Long-term survival after treatment in certified lung cancer centers and not certified hospitals: Results of a large German cohort study using clinical routine data

Abstract

Objective

Lung cancer represents the second most frequent tumor entity worldwide with an increasing number of patients treated in specialized centers. The aim of the WiZen study was to find out whether treatment at hospitals certified by the German Cancer Society (GCS) was associated with long-term survival benefits.

Methods

Data for this cohort study was derived from the largest German statutory health insurance (SHI) AOK, four regional population-based clinical cancer registries (CCR), and standardized hospital quality reports. The analyses were based on 173,999 incident lung cancer patients in the SHI dataset and 35,702 patients in the CCR dataset who received primary treatment for lung cancer (ICD-10-GM C33, C34, D38.1) between 2009 and 2017.

Results

Distributions of age, sex, comorbidities, and most tumor characteristics were similar among patients treated in certified and non-certified hospitals. The Kaplan-Meier estimator for 5-year overall survival was 28.0% for patients from certified and 16.9% from non-certified hospitals (SHI data; CCR data: 21.4% vs. 13.6%). Cox-regression adjusting for relevant confounders yielded hazard ratios of 0.97 (SHI data; 95%CI 0.94, 1.00) and 0.85 (CCR data; 95%CI 0.82, 0.88) for all-cause mortality. The adjusted hazard ratio for recurrence-free survival (CCR data, UICC stage I-III, R0 only) was 0.82 (95%CI 0.75, 0.90).

Conclusions

The presented analyses show that treatment in certified lung cancer centers is associated with relevant survival benefits and should therefore be supported.

 

MissingLinkText

Zusammenfassung

Ziel

Lungenkrebs stellt weltweit die zweithäufigste Tumorerkrankung dar und eine zunehmende Anzahl von Patienten wird in spezialisierten Zentren behandelt. Die WiZen-Studie untersucht, ob die Behandlung in von der Deutschen Krebsgesellschaft (DKG) zertifizierten Krankenhäusern mit langfristigen Überlebensvorteilen verbunden ist.

Methoden

Die Daten für diese Kohortenstudie stammen von der größten deutschen gesetzlichen Krankenversicherung (GKV) AOK, vier regionalen, bevölkerungsbezogenen klinischen Krebsregistern (KKR) und standardisierten Qualitätsberichten der Krankenhäuser. Die Analysen basieren auf 173.999 Patienten mit neu diagnostiziertem Lungenkrebs im GKV-Datensatz und 35.702 Patienten im KKR-Datensatz, die zwischen 2009 und 2017 eine Primärbehandlung ihres Lungenkrebses (ICD-10-GM C33, C34, D38.1) erhielten.

Ergebnisse

Ähnliche Verteilungen von Alter, Geschlecht, Begleiterkrankungen und den meisten Tumoreigenschaften wurden bei Patienten in zertifizierten und nicht-zertifizierten Krankenhäusern beobachtet. Der Kaplan-Meier-Schätzer für das 5-Jahres-Gesamtüberleben lag bei 28,0% für Patienten in zertifizierten und bei 16,9% in nicht-zertifizierten Krankenhäusern (GKV-Daten; KKR-Daten: 21,4% vs. 13,6%). Die Cox-Regression, adjustiert für relevante Störfaktoren, ergab Hazard Ratios von 0,97 (GKV-Daten; 95% CI 0,94, 1,00) und 0,85 (KKR-Daten; 95% CI 0,82, 0,88) für die Gesamtmortalität. Das adjustierte Hazard Ratio für das rezidivfreie Überleben (KKR-Daten, UICC-Stadium I-III, nur R0) lag bei 0,82 (95% CI 0,75, 0,90).

Schlussfolgerung

Die vorgestellten Analysen zeigen, dass eine Behandlung in zertifizierten Lungenkrebszentren mit relevanten Überlebensvorteilen verbunden ist und daher gefördert werden sollte.

 

MissingLinkText

Keywords

Certified cancer center - specialized treatment - lung cancer - cohort study - survival - quality of cancer care

Schlüsselwörter

Zertifiziertes Krebszentrum - spezialisierte Behandlung - Lungenkrebs - Kohortenstudie - Überleben - Qualität der Krebsversorgung

Introduction

With 2.21 million incident cases in 2020, lung cancer is globally the second most frequent malignancy [1]. This is reflected by the epidemiological situation in many countries like Germany: According to the latest national cancer report [2], 56,577 persons were diagnosed with lung cancer in 2022 and the incidence rates, especially among the female population, are still expected to rise.

With respect to the high complexity of lung cancer treatment [3] [4] [5], many countries have been developing programs to promote the establishment of specialized cancer centers [6] [7] [8] [9] [10]. In Germany, the German Cancer Society (GCS; German: Deutsche Krebsgesellschaft, DKG) has been offering organ-specific certifications programs since 2003 [11]. Currently, there exist 18 different GCS-certification programs which are also implemented in other European countries like Austria, Italy, and Switzerland. Among 1,960 GCS-certified centers in Germany [12], 79 are specialized in lung cancer and 3 are currently in the application process [13]. To achieve GCS-certification, a hospital must meet a wide range of requirements. These include structural measures such as regular interdisciplinary communication, consensus decision-making in structured tumor boards, and maintaining multi-professional outreach networks. Lung cancer centers must regularly demonstrate their expertise in structured survey forms regarding diagnostic procedures, radiation therapy, nuclear medicine treatments, drug-based and surgical tumor therapy, as well as palliative care. They undergo regular external audits, with their performance indicators published in publicly available annual quality reports by the GCS [14].

However, sound evidence for better patient outcomes in certified lung cancer centers compared to other hospitals is still missing due to a lack of studies on the topic. Previous studies with colorectal [15] [16] [17] and prostate cancer [18] [19] patients point towards better functional and prognostic outcomes after treatment in GCS-certified cancer centers but are subject to potential restrictions concerning their internal and external validity. To overcome issues like limited regional coverage, small sample sizes or insufficient information on patient- and hospital-level, the WiZen-study (German Innovation Fund, grant number 01VSF17020) has been initiated. Using medical routine data provided by the largest German statutory health insurance company and four large clinical cancer registries, it has been the goal of this large cohort study to compare patient survival after treatment in GCS-certified cancer centers and non-certified hospitals for eight different tumor entities. The present publication aims to provide an in-depth overview of the study’s lung cancer specific results.

 

MissingLinkText

Patients and methods

The WiZen-study

The WiZen-study was conducted as a set of retrospective cohort studies from July 1st, 2018, to August 31st, 2021 by four distinct institutions: Zentrum für Evidenzbasierte Gesundheitsversorgung (ZEGV)/ Hochschulmedizin Dresden, Tumorzentrum Regensburg (TZR), Arbeitsgemeinschaft Deutscher Tumorzentren e. V. (ADT), and Wissenschaftliches Institut der AOK (WIdO). Collaboration partners included GCS, Klinisches Krebsregister Dresden (KKRD), Klinisches Krebsregister Erfurt (KKE), and Klinisches Krebsregister für Brandenburg und Berlin (KKBB). The primary objective was to investigate whether receiving treatment at GCS-certified cancer centers for breast cancer, colorectal cancer, gynecological cancer, head and neck cancer, lung cancer, neurooncological tumors, pancreatic cancer, and prostate cancer is correlated with improved overall survival. To answer this question, data from different sources was analyzed: i) statutory health insurance (SHI) data provided Germany’s largest health insurance company “Allgemeine Ortskrankenkassen” (AOK), ii) clinical cancer registry (CCR) data, provided by four large population-based clinical cancer registries, iii) GCS-certification reports. More details about these data sources can be found in the original study protocol of the WiZen-study, which was registered on ClinicalTrials.gov (identifier: NCT04334239) and is also available online at the following link: https://innovationsfonds.g-ba.de/beschluesse/wizen-wirksamkeit-der-versorgung-in-onkologischen-zentren.111 and in existing publications [20] [21] [22] [23] [24] [25] [26] [27] [28]. The present publication focuses on the study’s lung cancer specific results.

 

MissingLinkText

Inclusion and exclusion criteria

All results reported in this paper refer to patients with diagnosis of incident lung cancer according to the ICD-10-GM codes C33 (malignant neoplasm of trachea), C34 (malignant neoplasm of bronchus and lung), and D38.1 (neoplasm of uncertain behavior of trachea, bronchus, and lung, SHI data only). Moreover, the following additional criteria had to be fulfilled to be included either in the SHI- or the CCR-based analyses: a) patients had to be at least 18 years of age at the time of diagnosis, b) the survival-time had to be greater than zero, c) no previous diagnoses of lung cancer were allowed (this information was directly available for the CCR data; concerning the SHI data, a patient was considered as incident between 2009 and 2017 only if there was no diagnosis of lung cancer between 2006–2008 following the guideline “good practice of secondary data analysis” [29]; for this reason patients with a cancer diagnosis between 2006 and 2008 were excluded both from the SHI and the CCR data based analyses), d) there had to be sufficient information concerning the certification status of the treating hospital.

In addition to this, the following criteria have been applied to the SHI-data: e) there had to be a continuous AOK-insurance over a patient’s entire observation period, f) there had to be at least one primary inpatient diagnosis according to one of the above mentioned diagnostic codes, g) patients treated in a hospital which became GCS-certified within one year subsequent to primary treatment (and, therefore, is likely to have already fulfilled the quality standards of certification, although it would be analyzed as part of the non-certified group) were excluded. Concerning the CCR-dataset, patients whose histological subtype was inconsistent with the tumor entities of interest (e. g. lymphoma or sarcoma) were additionally excluded.

 

MissingLinkText

Statistical analysis

Both data sources have their advantages: While SHI data e. g. provide comprehensive information about a patient’s comorbidities, CCR data can give detailed insights in tumor characteristics. However, since the overlap between the patient collectives in both data sources would have been comparatively small, it was decided to conduct the analyses separately and interpret the results together. Thus, the maximum amount of information could be extracted.

Patients with incident lung cancer were considered “certified lung cancer center patients” a) if primary tumor resection (documented by the OPS codes 5-320 – 5-325 and 5-327 – 5-329, together with a primary inpatient diagnosis according to ICD-10-GM C33, C34, D38.1) was performed in a certified lung cancer center or a directly associated hospital, or – in the absence of a documented primary resection – b) if the first lung cancer specific inpatient treatment (documented by a primary inpatient diagnosis according to ICD-10-GM C33, C34, D38.1) took place in a certified lung cancer center or a directly associated hospital.

The primary outcome assessed was overall survival, with recurrence-free survival (including local recurrence and recurrent distant metastases, analyzed based on CCR data only) as the secondary outcome. Each enrolled patient was considered at risk of death or tumor recurrence from the date of the index treatment (using SHI data) or diagnosis (using CCR data) onward. The follow-up duration extended until the occurrence of death or tumor recurrence. In instances where the specified outcomes did not occur, the patient's follow-up time was right-censored. The observation period for all patients concluded on December 31st, 2017. The survival time was expressed in years in all statistical analyses. To compare unadjusted survival rates between GCS-certified lung cancer centers and non-certified hospitals within the initial five years after the index treatment, the Kaplan-Meier method was employed.

To account for the potentially unbalanced distribution of important confounders, multivariable Cox-regression models were developed. In the CCR-based analyses, it was possible to adjust for age (categorized in groups based on epidemiological considerations: 18-59, 60-79, 80+), sex, year of diagnosis, histological subtype, UICC-stage, grade, lymphatic, and venous invasion. For the SHI-based analyses, the following covariates were included in the model: age (categorized in groups based on epidemiological considerations: 18-59, 60-79, 80+), sex, year of index treatment, distant metastasis, relevant comorbidities selected by a panel of independent clinical experts (categorized according to Elixhauser [30]), and hospital characteristics (case-loads, academic status, ownership). Moreover, a shared frailty term was included in the model to account for correlation between outcomes of patients treated in the same hospital [31].

All significance tests were conducted as two-sided tests with a significance level set at 0.05. The reported results include the associated p-value and/or the upper and lower boundaries of the 95% confidence interval. R, version 3.6.3, was utilized for the analyses based on SHI data. IBM SPSS 25 (IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY, USA: IBM Corp.) was employed for the analyses based on CCR data. The findings of this survey are presented in strict compliance with the Strengthening the Reporting of Observational studies in Epidemiology (STROBE) statement [32] (supplementary table S1, online). The study was also listed at ClinicalTrials.gov (identifier: NCT04334239).

 

MissingLinkText

Data protection and ethics

Information related to GCS-certification, patient details, tumor characteristics, and hospital attributes underwent pseudonymization at WIdO and the collaborating cancer registries. The pseudonymized data were subsequently analyzed at ZEGV (SHI) and TZR (CCR). The WiZen-study received approval from the ethics committee of TU Dresden (approval number: EK95022019). Data processing and analyses adhered to the principles outlined in the Declaration of Helsinki and the General Data Protection Regulation of the European Union.

 

MissingLinkText

 

MissingLinkText

Results

Inclusion process

Between 2009 and 2017, the SHI- and the CCR-dataset contained 304,751 and 36,069 patients, respectively, with the ICD-10-GM diagnoses C33, C34, or D38.1. After the application of all inclusion criteria, 173,999 patients (57.1%) from the SHI and 35,702 patients from the CCR-dataset (99.0%) met all eligibility criteria and were included in the study. ([Fig. 1]).

Fig. 1 Inclusion and exclusion of patients (a) SHI data, b) CCR data).

 

MissingLinkText

Share of patients treated in GCS-certified colorectal cancer centers

According to the SHI data, the share of patients treated in GCS-certified lung cancer centers was 3.9% in 2009. It gradually rose to 29.3% in 2017 ([Fig. 2]; analyzing the CCR data, a gradient from 0.9% in 2009 to 43.1% in 2017 was seen.). Moreover, the rate of center treatment over the whole observation period differed substantially between the different regions (defined by German federal states): It ranged between 1.9% and 64.1% (supplementary Table S2, online).

Fig. 2 Share of patients treated in GCS-certified lung cancer centers (SHI and CCR data).

 

MissingLinkText

Description of collectives

While age groups were distributed almost equally among the patients of certified and non-certified hospitals in the CCR-dataset, a slightly higher share of patients aged 80 and older was seen in the non-certified group of the SHI-dataset (16.5% vs. 11.5%, [Table 1]). An overview of Elixhauser comorbidities of the patients in the SHI dataset considered relevant by independent clinical experts is provided in online supplementary Table S3. Within the SHI dataset, more patients of non-certified hospitals already had distant metastases at the time of diagnosis ([Table 1]). Concerning further tumor characteristics, the CCR-dataset provides deeper insights: Again, there was a higher share of UICC stage IV tumors among patients of non-certified hospitals (49.7% vs. 43.8%) and patients of certified hospitals were more likely to have low stages (UICC I/0: 19.2% vs. 12.4%; [Table 2]). Generally, GCS-certified cancer centers had more hospital beds, were slightly less often privately owned, and had more frequently the status of a teaching or even university hospital than non-certified hospitals ([Table 3]).

Table 1 Patient characteristics (SHI and CCR data).

	SHI data				CCR data
treatment in GCS-certified centers	yes		no		yes		no
	n	%	n	%	n	%	n	%
sex
female	11,953	34.3	45,876	33.0	2,814	30.8	7,696	29.0
male	22,931	65.7	93,239	67.0	6,325	69.2	18,867	71.0
age
mean (median)	67.5 (68.0) years		69.1 (70.0) years		68.0 (68.8) years		68.4 (69.6) years
18 – 59	7,919	22.7	27,375	19.7	2,083	22.8	6,088	22.9
60 – 79	22,962	65.8	88,745	63.8	6,063	66.4	17,088	64.3
80+	4,003	11.5	22,995	16.5	993	10.9	3,387	12.8
distant metastasis
yes	13,513	38.7	68,427	49.2	4,003	43.8	13,191	49.7
total	34,884	100.0	139,115	100.0	9,139	100.0	26,563	100.0

Table 2 Tumor characteristics (CCR data).

Treatment in GCS-certified centers	yes		no
	n	%	n	%
histological entity
SCLC	1,366	14.9	4,433	16.7
NSCLC	7,773	85.1	22,130	83.3
UICC stage
I/0	1,757	19.2	3,294	12.4
II	870	9.5	1,681	6.3
III	2,202	24.1	4,715	17.8
IV	4,003	43.8	13,191	49.7
X	307	3.4	3,682	13.9
grade
G1	300	3.3	658	2.5
G2	2,491	27.3	6,343	23.9
G3/4	3,216	35.2	9,204	34.6
GX	3,132	34.3	10,358	39.0
lymphatic invasion
L0	1,701	18.6	3,292	12.4
L1	673	7.4	1,293	4.9
LX	6,765	74.0	21,978	82.7
venous invasion
V0	1,965	21.5	3,806	14.3
V1/2	402	4.4	751	2.8
VX	6,772	74.1	22,006	82.8
total	9,139	100.0	26,563	100.0

Table 3 Hospital characteristics (SHI data).

Treatment in GCS-certified centers	yes		no
	n	%	n	%
hospital beds
1 – 299	21	34.4	631	57.8
300 – 499	13	21.3	269	24.7
500 – 999	17	27.9	150	13.7
1000+	10	16.4	41	3.8
hospital ownership
public	20	32.8	399	36.6
non-profit	32	52.5	459	42.1
private	9	14.8	233	21.4
academic status
university hospital	6	9.8	25	2.3
teaching hospital	47	77.0	605	55.5
total	61	100.0	1,091	100.0

 

MissingLinkText

Overall survival, Kaplan-Meier analyses

The median follow-up time over all patients without an event regardless of certification status was 2.6 years (SHI data, CCR data 3.4). Within the SHI data, the 5-year Kaplan-Meier survival rate over all patients was 28.0% (95%CI 27.4%, 28.6%) for GCS-center patients and 16.9% (95%CI 16.7%, 17.1%) for patients from other hospitals ([Fig. 3a]). The difference between the two survival curves was highly significant (p<0.001). Similar results were seen in the CCR dataset: Here, the 5-year survival rates were 21.4% (95%CI 20.0%, 22.9%) for center and 13.6% (95%CI 13.1%, 14.1%) for non-center patients (p<0.001, [Fig. 3b] and supplementary table S4, online).

Fig. 3 Kaplan-Meier curves and Cox regression analysis for overall survival (OAS). (a) Kaplan-Meier survival curves OAS (SHI data). (b) Kaplan-Meier survival curves OAS (CCR data). (c) Unadjusted and adjusted hazard ratios with 95%CI for all-cause mortality (SHI and CCR data) following treatment in GCS-certified lung cancer centers compared to treatment in non-certified hospitals. * adjusted for age, sex, year of index treatment, distant metastasis, Elixhauser comorbidities, and hospital characteristics ** adjusted for age, sex, year of diagnosis, histological subtype, UICC-stage, grade, lymphatic, and venous invasion.

 

MissingLinkText

Overall survival, Cox-regression analyses

Within the SHI data, the unadjusted hazard ratio (HR) over all patients for all-cause mortality with shared frailty on hospital level was 0.90 (95%CI 0.87, 0.93) for treatment in a GCS-certified cancer center. After adjustment for age, sex, year of index treatment, distant metastasis, Elixhauser comorbidities, and hospital characteristics, it rose to 0.97 (95%CI 0.94, 1.00, [Fig. 3c]), missing significance slightly. The results for all covariates contained in the adjusted Cox-regression model can be found in online supplementary Table S5. Analyzing patients without distant metastases separately, significant advantages for center-based treatment were seen (HR 0.92, 95%CI 0.88, 0.96). These and further sensitivity analyses with stratification for age and sex can be found in online supplementary Table S6.

In the CCR data set, the unadjusted hazard ratio for all-cause mortality across all patients was 0.73 (95% CI 0.71, 0.75) when treated in a GCS-certified cancer center. Following adjustment for variables including age, sex, year of diagnosis, exact ICD-10-GM diagnosis, UICC stage, grade, lymphatic and venous invasion, this ratio changed to 0.85 (95% CI 0.82, 0.88, as shown in [Fig. 3c]; detailed results for all covariates in the adjusted Cox-regression model are available in supplementary Table S7, online). These findings indicate a statistically significant advantage in overall survival for those treated in a GCS-certified lung cancer center (p<0.001). This result remained stable in the sensitivity analyses: After the exclusion of patients with unknown UICC-stage, the hazard ratio for all-cause mortality was still 0.85 (95%CI 0.82, 0.88) in favor of treatment in GCS-certified cancer centers. Concentrating on stage I to III patients only, the effect was even more pronounced (HR 0.83; 95%CI 0.78, 0.88). Generally, the effect of center-treatment was stronger in lower (UICC I/0: HR 0.77, UICC II HR 0.70) than in higher stages (UICC III: 0.89, UICC IV: 0.97). Analyzing NSCLC- and SCLC patients separately, the hazard ratio for all-cause mortality was 0.84 (95%CI 0.81, 0.88) and 0.87 (95%CI 0.81, 0.94), respectively. The detailed results of these and further sensitivity analyses can be found in online supplementary Table S6.

 

MissingLinkText

Recurrence-free survival, Cox-regression analyses

Using the CCR data it was also possible to analyze recurrence free survival of R0-resected stage I – III patients; the adjusted hazard ratio for death or tumor recurrence was 0.82 (95%CI 0.75, 0.90). Further subgroup analyses concerning recurrence free survival can be found in online supplementary Table S8.

 

MissingLinkText

 

MissingLinkText

Discussion

Summary

The analyses based on SHI and CCR data yielded significantly higher survival rates for patients treated in GCS-certified lung cancer centers. After adjustment for important covariates like age, comorbidities, or the year of index treatment in the SHI-based analyses, the hazard of death was 5% lower in patients treated in certified lung cancer centers, although this difference missed the pre-defined significance level of 0.05 slightly. In the CCR-based analyses, the hazard of death after treatment at a GCS-certified lung cancer center was significantly lower by 15% after adjustment for a variety of patient- and tumor characteristics.

The observed divergence of the certification effect between the country-wide SHI-based analyses and the CCR-based analyses for four regions from the South-East is most likely to be attributed to regional heterogeneity on different levels: Analyzing epidemiological data from the observation period reveals substantial differences concerning incidence, mortality and relative survival rates for lung cancer between the different federal states [33]. For example, the 5-year relative survival rate of male lung cancer patients was 13.8% in the federal state of Mecklenburg-Vorpommern in 2009 compared to 25.2% in the federal state of Bavaria in 2017. At the same time, the rate of patients treated in certified institutions observed in the present study differed substantially between the federal states. This might be attributed to substantial structural differences between the federal states. While the accessibility of certified institutions is very high in urban states like Berlin or Hamburg, there are other states like Mecklenburg-Vorpommern or Bavaria with a high proportion of rural regions and a consecutively lower accessibility to certified institutions. Considerable socioeconomic differences between the different federal states might also play a role in this context. While certified cancer centers are already well established in the treatment of other tumor entities like breast or colorectal cancer, the number of lung cancer centers is comparatively low: In 2012 only 34 certified institutions existed, rising to a number of 49 in 2017 [13]. Currently, there are 79 lung cancer centers in Germany and the majority of lung cancer treatment still takes place outside of certified hospitals. Taking also into account the very low median life expectancy of lung cancer patients in general it becomes obvious that regional differences concerning epidemiological or structural aspects might have influenced the presented analyses substantially. Future research should therefore aim to analyze lung cancer treatment with special respect to regional disparities.

Notwithstanding this, the results of the present study show that after the exclusion of patients with distant metastases the survival benefit following center treatment was more pronounced and also significant in the SHI data-based analyses. This might be explained by the fact that prolonging survival might not always be the primary goal of the treatment of UICC-stage IV patients and the spectrum of possible treatment modalities is limited in the presence of distant metastases; at the same time, these patients represent approximately 50% of all lung cancer patients which could explain the absence of a significant survival benefit in the SHI-based main analysis.

 

MissingLinkText

Strengths and limitations

Unlike smaller previous studies on the topic, the WiZen-study’s analyses are based on a truly comprehensive patient cohort. In total, more than one million patients with eight tumor entities were included. While many studies exclude patients with unfavorable characteristics like high age or advanced tumor stages [34] [35], all these patients remained part of the study collective. Thus, the reported results are truly population-based. This is also reflected by the fact that the population characteristics (age-, sex-, and stage-distribution) and the observed survival rates reported in this paper are consistent with the figures presented in the national epidemiological cancer report of Germany [36].

Ideally, one would like to conduct a randomized controlled trial to analyze the effectiveness of center-based cancer treatment. Obviously, such a study design cannot be realized in this specific context: Patients select their hospital based on a variety of factors like referral by other health professionals, advice from fellow-patients, regional accessibility, individual preferences, and many more [37]. A patient’s socioeconomic status and her or his health literacy are also an important factor in this context. Subjecting people to a random allocation process would represent a severe restriction of their right of self-determination and has to be regarded as highly unethical. Therefore, the selected observational study design with independent standardized controlled prospective data collection in a large health insurance database and several comprehensive clinical cancer registries can be regarded as most adequate [25] [38]. However, future studies on the topic could try to reconstruct the patients’ and their treating ambulatory physicians’ preferences by means of artificial intelligence and implement them in the analyses [39].

With observational data and non-randomized group allocation, it is always important to correct for a potentially imbalanced distribution of potential confounders. Since data from two different sources was available, it was possible to integrate an extraordinarily large amount of different patient- and tumor-associated factors in the multivariable analyses. The reported results can be expected to reflect the cause-specific certification effect, although this assumption cannot be tested empirically. For certain variables, especially tumor characteristics in the CCR dataset, information was missing – predominantly in patients from non-certified hospitals. However, in the performed sensitivity analyses patients with incomplete information were excluded and the results remained stable.

To achieve reliable and valid results, the WiZen-study followed a conservative approach: In the detailed results of the SHI-based Cox-regression analysis (cf. supplementary table S5, online) one can see that the regression models attributed the certification effect partially to other confounders like hospital size (300-499 beds: HR 0.93, 500-999 beds: HR 0.86, 1000+beds: HR 0.87) , status as a teaching hospital (HR 0.99), or university hospital (HR 0.83), whose distribution is highly collinear to the status of certification. Due to the fact that the share of center-treatments rose over time, it is possible that adjusting for the year of index treatment eliminated part of the certification-effect, too. The center-effect might have been further under-estimated since patients treated in a hospital which forms part of an association with a GCS-certified cancer center have been regarded as center-treated although their treatment might still not have met center-standard.

 

MissingLinkText

The value of certification

The question whether concentrating (lung) cancer treatment in specialized centers or hospitals with large caseloads has been subject to international research for many years: In 2001, Bach et al. [40] analyzed over 2,000 patients from the US National Inpatient Sample and found that receiving a lung-resection in large-volume hospitals is associated with a 45%-reduction of postoperative complications, postoperative mortality (-50%), and better 5-year overall survival (+33%). Similar results were obtained from a national cohort analysis in England: In 2015, Møller et al. [41] showed that patients who received lung cancer surgery in a hospital belonging to the highest quintile concerning surgical volume “had about half the odds of death within 30 days than patients from the lowest quintile.” A recent nation-wide analysis from South Korea also showed that short- and long-term survival is significantly better for patients treated in high-volume hospitals [42].

While these findings generally support the results of this study, it has to be stated, though, that the GCS-certification goes far beyond a simple case-load based classification of hospitals and of course the treatment of lung cancer, especially of advanced stages, encompasses far more than just surgery. Therefore, it is very interesting to compare the results of this study to the findings of several studies from the US concentrating on the treatment in NCI (National Cancer Institute)-designated cancer centers, which have been established from 1971 onwards [9]. In 2005, Birkmeyer et al. [43] analyzed 63,860 patients aged 65-99 years with lung, esophageal, gastric, pancreatic, bladder, and colon carcinoma. They compared the patients from 51 NCI-designated cancer centers to patients from 51 high-volume hospitals without designation; in this retrospective cohort study treatment in designated centers was associated with lower adjusted surgical short-term mortality rates (pulmonary resection: 6.3% vs. 7.9%; p=0.010), although no significant difference in terms of long-term survival could be observed. In contrast to this, Onega et al. [44] found out that treatment for lung, breast, colorectal and prostate cancer in an NCI-designated cancer center is associated with “a significant reduction in the odds of 1- and 3-year all-cause and cancer-specific mortality”. In 2021, Okawa et al. [45] published their registry-based observational study about the situation in Japan: There also exist accredited high-capacity, high experience cancer care hospitals; treatment at these intuitions has been associated with the higher adjusted all-site 3-year survival probabilities (86.6%) in contrast to non-designated hospitals (78.8%). A systematic review informing the recently published “European Respiratory Society Guideline on various aspects of quality in lung cancer care” could also show that treatment in specialized lung cancer centers leads to lower 30- and 90-day morbidity and mortality and better long-term survival [46], which is absolutely in-line with the results of the present study.

Furthermore, it is essential to acknowledge that the landscape of metastatic lung cancer care has undergone a profound transformation in recent years following the introduction and widespread utilization of immunotherapy as well as second/third-generation targeted agents. It's noteworthy that these advancements occurred after the completion of data collection for the present study in 2017. Given this evolving scenario, it appears probable that the significance of certified lung cancer centers specializing in these innovative therapies will experience a notable escalation and the survival benefit associated with center-treatment will probably increase further. However, this assumption should be tested in a follow-up study on the topic based on recent clinical routine data.

In summary, the lung-cancer specific results of the Wizen-study presented in this publication contribute to a constantly growing international evidence base pointing towards the additional value of cancer treatment in designated centers. In concordance with research results from other countries with similar certification programs, the observed survival benefit associated with treatment in GCS-certified lung cancer centers seems to be not just a consequence of hospital size or large caseloads but also of structural and procedural improvements, like the reception of different consultations, early and dedicated image controls, availability of radiotherapy facilities, palliative care etc., induced by the certification. This information should be actively transported to patients, referring outpatient physicians, and decision makers. From a regulatory perspective, certification based on the criteria of the GCS could be defined as a structural requirement for billing corresponding healthcare services, leading to a potential shift of more treatment cases to certified centers.

 

MissingLinkText

 

MissingLinkText

Statement of Ethics

The WiZen-study was approved by the ethics committee of the TU Dresden (approval number: EK95022019, IRB 00001473, OHRP IORG0001076). Data processing and analyses was conducted in line with the Declaration of Helsinki and the General Data Protection Regulation of the European Union.

 

MissingLinkText

Registration

The study was listed at ClinicalTrials.gov (identifier: NCT04334239).

 

MissingLinkText

Data Availability Statement

The authors confirm that the data utilized in this study cannot be made available in the manuscript, the supplemental files, or in a public repository due to German data protection laws (‘Bundesdatenschutzgesetz’, BDSG).

 

MissingLinkText

Fundref Information

Gemeinsamer Bundesausschuss — 01VSF17020

 

MissingLinkText

 

MissingLinkText

Conflict of Interest

The authors declare that they have no conflict of interest.

• References

• 1 WHO-factsheet cancer: https://www.who.int/news-room/fact-sheets/detail/cancer accessed January 2023
• 2 https://www.krebsdaten.de/Krebs/SiteGlobals/Forms/Datenbankabfrage/datenbankabfrage_stufe2_form.html accessed Mar 20, 2025
• 3 Schenk EL, Patil T, Pacheco J. et al. 2020 Innovation-Based Optimism for Lung Cancer Outcomes. The Oncologist 2021; 26: e454-e472
• 4 Leitlinienprogramm Onkologie (Deutsche Krebsgesellschaft, Deutsche Krebshilfe, AWMF): Prävention, Diagnostik, Therapie und Nachsorge des Lungenkarzinoms, Langversion 2.1, 2022, AWMF-Registernummer: 020/007OL https://www.leitlinienprogramm-onkologie.de/leitlinien/lungenkarzinom/ accessed Aug 11, 2023
• 5 Mangiameli G, Cioffi U, Testori A. Lung Cancer Treatment: From Tradition to Innovation. Front Oncol 2022; 12: 858242
• 6 Bundesministerium für Gesundheit. Nationaler Krebsplan. Available from: https://www.bundesgesundheitsministerium.de/themen/ praevention/nationaler-krebsplan.html. accessed January 2023.
• 7 Griesshammer E, Adam H, Sibert NT. et al. Implementing quality metrics in European Cancer Centers (ECCs). World J Urol 2021; 39: 49-56
• 8 Wind A, Rajan A, van Harten WH. Quality assessments for cancer centers in the European Union. BMC Health Services Research 2016; 16: 474
• 9 https://www.cancer.gov/research/infrastructure/cancer-centers accessed January 2023
• 10 Kato M. Designated cancer hospitals and cancer control in Japan. J Natl Inst Public Health 2012; 61: 549-555
• 11 Brucker SY, Schumacher C, Sohn C. et al. Benchmarking the quality of breast cancer care in a nationwide voluntary system: the first five-year results (2003-2007) from Germany as a proof of concept. BMC Cancer 2008; 8: 358
• 12 https://www.krebsgesellschaft.de/deutsche-krebsgesellschaft/zertifizierung/zentrumssuche.html accessed January 2023
• 13 https://www.krebsgesellschaft.de/jahresberichte.html accessed January 2023
• 14 https://www.krebsgesellschaft.de/zertdokumente.html accessed Mar 20, 2025
• 15 Trautmann F, Reißfelder C, Pecqueux M. et al. Evidence-based quality standards improve prognosis in colon cancer care. European Journal of Surgical Oncology 2018; 44: 1324-1330
• 16 Völkel V, Draeger T, Gerken M. et al. Long-Term Survival of Patients with Colon and Rectum Carcinomas: Is There a Difference Between Cancer Centers and Non-Certified Hospitals?. Gesundheitswesen 2019; 81: 801-807
• 17 Richter M, Sonnow L, Mehdizadeh-Shrifi A. et al. German oncology certification system for colorectal cancer – relative survival rates of a single certified centre vs. national and international registry data. 2021; 6: 67-73
• 18 Butea-Bocu MC, Müller G, Pucheril D. et al. Is there a clinical benefit from prostate cancer center certification? An evaluation of functional and oncologic outcomes from 22,649 radical prostatectomy patients. World J Urol 2021; 39: 5-10
• 19 Kranz J, Grundmann R, Steffens J. Does structural and process quality of certified prostate cancer centers result in better medical care?. Urologe A 2021; 60: 59-66
• 20 Völkel V, Gerken M, Kleihues-van Tol K. et al. Treatment of Colorectal Cancer in Certified Centers: Results of a Large German Registry Study Focusing on Long-Term Survival. Cancers (Basel) 2023; 15
• 21 Hansinger J, Völkel V, Gerken M. et al. Endometrial Cancer – Long-Term Survival in Certified Cancer Centers and Non-Certified Hospitals: Comparative Analysis Based on a Large German Retrospective Cohort Study (WiZen). Geburtshilfe Frauenheilkd 2024; 84: 979-988
• 22 Roessler M, Schmitt J, Bobeth C. et al. Is treatment in certified cancer centers related to better survival in patients with pancreatic cancer? Evidence from a large German cohort study. BMC Cancer 2022; 22: 621
• 23 Schmitt J, Klinkhammer-Schalke M, Bierbaum V. et al. Krebserstbehandlung in zertifizierten versus nichtzertifizierten Krankenhäusern. Deutsches Ärzteblatt 2023; 120: 647-654
• 24 Schoffer O, Wimberger P, Gerken M. et al. Treatment in Certified Breast Cancer Centers Improves Chances of Survival of Patients with Breast Cancer: Evidence Based on Health Care Data from the WiZen Study. Geburtshilfe Frauenheilkd 2024; 84: 153-163
• 25 Bierbaum V, Schmitt J, Klinkhammer-Schalke M. et al. Assessment of the Potential of Concentrating Cancer Care in Hospitals With Certification Through Survival Analysis. Gesundheitswesen 2023; 85: S197-S204
• 26 Bobeth C, Tol KK, Rößler M. et al. [Methodology and Attribution Success of a Data Linkage of Clinical Registry Data with Health Insurance Data]. Gesundheitswesen 2023; 85: S154-S161
• 27 Bierbaum V, Bobeth C, Roessler M. et al. Treatment in certified cancer centers is related to better survival in patients with colon and rectal cancer: evidence from a large German cohort study. World J Surg Oncol 2024; 22: 11
• 28 Schoffer O, Schmitt J. WiZen in der Routineversorgung angekommen?. Forum 2024; 39: 449-453
• 29 Arbeitsgruppe Erhebung und Nutzung von Sekundärdaten der Deutschen Gesellschaft fur Sozialmedizin und Prävention. Arbeitsgruppe Epidemiologische Methoden der Deutschen Gesellschaft für Epidemiologie, Deutsche Gesellschaft für Medizinische Informatik Biometrie und Epidemiologie, Deutsche Gesellschaft für Sozialmedizin und Prävention: [Good practice of secondary data analysis, first revision]. Gesundheitswesen 2008; 70: 54-60
• 30 Elixhauser A, Steiner C, Harris DR. et al. Comorbidity measures for use with administrative data. Med Care 1998; 36: 8-27
• 31 Balan TA, Putter H. A tutorial on frailty models. Stat Methods Med Res 2020; 29: 3424-3454
• 32 von Elm E, Altman DG, Egger M. et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol 2008; 61: 344-349
• 33 GEKID-Länderatlas, available from: https://www.gekid.de/gekid-atlas/index.html#/survival accessed Aug 11, 2023
• 34 Kennedy-Martin T, Curtis S, Faries D. et al. A literature review on the representativeness of randomized controlled trial samples and implications for the external validity of trial results. Trials 2015; 16: 495
• 35 Habibzadeh F. Disparity in the selection of patients in clinical trials. The Lancet 2022; 399: 1048
• 36 Krebs in Deutschland für 2017/2018. 13. Ausgabe. Robert Koch-Institut (Hrsg) und die Gesellschaft der epidemiologischen Krebsregister in Deutschland e.V. (Hrsg). Berlin, 2021 ISBN 978-3-89606-309-0 DOI 10.25646/8353. : 172
• 37 Lipitz Snyderman AN, Fortier E, Li DG. et al. What do patients want to know when selecting a hospital for cancer care?. JCO 2018; 36: e18810-e18810
• 38 Pfaff H, Schmitt J. Shifting from Theoretical Best Evidence to Practical Best Evidence: an Approach to Overcome Structural Conservatism of Evidence-Based Medicine and Health Policy. Gesundheitswesen 2024; 86: S239-S250
• 39 Jochen Schmitt. Cancer Research Data Center – Conceptualization, Challenges, and Analysis Potential of Linkage of data close to care – AI-supported linkage and evidence of data from clinical cancer registries, SHI, and hospitals (DRKS00034650). 2024;
• 40 Bach PB, Cramer LD, Schrag D. et al. The influence of hospital volume on survival after resection for lung cancer. N Engl J Med 2001; 345: 181-188
• 41 Møller H, Riaz SP, Holmberg L. et al. High lung cancer surgical procedure volume is associated with shorter length of stay and lower risks of re-admission and death: National cohort analysis in England. European Journal of Cancer 2016; 64: 32-43
• 42 Kim BR, Sohn JY, Jang EJ. et al. Hospital case-volume and mortality after lung cancer surgery: A population-based retrospective cohort study. Lung Cancer 2022; 169: 61-66
• 43 Birkmeyer NJO, Goodney PP, Stukel TA. et al. Do cancer centers designated by the National Cancer Institute have better surgical outcomes?. Cancer 2005; 103: 435-441
• 44 Onega T, Duell EJ, Shi X. et al. Influence of NCI Cancer Center Attendance on Mortality in Lung, Breast, Colorectal, and Prostate Cancer Patients. Med Care Res Rev 2009; 66: 542-560
• 45 Okawa S, Tabuchi T, Nakata K. et al. Three-year survival from diagnosis in surgically treated patients in designated and nondesignated cancer care hospitals in Japan. Cancer Sci 2021; 112: 2513-2521
• 46 Blum TG, Morgan RL, Durieux V. et al. European Respiratory Society Guideline on various aspects of quality in lung cancer care. Eur Respir J 2022; 2103201

Publikationsverlauf

Eingereicht: 12. Dezember 2024

Angenommen: 16. April 2025

Artikel online veröffentlicht:
08. Juli 2025