Cost-Effectiveness of Diagnostic Approaches to Suspected Appendicitis in Children Jay Pershad, MD, Teresa M Waters, Eunice Y Huang, MD, MS, FACS
PhD,
Max R Langham Jr,
MD, FACS,
Tao Li,
MD, PhD,
Our group recently published a clinical pathway (Le Bonheur Clinical Pathway [LeB-P]) that used the Samuel Pediatric Appendicitis Score with selective use of ultrasonography (USG) for diagnosis of children at risk for appendicitis. The objective of this study was to model the cost-effectiveness of implementing the LeB-P compared with usual care. STUDY DESIGN: We constructed a decision analytic model comparing hospital costs for the following diagnostic strategies for suspected appendicitis: emergency department clinician judgment alone, USG on all patients, CT on all patients, overnight observation with surgical evaluation without studies, and the LeB-P. Prevalence of disease, outcomes probabilities, and hospital and professional costs for each option were derived from published literature, national cost data, and our previous study results. Cost-effectiveness was calculated using these 3 sets of parameters. RESULTS: In the base case model, USG was the preferred strategy over LeB-P and overnight observation with surgical evaluation without studies. Emergency department clinician judgment alone and CT were dominated by the other pathways, based on either lower diagnostic accuracy or increased costs. Compared with LeB-P, USG costs $337 less per patient evaluated, but increased the diagnostic error rate by 2%. Using LeB-P rather than USG would cost an institution an additional $17,206 to eliminate one misdiagnosis, which is known as the incremental cost-effectiveness ratio. CONCLUSIONS: Although performing USG on all children with suspected appendicitis was determined to be the most cost-effective strategy, using the Pediatric Appendicitis Score with selective use of USG (LeB-P) improved diagnostic accuracy at a moderate increase in cost and decreased CT use. (J Am Coll Surg 2015;-:1e9. 2015 by the American College of Surgeons) BACKGROUND:
appendicitis is important because of the increased morbidity, mortality, and costs associated with perforation.3,4 A missed diagnosis, with discharge of the child from the ED, delays treatment and is associated with subsequent longer lengths of stay and higher rates of perforation and complications, including readmission.4 Rates of perforated appendicitis in children are high, which also contributes to increased costs and accounts for a substantial portion of surgical care expenditure in pediatric hospitals.1,2 In 2011, thirty-four hospitals in the Pediatric Health Information System reported an estimated inpatient cost of nearly $150 million for care of children with appendicitis.5 In an attempt to improve overall diagnostic accuracy for pediatric appendicitis, investigators have prospectively evaluated the Samuel and Alvarado appendicitis scores in pediatric patients, and others have noted an increasing reliance on diagnostic imaging of children with abdominal pain.6-14 Computed tomography scans, the imaging modality of choice, have improved diagnosis of appendicitis15,16 and have been reported to be costeffective.17 As a result, use of CT for diagnosing pediatric
Appendicitis is the most common cause of surgical, nontraumatic abdominal pain among children presenting to the emergency department (ED), with some 80,000 cases reported in the United States each year.1,2 Diagnosis of appendicitis by clinical examination remains an art dependent on the skill of the clinical practitioner, who must differentiate early appendicitis from many other childhood illnesses causing abdominal pain.1,2 Early diagnosis of CME questions for this article available at http://jacscme.facs.org Disclosure Information: Authors have nothing to disclose. Timothy J Eberlein, Editor-in-Chief, has nothing to disclose. Presented at the Southern Surgical Association 126th Annual Meeting, Palm Beach, FL, November 30eDecember 3, 2014. Received December 12, 2014; Accepted December 15, 2014. From the Departments of Emergency Medicine (Pershad), Preventive Medicine (Waters, Li), and Surgery (Langham, Huang), University of Tennessee Health Science Center, Memphis, TN. Correspondence address: Eunice Y Huang, MD, MS, FACS, Department of Surgery, University of Tennessee Health Science Center, 51 N Dunlap St, P230, Memphis, TN 38105. email: [email protected]
ª 2015 by the American College of Surgeons Published by Elsevier Inc.
1
http://dx.doi.org/10.1016/j.jamcollsurg.2014.12.019 ISSN 1072-7515/14
2
Pershad et al
Cost of Diagnosing Pediatric Appendicitis
Abbreviations and Acronyms
ED ICER LeB-P OBS PAS USG
¼ ¼ ¼ ¼
emergency department incremental cost-effectiveness ratio Le Bonheur clinical pathway overnight observation with surgical evaluation without studies ¼ Pediatric Appendicitis Score ¼ ultrasonography
J Am Coll Surg
appendicitis has increased.18-23 The National Ambulatory Medical Care Survey data for patients younger than 19 years old presenting to a pediatric ED noted a rise in CT use from 0.9% in 1998 to 15.4% in 2008.21 In addition, several studies have suggested that the rate of CT use for the evaluation of appendicitis is higher at community hospitals compared with children’s hospitals, with one study reporting a 75% CT use rate at community hospitals, as compared with 26% at a children’s hospital.18,20,22
Figure 1. Le Bonheur Children’s Hospital clinical pathway for evaluation of suspected appendicitis. BMP, basic metabolic panel; CXR, chest x-ray; IVF, intravenous fluid; PAS, Pediatric Appendicitis Score; UA, urinalysis.
Vol.
-,
No.
-, -
2015
Pershad et al
Table 1. Parameters Used in Model, Including Appendicitis Prevalence, and Sensitivity and Specificity for Various Diagnostic Strategies Probability parameters
Prevalence of appendicitis Accuracy of Le Bonheur pathway Sensitivity Specificity Accuracy of abdominal CT Pooled sensitivity (meta-analysis) Pooled specificity (meta-analysis) Accuracy of ultrasonography Pooled sensitivity (meta-analysis) Pooled specificity (meta-analysis) Accuracy of emergency department clinician’s judgment Sensitivity Specificity Accuracy of overnight observation on surgical service Sensitivity Specificity
Point estimate, % (95% CI)
32
Reference 31 31
92 (83e98) 95 (89e98)
3
imaging modality (Le Bonheur clinical pathway [LeB-P]). Prospective evaluation of our pathway showed it to have high sensitivity and specificity for diagnosing pediatric appendicitis without increasing length of ED stay.31 The goal of this study was to compare the cost-effectiveness of the LeB-P against other available options for the workup of suspected pediatric appendicitis using standard methods that allow evaluation of both costs and clinical outcomes.
33
94 (92e97) 95 (94e97) 33
88 (86e90) 94 (92e95) 7
88 (83e93) 50 (42e59)
99 96
34-36 37,38
Overuse of abdominal CT results in increased overall health care costs, prolonged ED stays, and increased risk of radiation-induced malignancy.24,25 With heightened concern surrounding radiation exposure in children, ultrasonography (USG) has emerged as an increasingly popular first-line diagnostic modality, particularly at tertiary pediatric facilities, where pediatric ultrasonographers are readily available.15,18,26 However, USG, like clinical examination, is operator-dependent, and visualization of the appendix by USG can be variable, potentially leading to inconclusive studies.15,26-28 Unfortunately, negative appendectomy rates in children remain high, ranging from 4.4% to 13%, despite these efforts to improve diagnostic accuracy.2,18,19,29 There continues to be considerable variation in practice, resource use, and treatment-related costs associated with management of pediatric appendicitis, which translates to increased cost burden, especially in high-volume hospitals.4,30 Recently, our group prospectively evaluated a clinical pathway combining the Samuel Pediatric Appendicitis Score (PAS) with selective use of USG as the primary Table 2.
Cost of Diagnosing Pediatric Appendicitis
METHODS To model the cost-effectiveness of implementing the LeB-P for the evaluation of children with appendicitis, we generated a decision analytic model comparing LeB-P against 4 alternate diagnostic options in the workup of suspected appendicitis: clinical judgment of an experienced emergency medicine physician in conjunction with laboratory tests and/or abdominal radiographs, USG of the abdomen, CT of the abdomen and pelvis with intravenous contrast, and surgical evaluation with overnight hospital observation and no imaging (OBS). We hypothesized that the LeB-P would be the most cost-effective diagnostic option from a hospital perspective. After IRB approval was obtained for this study, a decision-tree model was created using TreeAge Pro 2014 decision analytic software (TreeAge Software, Inc). TreeAge Pro allows one to build large and complex decision trees to compare treatment strategies and explicitly examine the impact of uncertainty. Our model begins with a decision node and then offers branches for each treatment option associated with a specific health condition. The subtree for each treatment option follows the condition through treatment, including any number of possible outcomes. At each terminal node, a value is given for cost and effectiveness associated with those outcomes. The time horizon for our decision tree model was the patient’s acute illness period, based on the assumption that clinical outcomes from appendicitis will occur within a few days to weeks. Although a longer time horizon might have allowed us to incorporate the potential risk of radiation-induced cancer associated with CT use in our model, or late impact of delayed diagnosis on fertility and bowel obstruction, long-term data and utility weights for children (eg, use of quality-adjusted life-years) were insufficient to create an accurate model.
Baseline Costs of Care: Estimated Mean Inpatient Cost of Care Based on 2009 Kids’ Inpatient Database
Type of admission
Acute appendicitis (ICD-9 code 540.9) Perforated appendicitis (ICD-9 codes 540.0 or 540.1) Abdominal pain (ICD-9 codes 789.0x or 789.6x, no procedures performed, length of stay 1 day) KID, Kids’ Inpatient Database.
Mean cost from 2009 KID, $
Adjusted mean cost, $, 2012 value
6,757 11,801
8,429 14,721
2,670
3,331
4
Pershad et al
Table 3. Baseline Costs of Care: 2012 National Physician Fee Schedule Description of service
Emergency department visit, high complexity Observation/hospitalization, same date Ultrasound abdomen/back wall, limited CT abdomen/pelvis with contrast Laparoscopic appendectomy
Table 4. Baseline Costs of Care: Anesthesia Anesthesia services
AU
Cost, $
6
129.12
6 12
129.12 258.24
CPT code
Total RVUs
Professional fee, $
99285
4.94
168.15
99236
6.27
213.42
Billing conversion factor (per AU) is $21.52. AU, anesthesia base unit.
76775 74177 44960
3.23 10.51 17.74
109.94 357.74 603.83
Table 5. Baseline Costs of Care: 2012 Medicare Hospital Fees (Outpatient and Radiology Services)
Medicare conversion factor (per RVU) is $34.04. RVU, relative value unit.
The LeB-P algorithm was specified as follows: patients with a PAS of 1 to 3 were discharged from the hospital and received a follow-up phone call or were admitted to the pediatrics service with an alternate diagnosis. Patients with a PAS of 4 to 7 had a focused right lower quadrant USG performed after a period of observation and parenteral hydration in the ED. The duration of observation and decision to obtain the USG was left to the discretion of the treating clinician. If the USG was negative and there remained no continued suspicion of appendicitis, the patient was discharged from the ED with a followup phone call. If the USG was positive, surgical consultation was sought. For patients with a PAS of 8 to 10, a surgical evaluation was obtained. Computed tomography scans were obtained only at the request of the pediatric surgical consultant (Fig. 1). Baseline probabilities for our model, using data from our previously published study and other published studies on appendicitis prevalence using national databases, included an appendicitis prevalence rate of 32%, of which 72% were assumed to be acute (simple) and 28% perforated (complicated).5,31,32 Sensitivities and specificities for various strategies are provided in Table 1. The Table 6.
J Am Coll Surg
Cost of Diagnosing Pediatric Appendicitis
Anesthesia surgery lower abdomen (base) Anesthesia units (15-minute increments for total of 90 minutes as mean surgery length) Total anesthesia billing
Median cost, $
Outpatient and radiology services
CT abdomen with contrast Real-time ultrasonography of abdomen and pelvis Level 5 emergency department visit for abdominal pain
581 96 323
model also included the following potential outcomes for patients with abdominal pain: diagnosis and treatment of other causes of abdominal pain (true negatives), missed appendicitis that returned for care (false negatives), and negative appendectomies (false positives). The probability of each of these outcomes was calculated for each strategy in our model. Cost data were derived from two sources. The 2009 Kids’ Inpatient Database was used to obtain national estimates of mean hospital care costs (for each clinical scenario) for pediatric patients.39 Kids’ Inpatient Database, developed by the Agency for Healthcare Research and Quality as part of the Healthcare Cost and Utilization Project, is the only nationally representative sample of pediatric discharges with data on hospital use, outcomes, and charges, and is available every 3 years. As cost-tocharge ratios were not yet available for the 2012 database, 2009 data were used and adjusted to 2012 dollars by accounting for rate of inflation and rate of time preference per standard cost-effectiveness analysis techniques.40-42
Estimated Costs for the Different Strategies and Outcomes
Outcomes node
Patient with appendicitis (72% acute, 28% perforated) Patient with missed diagnosis, returns with complicated appendicitis Negative appendectomy Patient with no appendicitis, no imaging, discharged from ED Patient with no appendicitis who receives an USG, discharged from ED Patient with no appendicitis, work-up using Le Bonheur Pathway (previous study breakdown: 55% USG, 7% CT, 28% ED only, 10% overnight stay) Patient with no appendicitis, receives CT, discharged from ED Patient with no appendicitis, no imaging, discharged after observation stay of 1 day ED, emergency department; USG, ultrasonography.
Hospital cost, $
Professional fee, $
Cumulative cost, $
10,191 14,721 8,429 323 419
1,030 1,198 1,030 168 278
11,221 15,919 9,459 491 697
767 904
257 526
1,024 1,430
3,331
213
3,544
Vol.
-,
No.
-, -
2015
Pershad et al
Cost of Diagnosing Pediatric Appendicitis
5
Figure 2. Decision tree with imputed sensitivities and specificities and the resulting costs calculated by analytic software. AA, acute appendicitis; LeB, Le Bonheur; OBS, overnight observation with surgical evaluation without studies; USG, ultrasonography.
Encounters were identified using ICD-9-CM codes for the primary diagnosis of acute appendicitis (540.9), perforated appendicitis (540.0, 540.1), and abdominal pain (789.0x, 789.6x; Table 2). Professional fees were included in each strategy and calculated using national 2012 Medicare physician fee schedule amounts associated with the appropriate CPT codes (Table 3).43 Anesthesia fees were based on a mean of 90 minutes for the surgical procedure. Hospital facility
fees for outpatient and radiology services were obtained from the 2012 Medicare outpatient prospective payment system national payment rates (Tables 4 and 5).44 Hospital and physician costs for outcomes nodes in the decision tree were calculated using the cost data sources noted here (Table 6). Using data on the prevalence of appendicitis, data for both the sensitivity and specificity of each diagnostic algorithm and the treatment cost at each outcomes node, cost, and effectiveness were
6
Pershad et al
Table 7.
Results of Cost-Effectiveness Analysis
J Am Coll Surg
Cost of Diagnosing Pediatric Appendicitis
Strategy
Nondominated options USG all LeB pathway Observation stay All options referenced against baseline of “USG all” USG all LeB pathway CT all Clinician judgment Observation stay
Cost, $
Incremental cost, $
Effectiveness (accuracy)
Incremental effectiveness
3,325 3,662 5,857
337.24 2194.38
0.92 0.94 0.97
0.02 0.03
17,206 75,150
3,325 3,662 4,187 4,547 5,857
0 337.24 862.13 1221.87 2531.62
0.92 0.94 0.95 0.62 0.97
0 0.02 0.03 0.30 0.05
0 17,206 33,159 4,084 51,877
ICER
LeB, Le Bonheur; USG, ultrasonography.
calculated, as well as the incremental cost-effectiveness ratio (ICER) for each diagnostic strategy (Fig. 2). For pediatric appendicitis, the ICER was the change in cost divided by the change in effectiveness (or accuracy) rate associated with using one diagnostic strategy over the baseline strategy; that is, the incremental cost incurred to avoid one additional false-negative (delay in treatment) or false-positive case (negative appendectomy). We used “determinate dominance” to eliminate strategies. This means that if the worst outcomes of one strategy are better than the best outcomes of a second strategy, the second strategy is a dominated strategy and removed from consideration. Lastly, we calculated the “willingness-to-pay” amount that would be required to move the second most costeffective diagnostic pathway to the most cost-effective pathway, using the formula “net monetary benefit” ¼ ({effectiveness willingness-to-pay} cost). We then performed a sensitivity analysis to determine how the results of our decision analysis might change based on
changes in the parameters for the sensitivity and specificity for ultrasonography (using published sensitivity and specificity ranges for ultrasonography on pediatric patients), as this parameter was likely to be the most variable from institution to institution.45
RESULTS Under the assumptions of the base case model, nondominated strategies with the lowest to highest cost (in order) were: USG ($3,325, effectiveness ¼ 92%), LeB-P ($3,662, effectiveness ¼ 94%), and OBS ($5,857, effectiveness ¼ 97%). Clinical judgment cost $4,547 per patient treated, but was dominated by the other strategies due to an effectiveness/accuracy rate of 62%, the lowest of the strategies. Computed tomography was dominated by a blend of LeB-P and OBS because CT cost more ($4,187) than LeB-P ($3,662), with minimal improvement in effectiveness (CT effectiveness ¼ 95%, LeB-P effectiveness ¼ 94%), and OBS was considerably more
Figure 3. One-way sensitivity analysis of change in sensitivity of ultrasonography (USG) and its effect on net monetary benefit of the USG diagnostic pathway in comparison with the other pathways (using a willingness to pay of $18,000 to decrease one error in diagnosis). LeB, Le Bonheur.
Vol.
-,
No.
-, -
2015
Pershad et al
Cost of Diagnosing Pediatric Appendicitis
7
Figure 4. One-way sensitivity analysis of change in specificity of ultrasonography (USG) and its effect on net monetary benefit of the USG diagnostic pathway in comparison with the other pathways (using a willingness to pay of $18,000 to decrease one error in diagnosis). LeB, Le Bonheur.
effective (OBS effectiveness ¼ 97%), although at a higher cost ($5,857; Table 7). As a result, CT had a higher ICER ($33,159) when compared with the optimal strategy of USG, making this strategy not cost-effective. In comparison with USG, the ICER for LeB-P ($17,206) was lower and the ICER for OBS ($51,877) was higher than that of CT, but with a higher effectiveness rate. Within the nondominated options, USG was the preferred strategy over LeB-P and OBS, based on its having the lowest ICER, despite being 2% less effective than LeB-P and 5% less effective than OBS. Selecting the LeBP over USG for workup of suspected appendicitis would cost the institution an additional $17,206 (the ICER) to eliminate one misdiagnosis. We next examined the monetary investment required to move the LeB-P to a position where our analytic model would consider it to be the optimal diagnostic strategy because of better accuracy at comparable cost. Our net monetary benefit calculation showed that at a willingness to pay of $18,000 to decrease one diagnostic error, the LeB-P becomes the preferred diagnostic pathway over USG, with a higher net monetary benefit. At this willingness-to-pay threshold of $18,000, sensitivity analysis showed that the LeB-P was the most cost-effective diagnostic option over a wide range of USG sensitivities, from 59% to 90% (Fig. 3). Similarly, if the specificity of USG is
PhD,
Max R Langham Jr,
MD, FACS,
Tao Li,
MD, PhD,
Our group recently published a clinical pathway (Le Bonheur Clinical Pathway [LeB-P]) that used the Samuel Pediatric Appendicitis Score with selective use of ultrasonography (USG) for diagnosis of children at risk for appendicitis. The objective of this study was to model the cost-effectiveness of implementing the LeB-P compared with usual care. STUDY DESIGN: We constructed a decision analytic model comparing hospital costs for the following diagnostic strategies for suspected appendicitis: emergency department clinician judgment alone, USG on all patients, CT on all patients, overnight observation with surgical evaluation without studies, and the LeB-P. Prevalence of disease, outcomes probabilities, and hospital and professional costs for each option were derived from published literature, national cost data, and our previous study results. Cost-effectiveness was calculated using these 3 sets of parameters. RESULTS: In the base case model, USG was the preferred strategy over LeB-P and overnight observation with surgical evaluation without studies. Emergency department clinician judgment alone and CT were dominated by the other pathways, based on either lower diagnostic accuracy or increased costs. Compared with LeB-P, USG costs $337 less per patient evaluated, but increased the diagnostic error rate by 2%. Using LeB-P rather than USG would cost an institution an additional $17,206 to eliminate one misdiagnosis, which is known as the incremental cost-effectiveness ratio. CONCLUSIONS: Although performing USG on all children with suspected appendicitis was determined to be the most cost-effective strategy, using the Pediatric Appendicitis Score with selective use of USG (LeB-P) improved diagnostic accuracy at a moderate increase in cost and decreased CT use. (J Am Coll Surg 2015;-:1e9. 2015 by the American College of Surgeons) BACKGROUND:
appendicitis is important because of the increased morbidity, mortality, and costs associated with perforation.3,4 A missed diagnosis, with discharge of the child from the ED, delays treatment and is associated with subsequent longer lengths of stay and higher rates of perforation and complications, including readmission.4 Rates of perforated appendicitis in children are high, which also contributes to increased costs and accounts for a substantial portion of surgical care expenditure in pediatric hospitals.1,2 In 2011, thirty-four hospitals in the Pediatric Health Information System reported an estimated inpatient cost of nearly $150 million for care of children with appendicitis.5 In an attempt to improve overall diagnostic accuracy for pediatric appendicitis, investigators have prospectively evaluated the Samuel and Alvarado appendicitis scores in pediatric patients, and others have noted an increasing reliance on diagnostic imaging of children with abdominal pain.6-14 Computed tomography scans, the imaging modality of choice, have improved diagnosis of appendicitis15,16 and have been reported to be costeffective.17 As a result, use of CT for diagnosing pediatric
Appendicitis is the most common cause of surgical, nontraumatic abdominal pain among children presenting to the emergency department (ED), with some 80,000 cases reported in the United States each year.1,2 Diagnosis of appendicitis by clinical examination remains an art dependent on the skill of the clinical practitioner, who must differentiate early appendicitis from many other childhood illnesses causing abdominal pain.1,2 Early diagnosis of CME questions for this article available at http://jacscme.facs.org Disclosure Information: Authors have nothing to disclose. Timothy J Eberlein, Editor-in-Chief, has nothing to disclose. Presented at the Southern Surgical Association 126th Annual Meeting, Palm Beach, FL, November 30eDecember 3, 2014. Received December 12, 2014; Accepted December 15, 2014. From the Departments of Emergency Medicine (Pershad), Preventive Medicine (Waters, Li), and Surgery (Langham, Huang), University of Tennessee Health Science Center, Memphis, TN. Correspondence address: Eunice Y Huang, MD, MS, FACS, Department of Surgery, University of Tennessee Health Science Center, 51 N Dunlap St, P230, Memphis, TN 38105. email: [email protected]
ª 2015 by the American College of Surgeons Published by Elsevier Inc.
1
http://dx.doi.org/10.1016/j.jamcollsurg.2014.12.019 ISSN 1072-7515/14
2
Pershad et al
Cost of Diagnosing Pediatric Appendicitis
Abbreviations and Acronyms
ED ICER LeB-P OBS PAS USG
¼ ¼ ¼ ¼
emergency department incremental cost-effectiveness ratio Le Bonheur clinical pathway overnight observation with surgical evaluation without studies ¼ Pediatric Appendicitis Score ¼ ultrasonography
J Am Coll Surg
appendicitis has increased.18-23 The National Ambulatory Medical Care Survey data for patients younger than 19 years old presenting to a pediatric ED noted a rise in CT use from 0.9% in 1998 to 15.4% in 2008.21 In addition, several studies have suggested that the rate of CT use for the evaluation of appendicitis is higher at community hospitals compared with children’s hospitals, with one study reporting a 75% CT use rate at community hospitals, as compared with 26% at a children’s hospital.18,20,22
Figure 1. Le Bonheur Children’s Hospital clinical pathway for evaluation of suspected appendicitis. BMP, basic metabolic panel; CXR, chest x-ray; IVF, intravenous fluid; PAS, Pediatric Appendicitis Score; UA, urinalysis.
Vol.
-,
No.
-, -
2015
Pershad et al
Table 1. Parameters Used in Model, Including Appendicitis Prevalence, and Sensitivity and Specificity for Various Diagnostic Strategies Probability parameters
Prevalence of appendicitis Accuracy of Le Bonheur pathway Sensitivity Specificity Accuracy of abdominal CT Pooled sensitivity (meta-analysis) Pooled specificity (meta-analysis) Accuracy of ultrasonography Pooled sensitivity (meta-analysis) Pooled specificity (meta-analysis) Accuracy of emergency department clinician’s judgment Sensitivity Specificity Accuracy of overnight observation on surgical service Sensitivity Specificity
Point estimate, % (95% CI)
32
Reference 31 31
92 (83e98) 95 (89e98)
3
imaging modality (Le Bonheur clinical pathway [LeB-P]). Prospective evaluation of our pathway showed it to have high sensitivity and specificity for diagnosing pediatric appendicitis without increasing length of ED stay.31 The goal of this study was to compare the cost-effectiveness of the LeB-P against other available options for the workup of suspected pediatric appendicitis using standard methods that allow evaluation of both costs and clinical outcomes.
33
94 (92e97) 95 (94e97) 33
88 (86e90) 94 (92e95) 7
88 (83e93) 50 (42e59)
99 96
34-36 37,38
Overuse of abdominal CT results in increased overall health care costs, prolonged ED stays, and increased risk of radiation-induced malignancy.24,25 With heightened concern surrounding radiation exposure in children, ultrasonography (USG) has emerged as an increasingly popular first-line diagnostic modality, particularly at tertiary pediatric facilities, where pediatric ultrasonographers are readily available.15,18,26 However, USG, like clinical examination, is operator-dependent, and visualization of the appendix by USG can be variable, potentially leading to inconclusive studies.15,26-28 Unfortunately, negative appendectomy rates in children remain high, ranging from 4.4% to 13%, despite these efforts to improve diagnostic accuracy.2,18,19,29 There continues to be considerable variation in practice, resource use, and treatment-related costs associated with management of pediatric appendicitis, which translates to increased cost burden, especially in high-volume hospitals.4,30 Recently, our group prospectively evaluated a clinical pathway combining the Samuel Pediatric Appendicitis Score (PAS) with selective use of USG as the primary Table 2.
Cost of Diagnosing Pediatric Appendicitis
METHODS To model the cost-effectiveness of implementing the LeB-P for the evaluation of children with appendicitis, we generated a decision analytic model comparing LeB-P against 4 alternate diagnostic options in the workup of suspected appendicitis: clinical judgment of an experienced emergency medicine physician in conjunction with laboratory tests and/or abdominal radiographs, USG of the abdomen, CT of the abdomen and pelvis with intravenous contrast, and surgical evaluation with overnight hospital observation and no imaging (OBS). We hypothesized that the LeB-P would be the most cost-effective diagnostic option from a hospital perspective. After IRB approval was obtained for this study, a decision-tree model was created using TreeAge Pro 2014 decision analytic software (TreeAge Software, Inc). TreeAge Pro allows one to build large and complex decision trees to compare treatment strategies and explicitly examine the impact of uncertainty. Our model begins with a decision node and then offers branches for each treatment option associated with a specific health condition. The subtree for each treatment option follows the condition through treatment, including any number of possible outcomes. At each terminal node, a value is given for cost and effectiveness associated with those outcomes. The time horizon for our decision tree model was the patient’s acute illness period, based on the assumption that clinical outcomes from appendicitis will occur within a few days to weeks. Although a longer time horizon might have allowed us to incorporate the potential risk of radiation-induced cancer associated with CT use in our model, or late impact of delayed diagnosis on fertility and bowel obstruction, long-term data and utility weights for children (eg, use of quality-adjusted life-years) were insufficient to create an accurate model.
Baseline Costs of Care: Estimated Mean Inpatient Cost of Care Based on 2009 Kids’ Inpatient Database
Type of admission
Acute appendicitis (ICD-9 code 540.9) Perforated appendicitis (ICD-9 codes 540.0 or 540.1) Abdominal pain (ICD-9 codes 789.0x or 789.6x, no procedures performed, length of stay 1 day) KID, Kids’ Inpatient Database.
Mean cost from 2009 KID, $
Adjusted mean cost, $, 2012 value
6,757 11,801
8,429 14,721
2,670
3,331
4
Pershad et al
Table 3. Baseline Costs of Care: 2012 National Physician Fee Schedule Description of service
Emergency department visit, high complexity Observation/hospitalization, same date Ultrasound abdomen/back wall, limited CT abdomen/pelvis with contrast Laparoscopic appendectomy
Table 4. Baseline Costs of Care: Anesthesia Anesthesia services
AU
Cost, $
6
129.12
6 12
129.12 258.24
CPT code
Total RVUs
Professional fee, $
99285
4.94
168.15
99236
6.27
213.42
Billing conversion factor (per AU) is $21.52. AU, anesthesia base unit.
76775 74177 44960
3.23 10.51 17.74
109.94 357.74 603.83
Table 5. Baseline Costs of Care: 2012 Medicare Hospital Fees (Outpatient and Radiology Services)
Medicare conversion factor (per RVU) is $34.04. RVU, relative value unit.
The LeB-P algorithm was specified as follows: patients with a PAS of 1 to 3 were discharged from the hospital and received a follow-up phone call or were admitted to the pediatrics service with an alternate diagnosis. Patients with a PAS of 4 to 7 had a focused right lower quadrant USG performed after a period of observation and parenteral hydration in the ED. The duration of observation and decision to obtain the USG was left to the discretion of the treating clinician. If the USG was negative and there remained no continued suspicion of appendicitis, the patient was discharged from the ED with a followup phone call. If the USG was positive, surgical consultation was sought. For patients with a PAS of 8 to 10, a surgical evaluation was obtained. Computed tomography scans were obtained only at the request of the pediatric surgical consultant (Fig. 1). Baseline probabilities for our model, using data from our previously published study and other published studies on appendicitis prevalence using national databases, included an appendicitis prevalence rate of 32%, of which 72% were assumed to be acute (simple) and 28% perforated (complicated).5,31,32 Sensitivities and specificities for various strategies are provided in Table 1. The Table 6.
J Am Coll Surg
Cost of Diagnosing Pediatric Appendicitis
Anesthesia surgery lower abdomen (base) Anesthesia units (15-minute increments for total of 90 minutes as mean surgery length) Total anesthesia billing
Median cost, $
Outpatient and radiology services
CT abdomen with contrast Real-time ultrasonography of abdomen and pelvis Level 5 emergency department visit for abdominal pain
581 96 323
model also included the following potential outcomes for patients with abdominal pain: diagnosis and treatment of other causes of abdominal pain (true negatives), missed appendicitis that returned for care (false negatives), and negative appendectomies (false positives). The probability of each of these outcomes was calculated for each strategy in our model. Cost data were derived from two sources. The 2009 Kids’ Inpatient Database was used to obtain national estimates of mean hospital care costs (for each clinical scenario) for pediatric patients.39 Kids’ Inpatient Database, developed by the Agency for Healthcare Research and Quality as part of the Healthcare Cost and Utilization Project, is the only nationally representative sample of pediatric discharges with data on hospital use, outcomes, and charges, and is available every 3 years. As cost-tocharge ratios were not yet available for the 2012 database, 2009 data were used and adjusted to 2012 dollars by accounting for rate of inflation and rate of time preference per standard cost-effectiveness analysis techniques.40-42
Estimated Costs for the Different Strategies and Outcomes
Outcomes node
Patient with appendicitis (72% acute, 28% perforated) Patient with missed diagnosis, returns with complicated appendicitis Negative appendectomy Patient with no appendicitis, no imaging, discharged from ED Patient with no appendicitis who receives an USG, discharged from ED Patient with no appendicitis, work-up using Le Bonheur Pathway (previous study breakdown: 55% USG, 7% CT, 28% ED only, 10% overnight stay) Patient with no appendicitis, receives CT, discharged from ED Patient with no appendicitis, no imaging, discharged after observation stay of 1 day ED, emergency department; USG, ultrasonography.
Hospital cost, $
Professional fee, $
Cumulative cost, $
10,191 14,721 8,429 323 419
1,030 1,198 1,030 168 278
11,221 15,919 9,459 491 697
767 904
257 526
1,024 1,430
3,331
213
3,544
Vol.
-,
No.
-, -
2015
Pershad et al
Cost of Diagnosing Pediatric Appendicitis
5
Figure 2. Decision tree with imputed sensitivities and specificities and the resulting costs calculated by analytic software. AA, acute appendicitis; LeB, Le Bonheur; OBS, overnight observation with surgical evaluation without studies; USG, ultrasonography.
Encounters were identified using ICD-9-CM codes for the primary diagnosis of acute appendicitis (540.9), perforated appendicitis (540.0, 540.1), and abdominal pain (789.0x, 789.6x; Table 2). Professional fees were included in each strategy and calculated using national 2012 Medicare physician fee schedule amounts associated with the appropriate CPT codes (Table 3).43 Anesthesia fees were based on a mean of 90 minutes for the surgical procedure. Hospital facility
fees for outpatient and radiology services were obtained from the 2012 Medicare outpatient prospective payment system national payment rates (Tables 4 and 5).44 Hospital and physician costs for outcomes nodes in the decision tree were calculated using the cost data sources noted here (Table 6). Using data on the prevalence of appendicitis, data for both the sensitivity and specificity of each diagnostic algorithm and the treatment cost at each outcomes node, cost, and effectiveness were
6
Pershad et al
Table 7.
Results of Cost-Effectiveness Analysis
J Am Coll Surg
Cost of Diagnosing Pediatric Appendicitis
Strategy
Nondominated options USG all LeB pathway Observation stay All options referenced against baseline of “USG all” USG all LeB pathway CT all Clinician judgment Observation stay
Cost, $
Incremental cost, $
Effectiveness (accuracy)
Incremental effectiveness
3,325 3,662 5,857
337.24 2194.38
0.92 0.94 0.97
0.02 0.03
17,206 75,150
3,325 3,662 4,187 4,547 5,857
0 337.24 862.13 1221.87 2531.62
0.92 0.94 0.95 0.62 0.97
0 0.02 0.03 0.30 0.05
0 17,206 33,159 4,084 51,877
ICER
LeB, Le Bonheur; USG, ultrasonography.
calculated, as well as the incremental cost-effectiveness ratio (ICER) for each diagnostic strategy (Fig. 2). For pediatric appendicitis, the ICER was the change in cost divided by the change in effectiveness (or accuracy) rate associated with using one diagnostic strategy over the baseline strategy; that is, the incremental cost incurred to avoid one additional false-negative (delay in treatment) or false-positive case (negative appendectomy). We used “determinate dominance” to eliminate strategies. This means that if the worst outcomes of one strategy are better than the best outcomes of a second strategy, the second strategy is a dominated strategy and removed from consideration. Lastly, we calculated the “willingness-to-pay” amount that would be required to move the second most costeffective diagnostic pathway to the most cost-effective pathway, using the formula “net monetary benefit” ¼ ({effectiveness willingness-to-pay} cost). We then performed a sensitivity analysis to determine how the results of our decision analysis might change based on
changes in the parameters for the sensitivity and specificity for ultrasonography (using published sensitivity and specificity ranges for ultrasonography on pediatric patients), as this parameter was likely to be the most variable from institution to institution.45
RESULTS Under the assumptions of the base case model, nondominated strategies with the lowest to highest cost (in order) were: USG ($3,325, effectiveness ¼ 92%), LeB-P ($3,662, effectiveness ¼ 94%), and OBS ($5,857, effectiveness ¼ 97%). Clinical judgment cost $4,547 per patient treated, but was dominated by the other strategies due to an effectiveness/accuracy rate of 62%, the lowest of the strategies. Computed tomography was dominated by a blend of LeB-P and OBS because CT cost more ($4,187) than LeB-P ($3,662), with minimal improvement in effectiveness (CT effectiveness ¼ 95%, LeB-P effectiveness ¼ 94%), and OBS was considerably more
Figure 3. One-way sensitivity analysis of change in sensitivity of ultrasonography (USG) and its effect on net monetary benefit of the USG diagnostic pathway in comparison with the other pathways (using a willingness to pay of $18,000 to decrease one error in diagnosis). LeB, Le Bonheur.
Vol.
-,
No.
-, -
2015
Pershad et al
Cost of Diagnosing Pediatric Appendicitis
7
Figure 4. One-way sensitivity analysis of change in specificity of ultrasonography (USG) and its effect on net monetary benefit of the USG diagnostic pathway in comparison with the other pathways (using a willingness to pay of $18,000 to decrease one error in diagnosis). LeB, Le Bonheur.
effective (OBS effectiveness ¼ 97%), although at a higher cost ($5,857; Table 7). As a result, CT had a higher ICER ($33,159) when compared with the optimal strategy of USG, making this strategy not cost-effective. In comparison with USG, the ICER for LeB-P ($17,206) was lower and the ICER for OBS ($51,877) was higher than that of CT, but with a higher effectiveness rate. Within the nondominated options, USG was the preferred strategy over LeB-P and OBS, based on its having the lowest ICER, despite being 2% less effective than LeB-P and 5% less effective than OBS. Selecting the LeBP over USG for workup of suspected appendicitis would cost the institution an additional $17,206 (the ICER) to eliminate one misdiagnosis. We next examined the monetary investment required to move the LeB-P to a position where our analytic model would consider it to be the optimal diagnostic strategy because of better accuracy at comparable cost. Our net monetary benefit calculation showed that at a willingness to pay of $18,000 to decrease one diagnostic error, the LeB-P becomes the preferred diagnostic pathway over USG, with a higher net monetary benefit. At this willingness-to-pay threshold of $18,000, sensitivity analysis showed that the LeB-P was the most cost-effective diagnostic option over a wide range of USG sensitivities, from 59% to 90% (Fig. 3). Similarly, if the specificity of USG is
