Academic medical centers are integral to cancer care and research. However, most cancer patients receive treatment at non-academic medical centers within their local community. The present study revealed a high-level of guideline adherence of MTB recommendations at a representative non-academic medical center in Austria. Predictors of guideline non-adherence included older age at diagnosis, a higher ECOG status, colorectal cancer versus breast cancer, and a more recent MTB conference.

Adherence to guidelines has been associated with improved outcomes, and guideline adherence of MTB recommendations is considered a quality criterion for cancer care (Walraven et al. 2019). However, time pressure, a high number of cases, lack of MTB leadership, as well as lack of information including imaging, and pathologic workup, have been associated with low-quality decision making (Lamb et al. 2011a). On average, 3.2 min are spent on each case in MTBs (Lamb et al. 2011b). Interestingly, most literature on the quality of MTB recommendations originates from academic medical centers, while most patients receive treatment in non-academic medical centers (Wong et al. 2020). Furthermore, data indicate that increased travel time is associated with a worse quality of life, which potentially presents a barrier to the implementation of highly centralized care (Ambroggi et al. 2015; Bühn et al. 2020). Therefore, results from studies conducted in academic settings may not accurately reflect real-world experiences.

In the present study, guideline adherence was 78.2%, while studies from academic medical centers report adherence rates between 37 and 100% (Jaap et al. 2018; Walter et al. 2023; Krause et al. 2022; Brauer et al. 2017). However, differences in study designs, patient populations, guidelines, and others must be considered. Additionally, several predictors of guideline deviations were identified. Each additional year of age was associated with an approximately 2% higher probability of guideline deviation, while a one-point increase of ECOG status corresponded to a more than 60% higher probability of guideline deviation. Thus, both results indicate a clinically relevant difference. While performance status is frequently used as an indicator of a patient’s fitness for undergoing treatment, evidence suggests that it might be a suboptimal indicator (Sedrak et al. 2020). Additionally, randomized controlled trials, which form the foundation for guidelines typically include a highly selected population with low ECOG status, which might not reflect real-world patient cohorts (Booth and Tannock 2014). Our findings thus emphasize the lack of high-quality evidence for cancer treatment in the elderly (Sedrak et al. xxxx).

The strongest effect estimate was observed in patients suffering from colorectal cancer in contrast to breast cancer, resulting in a fivefold increase in the probability that the MTB recommendation would deviate from established guidelines. Several factors could be responsible for this finding. The Landesklinikum Baden-Moedling is certified as a breast cancer center by national authorities, potentially contributing to a higher guideline adherence in breast cancer cases. Interestingly, new study results, such as those from the RAPIDO trial in colorectal cancer, also led to guideline deviations, highlighting that deviations may be justified based on emerging evidence (Bahadoer et al. 2021).

Previous research from academic medical centers suggests that a higher tumor stage is associated with a higher probability of guideline non-adherence (Ronden et al. 2021; Hines et al. 2015). In our study, a higher UICC stage was associated with a 6% increased probability of guideline deviation. However, 95% confidence intervals were fairly evenly distributed around the null value. Future studies with a larger population are necessary to evaluate tumor stage as a predictor of guideline deviation.

Notably, a more recent case discussion at the MTB was associated with a higher probability of guideline deviation. The reasons for this are difficult to evaluate in a retrospective study; however, it is important to consider that the AWMF guidelines are not updated annually, while new evidence from clinical trials is continuously emerging. To address this, we opted for a relatively short eligibility time span when evaluating adherence, thereby reducing potential sources of heterogeneity. Interestingly, in the subgroup analysis focusing on breast cancer, only a higher ECOG status and older age at diagnosis remained independent predictors while a more recent MTB conference was not identified as a predictor.

Furthermore, previous reports have suggested that female sex is associated with guideline non-adherence of MTB recommendations (Walter et al. 2023). However, while significant in univariable analysis, after adjusting for potential confounding factors, no such association could be confirmed in our study. This might be due to different tumor entities.

Several strengths are worth mentioning. To the best of our knowledge, this is one of the first study specifically assessing guideline adherence of MTB recommendations and predictors of deviations in a non-academic medical center, thereby addressing a critical knowledge gap. Furthermore, a previously reported classification system was used to assess guideline adherence, enhancing validity and comparability of our results. Additionally, clinically relevant predictors of guideline non-adherence of MTB recommendations were identified using rigorous statistical methodology (Riley et al. 2022).

However, several caveats need to be acknowledged. Guideline adherence in this study was based on the AWMF guidelines, which are specific to German-speaking countries. In contrast, some guidelines, such as the NCCN guidelines, are updated more frequently and may better incorporate recent developments in the field. For instance, the S3 guideline for colorectal cancer was published in 2019, while the S3 guideline for breast cancer was last updated in 2021. This difference in the dates of the last updates might partially explain the higher probability of guideline non-adherence observed in colorectal cancer cases. Additionally, assessment of guideline adherence remains subjective. We tried to account for this by relying on the system reported by Krause et al. as well as requiring at least two reviewers to independently assess guideline adherence (Krause et al. 2022). Also, this study was performed in a Western European country with a universal healthcare coverage. Thus, the applicability of the results to other healthcare settings requires external validation.

Additionally, the present study only included colorectal and breast cancer cases, as these comprise the large majority of oncological surgeries at our hospital. Thus, the present results do not generalize to other malignancies, in particular cases of high-complexity oncological care such as esophageal and pancreatic cancer, for which a well-known volume-outcome relationship exists (Gooiker et al. 2011; Voeten et al. 2021).

Furthermore, no data on comorbidities, such as the Charlson Comorbidity Index (CCI), were available. This is noteworthy as previous studies have identified comorbidities as a relevant factor for guideline deviation (Krause et al. 2022). However, we tried to account for this limitation and incorporated performance status into our model, which has been shown to correlate with comorbidities (Grose et al. 2011). Also, as this was a retrospective study, selection bias cannot be ruled out. Only patients presented at the MTB were considered in the present study, no data were available on patients with breast or colorectal cancer who were not presented at the MTB. Additionally, no conclusions can be drawn on the effect of guideline adherent MTB recommendations on survival.

Finally, while this study took place at a non-academic medical center, it is part of an integrated care network. Thus, specialists at the MTB often join virtually from other centers. New possibilities in telemedicine might thus provide an opportunity for high-quality local cancer care. This is in line with data from the United States indicating higher guideline adherence at integrated network cancer programs compared to community cancer programs (Thiels et al. 2019). Some evidence points to improved outcomes after joining care networks with an academic-community hospital collaboration model (Tucker et al. 2020).