intervention site that a patient had a chief complaint of cough. The results demonstrated that antibiotic prescribing went from a baseline of 80% to 68.3% when paper CDS’s were given to patients. In addition, while CDS’s at various outpatient settings can help prevent unnecessary antibiotic prescriptions, at a national level the problem is still increasing (Gonzales et al., 2013). The authors concluded that providers may be worried about the potential of missing a diagnosis of pneumonia, which in certain circumstances can be life-threatening. Limitations to this study included the way in which opening the electronic CDS’s, known as Smart Sets, were tabulated; if a record was exited before being completed, reopening the record would register as an additional count. Therefore, the study could have overestimated the utilization of Smart Sets. Use of randomization contributed to the strength of this study. Gonzales et al. (2013) concluded that future studies could focus on CDS’s that do not compare with traditional implementation strategies.

A cluster randomized study done by Jenkins et al. (2013), used an intervention of a one-page decision support algorithm for providers and educational materials for patients in order to deter antibiotic prescribing for eight specific conditions, including acute bronchitis. The study included eight outpatient clinics in the Denver area that were randomized to intervention and control groups. Of the four intervention groups with a sample size of 46 providers, one site incorporated the clinical decision support algorithm into their electronic medical record and the others utilized various forms of hard copy pathways for examination rooms and provider workspaces. There was also a control group of 34 providers. There were two primary outcomes: the change over time in prescribing antibiotics for non-pneumonia acute respiratory infections, and the change over time in prescribing broad-spectrum antibiotics overall. Prior to initiating the study, the authors obtained approval from the Colorado Multiple Institutional Review Board. A quasi-experimental design was used, whereby the authors compared baseline data from two years prior to the intervention to a one-year implementation period. Data on adverse events was collected to ensure patient safety. As hypothesized by Jenkins et al. (2013), the proportion of antibiotics prescribed in the intervention group for non-pneumonia acute respiratory infections decreased by a statistically significant amount, 42.7% to 37.9% (p<.0001) while the control group was statistically unchanged, 39.8% to 38.7% (p=.25). With respect to outcome number two, there was a similar reduction in overall antibiotic prescribing in the intervention group, 26.4% to 22.6% (p<.0001), and no reduction of prescribing was seen in the control group. The second part of the study demonstrated clinical pathways might have improved providers’ diagnosing practices in addition to their prescription practices. For instance, previously a provider may have felt an inclination to diagnose pneumonia in order to justify an antibiotic prescription. Now with the support of the clinical pathways and patient education tools providers may have felt more empowered to diagnose the more appropriate condition of acute bronchitis. Jenkins et al. (2013) recommended study of the same intervention of clinical decision support algorithms over a longer period of time, and concluded that antibiotic use for acute bronchitis is “remarkably common” and future research should be conducted (Jenkins et al., 2014, p.8)

Litvin, Ornstein, Wessell, Nemeth, and Nietert (2012) conducted a study at nine primary care practices focused on facilitators and barriers to the adoption of a clinical decision support system. Litvin et al. (2012) used a larger site distribution than Jenkins et al (2013) did since their study was conducted in nine U.S. states. A total of 39 health care providers, including six nurse practitioners, agreed to use electronic CDS’s for a one-year period. The Institutional Review Board at the Medical University of South Carolina approved this study (Litvin et al., 2012). Diagnoses in the study included acute bronchitis as well as six other acute respiratory illnesses. A mixed-method analysis included both qualitative data from site visits as well as data extracted from the EMR. In a summary of the results, the CDS’s were used for 58.2% of adult patients with relevant diagnoses, for a total of 14,086 uses over the course of the study. In addition, identified facilitators and barriers of CDS use included concerns such as whether the providers were technically skilled in adopting and adjusting to this new digital support tool. Litvin et al. (2012) recommended that all non-physician staff, including nurse practitioners, also be trained in using EMR CDS’s.  A major factor shown to improve electronic CDS use was when medical assistants or nursing staff initiated the clinical pathway in examination rooms during patient visits. Non-physician provider CDS training, and customizing CDS’s to each practice are essential for CDS’s to be effective.  

Mangione-Smith et al. (2015) conducted a cross-sectional study on 1,285 pediatric patient visits that presented as acute respiratory tract infection (ARTI) over a 16-month period. Immediately after the encounter, providers completed a survey on diagnoses, treatments prescribed, presenting symptoms, and physical exam findings. Parents completed surveys on provider communication and care satisfaction. Positive treatment was classified as parents answering yes to “Did the doctor tell you things you can do to make your child feel better, for example, home remedies?” Negative treatment was classified as parents answering yes to “Did the doctor tell you antibiotics will not help your child get better” (Mangione-Smith et al, 2015, p.223). Overall, 34% of visits classified under ARTI, including both bacterial and viral etiologies, received antibiotics but the prescription rate was only 5% for viral ARTI visits. Patient encounters that utilized positive treatment recommendations were associated with a 52% decrease in antibiotic prescribing for viral ARTI’s. This improved to an 85% decrease in prescribing when a combined approach with negative and positive recommendations was used (Mangione-Smith et al., 2015).  The study also demonstrated a 16% greater chance of high parent satisfaction ratings if a combined positive and negative approach was used. Limitations to the study by Mangione-Smith et al. (2015) included a possible lack of generalizability, since the providers were from one geographic region (Seattle, Washington) and the patients were primarily non-Hispanic whites from highly educated parents with high economic status.

McCullough et al. (2014) conducted a cross-sectional study to identify the effect of clinical decision support on non-indicated antibiotic prescriptions. McCullough et al. (2014) used data from the 2006 through 2010 National Ambulatory Medical Care Surveys (NAMCS). NAMCS is a survey of non-federally employed, office-based providers of ambulatory care services.  According to McCullough et al. (2014), 3808 cases met the inclusion criteria using International Classification of Disease, ninth revision (ICD-9) codes 466, bronchitis not otherwise specified; 490, acute URI; and 465, acute nasopharyngitis. The results showed a 19% risk reduction of antibiotic prescribing when clinical support was used but an increase in use of broad-spectrum antibiotics. McCullough et al. (2014) also found that, despite national efforts aimed at reducing antibiotic prescribing, 40% of outpatient visits for acute bronchitis and URIs were given antibiotics. The study was limited in that the authors had access only to data entered on NAMCS and special patient conditions were not known. Strengths to this study included its large sample size.

A cross sectional study done by Vinnard et al. (2013) evaluated the effect of intensive intervention versus a mild intervention group on the rate of antibiotic prescribing for acute URIs using a sample size of 1344 patient visits. Intensive intervention was called “academic detailing”, wherein a pharmacist and hospital administrator presented published literature and provider-specific evaluations in an in-person session. The mild intervention consisted of provider approved patient educational mailed material on antibiotic prescribing. The study demonstrated a reduction in antibiotic prescribing for ARIs from 43% to 33% in the intensive group and no significant reduction in the mailing group (Vinnard et al., 2013). Patients with chronic bronchitis and emphysema were excluded from the study. In this study, the authors examined the rate of narrow-spectrum antibiotics versus broad-spectrum antibiotics such as azithromycin, clarithromycin and amoxicillin-clavulanate. Vinnard et al. (2013) found that academic detailing when combined with other educational interventions led to a reduction in use of broad-spectrum antibiotics.  Recommendations from Vinnard et al. (2013) included that hospitals should continue use of academic detailing and spend less time on direct patient mailings.

Rattinger et al. (2012) conducted a large retrospective study of 3831 patient encounters over a five-year period, from January 2002 through December 2006. The study was broken down into two time segments, before and after an EMR CDS implementation that examined fluoroquinolone and azithromycin prescriptions for ARI including acute bronchitis as well as pneumonia, sinusitis and non-specific ARI. The software used in the EMR verified the clinical condition and sought to identify circumstances where antibiotics may be warranted, such as temperatures greater than 38 degrees Celsius, pulse greater than 100, lung consolidation, or unrelenting symptoms greater than 7 days (Rattinger et al., 2012). When antibiotics were not indicated, providers were educated on how to maintain patient satisfaction without prescribing antibiotics. The rate of non-indicated prescriptions fell from 22% to 3%, one of the best results of any study included in this review. Strengths to this study were the large sample size and the five-year duration of study. This study was limited by its retrospective design, using a single site, and not employing a random allocation process that left it susceptible to bias (Rattinger et al., 2012).

Finally, Mainous, Lambourne, and Nietert (2012) performed a quasi-experimental study on 70 practices (nine intervention and 61 control) both before and after CDS implementation over a 15-month period. The EMR-based CDS intervention included prompts for recommended antibiotics once a diagnosis was made and for non-antibiotic prescriptions when they were not indicated. This strategy led to a reduction in non-indicated antibiotics of 16.6% compared to the control group. In addition, the rate of broad-spectrum antibiotics declined by an even more impressive 19.7%. The study by Manouis et al. (2012) was limited by the potential for ascertainment bias and the confounder that the study duration overlapped with the 2009 H1N1 epidemic. Mainous et al. (2012) recommended implementing an EMR strategy to continue to decrease broad-spectrum antibiotic prescriptions for ARI’s as an attempt to control antibiotic resistance. 

Comparison and Contrast of Studies

Four major themes that emerged from the studies were that antibiotic prescriptions are not indicated for acute bronchitis (without complications), inappropriate antibiotic administration contributes to resistant bacteria, broad spectrum antibiotics are of particular concern contributing to larger resistance, and CDS’s demonstrated various degrees of efficacy. Gonzales et al. (2013) stated there has been little change in the over-prescription of antibiotics for acute bronchitis despite evidence proving its ineffectiveness. Gibbs et al. (2014) and Ackerman et al. (2013) found high patient expectation was the most common contributor to antibiotic prescribing among providers. Gibbs et al. (2014) goes on to say, “educational meetings with providers are moderately effective in increasing knowledge but not effective in changing complex behaviors” (p.5). According to Ackerman et al. (2013), leading causes for non-indicated antibiotics are provider beliefs that patients will not be satisfied unless they receive antibiotics, there is insufficient provider time with patients, and providers are concerned about legal ramifications of not prescribing antibiotics if there is a bad outcome. Gibbs et al. (2014) also recommended the “need to continue to address inappropriate prescribing through patient and provider education and new interventions” (p.5).

Studies showed different CDS’s demonstrated varying degrees of efficacy in reducing antibiotic prescriptions. Paper interventions were found to be most helpful in the studies by Gonzales et al. (2013) and Ackerman et al. (2013), whereas Manouis et al. (2012) found that EMR-based prompts worked best at reducing the rate of broad-spectrum antibiotic prescription. In all of the studies, there was a statistically significant impact from CDS’s, but most still had a large persistent problem of non-indicated antibiotic prescriptions. Only Rattinger et al. (2012) nearly eliminated the problem, with a post-intervention prescription rate of only 3%. Jenkins et al. (2013) and McCullough et al. (2014) had conflicting results, as Jenkins demonstrated a reduction in broad-spectrum antibiotic prescriptions with a decision support tool, but McCullogh et al. (2014) found an increase in the rate of antibiotic prescription. This finding was counter-intuitive and may have been attributable to chance or an unknown confounder.

Strengths and Limitations

There were a variety of limitations in the sampled studies. One qualitative study used a survey that asked participants what strategies they thought might help reduce patient expectations for antibiotics, but did not actually test this theory (Gibbs et al. 2014). Another study was limited by the potential for ascertainment bias and was confounded by a duration that overlapped the 2009 H1N1 epidemic (Manouis et al. (2012). Other studies were limited by the sources from which they collected their data, such as McCullough et al. (2014), which had access only to data entered on NAMCS. Several of the studies were limited to certain geographic locations, participant demographics, and possibly lacked generalizability such as Mangione-Smith et al. (2015). However, the fact that there were several studies from different regions with the same results tends to counter this particular limitation. Jenkins et al. (2013) may have been impacted by the Hawthorne effect, where providers change their practice simply from knowing they were being studied. The providers in the control group, who did not receive clinical pathways and patient education tools, may have forgotten they were even being studied, since data was collected from an electronic database and not in-person.

Randomized and cross-sectional study techniques were applied in many of the studies, which led to increased strength and validity. Large sample sizes were used by Rattinger et al. (2012) and McCullough et al. (2014), among others. The main strength in this review was not in each individual study, but in the power of repetitive experimentation that showed similar results. Each study that directly examined the impact of CDS’s demonstrated that they have a measurable impact on the reduction of non-indicated antibiotics.

Controversies

There were no major controversies that emerged from the articles. Authors were free of conflicts of interest such as financial contributions from companies. One unspoken controversy, however, is that of publication bias. This literature review did not find any studies that demonstrated CDS’s had complete lack of reduction on the prescription of antibiotics for acute bronchitis. This may have been because authors who found CDS’s were ineffective did not have their studies accepted for publication.

Results

Multiple themes emerged from the review of the literature. The first theme that emerged among the studies was that antibiotics are not indicated for acute bronchitis and the majority of cases are viral in etiology. Another theme that emerged from the literature was that overprescribing antibiotics for acute bronchitis contributes to antibiotic resistance and increased health care costs. Ackerman et al. (2013) showed 97% of provider respondents from the study interventions believed antibiotic resistance was a major health threat and 93% of provider respondents believed overprescribing of antibiotics was a major reason for increased resistance. One of the largest declines in antibiotic prescriptions was from the study conducted by Rattinger et al. (2012). who saw a reduction in antibiotic prescriptions from 22% to 3 % when providers used EMR CDS’s.

For the purpose of this integrative review, the aim was to answer the question: Do clinical decision support systems reduce inappropriate antibiotic prescribing for acute bronchitis? The studies reviewed showed a consensus among the research that CDS’s have a measurable impact on reducing the rate of non-indicated antibiotic prescriptions for acute bronchitis. Gonzales et al. (2013), demonstrated that antibiotic prescribing went from a baseline of 80% to 68.3% when paper CDS’s were given to patients. National trends have shown that prescribing antibiotics for acute bronchitis has not changed, although there have been decreases in rates of prescribing antibiotics for other ARI’s. Major themes from the literature showed that combining patient and provider education is one of the most effective strategies to decrease antibiotic treatment. Some studies indicated that paper support tools were more effective than electronic prompts while others demonstrated they were equivalent (Ackerman et. al, 2013; Jenkins et. al, 2014). A key theme in the qualitative studies highlighted the way patient education was conducted, specifically that mixing positive and negative feedback was more effective in promoting patient satisfaction than either feedback method alone.

A major barrier found in prescribing antibiotics for acute bronchitis was that providers felt there was high patient expectation for an antibiotic prescription. In the study by Ackerman et al., (2013) 79% of providers that completed the survey responded that patients want and expect antibiotics and 72% of providers believed that patients would be dissatisfied if they did not receive antibiotics.

Discussion

Unnecessary antibiotic prescriptions for acute bronchitis are commonly prescribed in outpatient settings. This integrative review set out to answer the research question: In patients with acute bronchitis, do clinical decision support systems help reduce inappropriate treatment with antibiotics? The integrative review highlighted how CDS’s have a measurable impact on reducing the rate of inappropriate antibiotic prescription, but the degree of impact was shown to vary depending on the specific CDS used. Although the results answered the question, more research is still needed. Some studies showed that electronic CDS’s had a higher decline of antibiotic prescriptions whereas others showed papers-based CDS’s to be more beneficial. In the study by Rattinger et al (2012), the rate of inappropriate antibiotic administration fell from 22% to 3%. However, in the Vinnard et al (2013) study, the decline was less pronounced where prescriptions fell from 43% to 33% in the intensive intervention group and did not fall at all in the group that received a mailing brochure.  

A common theme among the studies showed patient and provider education on this topic should continue to be implemented. Validity was strong through all of the selected papers, as they were all randomized, controlled trials. In particular, Jenkins et al. (2013) demonstrated strong statistical conclusion validity with a p value of (<.0001) with respect to the question of overall antibiotic prescribing, and a high degree of power as a result of their large sample size (48,390 total patient visits). One threat to implementation of CDS’s in nurse practitioner practice demonstrated in some studies was that electronic CDS’s can be difficult to navigate and are time consuming.  Some studies may have potential issues with generalizability and reproducibility such as the ones done by Gonzales et al. (2013) and Jenkins et al. (2013). The one by Gonzales et al. (2013) was a cluster-randomized study that took place at various locations in central Pennsylvania and The Jenkins et al. (2013) study took place in specific clinics in Colorado. These studies should be reproduced across a larger area of the country in order to ensure that their results are generalizable.

Implications for Practice

Nurse practitioners and primary care offices should consider implementing a clinical decision support tool for the reduction of inappropriate antibiotic prescriptions. The exact clinical decision support tool that should be implemented would depend on the resources available and the patient education techniques that might already be in place. For instance, if the practice is already using an Electronic Medical Record as many practices are, implementing a program such as Smart Sets for acute bronchitis could be effective. If, on the other hand, the practice utilizes paper charts, they might want to implement the patient education posters that were used in studies such as the one done by Jenkins et al (2013). The only intervention demonstrated to be ineffective was mailings. Any of the other interventions would be possible as long as they were considered cost-effective and appropriate for the clinical context.  Continuing to support health care providers by making paper-based algorithms available for patients to visualize may help persuade patients that antibiotics are an unnecessary treatment for acute bronchitis. Since a common theme was the concern among providers for patient dissatisfaction, this may allow for continued education among patients at each encounter and an eventual decrease in antibiotic prescribing due to provider practice change. EMR’s allowed for better clinical reasoning among providers even though some providers found it to be more difficult than paper-based interventions since it encouraged an evidence-based approach.

Future Research

Further studies should be done to address continued inappropriate antibiotic prescribing.  For instance, since providers thought insufficient time with patients contributed to non-indicated antibiotic prescriptions (Ackerman et al., 2013), a study could be conducted which specifically compensated providers for extra time spent explaining to patients why antibiotics are ineffective in acute bronchitis and what alternative therapies could give better relief. The primary outcome of this study would be whether the extra time was able to reduce antibiotic prescriptions and increase patient satisfaction. The Ackerman et al. (2013) study showed that educational brochures and EMR’s made a minimal change in patient expectation on their own without clinician interaction and may be more effective in changing patient expectation if clinicians and patients review algorithms together. Litvin et al. (2012) recommended that all non-physician staff, medical assistants and nurses, be trained in accessing CDS’s, and showed that when medical assistants or nurses initiated opening the electronic CDS tool when patients were in exam rooms providers thought it was accessed easier and used more frequently. Future research could include studies that incorporated non-provider and provider CDS training to explore if the rate of using clinical pathways in CDS’s increased and if antibiotic prescriptions decreased with more CDS use. Gibbs et al., (2014) suggested Continuing Medical Education (CME) providers should experiment with multifaceted interactive campaigns to continue to improve practice. Nurse practitioners would benefit from CME courses to improve their practice and initiate provider change where needed. Vinnard et al. (2013) suggested further research that targeted behaviors of clinicians with high prescription rates. Also noted from other studies, future research could include more studies that focused on clinician and patient interaction with paper-based algorithms.

In conclusion to the research question, clinical decision support systems have been shown to be effective in improving provider and patient education, which addresses the practice gap in appropriate management of acute bronchitis. While they do not eliminate inappropriate antibiotic prescribing, they do promote a measurable reduction. Nurse practitioners are leaders in patient education and evidence-based care. In order to best serve their patients, nurse practitioners should continue to focus on the implementation and revision of effective CDS tools in order to be at the forefront of this important strategy to reduce inappropriate antibiotic prescription for acute bronchitis.

References

Ackerman, S., Gonazales, R., Stahl, M., Metlay, J. (2013). One size does not fit all: Evaluating and intervention to reduce antibiotic prescribing for acute bronchitis. BMC Health Services Research, 13(462), 2-9. Retrieved from: http://www.biomedcentral.com/1472-6963/13/462/

Centers for Disease Control and Prevention. (2015, April). Bronchitis (chest cold). Retrieved from: http://www.cdc.gov/getsmart/community/for-patients/common-illnesses/bronchitis.html

File, T. (2015). Acute bronchitis in adults. Up to date. Retrieved from: http://www.uptodate.com/contents/acute-bronchitis-in-adults

Gibbs, R., Wieber, C., Myers, L. & Jenkins, T. (2014). A continuing medical education campaign to improve use of antibiotics in primary care. Journal of Biomedical Education. 2014, 1-6. Article ID 537681. Retrieved from: http://dx.doi.org/10.1155/2014/537681

Gonzales, R., Anderer, T., McCulloch, C., Maselli, J., Bloom, F., Graf, T., Stahl, M.& Yefko, M. (2013). A cluster randomized trial of decision support strategies for reducing antibiotic use in acute bronchitis. JAMA Internal Medicine, 81(8), 521-526. doi:10.1001/jamainternmed.2013.1589. Retrieved from: http://archinte.jamanetwork.com/article.aspx?articleid=1556795

Jenkins, T., Irwin, A., Coombs, L., DeAlleaume, L., Ross, S., Rozwadoski, J., Webster, B., Dickinson, M., Sabel, A., MacKezie, T., West, D. & Price, C. (2013).  Effects of clinical pathways for common outpatient infections on antibiotic prescribing. The American Journal of Medicine, 126, 327-335. Retrieved from: http://dx.doi.org/10.1016/j.amjmed.2012.10.027

Kaminski, J. (2011). Diffusion of innovation theory. Canadian Journal of Nursing Informatics, 6 (2),Retrieved from: http://cjni.net/journal/?p=1444

Litvin, C., Ornstein, S., Wessell, A., Nemeth, L. & Nietert, P. (2012). Adoption of a clinical decision support system to promote judicious use of antibiotics for acute respiratory infections in primary care. International Journal of  Medical Informatics, 81, 521-526. oi:10.1016/j.ijmedinf.2012.03.002

Mainous, A., Lambourne, C. & Nietert, P. (2012). Impact of a clinical decision support system on antibiotic prescribing for acute respiratory infections in primary care: quasi-experimental trial. Journal of American Medical Informatics Association, 20, 317-324. doi:10.1136/amiajnl-2011-000701

Mangione-Smith, R., Zhou, C., Robinson, J., Taylor, J., Elliot, M. & Heritage, J., (2015) Communication practices and antibiotic use for acute respiratory tract infections in children. Annals of Family Medicine, 13, 221-227. doi: 10.1370/afm.1785.

McCullough, J., Zimmerman, F., Rodriguez, H. (2014). Impact of clinical decision support on receipt of antibiotic prescriptions for acute bronchitis and upper respiratory tract infection. Journal of American Medicine Informatics Association,  21, 1091-1097. doi:10.1136/amiajnl-2014-002648

National Institute of Health. (2015). Acute bronchitis.  Retrieved from: https://www.nlm.nih.gov/medlineplus/acutebronchitis.html#summary

Rattinger, G., Mullins, C., Zuckerman, I., Onukwugha, E., Walker, L., Gundlapalli, A., Samore, M. & DeLisle, S. (2012). A sustainable strategy to prevent misuse of antibiotics for acute respiratory infections. Plus One, 7(12), e51147. doi:10.1371/journal.pone.0051147

Sauver, J., Warner, D., Yawn, B., Jacobson, D., McGree, M., Pankratz, J., Melton, J., Roger, V., Ebbert, J. & Rocca, W. (2013). Why patients visit their doctors: assessing the most prevalent condition in a defined American population. Mayo Clinic Proceedings. 88 (1),56-57.  Retrieved from:
http://www.mayoclinicproceedings.org/article/S0025-6196(12)01036-1

Vinnard, C., Linkin, D., Lokalio, A., Leonard, C., Teal, V., Fishman, N. & Hennessy, S. (2013). Effectiveness of interventions in reducing antibiotic use for upper respiratory infections in ambulatory care practices. Population Health Management, 16, 22-27. DOI: 10.1089/pop.2012.0025

Biography

Angela Terry, RN, FNP student is originally from Auburn, Massachusetts. She graduated from the University of Massachusetts Boston with her BSN and has been practicing as a Registered Nurse since 2009. She is currently a full time family nurse practitioner student at MCPHS University and due to graduate in May 2016. Angela worked as an emergency room registered nurse for five years at the University of Massachusetts Medical Center and 2 years at Massachusetts General Hospital in Maxillofacial surgery prior to attending nurse practitioner school. She now focuses her time on her studies and her two-year-old son.