Measuring Provider Compliance with an Institution-Based Clinical Pathway for Diverticulitis Using Radiologic Criteria

Abstract

Background:

Diverticulitis remains a common disease encountered in the acute care setting. Management strategies have been developed to guide treatment decisions based on imaging. By using a multi-faceted clinical pathway approach, a standardized method of diagnosing and categorizing disease severity can be performed in order to guide appropriate management. This study evaluated provider compliance with an institutional clinical pathway designed to guide management of diverticulitis.

Methods:

An institutional clinical pathway was developed to manage diverticulitis, including radiologic classification, primary service line assignment, interventional strategies, and antimicrobial treatment. To assess provider compliance with the algorithm, we queried the institutional acute diverticulitis database for patients admitted from May 19, 2016 to February 8, 2017, which identified 83 patients. Provider compliance with the pathway was assessed using subgroup analysis of radiologic documentation (modified Neff [mNeff] classification), primary service assignment, and interventions (i.e., interventional radiology [IR] and antimicrobial agents).

Results:

The cohort represented a diverse group of mNeff classifications, predominantly Stage 0. Surgical interventions occurred in 10.8% of the cohort. Antimicrobial agents were administered to 88.0% and 78.3% of the outpatients and inpatients, respectively. Patients received a total duration of antimicrobial therapy (mean ± standard deviation [SD]) of 10.2 ± 5.1 days. Overall compliance occurred in 10% of the patients. Compliance with radiologic documentation, antimicrobial choice, and antimicrobial duration were 90.4%, 20.5%, and 69.9%, respectively.

Conclusions:

Overall compliance with the clinical pathway was poor, except as it related to compliance with radiologic documentation, appropriate assignment to surgical service line, and antimicrobial duration. These results suggest areas for future improvement to augment compliance with the clinical pathway.

Diverticulosis is a condition characterized by the outpouching of the colonic mucosa and submucosa through the outer colonic layers where the vasa recti penetrate through the colonic wall [1]. This process is thought to be caused by the increased intra-luminal pressure within the bowel and has been attributed to factors such as age, constipation, and the Western diet. When these colonic outpouchings become obstructed, progressive venous congestion leads to eventual necrosis and perforation of the colonic wall resulting in diverticulitis [2]. Diverticulitis can be described as a spectrum of pathologic stages ranging from a localized, contained focus of pneumoperitoneum to frank peritonitis caused by contamination from intra-luminal contents. On routine colonoscopy, the prevalence of diverticular disease has been noted in 30% of patients aged 50–59 years old and more than 70% in patients aged 80 years and older [3]. Diverticulitis has been noted to occur in 20%–25% of patients with diverticular disease [4].

The clinical staging of diverticulitis has been defined classically using the Hinchey classification, which is primarily a surgical grading system to identify patients likely to require operative management [5]. The radiologic staging of diverticulitis has been defined using the modified Neff (mNeff) classification. The mNeff classification includes the following: Stage 0, uncomplicated diverticulitis with adjacent wall thickening; Stage 1, localized complicated diverticulitis (Stage 1a, localized pneumoperitoneum in the form of air bubbles and Stage 1b, abscess less than 4 cm); Stage 2, complicated diverticulitis with pelvic abscess greater than 4 cm; Stage 3, complicated diverticulitis with distant abscess; and Stage IV, complicated diverticulitis with abundant pneumoperitoneum and/or intra-abdominal free fluid [4]. With utilization of the mNeff classification as a standardized diagnostic tool, patients can be categorized by their pathologic stage of diverticulitis on presentation. Management strategies have been developed to guide treatment decisions based on this imaging classification and specific patient characteristics [6].

At our institution, a clinical pathway was developed in order to provide treatment guidelines based on initial mNeff classification. The purpose of this study was to evaluate provider compliance with our institutional-based clinical pathway.

Patients and Methods

Study design

An institutional clinical pathway was developed to manage diverticulitis, including radiologic classification, primary service line assignment, interventional strategies, and antimicrobial treatment. Figure 1 illustrates the process for managing diverticulitis and each decision point represents an opportunity for compliance measurement. To assess provider compliance with the algorithm, we queried the institutional acute diverticulitis database for patients admitted from May 19, 2016 to February 8, 2017, which identified 83 patients. Provider compliance with the pathway was assessed using subgroup analysis of radiological documentation of mNeff classification, primary service assignment, and interventions (i.e., interventional radiology [IR] and antimicrobial agents). Appropriateness of antimicrobial choice included selected agent based on severity (i.e., no antimicrobial, intravenous vs. oral, and narrow vs. broad-spectrum). Acceptable antimicrobial duration was up to 10 days for patients with no source control procedure and four days post-intervention for patients with a source control procedure performed. The study was approved by the John Peter Smith Institutional Review Board.

FIG. 1.

Process map for management of acute diverticulitis.

Statistical analysis

Patient descriptive statistics were performed. Population characteristics of interest include demographics and medical history including comorbidities, complications, and treatment for diverticulitis. Overall compliance was measured as percentage reflecting patients meeting all criteria described in study design and subgroup analyses were performed to quantitate percent compliance with radiologic, antimicrobial utilization, and patient disposition. Statistics were performed using SAS^® 9.4 (SAS Institute, Cary, NC).

Results

Baseline patient characteristics

Table 1 demonstrates baseline characteristics for the 83 patients diagnosed with diverticulitis. The majority of patients were female with an average age (± SD) of 49.2 ± 11.9 years. Comorbidities were documented in 40% of the cohort, with diabetes mellitus or underlying heart disease representing the most common conditions.

Table 1.

Baseline Characteristics

Characteristic
Age, years, mean ± SD	49.2 ± 11.9
Gender, no. (%)
Male	37 (44.6)
BMI, kg/m², mean ± SD	31.0 ± 7.9
Race, no. (%)
Caucasian	35 (42.2)
African American	14 (16.9)
Asian	1 (1.2)
Other	33 (39.8)
Hispanic, no. (%)	38 (45.8)
Comorbid conditions, no. (%)
Diabetes mellitus	18 (21.7)
Heart disease	12 (14.5)
End-stage renal disease	2 (2.4)
Irritable bowel disease	0 (0)
Alcohol/drug addiction	7 (8.4)
Liver disease	3 (3.6)
Neoplastic disease	5 (6.0)
Active chemotherapy	1 (1.2)
Organ transplant	0 (0)
Chronic steroids	0 (0)
Patient consents to outpatient therapy, no. (%)	54 (65.1)

SD = standard deviation; BMI = body mass index.

Baseline disease characteristics

Diverticular disease-specific characteristics are displayed in Table 2. Of the patients included, 27.7% had a previous history of diverticulitis. The cohort represented a diverse group of mNeff classifications with the following distribution: Stage 0 (43.4%), Stage 1 (18.1%), Stage 1a (13.3%), Stage 1b (7.2%), Stage 2 (7.2%), Stage 3 (0%), and Stage 4 (1.2%) (Table 2). Radiologic documentation and primary service assignment occurred in 90% of patients facilitating clinical pathway compliance (Table 3). Nearly one-quarter of the cohort had some complication associated with diverticulitis with concomitant abscess being the most common (Table 2).

Table 2.

Disease Characteristics

Characteristic
Previous history of diverticulitis, no. (%)	23 (27.7)
Previous abdominal surgery, no. (%)	46 (55.4)
mNeff Stage, no. (%)
Stage 0	36 (43.4)
Stage 1	15 (18.1)
Stage 1a	11 (13.3)
Stage 1b	6 (7.2)
Stage 2	6 (7.2)
Stage 3	0 (0)
Stage 4	1 (1.2)
Not identified	8 (9.6)
Complications associated with diverticulitis, no. (%)	20 (24.1)
Abscess	11 (13.3)
Sigmoid perforation	4 (4.8)
Abscess and sigmoid perforation	5 (6.0)

mNeff = modified Neff classification.

Table 3.

Clinical Pathway Compliance

Characteristic	% (n/N)
Overall compliance	9.6 (8/83)
Antimicrobial
Antimicrobial choice	20.5 (17/83)
Duration
Source control	33.3 (3/9)
No source control	74.3 (55/74)
mNeff
Documentation	90.4 (75/83)
Primary service assignment
Non-surgical primary team mNeff Stage 0	100.0 (36/36)
Non-surgical primary team mNeff Stage 1 or 1a	88.5 (23/26)
Surgical primary team mNeff Stage 1b	16.7 (1/6)
Surgical primary team mNeff Stage 2–4	100.0 (7/7)
IR drainage for mNeff Stage 1b	16.7 (1/6)

mNeff = modified Neff classification; IR = interventional radiology.

Clinical pathway compliance

Table 3 demonstrates the overall compliance and subgroup compliance rates for the clinical pathway. Overall compliance was low, driven primarily by antimicrobial choice.

Patients were assigned to clinical service based on mNeff classification: the non-surgical primary team (mNeff 0, mNeff 1, or 1a) or the surgical primary team (mNeff 1b or mNeff 2–4, Fig. 1). Surgical interventions were performed in approximately 10% of the cohort; specific surgical interventions are listed in Table 4. Despite the recommendation, only one of six qualifying patients received percutaneous drainage for mNeff 1b (16.7%; Table 3).

Table 4.

Surgical and Antimicrobial Management

Characteristic
Surgical interventions, no. (%)	9 (10.8)
IR drain	5 (6.0)
IR drain and OR with ostomy	1 (1.2)
IR drain and OR without resection or ostomy	1 (1.2)
OR with resection and ostomy	2 (2.4)
Antimicrobial therapy
Outpatient, no. (%)	73 (88.0)
Duration of antimicrobial therapy, days, mean ± SD	9.3 ± 3.6
Appropriate antimicrobial therapy, no. (%)	32 (43.8)
Antimicrobial, no (%)
Ciprofloxacin + metronidazole	30 (42.3)
Amoxicillin-clavulanate	28 (39.4)
Amoxicillin-clavulanate + metronidazole	2 (2.8)
Other	11 (15.5)
Inpatient, no. (%)	65 (78.3)
Duration of antimicrobial therapy, days, mean ± SD	3.0 ± 3.0
Appropriate antimicrobial therapy, no. (%)	36 (55.4)
Antimicrobial, no (%)
Ceftriaxone	37 (56.9)
Metronidazole
IV	47 (72.3)
PO	10 (15.4)
Piperacillin-tazobactam	20 (30.8)
Ciprofloxacin
IV	12 (18.5)
PO	7 (10.8)
Vancomycin	5 (7.7)
Cefepime	2 (3.1)
Meropenem	1 (1.5)
Overall antimicrobial
Duration of antimicrobial therapy, days, mean ± SD	10.2 ± 5.1

IR = interventional radiology; SD = standard deviation; IV = intravenous; PO = oral.

Thirty-six patients were all assigned correctly to the non-surgical primary team for mNeff Stage 0, which had 100% compliance in primary service assignment. The next highest rates of compliance with service assignment occurred in the mNeff Stage 1 or 1a subgroup (88.5%) and mNeff Stage 2–4 subgroup (100%). Lower compliance occurred in the mNeff Stage 1b subgroup; the institutional clinical pathway recommends percutaneous drainage of mNeff Stage 1b when feasible (Fig. 1).

Antimicrobial management characteristics are displayed in Table 4. A large proportion of patients received antimicrobial therapy, 88.0% and 78.3%, during outpatient and inpatient setting, respectively. The most common antimicrobial agents prescribed for outpatient therapy included ciprofloxacin plus metronidazole or amoxicillin-clavulanate, whereas, ceftriaxone plus metronidazole, piperacillin-tazobactam or ciprofloxacin plus metronidazole were used most commonly in the inpatient setting. Appropriateness of antimicrobial therapy varied from outpatient to inpatient setting with a higher compliance in the inpatient setting.

Patients received a total duration of antimicrobial therapy (mean ± SD) of 10.2 ± 5.1 days (Table 4). Despite choosing antimicrobial agents discordant with the protocol, the majority of patients received the appropriate antimicrobial duration. Compliance for antimicrobial duration without source control was higher (74.3%) than patients who received source control (33.3%; Table 4).

Discussion

This study evaluated provider compliance with an institutional clinical pathway for the management of diverticulitis based on a radiologic classification. The study revealed that the radiologic classification of diverticulitis seen on presentation represented the highest area of pathway compliance, whereas, provider compliance remained poor with antimicrobial selection. These discrepancies in compliance demonstrate the need for further assessment and development of methods for ensuring compliance with the multi-step algorithm.

The mNeff classification provides an efficient methodology for classifying patients radiologically that then differentiates the management of diverticulitis [4,6]. The clinical pathway developed remains contingent upon the radiologic reporting of mNeff classification to provide downstream treatment decisions. Thus, for this clinical pathway to be 100% effective, a compliance rate of 100% is necessary. Only approximately 90% of the patients had a mNeff score documented hindering the provider's ability to follow the algorithm and thus lowering overall compliance. Over time the compliance with mNeff documentation improved to 100% (data not shown) secondary to ongoing collaboration between the radiology and surgical departments.

Of the 75 patient cases with a documented mNeff classification, the majority of patients presented as a Stage 0 diverticulitis encompassing nearly half of our total patient population studied. This group of patients per the algorithm may qualify for no antimicrobial therapy, which represents a novel approach at our institution. Elimination of antimicrobial therapy in this group remains contingent on the patient meeting several criteria (Fig. 1). A large majority of the cohort had significant underlying comorbidities (40% of those studied), whereas approximately a quarter had recurrent diverticulitis, and almost a third were deemed unreliable as a result of not consenting for outpatient treatment; these factors would result in these patients being treated with antimicrobial agents. Further investigation with a larger cohort will be required to determine the overall impact of an antimicrobial-free approach to the management of mild diverticulitis, especially in a more heterogeneous population seen at our institution; such studies would need to include outpatient follow up information to determine treatment failures in this arm. The majority of the published data apply to more homogeneous populations outside of the United States [7] so additional study in these more complex patient populations is warranted.

Another point of emphasis of the clinical pathway was to promote the disposition of patients with higher mNeff classifications to surgical service lines to ensure more uniform management and assessment for possible surgical intervention. In addition, patients not likely requiring surgical intervention (i.e., mNeff Stage 0 and Stage 1 or 1a) were assigned to non-surgical services to reserve resources. These patients were either discharged directly home or held in an observation unit and then discharged if they did not require a higher level of care such as intravenous antimicrobial agents. A total of 8.4% of patients presented as a Stage 2–4 mNeff classification and received appropriate assignment to a primary surgical team for overall management reaching 100% compliance. This appropriate assignment remains key to addressing the need for surgical source control. In patients who were classified as Stage 1b, provider compliance was lower measuring 16.7% both for service assignment and percutaneous intervention. The Stage 1b represents a small number of patients and poor compliance may have resulted in the transition of assigning patients to medical rather than the surgical service. It could have also resulted from an inability to access a drainable collection through interventional radiology. As for mNeff Stage 0 patients, 100% were assigned appropriately to non-surgical services, which potentially minimizes unnecessary surgical consultation and allocates resources to appropriate cases. This approach may prove to have positive far reaching consequences, especially if it is applied at institutions with limited general surgery coverage for emergency departments.

In our study, the most common outpatient antimicrobial regimen included ciprofloxacin plus metronidazole or amoxicillin-clavulanate. In a study by Moya et al. [2], the most commonly used antimicrobial agents used in the outpatient management of uncomplicated acute diverticulitis included these same antimicrobials for a 7- to 10-day course [2]. On the inpatient side the most common antimicrobial selection was ceftriaxone plus metronidazole as recommended per the clinical pathway. This follows recommendations for intravenous management of acute diverticulitis closely [8,9]. Nearly one-third of the patients did receive broader spectrum therapy with piperacillin-tazobactam resulting in inappropriate antimicrobial selection. In addition, non-approved antimicrobial agents were provided (e.g., vancomycin and cefepime) and utilization of fluoroquinolone (FQ) regimens in non-β-lactam allergic patients occurred. These deviations led to an overall low compliance rate of approximately 20%. Unknown factors may have resulted in this low compliance rate including broad-spectrum coverage resulting from clinical decompensation and historic preference (i.e., FQ therapy). We hypothesize that even though the antimicrobial agents utilized may provide effective, yet overly broad coverage, the deviation from protocol and assessed non-compliance may not reflect differences in overall outcomes. Both the average outpatient and total antimicrobial durations approximated the normal duration recommended [8,9]. Compliance with antimicrobial duration measured approximately 70% at our institution. Deviations occurred commonly because of prolonged therapy post-source control and in patients receiving a day or two more past the recommended 10 days in patients without source control. The poor compliance for antimicrobial duration post-source control may result from the novelty of this approach (i.e., recent release of randomized control trial and guidelines [9,10]) and patient-specific factors unrecognized in the study.

The present study has several limitations. First, the cohort represents a small sample size limiting the overall generalizability to the entire diverticulitis population. Second, the assessment of compliance near the adoption of the protocol limits the ability to determine optimal compliance and potential for waning compliance over time. Third, the overall compliance definition relied on strict compliance to all portions of the pathway assessed. This resulted in a low compliance rate. Subgroup analysis did provide areas with acceptable compliance with low compliance in antimicrobial therapy hindering overall compliance. Last, the cohort had small numbers of patients with Stage 1b or above limiting our assessment of these patient populations. Because these are managed by surgical service lines the compliance may have improved as a result of the pathway being championed by the surgical service and the ability to educate a smaller and more homogenous group of providers.

The measurement of provider compliance with an institutional clinical pathway provides information to help better standardize care in the inpatient and outpatient setting with the downstream goal of improving overall patient care and outcomes. Limited compliance by providers as seen in our study can possibly be attributed to poor provider understanding of the clinical pathway, as well as, individual provider bias when guiding treatment. Our study further identifies areas in which provider compliance can be improved and offers areas for continued focus and study, especially the ramifications of pathway implementation and compliance on treatment outcomes.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Feuerstein

, Falchuk

. Diverticulosis and diverticulitis. Mayo Clin Proc, 2016; 91:1094–1104.

Moya

, Bellon

, Arroyo

, et al. Outpatient treatment in uncomplicated acute diverticulitis: 5-year experience. Turk J Gastroenterol, 2016; 27:330–335.

Everhart

, Ruhl

. Burden of digestive disease in the United States Part II: Lower gastrointestinal disease. Gastroenterology, 2009; 132:741–754.

Mora López

, Flores Clotet

, Serra Aracil

, et al. The use of the modified Neff classification in the management of acute diverticulitis. Rev Esp Enferm Dig, 2017; 109:328–334.

Hinchey

, Schaal

, Richards

. Treatment of perforated diverticular disease of the colon. Adv Surg, 1978; 12:85–109.

Mora Lopez

, Serra Pla

, Serra-Aracil

, et al. Application of a modified Neff classification to patients with uncomplicated diverticulitis. Colorectal Dis, 2013; 15:1442–1447.

Mocanu

, Dang

, Switzer

, et al. The role of antibiotics in acute uncomplicated diverticulitis: A systematic review and meta-analysis. Am J Surg [Epub ahead of print; DOI: 10.1016/j.amjsurg.2018.01.039.]

Byrnes

, Mazuski

. Antimicrobial therapy for acute colonic diverticulitis. Surg Infect, 2009; 10:143–154.

Mazuski

, Tessier

, May

, et al. The Surgical Infection Society revised guidelines for the management of intra-abdominal infection. Surg Infect, 2017; 18:1–76.

10.

Sawyer

, Claridge

, Nathens

, et al. Trial of short course antimicrobial therapy for intra-abdominal infection. N Engl J Med, 2015; 372:1996–2005.