September 2019 HSC Section 1 Congenital and Pediatric Problems

Am J Otolaryngol 39 (2018) 418–422

L.K. House et al.

invasive technique, some authors are recommending up front adenoi- dectomy with BCD in order to maximize disease resolution while minimizing risk [ 10 ]. As the cost of healthcare rises, the appropriate allocation of re- sources becomes more important. However, there is a strong desire to maintain or even improve outcomes while decreasing cost. Cost e ff ec- tiveness analysis is an important tool to evaluate the gains in health relative to the costs of various treatment strategies. Our study examines the cost e ff ectiveness of adenoidectomy alone versus adenoidectomy and upfront BCD for the management of pedia- tric CRS. A decision tree analysis was created to determine the incremental cost e ff ectiveness ratio of treating a hypothetical pediatric patient with known CRS who has failed maximal medical management. A literature search was performed using the PubMed database, using pediatric si- nusitis treatment, adenoidectomy, balloon sinuplasty, and balloon ca- theter dilation as keywords. We eliminated articles with low power or small sample sizes. Data was tabulated from the literature regarding the e ff ectiveness of each procedure, which was used to estimate the success rate of each procedure. The values used and their sources are listed in Table A . An average of these statistics was utilized for the outcomes of BCD, adenoidectomy, and FESS, by averaging all of the success rates from the sources listed in Table A [ 9 – 18 ]. The success rate used for adenoidectomy and BCD was assumed to be at least as successful as BCD, therefore the success rate for BCD was used. Initial treatment arms include adenoidectomy alone versus ade- noidectomy with up-front BCD. Three separate decision trees were created ( Figs. A1 – A3 ). In the fi rst, a patient who fails either adenoi- dectomy alone or adenoidectomy with BCD will proceed to FESS. In the second, a patient who fails adenoidectomy alone or adenoidectomy with BCD will proceed to BCD or repeat BCD. In the third, a patient who fails adenoidectomy will go on to BCD, and a patient who fails ade- noidectomy with BCD will be observed. The study makes several assumptions, which are listed in Table B . The study assumes 100% prevalence of disease in the patients who are treated. All workup and medical management, including computed tomography (CT) scanning, allergy testing, cultures, and antibiotic therapy are excluded from the cost analysis. For simplicity and because of lack of data in dropout rates, the study also assumes that all patients will be taken either to resolution or to the end of the treatment arm. The University of Mississippi Medical Center professional hospital and billing departments provided cost data for surgical treatments, which was based on average Medicaid reimbursements. Both profes- sional and hospital fees were included in the study. All preoperative workup costs were excluded from analysis. The costs are listed in Table C , rounded to the nearest US dollar. The cost of BCD included only the cost of treating the bilateral maxillary sinus ostia. The cost of FESS included the cost of treating bilateral maxillary, as well as anterior ethmoid sinuses. The incremental cost e ff ectiveness ratio (ICER) was calculated for adenoidectomy plus up-front BCD against adenoidectomy alone. The ICER is a commonly used equation in healthcare to guide decision- making and represents the ratio of di ff erence in cost to the di ff erence in 2. Materials and methods

e ff ectiveness.

= − − ICER (C C )/(E E ) 1 2 1 2 “ C ” and “ E ” are the expected cost and e ff ectiveness of the respective testing strategies. Generally, the C 2 and E 2 represent a control group. In this study, the control is the arm that uses a stepwise approach, starting with adenoidectomy alone. The units are cost per percent e ff ectiveness, which allows the user to compare costs without compromising health- care. A higher number indicates that the procedure is more expensive per unit of e ff ectiveness (see Table D ). A sensitivity analysis was also performed. A sensitivity analysis is used to determine how di ff erent values of the independent variables within a study will impact the results. Cost e ff ective analyses are based on certain assumptions, which are not always accurate. Sensitivity analysis is a formal way to measure the weaknesses or uncertainties of the study. Our variables were changed su ffi ciently to either a ff ect the outcome or were taken to their limit (i.e. 100% e ff ectiveness). The cost of CT scan following a “ failed ” fi rst procedure was also included in the sensitivity analysis to determine if this would change the outcome. The threshold value represents the value at which the ICER outcome changes. This allowed for discrepancies in success rates from the lit- erature, as well as di ff erences in cost data from various institutions and for various insurance carriers. Adenoidectomy as the sole fi rst procedure was found to be more cost e ff ective in all three decision trees. The weighted costs in Tree 1, where those with symptoms after the initial treatment went on to have FESS, were $4349 and $6792 for adenoidectomy and adenoidectomy plus BCD, respectively (see Table D ). The adenoidectomy plus BCD arm was 0.03% more e ff ective in the end (98% vs. 95%), but with an 81, 431 incremental cost. This amounts to $81,431 for each additional patient e ff ectively treated with balloon sinuplasty. The weighted costs are only slightly di ff erent for the second and third trees. The adenoi- dectomy arm has a weighted cost of $4133 and $4100 for trees 2 and 3, and adenoidectomy plus BCD costs $6705 and $5835 in trees 2 and 3. The ICER remains high in tree 2 at 51,445. For tree 3, the ICER is − 17,189, indicating that this treatment paradigm is less e ff ective overall. The sensitivity analysis found that when the values for the in- dependent variables were changed within reasonable limits, the ICER could be changed for most of the variables (see Table E ). In terms of cost, neither reasonably increasing the charge for adenoidectomy nor decreasing the charge for BCD changes the outcome in any of the de- cision trees. The threshold value for decreasing the cost of BCD is over $2000 for tree 1 and nearly $2000 for tree 3. The threshold value for increasing the cost of adenoidectomy to make it less cost e ff ective is over $4000 for trees 1 and 2 and nearly $4000 for tree 3. Shifts as large as these are unlikely to occur. We attempted to add in the cost of computed tomography after failure of the initial intervention to try to increase the cost e ff ectiveness of up-front BCD, but the threshold value was approximately $9000 per CT scan. Di ff ering e ff ectiveness was also evaluated in the sensitivity analysis. Adenoidectomy would have to be less than 20% e ff ective in order to be less cost e ff ective than the combined procedure. A success rate this low is not supported by the literature. Varying the e ff ectiveness of BCD in the sensitivity analysis never reached a threshold value making BCD more cost e ff ective and is listed as not feasible in Table E . 3. Results

Table A Parameters used in the decision tree analysis.

4. Discussion

Success rate used

Sources

Adenoidectomy

0.55 0.82 0.88 0.82

9,10,11,12,13

Chronic rhinosinusitis in the pediatric population remains a major source of morbidity for these patients and continues to be a challenging problem for pediatric otolaryngologists. Because of concern for com- plications, both immediate and future, surgery is often o ff ered after

Adenoidectomy plus BCD

10,14,17

FESS BCD

13,15,16,18

10,14,17

129