Use of the DMAIC Lean Six Sigma quality improvement framework to improve beta-lactam antibiotic adequacy in the critically ill

Abstract

Beta-lactam antibiotics are widely used in the intensive care unit due to their favorable effectiveness and safety profiles. Beta-lactams given to patients with sepsis must be delivered as soon as possible after infection recognition (early), treat the suspected organism (appropriate), and be administered at a dose that eradicates the infection (adequate). Early and appropriate antibiotic delivery occurs in >90% of patients, but less than half of patients with sepsis achieve adequate antibiotic exposure. This project aimed to address this quality gap and improve beta-lactam adequacy using the Define, Measure, Analyze, Improve, and Control Lean Six Sigma quality improvement framework. A multidisciplinary steering committee was formed, which completed a stakeholder analysis to define the gap in practice. An Ishikawa cause and effect (Fishbone) diagram was used to identify the root causes and an impact/effort grid facilitated prioritization of interventions. An intervention that included bundled education with the use of therapeutic drug monitoring (TDM; i.e. drug-level testing) was projected to have the highest impact relative to the amount of effort and selected to address beta-lactam inadequacy in the critically ill. The education and TDM intervention were deployed through a Plan, Do, Study, Act cycle. In the 3 months after “go-live,” 54 episodes of beta-lactam TDM occurred in 41 unique intensive care unit patients. The primary quality metric of beta-lactam adequacy was achieved in 94% of individuals after the intervention. Ninety-four percent of clinicians gauged the education provided as sufficient. The primary counterbalance of antimicrobial days of therapy, a core antimicrobial stewardship metric, was unchanged over time (favorable result; P = .73). Application of the Define, Measure, Analyze, Improve, and Control Lean Six Sigma quality improvement framework effectively improved beta-lactam adequacy in critically ill patients. The approach taken in this quality improvement project is widely generalizable to other drugs, drug classes, or settings to increase the adequacy of drug exposure.

Introduction

Antibiotic therapy is lifesaving for critically ill patients with infections. To be effective, antibiotics must be initiated as soon as an infection is suspected (referred to as “early”) to prevent the development of life-threatening organ dysfunction (i.e. sepsis) [1]. In addition to timeliness, antibiotic selection must effectively target the suspected or documented pathogen(s) (referred to as “appropriateness”) and be delivered at a dose that achieves adequate drug concentrations at the site of infection (referred to as “adequacy”). Existing reports suggest that in >90% of cases, antibiotics given to patients with sepsis meet the requirements for early and appropriate delivery [2, 3]. Considerably fewer patients achieve adequate antibiotic therapy due to pharmacokinetic/pharmacodynamic derangements in critical illness [4].

Beta-lactam antibiotics are used ubiquitously in the intensive care unit (ICU) for the treatment of infections due to their favorable effectiveness and safety profiles [5]. We recently demonstrated that only 30% of critically ill patients reached the target concentrations of cefepime, a broad-spectrum beta-lactam antibiotic, in the first 24 h of sepsis (the “golden hours”) [6]. Similarly inadequate concentrations of piperacillin/tazobactam have been found in the first 24 h of sepsis [7]. There is a need to address the quality gap associated with beta-lactam adequacy in the critically ill. A quality framework methodology like Lean Six Sigma could be used to systematically evaluate beta-lactam adequacy locally and take actions needed to address gaps in quality in the care of critically ill patients.

Lean Six Sigma is a performance improvement methodology that has demonstrated effectiveness in healthcare settings [8, 9]. Lean Six Sigma aims to reduce waste and implement lean thinking [9, 10]. The Define, Measure, Analyze, Improve, and Control (DMAIC) process is among the most used Lean Six Sigma models for systematically and strategically achieving improved performance [8, 10]. DMAIC begins with defining a project’s goals and proceeds with identifying the root causes of a problem. Solutions are then considered, implemented, and evaluated for effectiveness [8]. The Lean Six Sigma DMAIC framework has been applied to initiatives aiming to improve infection control [11–13], duration of antibiotic therapy in pneumonia [14], and time to antibiotic therapy in pediatric sepsis [15]. One study evaluated its role for improving dosing and monitoring of gentamicin, a narrow therapeutic window aminoglycoside antimicrobial [16]. The present quality initiative builds on these existing data and leverages the DMAIC framework to address the gap in quality observed with beta-lactam adequacy.

The overall objective of the project was to improve beta-lactam adequacy in critically ill patients using the DMAIC model as part of the Lean Six Sigma methodology. We aimed to improve beta-lactam adequacy without adversely impacting antimicrobial days of therapy (DOT), an antimicrobial stewardship metric that characterizes antibiotic drug usage.

Methods

Study context and design

Mayo Clinic in Rochester, MN, is an academic medical center with approximately 200 adult ICU beds and 15 000 ICU admissions annually. During the quality improvement project period from 2022 to 2023, beta-lactam antibiotic selection, dose, and monitoring were determined by a multidisciplinary critical care team inclusive of clinical pharmacists. Antibiotic dose decisions were guided by general clinical decision support (e.g. renal dosing recommendations) and a standardized institutional antimicrobial dosing guide [17].

To improve beta-lactam adequacy, a multidisciplinary steering committee was formed, which included ICU pharmacists, infectious diseases (ID) pharmacists, and ID physicians. The steering committee oversaw and implemented process improvements in the ICU setting. The Lean Six Sigma DMAIC framework was followed [8, 10]. The Mayo Clinic Institutional Review Board approved the project with a waiver of informed consent (IRB 22–004273, 22–002300), and the findings were reported in accordance with the SQUIRE statement [18].

Define

The steering committee completed a stakeholder analysis with the staff involved in beta-lactam utilization. Pharmacists were identified as key stakeholders [19], with physicians, advanced practice providers, trainees, nurses, and laboratory personnel also having a role. Information collected from stakeholders and published literature was used to define the gap in practice. Project scope and target patient population were identified. Although published guidelines favor universal efforts to improve beta-lactam adequacy in all critically ill patients [1, 20, 21], there are both clinical and practical reasons that limit the viability and utility of this approach. At Mayo Clinic, to manage the scope of the quality improvement project, the steering committee focused on select patients with profound pharmacokinetic/pharmacodynamic derangements, including those treated with extracorporeal membrane oxygenation (ECMO), kidney replacement therapy, or at extremes of weight (weight >120 kg or <40 kg; body mass index <18.5 kg/m² or >40 kg/m²) [22–24]. The scope involved the three most used beta-lactam antibiotics in the critical care setting, cefepime, piperacillin/tazobactam, and meropenem. These antibiotics are commonly prescribed in the ICU because of their general tolerability and broad spectrum of activity.

Measure

The primary outcome measure, beta-lactam adequacy, was defined as the proportion of patients that achieved desired beta-lactam pharmacokinetic/pharmacodynamic targets (≥8 mg/l for cefepime, ≥16 mg/l for piperacillin, and ≥2 mg/l for meropenem) [25]. Achievement of these targets has been associated with a greater likelihood of treatment success without the need for elongated antibiotic courses, or the occurrence of complications like longer lengths of stay or mortality [26]. Pharmacokinetic data from Mayo Clinic indicated that only 30% of ICU patients treated with these broad-spectrum beta-lactams achieved these drug concentrations in the first 24 h of critical illness [6]. In the published literature, between 10% and 45% of patients achieve adequate beta-lactam concentrations in the first 24 h of critical illness, and the proportion increases with longer durations of therapy [23, 27]. Importantly, these baseline estimates for the primary quality improvement measure were the best available at the time but were not directly reflective of data from the select patient population chosen by the steering committee to target (i.e. ECMO, kidney replacement therapy, or at extremes of weight). We secondarily considered overall target attainment, which balances effectiveness with safety. Target attainment was achieved when levels were between 8 and 50 mg/l for cefepime, 16 and 150 mg/l for piperacillin, and 2 and 30 mg/l for meropenem. Beta-lactam target attainment estimated from the published literature in the absence of local data for this specific metric ranges from 25% to 75% [28].

The primary balancing measure selected was DOT per 1000 patient-days, an antimicrobial stewardship metric that characterizes drug usage, for cefepime, meropenem, and piperacillin–tazobactam. The baseline for the balancing measure was 279 DOT/1000 patient-days for the three drugs monthly averaged from March through May 2022 (N = 527). Antimicrobial DOT is a useful measure to track antimicrobial stewardship efforts and a quality metric to assess if the solution (undesirably) increased overall usage [29].

Analyze

Potential contributors to suboptimal use and inadequacy of the three beta-lactams of interest in the critically ill were identified through the use of a structured Ishikawa cause and effect (Fishbone) diagram completed by subject matter experts under the categories of process, people, resources, equipment, policy, and skill-related causes (Fig. 1).

Figure 1.

Ishikawa cause and effect (Fishbone) diagram. An Ishikawa cause and effect (Fishbone) diagram was used to identify the root causes contributing to the gap in quality, suboptimal use of beta-lactams in the critically ill, as part of the “Analyze Phase” of the Lean Six Sigma DMAIC framework. Abbreviation: ICU, intensive care unit.

Selecting an intervention

From the identified causes from the Fishbone diagram, interventions were proposed to address the gap in quality. An impact/effort grid was developed for prioritization. Three subject-matter experts independently scored and ranked the impact and effort required to address the cause(s) on a scale of 1 (minimal) to 10 (high). These scores were then averaged and plotted on an impact/effort grid. An individualized medicine approach that included a bundled educational intervention and use of therapeutic drug monitoring (TDM; i.e. drug-level testing) was projected to have the highest impact on addressing the quality gap relative to the amount of effort needed to implement the program. Clinician education coupled with beta-lactam TDM has the potential to improve the timely administration of appropriate and adequate beta-lactam antibiotics [28].

Improve

To understand the impact of the proposed quality improvement plan, key stakeholders were surveyed, including pharmacists, physicians, advanced practice providers, trainees, nurses, and laboratory personnel. Baseline familiarity with key concepts regarding beta-lactam utilization was queried, along with perceptions about the implications of beta-lactam TDM implementation. Specific survey items directed at pharmacists probed operational considerations and barriers to implementation due to their prominent role in the process [19].

All necessary organizational and departmental approvals were gathered prior to implementation. The multimodal educational program included print materials, in-person lectures, and an asynchronous web-based eLearning module with knowledge checks and applied case examples. The repetitive exposure to key concepts in the suite of educational resources was designed to facilitate individual internalization of beta-lactam utilization key concepts. Individual internalization, or the work clinicians do to understand, agree to, and adopt a new practice, was identified in our formative work as a key factor in beta-lactam optimization efforts [19].

The process of developing the TDM program has been described in detail [24]. In brief, institutional support was secured for the development and validation of cefepime, piperacillin, and meropenem assays that could be processed in our institutional laboratory [30]. These tests were performed once daily 5 days per week (Monday through Friday). A process map was adapted from existing workflows for other antibiotics (e.g. vancomycin) to delineate key steps and stakeholder responsibilities (Fig. 2). Briefly, patients selected as preferred candidates for beta-lactam TDM underwent two-sample monitoring (peak/trough) at steady state to establish an individual pharmacokinetic profile. Tests were performed once daily, 5 days per week; thus, the turnaround time was between 12 and 24 hours for most samples. TDM results were analyzed using developed clinical tools, including an Excel-based institutional calculator. In cases of nontarget levels, the calculated pharmacokinetic parameters were used to tailor the drug therapy regimen. Documentation templates were provided to end users to report results and future monitoring needs [24]. Each ICU identified a “Pharmacy Clinical Champion” for the project, who was a pharmacist who received additional training in the process and served as a practice expert for other end users.

Figure 2.

Beta-lactam TDM process map. A beta-lactam TDM process map was developed to visually depict each stakeholder action involved and delineate the responsibilities of ordering levels, collecting blood samples, interpreting results, and performing dose adjustments as part of the “Improve Phase” of the Lean Six Sigma DMAIC framework. Abbreviations: APP, advanced practice provider; BL TDM, beta-lactam TDM; EMR, electronic medical record.

One month in advance of “go-live,” beta-lactam education was deployed to all end users. Clinical pearls and additional resources were made available approximately every 2 months after “go-live.” The beta-lactam TDM program commenced in June 2022 and was implemented in a phased fashion over 6 months. Each phase of implementation expanded patient candidacy and used objective criteria to identify patients eligible for beta-lactam TDM at that specific time. A structured stepwise implementation process was selected to streamline identification of TDM candidates and limit the clinical and laboratory burden of a new process. Continuous improvement throughout the implementation occurred in response to end-user feedback.

End-user feedback using a survey similar to the baseline evaluation was evaluated at 1 month after “go-live.” Beta-lactam adequacy, overall target attainment, and DOT/1000 patient-days were evaluated at baseline and 3 months after “go-live.” Statistical analysis was performed in BlueSky R (V10.3). Two-way comparisons of proportions were evaluated with the Chi-squared test, where a P < .05 was considered to be statistically significant.

Results

In the 3 months after “go-live,” 54 episodes of beta-lactam TDM occurred (49 initial TDM occasions and five repeat TDM evaluations during the same antibiotic course) in 41 unique patients (Table 1). The 49 initial TDM occasions were used for summarizing results. The indications for TDM were extremes of weight (N = 14; 28%), ECMO (N = 13; 27%), and continuous renal replacement therapy (N = 22; 45%). The ID service was consulted in 27 (56%) cases. The antibiotics monitored included cefepime (N = 14; 29%), piperacillin–tazobactam (N = 21; 43%), and meropenem (N = 14; 29%).

Table 1.

Characteristics of unique patients with beta-lactam TDM.

Characteristic	Patients (N = 41) N (%)a
Age at ICU admit, years	57 ± 18
Sex, female	22 (54)
Race
White	30 (73)
Black or African American	2 (5)
Asian	2 (5)
Other	5 (12)
Unknown	2 (5)
Weight (kg) at initiation of first beta-lactam antibiotic course of interest	108 ± 50
BMI, kg/m2 at initiation of first beta-lactam antibiotic course of interest	38 ± 16
BMI >30 kg/m2	27 (66)
Hospitalization parameters
Time to ICU admission from hospital admission, days; median [interquartile range (IQR)]	0.5 (0, 2)
Hospital length of stay, days; median (IQR)d	48 (26, 60)
ICU admitting diagnosis type
Non-operative (N = 21)	21 (51)
Cardiovascular	5 (12)
Genitourinary	0 (0)
Respiratory	10 (24)
Gastrointestinal	1 (2)
Other	5 (12)
Operative (N = 20)	20 (49)
Vascular	2 (5)
Cardiac or thoracic	14 (34)
Gastrointestinal	0 (0)
Trauma	2 (5)
Other	2 (5)
Charlson Comorbidity Index	4 ± 4
APACHE III scoreb	119 ± 35
SOFA scorec	12 ± 4
Survival to discharged	28 (74)
Readmission within 30 dayse	7 (25)

BMI, body mass index; APACHE III score, Acute Physiology and Chronic Health Evaluation score; SOFA score, Sequential Organ Failure Analysis score.

^aData reported as mean ± standard deviation (SD) or N (%) unless otherwise specified.

^bAvailable for 24 patients; calculated using parameters within the first 24 h of ICU admission.

^cAvailable for 39 patients; calculated using parameters within the first 24 h of ICU admission.

^dThree patients were still hospitalized at the time of analysis; therefore, the denominator for this variable is 38 unique patients.

^eAmong the 28 patients who were discharged from the hospital alive within the evaluation time frame.

By 3 months, the primary quality outcome of beta-lactam adequacy (≥8 mg/l for cefepime, ≥16 mg/l for piperacillin, and ≥2 mg/l for meropenem) was 94%. Although it was difficult to estimate the baseline measure for this specific population, the observed adequacy was significantly better than the estimated plausible baseline values taken from the literature, which were 30% (P < .001), 45% (P < .001), or 75% (P = .01) [6, 23, 27, 28]. In the secondary evaluation of the overall target attainment, which balances effectiveness with safety (cefepime 8–50 mg/l, piperacillin 16–150 mg/l, and meropenem 2–30 mg/l), beta-lactam concentrations were within the target in 36 cases (73%). Approximately 40% of the patients experienced a dose adjustment in response to beta-lactam TDM, of which three-fourths were de-intensification (dose decrease or interval elongation). There was no clinical evidence of beta-lactam neurotoxicity observed (e.g. drug-associated encephalopathy, seizures, and myoclonus) [31], as identified through clinical progress notes.

Sixty-six pharmacists in critical care and ID were surveyed at baseline and 1 month after “go live.” Fifty-five (83%) individuals responded to baseline surveys. Twenty-nine (44%) individuals responded to the follow-up survey 1 month after “go live.” The majority of pharmacist survey respondents primarily practiced in critical care (60% of respondents) and had 1–5 years of practice experience after terminal training (42% of respondents). Survey results demonstrated a 68% absolute improvement in the percentage of pharmacists who felt comfortable recommending beta-lactam TDM for critically ill patients. There was a 20% absolute increase in the perception that beta-lactam TDM was worthwhile, 50%–80% improvement in familiarity with operational considerations. Ninety-four percent of the respondents gauged the education as sufficient (Table 2). These results indicated that the bundled education intervention successfully increased knowledge, comfortability, and confidence among pharmacists with beta-lactam TDM. The survey results of physicians, advanced practice providers, and nurses all demonstrated a favorable improvement in beta-lactam knowledge and interest after education (Supplementary Tables S1–S4).

Table 2.

Survey results of pharmacists: baseline and post (1 month after “go-live”). Survey results of pharmacists, the key stakeholders, evaluated at baseline (N = 55 respondents out of a total of 66 surveyed) and at 1 month after “go-live” (N = 29 respondents out of a total of 66 surveyed) demonstrated favorable improvements in all items assessed indicating that the bundled education intervention successfully increased knowledge, comfortability, and confidence among pharmacists with beta-lactam TDM.

	Agree		Neutral		Disagree
Survey question	Baseline (%)	Post (%)	Baseline (%)	Post (%)	Baseline (%)	Post (%)
I can explain the scientific rationale for beta-lactam TDM.	65	100	11	0	24	0
Beta-lactam TDM will improve patient outcomes through personalizing drug therapy.	65	89	33	11	2	0
Implementing beta-lactam TDM will be/was worthwhile.	60	80	38	20	2	0
I want beta-lactam TDM to be implemented. (Baseline survey only)	60	Not applicable	33	Not applicable	7	Not applicable
I understand my role in beta-lactam TDM.	56	100	9	0	35	0
I have the required skills to implement beta-lactam TDM.	36	100	18	0	45	0
I know how to select patients most likely to benefit from beta-lactam TDM.	44	94	11	3	45	3
I feel comfortable recommending beta-lactam TDM for infected critically ill patients.	29	97	24	3	47	0
I am familiar with the preferred number and timing of blood samples for beta-lactam TDM.	11	94	13	6	76	0
I can accurately apply the results of a beta-lactam drug level test to make treatment decisions.	24	10	16	0	60	0
Beta-lactam TDM education provided was sufficient for me to implement beta-lactam TDM into my clinical practice. (Post survey only)	Not applicable	94	Not applicable	3	Not applicable	3
I would recommend the beta-lactam TDM education to others. (Post survey only)	Not applicable	94	Not applicable	6	Not applicable	0

The primary balancing measure was 283 DOT/1000 patient-days (N = 629) after the intervention was introduced. No significant difference was observed from baseline (P = .73).

Control

The summary of interventions, measured results, and lessons learned were shared with project sponsors and applicable stakeholders at institutional meetings. The results were used to justify expanded patient candidacy to include other individuals at high risk for deranged pharmacokinetic/pharmacodynamic profiles, drug resistance, or drug toxicity. An electronic health record-embedded pharmacokinetics calculator based on the prior Excel version has been developed to improve clinical workflow efficiency.

Antimicrobial stewardship pharmacists were selected as the long-term operational owners of the beta-lactam optimization program and handoff occurred from the project team. An electronic health record-based dashboard was created to monitor target attainment and program performance (Supplementary Fig. S1). Operational owners will monitor target attainment quarterly using a run chart. A sustained target attainment <75% will be flagged and interrogated for next steps.

Discussion

Statement of principal findings

This quality improvement project successfully improved the adequacy of beta-lactam therapy without potentiating antibiotic overuse. Beta-lactam adequacy was defined as the proportion of patients that achieved the desired beta-lactam pharmacokinetic/pharmacodynamic targets. In the literature, greater beta-lactam adequacy has been associated with improved treatment of infections, microbial eradication, decreased risk for drug toxicity, and ultimately better survival [28]. Improving the utilization of beta-lactams could also prevent or slow the development of antibiotic resistance, which positively impacts patient outcomes and limits the spread of resistance in our communities.

Interpretation within the context of the wider literature

In the present study, we adopted an educational intervention and TDM to improve beta-lactam adequacy, which has been shown to be suboptimal in the critically ill [6]. After “go live,” the beta-lactam target attainment was >90%. The bundled intervention makes it difficult to parse which component was most impactful. Most of the serum concentrations performed for TDM during the follow-up interval were initial rather than repeat concentrations. Initial concentrations would be most directly affected by the educational component of the intervention rather than the TDM. Nevertheless, we surmise that TDM may have had a clinical impact, given that 40% of individuals underwent a dose adjustment in response to the results, which suggests perceived utility by clinicians. Although two recent randomized controlled trials suggested no improvement in ICU length of stay or degree of organ dysfunction with beta-lactam TDM-guided therapy [32, 33], systematic reviews indicate consistent improvement in pharmacokinetic/pharmacodynamic target attainment and an overall signal toward improved clinical and microbiologic cure [28]. To further enhance up-front dose optimization in critically ill patients, model-informed precision dosing with TDM may be warranted [34].

Although the frequency is growing, beta-lactam individualization using tools like TDM remains uncommon, especially in the USA [35]. Multiple reasons exist for this and likely relate to organizational and individual barriers to implementation [19]. Introduction of a beta-lactam individualization program can be complex and involves critical infrastructure (e.g. electronic health record, clinical decision support resources, and laboratory assays) and end-user training and education [36, 37]. A project steering committee, a detailed charter guided by quality improvement methods, proactive engagement from multidisciplinary stakeholders and identification of needed resources, phased implementation, and responsiveness to continuous feedback, each facilitated uptake and success of the project.

Implications for policy, practice, and research

Beta-lactam optimization is an area of great interest to the clinical and research community. The Surviving Sepsis Campaign guidelines recommend “optimizing dosing strategies of antimicrobials based on accepted pharmacokinetic/pharmacodynamic principles and specific drug properties” [1]. Multiple methods have been and continue to be considered to optimize beta-lactam therapy, including extended or continuous infusion of beta-lactam therapy, model-informed precision dosing, and TDM [32, 33].

Strengths and limitations

A main strength of this study included successfully leveraging the Lean Six Sigma DMAIC framework to approach beta-lactam optimization through a quality improvement lens. Numerous potential root causes were identified, and the intervention to address the quality gap was selected to balance impact with effort. The approach selected mimics that used to optimize other antimicrobials like aminoglycosides and can be widely applied [16].

Limitations of this project must be acknowledged. As a quality improvement project with a before and after design, we are not able to ascribe causality of the findings to the intervention we selected. Nevertheless, no other quality improvement projects devoted to improving beta-lactam therapy in the critically ill were concurrently ongoing at the site, and thus we surmise the observed impact was from our dedicated efforts. Baseline data for the primary metric was not taken directly from the patient population selected by the steering committee. For clinical and feasibility reasons, the patient population selected for the intervention was highly restrictive (i.e. use of ECMO, kidney replacement therapy, or patients at extremes of weight). The uniqueness of this population is evident by the prolonged hospital length of stay (median of 48 days) and high severity of illness scores. Not unexpectedly, given the criteria for beta-lactam TDM use, 34% of the sample was cardiothoracic surgery patients requiring extracorporeal support. Sensitivity analyses of varied thresholds of baseline adequacy based on local data and the published literature were probed, and demonstrated the robustness of the findings.

The project was made possible by existing resources and infrastructure that may otherwise not be available at all centers. Mayo Clinic has an existing electronic health record, an embedded multidisciplinary team infrastructure in the critical care environment, access to quality improvement expertise, and the ability and resources to develop and validate local specialty assays with adequate turnaround time to ensure clinical applicability. While beta-lactam assays are not widely available, the proposed approach provides infrastructure for a quality improvement project that can be used to improve beta-lactam adequacy through a variety of mechanisms, including use of prolonged infusion therapy [38], model-informed precision dosing [34], or TDM with internal or external assays (i.e. by send out). Preparation and process mapping for such programs can use a similar approach to that undertaken in this project but must consider unique nuances of possible interventions when assessing the clinical utility in dynamic critically ill patients (e.g. delayed turnaround time for external assay results) [36]. Moreover, the quality improvement methods herein described can be applied to other settings (e.g. outside of the ICU) or other drugs or drug classes (e.g. azole antifungals).

Application of the Lean Six Sigma DMAIC framework provided the needed structure to the quality improvement effort. Among the lessons learned, feedback from pharmacists, the key stakeholders involved in the day-to-day operation of the process, was critical as it allowed for early identification of process gaps and allowed the team to proactively develop guidance. Second, the semi-controlled stability of using a limited-scope population at the outset of the program provided a useful context to establish clinical utility while building clinician understanding during the improve phase. Third, developing a mechanism for real-time feedback throughout implementation using a web-based tool facilitated identification of thematic issues and streamlined resolution.

Conclusion

This quality improvement initiative demonstrated how education and a beta-lactam TDM program can be utilized to improve beta-lactam adequacy in the critically ill. Use of the Lean Six Sigma DMAIC framework led to a theoretically based, coordinated approach to address the identified quality gap. Insights gained from this beta-lactam quality improvement effort can be extrapolated to other settings beyond critical care.

Authors contributions

All authors take responsibility for the contents of the submitted manuscript. Rebecca J. Wessel (drafted the manuscript). Christina G. Rivera, Sara E. Ausman, Nathaniel Martin, Shienna A. Braga, Natalie T. Hagy, Lindsay N. Moreland-Head, Omar M. Abu Saleh, Ognjen Gajic, Paul J. Jannetto, and Erin F. Barreto (provided critical revision of the manuscript).

Supplementary data

Supplementary data are available at IJQHC online.

Conflict of interest

E.F.B. consults for Wolters Kluwer and Baxter Health (unrelated). All other authors declare that they have no competing interests or relevant conflicts of interest to disclose.

Funding

This work was supported in part by the Mayo Clinic Division of Public Health, Infectious Diseases and Occupational Medicine and the National Institute of Allergy and Infectious Diseases of the National Institutes of Health under Award Number (K23AI143882 to E.F.B.). The funding source had no role in study design; data collection, analysis, or interpretation; writing the report; or the decision to submit the report for publication. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health.

Data availability

Not applicable.