Our study results show that LA has a high probability of producing higher NMB than OA and NOM in patients with uncomplicated AA with a willingness to pay less than 1 GDP capita.
The lower cost of LA could explain this higher benefit. Previous studies have shown that LA might have lower costs than the open approach (5, 33–36). This reduced cost is explained by the shorter length of hospital stay, quicker postoperative recovery, and the lower probability of postoperative complications. In this aspect, the results of our study are consistent with those reported in the literature.
Likewise, our model revealed that the costs of LA are lower than NOM. This result differs from other published economic evaluations, showing that NOM is cost saving compared to surgical alternatives. This difference could be explained by the different time horizons of these studies. Sippola et al. followed some participants in the APPAC study for one year (3). On the other hand, Wu et al. designed a decision model with a 1-year time horizon (25). By adopting a shorter time horizon, the costs of patients who will recur after the first year would not be estimated. Salminen et al. demonstrated that the percentage of people who underwent appendectomy could increase from 27–40% after the first year (13). Of these patients, almost 10% might recur with complicated appendicitis, which implies higher costs subsequently. Sceats et al. revealed that NOM was associated with higher abscess rates, readmission, and higher overall care costs (37). As long as the appendix remains, the risk of recurrence is latent.
LA might represent significant savings for the health care system. Our model estimated a mean value of more than 2 million savings for every 100 000 LA performed instead of OA and more than 1 million compared to NOM. The probability of achieving these savings is greater than 64%. This saving would represent 2.3% and 1.4% respectively of the total health expenditure for many people in Colombia (35). From a financial perspective, LA is the best option for the health system.
A meaningful reduction in the cost of the NOM could become a more efficient alternative than LA. According to our model, this objective would be attained by reducing NOM costs by 55%. In this case, the main expense was represented by the antibiotic regimen. Potentially, the use of other lower-cost antibiotic regimens might improve the efficiency. Unfortunately, the lack of long-term comparative studies of antibiotic regimens is a limitation.
On the other hand, NOM provided a minimal increase in QALYs than the surgical alternatives. Our model revealed that NOM provided 0.003 and 0.0086 QALYs more than LA and OA, respectively. That marginal health benefit is equivalent to having 1 and 3.1 days in perfect health compared to LA and OA. Among the factors that could explain these differences are the lower probability of adverse events, less pain, and shorter absence from work than surgical options. In addition to a significant proportion who do not present symptoms or recurrence (38). Another reason is the surgery could affect the quality of life for up to a month after being carried out (39).
Despite the above results, NOM of uncomplicated appendicitis might represent an efficient alternative in some situations. The first situation would be in case of lack of surgeons in some regions. In this case, the additional costs of transporting the patient to a center would result in LA not being the most efficient alternative. Also, the delay in the treatment would increase the risk of complications. The second situation is when the only option available is OA. This situation is frequent in hospitals where there is neither the technology nor the surgeons trained to perform LA. Two other possible situations would be in patients with high surgical risk or a collapse of the health system due to extreme public health situations (e.g., COVID). However, these two situations were not evaluated by our model.
Our economic evaluation has several limitations. Firstly, we considered a time horizon of five years only. This consideration might underestimate costs or overestimate QALYs in NOM because subsequent events are not quantified. Nevertheless, the incidence of this disease tends to decrease in the older population (40, 41). Secondly, our evaluation did not include an analysis by the patients' subgroup. Possibly, NOM may be a more efficient alternative in patients with high surgical risk. To incorporate heterogeneity in baseline health states, we performed a patient-level simulation with different utilities. In a particular manner, these microsimulations could partially reflect worse health states due to comorbidities. Finally, our model did not consider LA's learning curve and its impact on costs, postsurgical complications, or conversion to OA. The model also did not value capital costs (i.e., fiber optics, monitors) or some of the supplies needed to perform LA. Published studies show that these factors can increase costs and, therefore, could decrease this surgical approach's efficiency (42, 43).
Several issues remain to be clarified. The first one is about the probability of recurrence after five years and the impact on the quality of life and costs. On the other hand, it is mandatory quantifying the alternatives' efficiency in populations with high surgical risk or difficulty (e.g., obesity, pulmonary diseases). Finally, it is necessary to incorporate into the new evaluations the effect on the costs of the LA learning curve's efficiency.