Modeling the Cost of Management Options for small benign non-functional adrenal incidentaloma: A Decision Tree Analysis

Objective Using decision analysis to determine the economic benefits of surgery versus surveillance for patients with small benign non-functional adrenal incidentaloma. Methods A detailed model was developed that integrates projected costs for a few possible treatment outcomes. Starting values and probabilities were derived from a comprehensive meta-analysis of 1298 patients. Hypothesis testing was performed using sensitivity analysis. Results The model predicts a cost premium for follow-up compared with surgery. The tornado analysis identifies the cost of follow-up for once as the most important variables for our model. One-way sensitivity analysis demonstrated that cost equivalence could be met with ¥4359.5 for once follow-up, keeping other factors constant. Conclusion Our models of treatment for small benign non-functional adrenal incidentaloma showed significant cost differences for follow-up versus immediate surgery. The cost of follow-up for once featured threshold values.


Introduction
Adrenal incidentaloma (AI) is the term for any unsuspected lesion of the adrenal gland found by radiological imaging, usually done for other indications. Since AI is a clinical dilemma, several organizations including the National Institutes of Health (NIH), the European Society of Endocrinology 1 (ESE) and the European Network for the Study of Adrenal Tumors (ENSAT) published guidelines to improve the evaluation and management of these patients. Although it is consistently recommended that hyper-functional tumor and suspicious malignancy need surgery, the management of patients with small benign non-functional adrenal incidentaloma remains controversial and some experts recommend surgery rather than surveillance [2][3][4] . Contrast to temporary guidelines, in our previous report 5 , silent malignancy and pheochromocytoma were not rare; also some studies demonstrated that follow-up was associated with a higher incidence of metabolic syndrome 6 . As small benign non-functional adrenal incidentaloma within initial assessment may progress to hyper-functional, larger size or malignancy plus anxiety is burdensome for some patients, both in economic and psychology, surgery could be anther choice besides surveillance.
Here we develop a detailed mathematical decision analysis model comparing the projected management costs for small benign non-functional adrenal incidentaloma. It is built on a meta-analysis 7 and reflects a multi-institutional clinical experience. The model incorporates a variety of clinical outcomes downstream from the original clinical decision ( Figure 1), given that these events would potentially influence costs.

Methods
Decision analytic models are methods of estimating and calculating outcomes by identifying the clinical question, disaggregating the problem into discrete units to include all reasonable choices and consequences, and assigning probabilities and value to the various events and outcomes. The decision model was designed and analyzed using TreeAge Pro software with the Healthcare suite (available http://www.treeage.com).
Surveillance versus surgery of small benign non-functional adrenal incidentaloma constituted the arms of the decision tree ( Figure 1).

Model assumptions
According meta-analysis from Loh HH 7 , we assume 1.5% of cases end up surgery for developing Cushing's syndrome or pheochromocytoma at the third year of follow-up, and the same percentage of cases accept surgeries at the fourth year. The rest of others complete the five-year follow-up without surgery.

Costs, charges, and time values
The surgeon's fee and anesthesiologist's fee, as well as the charges of pre-operation stay, post-operation stay and follow-up, were all derived from Zhuhai Peoples' hospital of Guangdong province in China in 2018. (Table 1) The costs for each treatment arm were calculated as the average costs for all patients in each arm and were cumulative from the beginning until the end of follow-up or surgery.
The annual inflation rate is not constant, we incorporate the approximate average number, 5%, into the cost.

Sensitivity analysis
To address the effect of the individual variables on our model and test the validity of our model assumptions, we performed 1-way sensitivity analysis on all model parameters. This was performed by individually varying each variable for a range of values, keeping all other variables constant. If the resulting costs were unchanged, the analysis was considered "insensitive" to the tested variable. Otherwise, we calculated the threshold(ie, "crossover point") at which a given value for the variable in question would change which treatment arm featured the lowest overall costs.
The variables that featured threshold values on 1-way analysis were also used for 2-way sensitivity analysis, in which 2 variables were varied simultaneously, and the thresholds were measured.

Base-case analysis
The base case analysis demonstrated follow-up featured the lowest overall cost per patient (¥7213) compared with immediate surgery (¥24691) (Figure 2).

Sensitivity analysis
The tornado analysis, which included 1-way sensitivity analysis of all parameters, was used to identify the most important variables for our model. The cost of follow-up for once was the most important because varying this parameter during the sensitivity analysis resulted in the greatest effects on costs (wider bars in the tornado diagram); other values that had a minimal effect on the analysis (Figure 3).
To identify the cost of follow-up for once resulting in cost equivalence between follow-up and immediate surgery, we performed a one-way sensitivity analysis. This analysis demonstrated that cost equivalence could be met with ¥4359.5 for once follow-up, keeping other factors constant (Figure 4).
Because the cost of follow-up for once was the only parameter demonstrating threshold value, we focused the 2-way sensitivity analysis on this set of parameters. Using two-way sensitivity analysis, cost of follow-up for once and length of post-surgery stay were varied to evaluate how these factors affected the choice of the least costly therapy option. For the majority of the cost of follow-up for once and length of post-surgery stay, the followup shows a clear cost advantage( Figure 5).

Discussion
When facing small benign non-functional adrenal incidentaloma, patients and urologists must decide between follow-up and surgery. This decision hinges on the critical issues of effectiveness, morbidity and cost. Both options have good outcomes. Only Belmihoub 8 reported one case that benign adrenal incidentaloma progressed adrenocortical carcinoma. Loh HH 7 reported that no surgery is done due to primary adrenal malignancy during follow-up. Likewise, the morbidity of surgery is mild 9-12 . Cost may, therefore, play an important role in directing clinical decisions in this disease. A detailed decision-tree model was developed for projecting the management costs for small benign non-functional adrenal incidentaloma. The model's advantage lies in the ease with which various hypotheses can be rapidly tested through the use of sensitivity analysis.
The current model includes treatment-related morbidity because the various cost derived from the actual charge from patients and/or medical insurance bureau. Three of 21 patients had surgery complication, which is similar to other reports 9-12 . One patient complicated major hemorrhage and transferred to open surgery, another one had urinary infection and fever, the third one had pneumonia after surgery. Although these three patients stayed longer than the average length of stay and cost more than the others, they were discharged with good health finally.
As a model based on temporary data, limitations are inevitable. The mortality difference between surveillance and immediate surgery is seemed to be minor, and the use of mortality as a measure of effectiveness is unavailable. Estimates based on such small and controversial differences in morbidity would mistake the results. And cost modeling is correlated with local environment, and utility of this model should be adjusted to other centers provided that institution-specific cost components are modified. Finally, because we examined the costs from the perspective of the medical establishment, indirect costs, such as psychological factor, were not included. No published work is available regarding the indirect costs of surveillance versus immediate surgery. This is an area worthy of additional study.

Conclusions
Our models about the management of small benign non-functional adrenal incidentaloma showed significant cost differences for follow-up vs immediate surgery. The cost of followup for once featured threshold values.

Competing interests
The authors declare that they have no competing interests.

Ethics approval and consent to participate
Due to the retrospective nature of this study, ethics approval by our Institutional Review Board was obtained and the data were used confidentially for research work. Informed consent was obtained by every patient when they referred to the hospital.

Declaration
The dataset supporting the conclusions of this article is included within the article and its additional files.