Clinical Scoring Systems in the Management of Suspected Appendicitis in Children

Abdominal pain is a common problem in children presenting to the Emergency Department (ED) and though the differential diagnosis is expansive, appendicitis is the most common surgical emergency of childhood. While many children present with classical findings of right lower quadrant (RLQ) pain associated with nausea or vomiting and fever, subtle features and difficult examinations can make identifying appendicitis in a child challenging, leaving Health Care Providers struggling to distinguish this surgical emergency from less urgent conditions. Appendicitis is a progressive condition making early recognition essential in limiting morbidity and mortality. While some suggest Diagnostic Imaging (DI) as a routine screen for all children with abdominal pain, ED wait times, fiscal restraints and increasing concern related to radiation exposure require a more prudent, selective approach to identifying the child with suspected appendicitis. Clinical Scoring Systems (CSSs) have been developed to assist clinicians in appropriately stratifying a child’s clinical risk of having appendicitis. This chapter reviews the literature and reports on the experience of a tertiary care Pediatric Emergency Department (PED) in incorporating a clinical score into a Clinical Pathway in order to stratify children into High/Moderate/Low risk for appendicitis, thus guiding management and departmental patient flow.


Why use Clinical Scoring Systems?
Making wise, educated decisions is the cornerstone of good medical practice and often involves estimating the probability of an event. Inherent to all medical decisions is an assessment of potential risk and benefit. Risk tolerance within a clinical setting is dependent on the key stakeholders involved, for example the health care providers, the patients, the general public, the health care organization and policy makers. For a clinician, factors such as personality traits, quality and quantity of practice, experience with recent adverse events or near misses, fears of litigation and external stressful events may impact risk tolerance. Risk assessment requires at least some basic knowledge of statistics, though mastery is far from needed. While terms such as sensitivity and specificity are familiar to many medical staff they are not as useful as other concepts. An understanding of Pre-and Post-Test likelihoods, Positive-and Negative-Predictive Values (PPV, NPV), Positive-and Negative-Likelihood Ratios (PLR, NLR) and Accuracy impact on the interpretation of results. There are significant variations in clinical practice and outcomes, at national, regional and even local levels in a number of conditions and appendicitis is amongst them. A number of studies have demonstrated practice variation, as well as the impact of variation on clinical outcome measures (Chang, Ng, Y.-C. Chen, J.-C. Chen, & Yen, 2010) (Goldman et al., 2009) (Plint et al., 2004) (Richer et al., 2010) (Jain, Elon, Johnson, Frank, & Deguzman, 2010). While practice variation results in patient outcome differences, standardization of practice based on the best evidence can result in improved care (Eitel, Rudkin, Malvehy, Killeen, & Pines, 2010). Numerous studies have demonstrated the efficacy of Evidence Based Clinical Algorithms (EBCA) such as pathways and protocols in reducing delays in time-sensitive medication administration, reducing unwarranted radiation exposure and reducing mortality (Rivers et al., 2001) (Francis 2010, Osmond 2010. Integrating CSSs into EBCA is key to standardizing patient care in an effort to improve global and individual health outcomes.

The literature search strategy
To obtain complete information related to CSSs for suspected appendicitis in children, a formal literature review of common scientific databases was performed by the Health Information Network Calgary, University of Calgary, Alberta, Canada. Multiple databases were searched including Medline, Embase, Cochrane, PubMed, CINAHL Plus and Academic Search Complete. The following terms were utilized in the search: appendicitis, acute appendicitis, clinical decision rule, clinical prediction rule, prediction, score, and risk stratification. Search strategies limited the results to those published between 1980 and 2011 and included all children aged 0 -18 years (infant, preschool, school aged, adolescent, all child). Abstracts of the above search strategy were reviewed, refining the final manuscript database to those relevant to the current topic. Two hundred sixty six references were reviewed. Thirty-six articles were retrieved for inclusion in this review. Reference lists of these manuscripts were examined and any additional citations relevant to the topic were added.

Clinical Scoring Systems for suspected appendicitis in children
Over the last 3 decades, a number of CSSs have been developed to assist the clinician in assessing patients presenting with abdominal pain and suspected appendicitis. Several of these scores were specifically derived for children, while others were developed for adults or mixed populations and subsequently validated in children. The best known, such as the Alvardo Score and the Pediatric Appendicitis score, have been studied at length. Lesser known scores such as Kharbanda's Low Risk Score, the Lindberg Score, and the Ohmann Sore, among others, are listed in Table 2. Most Scoring Systems include a combination of Historical, Clinical and Laboratory measures. Each of these scores will be reviewed in detail.
Alvarado Score (MANTRELS) Pediatric Appendicitis Score (Samuel) Low Risk for Appendicitis Score (Kharbanda) Lintula Score Eskelinen Score Fenyo -Lindberg Score Ohmann Score Christian Score RIPASA Score Table 2. Clinical Scoring Systems used in the Diagnosis of Appendicitis in children Care is needed when evaluating studies of CSSs. Several studies include children and adults and some have few children. Populations may also differ; some studies include all-comers to the ED with abdominal pain whilst others include only those with suspected appendicitis, still others include those in whom a surgical consult was obtained and finally, some are limited to those children who had an appendectomy. Additionally, the medical specialty, level of training and experience of the staff performing Score assessments may also have a significant impact on generalizability (Emergency Physician vs. Surgeon, Senior Trainee vs. Attending Staff). Those studies that enroll prospectively are obviously more robust than retrospective analyses. And finally, some studies use modifications to a Score criteria (e.g. dropping a criteria) or to the threshold level (e.g. standard Alvarado Score threshold for probable appendicitis is ≥7, but some studies use a cutoff of 6). In light of some of these potential biases, Ohmann et al re-evaluated data from 10 published CSSs for appendicitis to determine their performance in meeting predefined quality criteria. Subsequently, they prospectively collected data on 1254 patients with acute abdominal pain from 6 different sites in order to evaluate the same Scores. The predefined quality criteria included a) initial negative appendectomy rate < 15%; b) potential perforation rate < 35%; c) initial missed perforation rate < 15% and d) missed appendicitis rate < 5%. Four of the original derivation studies met at least one quality criteria, however, when applied to the prospective evaluation, none of the 10 Scores were successful in meeting the predefined quality criteria. Ohmann et al concluded that significant bias existed in the derivation of the Scores, as mentioned in the paragraph above. (Ohmann, Yang, & Franke, 1995)

The Alvarado Score (MANTRELS)
In 1986, Alvarado published what is now one of the most well-known and studied appendicitis scores (Alvarado, 1986). This retrospective study of 305 patients admitted for suspected appendicitis evaluated common clinical and laboratory findings in relation to pathologically proven acute appendicitis. 277 patients were eligible for analysis.

www.intechopen.com
Eight criteria were chosen for inclusion in the diagnostic score, weighted to represent joint probability of disease. The Diagnostic criteria for the Alvarado Score are shown in Table 3. Right Lower Quadrant (RLQ) Pain and a Left Shift were found to be the most prevalent, thus receiving 2 points each, while each of the remaining criteria were attributed 1 point. This initial study included both adults and children, with an age range of 4 to 80 years (mean 25.3). An Alvarado Score of ≥7 was considered high risk for appendicitis. Though not explicitly stated in the study, this threshold value had a sensitivity of 81% and a specificity of 74%. Several elements of the score have been criticised, particularly the threshold for fever (37.3 C) and the availability of peripheral cell count differentials at some health centres, prompting some investigators to modify the score (see below).  Table 3. Comparison of Diagnostic Criteria between the Alvarado Score (MANTRELS) and the Pediatric Appendicitis Score (Samuel Score) Numerous studies have examined the Alvarado Score, particularly in children. (Table 4) Bond et al prospectively studied 187 children aged 2 -17 years with suspected appendicitis, of which 143 were admitted. Using Alvarado's cutoff score of 7 to indicate the need for surgery, the authors found a sensitivity and specificity of 90% and 72% respectively, with a negative appendectomy rate of 17%. Lower cutoff scores (5 or 6) demonstrated improved sensitivity, but corresponding reductions in specificity, as expected. Subgroup analysis showed the score to be least accurate in preschool children, corresponding to the clinical experience of many health care providers, though overall numbers in this age group were limited. The authors concluded that the Alvarado Score failed to achieve their predetermined standard for accuracy, however, this was set quite high at 99.5% sensitivity. (Bond, Tully, Chan, & Bradley, 1990) A retrospective study of children under 14 years by Hsiao et al confirmed Alvarado's data showing that RLQ tenderness and a left shift were the most prevalent signs in those with pathologically proven appendicitis. Children with Alvarado Scores ≥7 were statistically more likely to have appendicitis than controls. Overall sensitivity and specificity for an Alvarado Score >=7 were 60% and 61% respectively. (Hsiao, Lin, & D.-F. Chen, 2005)   CT evaluation. High sensitivity and specificity were maintained, at 100% and 97% respectively, suggesting that surgical intervention was best suited to children with an Alvarado Score of 8-10, while those with scores of 5-7 be further evaluated with imaging studies (Rezak, Abbas, Ajemian, Dudrick, & Kwasnik, 2011). In a mixed pediatric-adult population, Owen et al prospectively evaluated 215 patients, 70 of whom were children. In this pediatric subgroup analysis, sensitivity and specificity were 93% and 81% (Owen 1992). Shreef et al recently performed a dual-centre prospective study, reviewing 350 children aged 8 to 14 years. Interestingly, their reported statistical analysis was based on an Alvarado threshold of 6, and was based upon 2 different outcomes; 1) performance of appendectomy and 2) histology. Using the standard threshold of 7 and including all comers related to histologic diagnosis, the sensitivity and specificity were 86% and 83% respectively (Shreef, Waly, Abd-Elrahman, & Abd Elhafez, 2010). Several attempts have been made to modify the Alvarado Score to improve its accuracy. Macklin et al sought to simplify the Alvarado Score by eliminating the criteria for left shift (Modified Score total 9), as done by Kalan in a mixed adult/pediatric study. Children aged 4-14 years were enrolled, demonstrating sensitivity and specificity of 76.3% and 78.8% respectively using a cutoff score of 7 or higher to predict histological appendicitis. Kalan's study was limited to 11 children, all of which had modified Alvarado Scores >=7 and corresponding appendicitis. Obviously these numbers are too small to draw any conclusions (Macklin, Radcliffe, Merei, & Stringer, 1997)(Kalan 1994). Sooriakumaran et al further modified the score by decreasing the value of leukocytosis, to make a total score of 8. This score was then compared to clinical assessment by Emergency Physicians, and found wanting. However, one must be cautious, as only 3 children were included (!) , and, due to the change in total score, the threshold value was tested at 5 (Sooriakumaran, Lovell, & Brown, 2005). Significant changes to the Alvarado Score were suggested by Impellizzeri et al. who studied 156 children aged 2-17 years, replacing anorexia with an elevated fibrinogen level (>400mg/dL), changing migration of pain to length of pain (although not defined), combining RLQ pain and rebound into one criteria, and decreasing the temperature cutoff to 37 C. Of note, the diagnosis of appendicitis was made on surgical report, not pathologic diagnosis. The authors suggest the above modifications would have decreased admission rates by 15% (Impellizzeri et al., 2002).

The Pediatric Appendicitis Score (Samuel Score)
Madan Samuel introduced the Pediatric Appendicitis Score (PAS) in 2002. A theoretical advantage to the PAS exists for 2 reason; 1) data was prospectively collected, and 2) the score was specifically derived in a population of children (aged 4 -15 years). The PAS has been subject to multiple subsequent validation and comparison studies. Evaluating 1170 children with suspected appendicitis, Samuel compared historical, clinical and laboratory features in children with appendicitis (n=734) and those without appendicitis (n=436). Using stepwise multiple linear regression, 8 variables were included in a diagnostic model out of 10 points, with greater weight attributed to RLQ pain and maneuvers eliciting rebound tenderness (cough/percussion). Diagnostic criteria for the PAS are shown in Table 3. Samuel concludes that a score of 6 or greater shows a high probability of acute appendicitis. (Samuel, 2002) Unlike the Alvarado Score, there have been no attempts to modify the PAS. However, multiple studies have sought to prove its validity (Summary provided in Table 5 (2) while demoting leukocytosis to a value of 1. This "new" modified score, however, seems to have the exact criteria of the PAS (Samuel combined cough/percussion tenderness with hopping tenderness because of good correlation and also promoted this elements value), and therefore could be considered in the PAS group (Shera, Nizami, Malik, Naikoo, & Wani, 2010).

Comparison of the Alvarado Score and the Pediatric Appendicitis Score
Upon reviewing Table 3 one will notice the similarities between the Alvarado Score and the PAS. However, several differences exist between the two. These include the following: 1. The Alvarado Score was derived in a mixed pediatric/adult population (aged 4 -80 years) and subsequently validated in children. The PAS was derived in children (aged 4 -15 years). 2. The Alvarado Score was derived retrospectively and subsequently validated both retrospectively and prospectively. The PAS was derived prospectively and has been validated as such. 3. The Alvarado Score specifically defined elevated temperature as ≥37.3 C, while the PAS does not define pyrexia. 4. The Alvarado Score specifically defined neutrophilia as > 75%, while the PAS does not define neutrophilia (similarly most subsequent studies utilize > 75%). 5. The weighted criteria differ. Alvarado emphases leukocytosis, while Samuel places higher value on rebound tenderness. Given the above differences, can one choose which score is better? Three well-designed prospective studies have performed head-to-head comparison of the Alvarado Score and the PAS.
Schneider et al enrolled 755 children aged 3 to 21 years who were evaluated by their surgical team for suspected appendicitis. Alvarado Scores and PAS were calculated on 588 participants with complete data. Overall, the PAS was more sensitive (82% vs. 72%) while the Alvarado Score was more specific (81% vs. 65%) in this population. Negative-and Positive-Predictive values were similar between groups (85 vs. 88% and 65 vs. 54% for Alvarado vs. PAS respectively). However, the Alvarado Score had a better Positive Likelihood ratio (3.8 vs. 2.4). ROC curves were similar between the two scores. Unfortunately, this study included patients up to 21 years of age, which may have improved the diagnostic accuracy of the score in this population, though the number of patients over 17 years was not large. Interestingly, the Positive Predictive Value of the Alvarado Score in children < 10 years was diminished (65% vs. 58%) (Schneider, A. Kharbanda, & R. Bachur, 2007).

Study
Year   (Escribá, Gamell, Fernández, Quintillá, & Cubells, 2011). Using a slightly different approach, Mandeville enrolled 287 of 487 children aged 4 to 16 years with a clinical diagnosis of appendicitis in whom 155 had pathologically proven appendicitis. Similar to Schneider's results, the PAS was more sensitive (88% vs. 76%) while the Alvarado Score was more specific (72% vs. 50%). ROC curves were once again similar, yet somewhat lower than the two studies described above (PAS -0.78, Alvarado -0.78). When stratified by sex, both Scores had slightly improved sensitivities in boys (Mandeville, Pottker, & Bulloch, 2010). The authors of these three prospective comparison studies concluded that there was insufficient evidence to favor one CSS over the other. Caution was stressed, suggesting that neither score was sensitive nor specific enough to be used as a stand-alone diagnostic test; further investigations such as Computed Tomography (CT) or Ultrasonography (US) were encouraged to complete the evaluation for intermediate-risk children.
The Alvarado Score and the PAS both make use of several key features of CSSs. The criteria are easy to elicit, each criteria is dichotomous (Yes/No), and the Score is easy to calculate. Overall, the PAS appears to be a more sensitive tool, while the Alvarado Score is more specific.

The Low Risk for Appendicitis Score (Kharbanda)
Increased ED wait times, hospital over-crowding and concerns related to radiation exposure from imaging studies have put pressure on clinicians to quickly and accurately decide which children with abdominal pain should be admitted and observed or discharged without a CT evaluation. Kharbanda et al derived and validated a score to do just that; identify children at low risk for appendicitis. Kharbana et al prospectively enrolled 767 children aged 3 to 18 years with suspected appendicitis who were evaluated by a surgeon. Of these 767, 601 were included (425 derivation set, 176 validation set). Using logistic regression 6 weighted predictors of appendicitis were determined for a total score of 14. (Table 6) Children with a score of <=5 were highly unlikely to have appendicitis (sensitivity 99%, NPV 98%, NLR 0.032 during derivation, 96%, 96% and 0.102 for validation set)

Diagnostic Criteria Value
Absolute Neutrophil Count >6.75 6 Rebound pain or pain with percussion 2 Unable to walk, or walks with a limp 1 Nausea 2 History of migration of pain to RLQ 1 History of focal RLQ pain 2 Total 14 In addition to creating the Low Risk Score, the Kharbanda study was novel in that it created a low risk decision tree using recursive partitioning. During derivation, the rule was perfectly sensitive, with a NPV of 98% and a NLR of 0. Validation demonstrated a sensitivity of 98%, NPV of 98% and NLR of 0.058. The Low Risk decision tree in shown in Figure 1

The Lintula Score
The Lintula Score relies on clinical data alone. There are no laboratory results required. Using a nice 2-phased approach, Lintula et al first prospectively evaluated 35 clinical variables to derive a score in 127 children aged 4 to 15 years (Score criteria are found in Table 7). Subsequent prospective validation of the score was performed on a similar sample of 109 children. The Lintula Score has a maximum value of 32. A high risk threshold was established at >=21, while low risk was <=15 (Lintula, Pesonen, Kokki, Vanamo, & Eskelinen, 2005).
Four years later, Lintula tested the score in a prospective randomized parallel design. Children aged 4 to 15 years randomized to assessment by score had surgical intervention based on the score result while intervention for those randomized to "no-score" was based on overall clinical and laboratory assessments by the surgeon. Assessments were made at three-hour intervals until a decision to operate or discharge home was established. Of note, imaging studies were not used in either group. Use of the Lintula Score resulted in a significantly higher accuracy (92% vs. 80%) and a lower rate of negative appendectomies. The 2 groups showed no significant difference in sensitivity, however specificity was improved in the Lintula Score group (88% vs. 67%) (Lintula, Kokki, Kettunen, & Eskelinen, 2009 Table 7. Diagnostic Criteria for the Lintula Score

The Eskelinen Score
The Eskelinen Score is relatively complex to perform, (requiring factor multiplication) and was originally designed for use within a computer program. was 0.91. The authors further calculated these statistical variables using thresholds ranging from 50 to 60, and determined 57 to be the most favorable in their population. (Sitter, Hoffmann, Hassan, & Zielke, 2004)

The Fenyo-Lindberg Score
This score appears to be one of the most complex, incorporating criteria with multiple levels of response that both add to and subtract from the total score. (Diagnostic Criteria found in Table 9) In 1987, Fenyo prospectively evaluated 259 adult patients with suspected appendicitis. The resulting score was further validated in 830 patients, of which 256 had proven appendicitis. Sensitivity, Specificity, PPV and NPV were 90%, 91%, 83% and 95% respectively. (Fenyo 1987) Fenyo and Lindberg prospectively validated their score in 1167 patients with suspected appendicitis. Of these, 392 had histologically proven appendicitis. Using the standard threshold score of -2 to predict appendicitis, the sensitivity was 73% and specificity was 87%, notably less than in the original study. Of note, this study made use of 2 different settings, a district and a university hospital. 30% of the patients included from the University hospital were children (age unknown) (Fenyö, Lindberg, Blind, Enochsson, & Oberg, 1997). Once again, modification to a CSS was studied. The following changes were made by Dado et al: a) increased WBC cut-off values (<12, 12-20, >20), b) altered values of migratory pain (Yes +4, No -11), c) insertion of elevated temperature >37.5 C (Yes +7, No -9), and d) removal of aggravation by coughing. 197 children aged 2 to 17 years were retrospectively stratified into 3 risk groups using the modified Lindberg Score. Sensitivity and Specificity were 86% and 87%, with an excellent PPV of 96%, but only a modest NPV at 69% (Dado et al., 2000).

The Ohmann Score
In 1999, Ohmann prospectively validated his own score in a multi-centre, multi-phase trial (Diagnostic Criteria are found in Table 10). Subjects evaluated during phase 1 (n=870) received surgical intervention based on surgeon assessment, while those in phase 2 (n= 614) received computer-assisted diagnostic support using the Ohmann Score. Children less then 6 were excluded from the study, overall pediatric numbers were not published. The authors found a statistically significant improvement in specificity, PPV and accuracy in the phase 2 Score group, along with a decrease in the number of delayed diagnoses (defined as appendectomy on the second day after admission or later) (Ohmann 1999). Several studies have evaluated the Ohmann Score. In a large study of 2359 subjects (age 0 -95 years) Zielke compared the score to clinical assessments. Overall accuracy using the Ohmann Score was found to be better than junior surgical staff, with a sensitivity, specificity, PPV, NPV and accuracy of 63%, 93%, 77%, 86% and 84%. However, it was not found to be better than senior surgical staff assessments. (Zielke et al., 1999) Table 10. Diagnostic Criteria of the Ohmann Score

The Christian Score
Probably the simplest of the group, the Christian Score uses a mere 5 criteria (Diagnostic Criteria are found in Table 11). The case group of 58 subjects with suspected appendectomy had surgical intervention if >=4 criteria were met. Fifty-nine appendectomy controls had www.intechopen.com intervention based solely on surgical staff assessment. Ages ranged from 7 to 56 years. The negative appendectomy rate was significantly less in the Score group than that of the controls (6.5% vs. 17%). This is a rather simple score, which unfortunately does not to appear to have been validated or assessed in a pediatric specific population, but probably should be (Christian 1992).
Abdominal pain on history, occurring within 48 hours of presentation Vomiting -one or more episode

RLQ tenderness on examination
Low grade fever -defined as <=38.8 C Polymorphonuclear leukocytosis -define as WBC > 10 000 AND neutrophils > 75% Table 11. Diagnostic Criteria for the Christian Score

The RIPASA Score
What is probably the newest member to the group of appendicitis scores is the RIPASA Score, named after its hospital of origin in Brunei. A mixed population of 400 adults and children who had an appendectomy were retrospectively identified, the records of 312 were used to derive the score. Individual criteria were weighted (0.5, 1, 2) based on probabilities and a panel of staff surgeons. The resulting maximal RIPASA score is 16 (diagnostic criteria are found in Table 12); a threshold of 7.5 proving a sensitivity of 88% and specificity of 67% PPV and NPV were 93% and 53%, while accuracy was 81%. Using the score, an absolute reduction in negative appendectomies of 9% would have occurred.  Chong et al continued to evaluate their new score by prospectively enrolling 200 adults and children in a comparison of the RIPASA and Alvarado Scores. In this group of patients, the RIPASA was statistically superior to the Alvarado Score in Sensitivity (98% vs. 68%), NPV (97% vs. 71%) and accuracy (92% vs. 87%). Specificity, PPV and negative appendectomy rates were similar between the 2 scores. (Chong 2011)

Other Scores
Several other CSSs have been developed for patients with suspected appendicitis, but do not appear to have been formally evaluated in children and as a result are not further discussed in this chapter. Some of these include the Teicher Score, Arnbjornsson Score, Izbicki Score, and DeDombal Score.

Clinical Scoring Systems in practice: Experience of the Alberta Children's Hospital Pediatric Appendicitis Pathway
The real test of a CSS is whether it works in practice. Here I report our experience at the Alberta Children's Hospital (ACH) during a Quality Improvement process from 2006-2011. The Alberta Children's Hospital is a tertiary referral centre for children aged 0 -18 years, serving a population of approximately 1.8 million in southern Alberta, western Saskatchewan and eastern British Columbia, the 3 western-most provinces of Canada. The Pediatric Emergency Department (PED) at ACH has an annual census of approximately 60 000 visits, and the surgical staff perform approximately 350 acute appendectomies each year. Following several highly publicized adverse outcomes surrounding appendicitis in both children and adults in the former Calgary Health Region (now Alberta Health Services -Calgary and Area), a formal safety review was conducted. Early diagnosis and standardization of care were determined to be of utmost importance. As a result, Clinical Pathways were developed for both adults and children. Early diagnosis remains a significant challenge for the pediatrician. After reviewing the literature related to CSSs, the ACH Pediatric Appendicitis Committee agreed to incorporate a score into the pathway development to assist in standardization of assessment, investigation and inter-disciplinary communication. The Alvarado Score and PAS were felt to have similar qualities and to be the most thoroughly evaluated of the CSSs in children with acceptable performance for risk stratification. Since a number of staff groups are employed in both pediatric and adult hospital settings one consistent CSS was felt to be optimal. The Alvarado Score was incorporated into both the adult and pediatric Appendicitis Pathways for the region. The Pediatric Appendicitis Pathway (Figure 2) uses the Alvarado Score in 2 different ways; the first is a novel departmental flow advancement through a screening tool that initiates Advanced Nursing Directives (ANDs); the second, is a risk stratification tool for physician decision making. Since ED assessment and management are a team effort, we felt it was vital for both nursing and medical staff to use an assessment tool with common features.

Incorporating Clinical Scoring Systems into Advanced Nursing Directives
Advanced Nursing Directives are used in the ACH PED to improve patient flow and reduce waiting times. ANDs are not simple nursing protocols for administering antipyretics for children with fever or non-steroidal anti-inflammatory drugs for musculo-skeletal pain. Nor www.intechopen.com are they meant to formally diagnose disease. Rather they are a recognition that skilled nurses are able to identify certain common disease processes (in this case appendicitis), that some of these disease presentations have common investigations and/or management processes, and that empowering pediatric emergency nurses in the frontline can expedite patient care. The purpose of an AND, therefore, is 1) to assist pediatric emergency nurses in identifying children who would likely need further investigation and (2) to empower pediatric emergency nurses to initiate investigations and management before pediatric emergency physician assessment (deForest & . (Thompson 2010a) A number of ED practices, such as initiating intravenous access and fluid management, are commonly performed by nurses caring for adults but are less widely performed in children. Pediatric centres rightly try to limit potentially painful procedures unless they are absolutely required and our AND aims to do precisely this. The important components of any AND include standardized assessment measures using set criteria, a defined care plan if criteria are met, and the option to seek assistance when necessary. Validated CSSs are ideal for integration into an AND. It must be recognized that different clinical settings may be more appropriate for the use of ANDs. For some health care centres, implementing ANDs into departmental flow may stretch beyond normal nursing practice. For others, it may simply be a matter of formally documenting a process already in place. The Alvarado Score utilizes both clinical and laboratory variables but at the initial assessment triage of a child with abdominal pain laboratory results are rarely available. Our AND ( Figure 3) uses a modification of the original score, leaving out laboratory criteria, and increasing the cutoff value of elevated temperature from 37.3 C to 38.0 C (as a fever is defined as temperature > 38 C in our department). The remaining historical and clinical variables are evaluated by the nursing staff and recorded as a dichotomous variable, either Yes or No. If overall AND criteria are met, the nursing staff are empowered to initiate intravenous access, obtain blood and urine samples for laboratory assessment and give a bolus of crystalloid fluid. These processes occur prior to physician assessment. The objectives of the AND are to identify children with suspected appendicitis earlier in their health care visit, to decrease the time to obtain laboratory results, to identify potential confounding diagnoses early on (i.e. urinary tract infection, pregnancy) and to prepare the child for potential diagnostic imaging. Preliminary data demonstrates accuracy of our nurses in predicting appendectomy using the AND is similar to the previously published data from the Alvarado Score studies discussed above (Sensitivity 72%, Specificity 72%, NPV 91%, PPV 40%, accuracy 72%. (Thompson 2010b) 6.2 Incorporating the Alvarado Score for medical decision making in the ACH PED It is well recognized that some children presenting to the ED clearly require surgery for acute appendicitis. However, over the last 2 decades, there has become increased reliance on Diagnostic Imaging modalities (DI) to confirm or rule out appendicitis and potentially provide alternate diagnoses (particularly in post-menarchal girls). Given the availability of DI including Ultrasonography (U/S) and Computed Tomography (CT), and the relatively high sensitivity and specificity of these tests, they are often requested by the surgical team in order to improve diagnostic accuracy and decrease the rate of negative appendectomy. However, given recent concerns related to radiation exposure in children (Brenner & Hall, 2007), as well as overcrowding in many EDs and DI departments leading to delays in imaging acquisition, a more responsible approach to risk stratification is required.  High risk patients (Alvarado Score ≥7) are evaluated by the surgical team after consultation by the ED staff without the need for imaging. Similarly, children at low risk for appendicitis with an Alvarado Score of ≤4, are evaluated by the ED staff for alternate diagnoses or managed with watchful waiting in the home setting, to return should the child's condition worsen. Those children at moderate risk for appendicitis  or those with high risk for alternate diagnoses (post-menarchal females) are most likely to benefit from imaging studies.

Nursing Protocol for the Child w ith Suspected Appendicitis
Purpose: For Emergency Nursing staff to initiate investigations and treatment for the patient presenting in the Emergency Department with signs and symptoms of ap pendicitis prior to an assessment by an Emergency Physician.

For this protocol to be initiated:
The patient must have ONE of the following clinical signs: YES NO  Any tenderness in the right lower quadrant with palpation by examiner OR  Rebound tenderness in the right lower quadrant (eg. Positive Jump Test/Positive Pothole Test) ! ! The patient must also have 3 OR MORE of the following screening criteria: YES NO  Any complaint of right lower quadrant pain by patient  Nausea and/or vomiting  Decreased appetite (anorexia)  Elevated temperature and/or history of ( " 38.0C)

Does this patient meet the above clinical criteria? ! YES ! NO
If a patient meets the above screening criteria, an Emergency Depar tment Nurse is able to perform the following procedures prior to Emergency Department Physician assessment/orders:  IV access with double lumen "Y" connector  Obtain bloodwork (glucometer check, CBC/differential, electrolytes)  Initiate a 20cc/kg bolus of 0 .9% NaCl (maximum 1 litre), then run 0.9% NaCl at maintenance rate:  Collect a midstream urine for urine dip/R&M, send for culture if urine dip is positive . For all female patients " 10 years of age, a point of care #-HCG test should also be performed. Inform patient not to void after initial urine sample in case of need for full bladder for abdominal ultrasound.  Ensure patient is NPO A maximum of 2 IV attempts will be made prior to the Physician's assessment. The Nurse will communicate with the Charge Nurse to ensure that the patient is prioritized appropriately prior to initiation of this protocol.

Patient Information Label
Calculating IV Maintenance Rate: Example Calculation for patient weighing 27kg: 4 ml/kg/hr for first 10 kg of body weight 4mlx10kg= 40ml/kg/hr + 2 ml/kg/hr for next 10 kg +2mlx10kg= 20ml/kg/hr + 1 ml/kg/hr for the remainder +1mlx7kg= 7ml/kg/hr = 67ml/kg/hr maintenance rate Our post -implementation data has shown that over 40% of appendectomy patients went to the Operating Room (OR) without any imaging studies, a reflection of high risk stratification related to incorporating the Alvarado Score into our pathway. (Thompson 2010c)

Implementation and measurement
While it is well known that incorporating carefully developed CSSs into practice improves patient care and departmental processes, the optimal method of implementation is yet to be determined. Many local, national and international medical organizations have developed strategies related to implementation science and knowledge management/translation. Cognitive, social, motivational and organization factors all influence knowledge uptake and use (Gaddis, Greenwald, & Huckson, 2007).
Realistically, it is difficult to achieve 100% uptake of CSSs. Careful planning, with input from all key stakeholders is vital. Introducing new system processes for the care of the child with suspected appendicitis has a multidisciplinary impact. It is highly advisable to solicit representative input from Emergency Medicine, Surgery, Nursing, Diagnostic Imaging, Infectious Disease, Anesthesia, Pharmacy as well as unit managers of Emergency Department, Operating Room, and clinical wards. In order to optimize the potential buy-in from these key stakeholders, departmental leaders would be wise to identify specific outcomes measures ("key wins") geared to each discipline that they will target, for example reduced ED and post-operative lengths of stay. Donabedian identifies 3 quality measurement pillars. These include structural measures (factors that are present prior to a client visit), process measures (factors occurring during the client visit) and outcome measures (factors occurring after the client visit). Ideally, these outcomes are easily measurable, and within attainable reach (Donabedian, 1992) (Schiff & Rucker, 2001). The statistical Methods for measuring change related to implementation of CSSs are beyond the scope of this chapter. Interested readers are encouraged to review the literature on Quality Assessment and Measurement.

Conclusions
Due to the often-difficult task of the early identification of appendicitis in children, the development of CSSs has increased over the last 3 decades. While most clinicians caring for children with suspected appendicitis are well versed in regard to the Alvarado Score and the Pediatric Appendicitis Score, many other models have been developed. Overall, these scores have been shown to improve clinical and process outcomes including reduced negative appendectomy rates, reduced radiation exposure from unwarranted DI studies, and reduced missed diagnoses. However, one must remain optimistically cautious; to date these Scores have yet to demonstrate a sensitivity or specificity sufficient enough to recommend their use beyond a calculated risk stratification (low, moderate or high). Even with the abundance of literature regarding CSSs related to appendicitis in children, the need for well-designed, prospective studies to further validate the scores, evaluate implementation strategies and assess impact provides ample opportunity for future research. Due to the vast number of CSSs and the significant variability in the quality and quantity of validation studies, implementing Clinical Scores into practice can be challenging for individual clinicians. Departmental leaders should therefore carefully consider incorporating CSSs into locally driven Evidence Based Clinical Algorithms. www.intechopen.com