Secondary Logo

Journal Logo

CASES OF NOTE

Use of a Clinical Decision Support Tool to Improve Guideline Adherence for the Treatment of Methicillin-Resistant Staphylococcus aureus

Skin and Soft Tissue Infections

Carman, Margaret Jean DNP, MSN, RN, CEN, ACNP; Phipps, Julie MSN, RN; Raley, Jennifer MD; Li, Suling PhD; Thornlow, Deirdre PhD, RN, CPHQ

Author Information
Advanced Emergency Nursing Journal: July/September 2011 - Volume 33 - Issue 3 - p 252-266
doi: 10.1097/TME.0b013e31822610d1
  • Free

Abstract

SKIN AND SOFT TISSUE INFECTIONS (SSTIs) attributed to community-acquired methicillin-resistant Staphylococcus aureus (CAMRSA) have been identified as a distinct and emerging phenomenon since the early 1990s. The organism is extremely virulent and is gaining prevalence, increasing from 59% among all SSTIs in 2004 to 79% in 2009 (Shapiro, Raman, Johnson, & Piehl, 2009). These infections typically present as abscess formation and are frequently referred to as a “spider bite.” Costly and significant complications may include recurrent abscess, necrotizing fasciitis, sepsis, and methicillin-resistant Staphylococcus aureus (MRSA) pneumonia, each of which increases the importance of appropriately managing this population. Care should consist of appropriate antibiotic coverage, consistent wound culturing, and educating patients on strategies for infection control (Dellit, 2007) (Figure 1).

Figure 1
Figure 1:
North Carolina consensus guideline for management of suspected community-acquired Staphylococcus aureus (CA-MRSA) skin and soft tissue infections (SSTIs).

Microbial resistance to methicillin was identified within 2 years of its introduction in 1959; MRSA infections were generally considered to be hospital acquired until 1982, when a genetically distinct form of the bacteria was discovered in specific high-risk groups. Persons using illicit drugs, prisoners, and individuals participating in team sports were initially identified as high risk until the mid-1990s when risk was more broadly defined to include persons sharing a living space with multiple household members and those living in close contact with another individual having a history of abscess (Cohen, 2007).

The MRSA infections present challenges to medical providers. Components of care evolve and require new or modified interventions as clinical manifestations of CAMRSA are affected by several factors, including genetic makeup.

For example, organism virulence factors regulated by the mecA gene are distinct from hospital-acquired MRSA (Miller et al., 2007), promoting neutrophil lysis and exhibiting initially as 1–2-mm cutaneous pustules. These pustules frequently develop into large subcutaneous abscesses, prompting patients to present to the emergency department (ED) for definitive care. Given the prevalence of MRSA, it is reasonable to assume that persons presenting with this complaint should be empirically covered for the infection (Gorwitz, 2006).

Incision and drainage has traditionally been recognized as the standard treatment for abscesses; this has not changed (Dellit, 2007). Options for empiric antibiotic therapy to treat abscesses complicated by systemic illness or cellulitis must consider cost, availability on the institution's formulary, and changing resistance patterns. Guidelines for the treatment of SSTI, presumably caused by CAMRSA based on prevalence studies, are available through the Centers for Disease Control and Prevention (CDC; Gorwitz, 2006), the Infectious Diseases Society of Washington (Dellit, 2007), and the North Carolina Statewide Program for Infection Control and Epidemiology (SPICE). The North Carolina Consensus Guideline (NC Consensus Guideline) for Management of Suspected CAMRSA SSTIs is consistent with CDC and Infectious Diseases Society of Washington recommendations and is based on local antibiogram data (SPICE, 2007).

Despite the availability of clinical guidelines specific to the population of interest, medical providers continue to exhibit variation in the care of patients presenting to the ED for outpatient management (Moran et al., 2006), including inconsistently prescribing empiric antibiotic therapy to cover for MRSA, sporadic culturing of wounds requiring drainage, and infrequently educating patients on methods for limiting spread of the infection.

Clinical decision support (CDS) tools have been used to increase provider awareness and utilization of established guidelines in the clinical setting (Christakis, 2001; Garg et al., 2005). For this study, researchers examined whether the development of a CDS tool based on current guidelines for the outpatient management of CAMRSA-related abscesses, and interfaced with the documentation and order-entry process would improve prescriptive practice in an ED setting.

REVIEW OF EVIDENCE

The clinical significance of CAMRSA has drawn increasing attention as the organism has become better defined and clinically evident. Research has focused on characterization and prevalence data unique to the organism (Daum, 2007; Dellit, 2007; Kowalski, Berbari, & Osmon, 2005; Moran, 2006; Sattler, Mason, Jr, & Kaplan, 2002), specific interventions for the care of MRSA SSTIs (Miller et al., 2007; Shapiro et al., 2009; Wells, Mason, Roarty, & Dooley, 2009), and the effects of decontamination techniques to limit colonization and recurrent infection (Ellis et al., 2007; McConeghey, Mikolich, & LaPlante, 2009).

Moran et al. (2006) studied 422 patients with SSTIs from a diverse sample of eleven university-affiliated EDs across major cities in the United States. The MRSA SSTI was found to be significantly more prevalent than any other etiology of SSTI at that time. Recommendations included consistent wound culturing and modifying empiric antibiotic therapies directed at the resistant organism. Demographic data also reinforced the presence of risk factors in patients found to have CAMRSA infection, including recent antibiotic therapy, report of “spider bite” as the causative mechanism, history of MRSA infections, and close contact with patients having SSTIs.

The question of appropriate selection of empiric antibiotic therapy is complex. Medications traditionally effective in treating SSTI have previously responded to β-lactams. By definition, CAMRSA is insensitive to β-lactam antibiotics, whereas SSTIs caused by streptococcus or other atypical organisms are not covered by sulfonamides.

Wells et al. (2009), using a cross-sectional design, examined 751 patients with SSTI for treatment failure, excluding those with abscess formation, surgical intervention, or admission at the time of initial visit. Clinicians evaluating patients with SSTIs were unable to differentiate CAMRSA from those sensitive to β-lactam therapies. This supports the recommendation for universal culturing as well as the potential benefits of a tool to guide empiric treatment of CAMRSA to patients presenting for the complaint of SSTI.

Although both CDC and NC Consensus Guidelines recommend that all abscesses be cultured, no studies examine the value of culturing or whether the information yielded is more helpful for regional surveillance or individual case follow-up. A small study conducted at the institution targeted for this project identified a prevalence rate of 75.9% MRSA (n = 229) isolates in pediatric patients presenting to the health system over a 7-month interval (Shapiro et al., 2009), thus supporting recommendations (Gorwitz, 2006) for universal culturing based on resistance and sensitivity results within the population. Limitations included retrospective review and lack of genotyping to differentiate hospital-acquired infection from CAMRSA, but the study indicated that culturing the wound and identifying risk factors specific to the population of interest might increase the rate of successful treatment.

Miller et al. (2007) studied 280 patients with SSTI, finding that strains of beta-lactam sensitive staphylococcal SSTI's could not be reliably differentiated from those that were resistant, based on clinical examination or epidemiologic risk factors. This, in addition to evolving resistance patterns, further supports the recommendation for universal culturing of wounds requiring surgical drainage.

The availability of antibiogram data to increase provider awareness of local resistance patterns may affect prescriptive practice in the management of CAMRSA SSTIs (McGregor, Dumyati, Casiano-Colón, Chang, & Klevens, 2009). In a questionnaire-guided study of 24 pediatric and family physicians in Monroe County, New York, authors examined changes in provider decision making based on the availability of local prevalence data for MRSA. Although limited by a low response rate, the investigators demonstrated a change in prescriptive behaviors and provider interest in obtaining information for clinical decision making.

Several large, well-designed studies have been conducted to examine the effects of decolonization of MRSA on recurrent abscess formation or transmission to other individuals.

The use of intranasal Bactroban (Mupirocin) was found to be transiently effective in decolonizing subjects but had no long-term effects on the rate of infection or colonization of other individuals (Ellis et al., 2007). Small studies conducted in the inpatient setting indicate that the use of chlorhexidine scrubs may effectively reduce the recurrence of MRSA infections, but no large studies support universal recommendation of specific protocols for decontamination in the outpatient setting. The use of hypochlorite baths has also been studied on a very limited basis with no evidence to support a change in clinical practice (Fisher, Chain, Hair, & Cunnion, 2008).

Both CDC and North Carolina Guidelines for the treatment of MRSA-related skin infections indicate that although it is not practical to recommend decontamination of all patients, it is reasonable to assess risk and make recommendations for patients at high risk for recurrence (McConeghey et al., 2009). “Skin antisepsis with chlorhexidine or other agents may be used,” but routine decolonization is not recommended for those at low risk.

In a systematic review of CDS applications for improving guideline adherence and clinical outcomes, Garg et al. (2005) found that the use of automated prompts significantly improved provider performance. The authors reported a 64% improvement in diagnosis, preventive care, disease management, and drug dosing and prescribing.

Christakis (2001) looked more specifically at the use of CDS tools to improve guideline adherence in prescribing antibiotics for otitis media by using automated prompts. This was a well-designed, randomized controlled trial closely paralleling the goals of the current project. Use of CDS prompts improved prescriptive practice directed at the duration of therapy by 10% (p < 0.01), whereas the intervention showed a trend in the increase of treatment with antibiotics.

In summary, the North Carolina Guideline for Management of Suspected Community-Acquired methicillin-resistant Staphylococcus SSTIs (NC Consensus Guideline) is consistent with current evidence supported in the literature. The guideline reflects local prevalence and sensitivity data and aligns with guidelines proposed by the CDC and Infectious Diseases Society of America. We hypothesize that implementation of a CDS tool will promote adherence with the NC Consensus Guideline and improve outcomes for patients presenting to the ED for SSTIs with a suspected etiology of MRSA.

PROJECT DESIGN

Purpose

The intent of this project was to determine whether implementation of an automated response tool embedded into the electronic medical record system at a large, urban, university-affiliated hospital located in North Carolina would improve provider adherence to the NC Consensus Guideline in the treatment of patients presenting to the ED with subcutaneous abscess formation. Indicators of guideline adherence include prescription of empiric antibiotic therapies consistent with local prevalence data, universal wound culturing, and assessment of patients for risk of recurrent infection and decontamination.

Setting

The health system targeted includes one level-one trauma center with separate adult and pediatric EDs, a smaller community hospital ED, and two freestanding EDs. All sites are medically staffed by a private group practice, with 71 board-certified emergency physicians and 30 mid-level providers. More than 200,000 patients received care across the various EDs in 2009.

HMED (Allscripts, Chicago, IL) has been the electronic documentation system used in the organization's EDs for the past 10 years. The new 6.3 version of HMED software was released in December 2009. Implementation of the tool was delayed to allow a period of user adjustment to the new system, and this project served as the inaugural use of many HMED 6.3 features, which allowed customization and automated prompts to guide provider decision making.

Design

The project was conducted using a quasiexperimental, pretest–posttest design with convenience sampling. We received investigational review board exemption from the health system where the project was implemented, as well as the university overseeing the clinical project.

Data extracted from the electronic medical record (EMR) included antibiotics prescribed, wounds cultured, and prescription for chlorhexidine scrub as a marker for risk assessment. These variables were measured at 6 and 12 weeks after the introduction of the CDS tool. Provider demographic characteristics were obtained through survey responses. SPSS version 18.0 analysis (Chicago, IL) was used for the statistical analysis.

Baseline data were obtained for a 6-month period 1 year before implementation, with a 6-month hiatus to allow for a more smooth transition after the introduction of the new EMR. Subsequent measurements were taken at 6 and 12 weeks postimplementation.

MEASURES

Patient Data

All patients presenting to the EDs with a chief complaint of abscess during the study period were included in the sampling frame. Patients with animal bites and oral or dental infections were excluded from the study. All personal patient identifiers were removed before compiling Excel spreadsheets for analysis. Patient demographics included age and gender.

Guideline Adherence

The CDS tool included a history and physical examination (H&P) template, abscess order-entry form (Figure 2), and prebuilt discharge pathway to guide prescriptions, follow-up, and discharge instructions specific to the care of MRSA abscesses. When using the tool, providers are queried regarding presence of cellulitis, size of abscess, and systemic illness to determine whether antibiotic coverage would be appropriate. If none of the criteria are met, an automated message indicates that incision and drainage would be sufficient per the guideline. If a “yes” response is provided, automated prompts are generated to suggest empiric antibiotic coverage for MRSA.

Figure 2
Figure 2:
WakeMed hospital adult abscess order entry form.

The anatomical sites for wound culturing are included in the tool, thus facilitating ease in the order-entry process. This eliminates the need for the provider to enter three different screens to order the test.

At the time of disposition, an automated prompt appears on all charts, indicating a diagnosis of abscess, querying the provider to indicate risk factors for recurrent MRSA abscess. Cases in which risk are identified prompt the provider with a recommendation to prescribe chlorhexidine scrubs at discharge.

A time series analysis using individual regression was used to examine changes in guideline adherence and targeted MRSA therapy over time (Table 1). The data were then collated, and chi-square analysis were conducted to determine whether changes in adherence were directly related to use of the tool (Tables 2 and 3).

Table 1
Table 1:
Logistic regression examining changes in guideline adherence over time, with consideration of age and gender as confounding variables for adherence (n = 873, p < .05, df = 1)a
Table 2
Table 2:
t Tests for interval-level data
Table 3
Table 3:
Chi-square analysis for categorical data

Provider Awareness

We used the CDC-Rochester Survey to examine provider awareness of local prevalence and prescriptive practices. The survey was released to providers via Survey Monkey (Palo Alto, CA) during a 1-week period immediately before implementation and at the 6- and 12-week data collection points. The survey was administered anonymously through the office staff, and the lead investigator was blinded to all identifying information.

On the basis of responses from providers indicating that there was some attribution of frustration to the tool, rather than with the new EMR system features, we added five questions from the Questionnaire for User Interface Satisfaction (QUIS) survey to obtain more focused feedback from the providers regarding the use of the decision support application, rather than the overall system functions.

Independent t tests were used to compare provider responses to the Rochester Survey between baseline and 6-week surveys; low response rates at week 12 precluded the researchers ability to analyze these data. The researchers included descriptive analysis of the QUIS portion of the survey (Tables 4 and 5). Provider characteristics were compared by using paired t tests for each of the three time intervals.

Table 4
Table 4:
Descriptive statistics for Questionnaire for User Interface Satisfaction survey responses, week 6 (n = 16)
Table 5
Table 5:
Descriptive statistics for Questionnaire for User Interface Satisfaction survey responses, week 12 (n = 6)

Project Adoption

Use of the form was key in determining the impact of the CDS tool on care of patients presenting to the ED with abscess.

Although adherence to the guideline across time intervals was the key indicator of success, providers had the option to opt out from using one or all forms. Frequency data were obtained to determine the rate and adoption of tool use.

Implementation

Project information was disseminated through e-mail and verbal presentation at the department meeting immediately preceding implementation. Any pertinent suggestions recommended by related services, including general surgery and pediatric infectious disease, were made with minor revisions to the CDS tool.

An Internet-based informational presentation with visual screen shots and demonstration of the tool was recorded and e-mailed to all providers 3 days before implementation. Questions were directed to the primary investigator, who remained on call by cell phone and e-mail through the duration of the project. The ED technological support systems remained in place on a 24-hour basis, and systems issues regarding the tool were referred back to the administrator, who served on the project committee. Providers were encouraged to send comments and recommendations for modification of the tool to improve user friendliness and adoption.

RESULTS

The final sample size included 873 patient encounters.

Appropriate antibiotic coverage for patients with MRSA increased significantly during the course of this project, increasing from 86.8% at baseline to 96.7% at 12 weeks, with the greatest increase during the initial 6-week interval. Providers were 2.41 times more likely to order empiric therapy appropriate for the treatment of MRSA for each 6-week period in the course of the study.

The rate of orders for wound culture decreased from baseline to the 6-week interval by 31%. There was a sharp recovery at 12 weeks (Figure 2), but in general, the project appears to have had a negative effect on culture surveillance efforts.

Prescriptions for chlorhexidine increased significantly, with an odds ratio of 2.48 (p < .001) over each 6-week interval. This indicates not only that providers had assessed risk of recurrence but also that there was an increased awareness of the need to inform and educate patients about infection control measures. Neither age nor gender significantly impacted guideline adherence over the course of this study (Table 1).

Participation in the provider survey was limited. With more than 100 potential respondents, approximately 20% responded at the baseline and at the 6-week intervals, yet only 6% responded at 12 weeks. Analysis to compare provider characteristics demonstrated no significant differences between those responding across time intervals in terms of age, gender, years in practice, years practicing in the state of North Carolina, or years practicing within the physician group (Tables 7 and 8). Demographic data collected via the survey cannot be correlated with adherence data, given the limited provider sample size. Responses for the QUIS demonstrated a wide range across the nine-point Likert-type scale on each of the indicators, with a more narrow range in responses at the 12-week interval. Generally, the mean scores improved slightly, meaning that over time, providers experienced less frustration with the tool.

Per the Rochester Survey, no significant change was found in provider perspectives on MRSA prevalence, recommendations for universal wound culturing, or efforts to assess and educate patients regarding decontamination techniques to possibly contain MRSA infections. Findings do indicate that the providers would consider using a CDS tool for selection of appropriate antibiotic options.

Table 6
Table 6:
Relationship between guideline adherence and form use (n = 538)
Table 7
Table 7:
Differences in provider demographic characteristics for Rochester survey responses at baseline (Group 1, n = 20) and 6 weeks (Group 2, n = 16), df = 15
Table 8
Table 8:
Differences in provider demographic characteristics for Rochester Survey responses at 6 (Group 1, n = 16) and 12 weeks (Group 2, n = 6, df = 5)

Although provider adherence improved dramatically for two of the three outcomes measured over the course of the study, use of the CDS tool significantly impacted only one of the three outcome measures: prescription of chlorhexidine for decontamination purposes (Table 6). Although provider performance improved regarding choice of antibiotic agent, this did not correlate with CDS utilization. There was also no direct relationship in the use of the tool and the decreased rate of wound cultures. Use of the form declined over the course of the project (Figure 3).

Figure 3
Figure 3:
Graphic representation of changes in wound culture order entry over time.

DISCUSSION

Providers were more than twice as likely (odds ratio, 2.41; p = 0.000) to prescribe antibiotic coverage appropriate for the treatment of MRSA at the conclusion of the project. It is unclear what prompted the improvement, given low utilization of the CDS tool and an insignificant relationship between tool use and guideline adherence. The improvement may have resulted from a heightened awareness of the guideline due to the dissemination process, or we may simply have experienced a Hawthorne effect as the providers were aware of the study. It is encouraging that patient care benefited from this initiative; further refinement of the tool may lead to increased efficiency in the order-entry process and sustained improvement in practice.

Some sentiment existed among providers that culturing was costly and unnecessary, given the high prevalence of MRSA, and that all patients should be empirically covered. Discussion of increasing resistance to clindamycin and emerging resistance to Bactrim may improve awareness of the importance of surveillance. There is room for improvement in this aspect of the project, and with further provider education and refinement of the tool, we may reach our goal of a 100% culture rate.

One clinical issue that was not considered during the project was the clarification of the use of specific antibiotic considerations, given the location of the abscess. Although the researchers excluded oral abscesses or those resulting from animal bites, it is possible that providers deviated from the guideline based on concerns that the antibiotics recommended would not treat wounds prone to non-MRSA infection. Adding more site-specific cultures on the order-entry form may help to pinpoint those cases in which the guideline may not provide comprehensive options for care. This should be tracked by a review of both culture results by anatomical site and the number of abscesses requiring a change in antibiotic therapy based on those results.

Identification of patients at risk for recurrent abscess, and patient education in terms of infection control, improved as represented by the increase in chlorhexidine prescriptions from zero at baseline to 7.3% at the conclusion of the study. This again indicates that providers were more than twice as likely (odds ratio, 2.48; p < 0.001) to assess their patients' risk for recurrence over the course of the project. Form use was significantly associated with an improvement in guideline adherence for this particular recommendation.

Although the evidence does not support routine decolonization of all patients to prevent future CAMRSA infections, ambiguity persists regarding the use of antiseptic solutions to limit contagion. Inclusion of chlorhexidine in the NC Consensus Guideline as an option for patients at high risk may have indicated to providers that this is a safe component of care. Chlorhexidine is now included on several “four dollar lists” at local pharmacies; this is indicated on the discharge pathway and may have affected prescriptive decision making. Opportunity exists for further study on the effectiveness of patient education techniques, use of various antiseptic agents, and long-term surveillance to determine whether chlorhexidine use will decrease transmission and recurrence.

Although providers inconsistently used the CDS tool over the course of the study, they indicated that they would consider a CDS tool to assist with appropriate antibiotic selection. Indeed, provider selection of appropriate antibiotics improved throughout this study, as did provider assessment of risk for recurrence. The tool lacked sufficient automaticity, requiring that providers choose the tools over the more generic default template. We believe that automating form use by linking the CDS tool to the chief complaint would be one way to improve provider use of the form.

Although the tool theoretically condensed the workload from three windows into one, providers had to consciously select it. Improved usability and integration of the tool into the natural workflow may improve adoption.

Providers also indicated some willingness to order cultures when caring for their patients with abscess, as recommended in the guideline. Use of automated orders for culture, requiring that the provider enter only anatomic location, may improve the rate of culture surveillance. In addition, provider use of alternative antibiotics in addition to or in place of recommended therapies for MRSA could be tracked for correlation with patient outcomes.

Addition of the QUIS survey at 6 weeks helped to identify factors that might have made the tool more user friendly in terms of spatial presentation, automaticity, and content. The QUIS questions were condensed into five general questions to gauge user perceptions regarding the tool's convenience and usefulness. Information gained through the QUIS survey would be helpful in guiding future modifications. We recommend intermittently obtaining provider feedback by using the QUIS questions to tailor tool features to the needs of the providers.

And finally, providers remain concerned about adopting a tool that did not allow them to exercise judgment based on their medical expertise. Although modifications should include automated use of the tool for clinical decision making, it is equally important for providers to be able to customize generic order sets to address the needs of individual cases. The researchers will track deviations from the guideline to determine whether these are justifiable, based on clinical findings, and to determine whether additional items should be added to the order-entry form.

Limitations

The researchers were limited in their ability to examine changes in individual provider behavior over time, as provider demographics were collected via survey data rather than through manual or electronic review. Future studies should link provider practice patterns with guideline adherence.

Adoption of the CDS tool appears to have been limited by several factors. The timing of the intervention, given recent deployment of the new EMR system, appears to have created confusion between systems issues and user friendliness of the tool. Providers also reported being overwhelmed with information regarding the features of the new EMR, in addition to learning how to use the CDS tool. Although new features of HMED 6.3 allowed for the creation of the tool, periods of downtime, system crashes, and disruptions in communication led to significant end-user frustration. We will improve the automaticity and appearance of the order-entry forms, and once the new EMR system is more established, the researchers will reintroduce and reevaluate the usefulness of the CDS tool.

Evidence to support the use of CDS tools indicates that automated prompts increase successful implementation, yet the barriers to adoption of the CDS tool parallel those mentioned by Garg et al. (2005). Although the design of the form simplified order entry from the use of three screens into one, the provider had to select the form rather than enter into the generic H&P format. In addition, the generic H&P option was displaced on the screen by the CDS tool, which led to some frustration in instances when the chief complaint was not abscess. Willingness to try a new method of documentation may have been impacted by each of these factors.

Communication overload may have affected provider survey response rates. E-mail is the common communication for this provider practice; it was assumed that e-mail would be the optimal method for reaching the group, yet with daily updates on various issues, the volume of mail became overwhelming, and many providers admittedly did not read all transmissions. In the future, using alternative and complementary modes of communication may be helpful. Coordinating care recommendations sent to the group may improve consistency in practice and decrease confusion as well.

CONCLUSION

This project explored the feasibility of embedding a CDS tool into the researchers electronic medical record. The goal was to improve provider adherence to clinical guidelines, to facilitate appropriate documentation, and improve patient outcomes.

The project was associated with improved prescription rates for MRSA-appropriate antibiotics and increased recommendations for the use of chlorhexidine in high-risk patients; yet, the rate for culturing abscess decreased significantly during the time in which the study was conducted. Providers indicate that they are willing to use tools that support their clinical decision making and ease their documentation workload. Given the significant relationship between use of the tool and adherence to recommendations for infection control, future efforts to improve care through the development of CDS tools are promising.

Further study is needed to examine what features of the tool are barriers to adoption for mainstream use and how these may be corrected. Providers should be engaged in this process, and alternative methods for disseminating information about the tool should be considered.

The assessment of organizational readiness for this change had focused largely on the providers, yet the final product may have improved if technical support staff had been given additional time to become acclimated to the new EMR features. Pop-up menus and automated links between forms had not been available in previous versions of the EMR. This project was the first effort to use the new features in assisting with CDS, and expertise in designing the tools will improve with familiarity. Future efforts should include assessment of readiness not only for adopters but also for those assisting in the creation of the tool.

As an academic practice, this physician group is accustomed to intellectual debate and individual interpretation of the available evidence. This includes appropriate management of SSTIs suspicious for CAMRSA infection. Discussion continues regarding the benefit of treating patients with systemic antibiotics versus only incision and drainage, the cost-to-benefit ratio of culturing purulent wounds, and the lack of clear-cut evidence to support the use of chlorhexidine in patients with suspected CAMRSA infections. Physicians have cited discrepancies between various guidelines. Culturally, the group was unreceptive to a tool that limited their ability to individualize clinical decisions. Further development of the tool must allow providers to tailor care according to their professional beliefs while utilizing the evidence-based guideline as a resource. Deviations from guidelines will be fedback to clinicians as part of an ongoing quality improvement process.

Next steps include refining the current tool to increase user friendliness in terms of automaticity, overall appearance, and consideration of user needs in the clinical work setting. We intend to take the lessons learned from this project and apply them when developing future CDS tools that facilitate evidence-based care for other specified conditions.

REFERENCES

SPICE N.C. (2007). North Carolina Consensus Guideline for Management of Suspected Community-Acquired Staphylococcus aureus (CA-MRSA) Skin and Soft Tissue Infections (SSTI's). Retrieved September 6, 2010, from North Carolina Division of Public Health: http://www.unc.edu/depts/spice/CA-MRSA-2007-03-05.pdf
Christakis D. (2001). A randomized controlled trial of point-of-care evidence to improve the antibiotic prescribing practices for otitis media in children. Retrieved September 20, 2009, from http://www.pediatrics.org/cgi/content/full/107/2/e15
Cohen P. (2007). Community-acquired methicillin-resistant Staphylococcus aureus skin infections. American Journal of Clinical Dermatology, 8 (5), 259–270.
Daum R. (2007). Skin and soft tissue infections caused by methicillin-resistant Staphylococcus aureus. New England Journal of Medicine, 357, 380–390.
Dellit T. D. (2007). Guidelines for evaluation and management of community-associated methicillin-resistant Staphylococcus aureus skin and soft tissue infections in outpatient settings. Seattle and King County, WA: Infectious Diseases Society of Washington.
Ellis M. G., Griffith M. E., Dooley D. P., McLean J. C., Jorgensen J. H., Patterson J. E., Hospenthal D. R. (2007). Targeted Intranasal Mupirocin to Prevent Colonization and Infection by Community-Associated Methicilllin-Resistant Staphylococcus aureus Strains in Soldiers: A Cluster Randomized Controlled Trial. Antimicrobial Agents and Chemotherapy, 51, 3591–3598.
Fisher R., Chain R. L., Hair P. S., Cunnion K. M. (2008). Hypochlorite killing of community-acquired methicillin-resistant Staphylococcus aureus. Pediatric Infectious Disease Journal, 27, 934–935.
Garg A., Adhikari N. K., McDonald H., Rosas-Arellano M. P., Devereaux P. J., Beyene J., Haynes R. B. (2005). Effects of computerized clinical decision support systems on practitioner performance and patient outcomes: A systematic review. JAMA, 293, 1223–1238.
Gorwitz R. J. (2006). Strategies for clinical management of MRSA in the community: Summary of an experts' meeting convened by the Centers for Diease Control and Prevention. Washington, DC: Centers for Disease Control and Prevention.
Kowalski T. B., Berbari E. F., Osmon D. R. (2005). Epidemiology, treatment, and prevention of community-acquired methicillin-resistant Staphylococcus aureus infections. Mayo Clinic Proceedings, 80, 1201–1208.
McConeghey K., Mikolich D. J., LaPlante K. L. (2009). Agents for the decolonization of methicillin-resistant Staphylococcus aures. Pharmacotherapy, 29, 263–280.
McGregor J. D. C., Dumyati G., Casiano-Colón A. E., Chang P. J., Klevens R. M. (2009). Usefulness of antibiogram surveillance for methicillin-resistant Staphylococcus aureus in outpatient pediatric populations. Diagnostic Microbiology and Infectious Disease, 64, 70–75.
Miller L. P. R., Perdreau-Remington F., Bayer A. S., Diep B., Tan N., Bharadwa K., Spellberg L. (2007). Clinical and epidemiological characteristics cannot distinguish community-aquired methicillin-resistant nStaphylococcus infection from methicillin-susceptible S. aureus infection: a prospective investigation. Clinical Infectious Diseases, 44, 471–482.
Moran G. K., Krishnadasan A., Gorwitz R. J., Fosheim G. E., McDougal L. K., Carey R. B. Emergency ID Net Study Group (2006). Methicillin-resistant S. aureus infections among patients in the emergency department. New England Journal of Medicine, 355, 666–674.
Sattler C. M., Mason E. O. Jr, Kaplan S. L. (2002). Prospective comparison of risk factors and demographic and clinical characteristics of community-acquired methicillin-resistant versus methicillin-susceptible Staphylococcus infection in children. Pediatric Infectious Disease Journal, 21, 910–917.
Shapiro A. R., Raman S., Johnson M., Piehl M. (2009). Community-acquired MRSA infections in North Carolina children: Prevalence, antibiotic sensitivities, and risk factors. North Carolina Medical Journal, 70, 102–107.
Wells R.M., Mason P., Roarty J., Dooley M. (2009). Comparison of initial antibiotic choice and treatment of cellulitis in the pre- and post-community-acquired methicillin-resistant Staphylococcus aureus eras. American Journal of Emergency Medicine, 27, 436–439.
Keywords:

abscess; clinical decision support; methicillin-resistant Staphylococcus aureus; MRSA

© 2011 Lippincott Williams & Wilkins, Inc.