A case-mix model for monitoring of postoperative mortality after surgery for lung cancer

doi:10.1016/j.lungcan.2005.08.007

Lung Cancer

Volume 51, Issue 1, January 2006, Pages 123-129

https://doi.org/10.1016/j.lungcan.2005.08.007 Get rights and content

Summary

Background

Postoperative mortality (POM) after surgery for lung cancer has been proposed as a performance indicator. Information on the size of the risk and its prognostic factors is needed to serve as a reference standard.

Patients and methods

Electronic records from the Rotterdam Cancer Registry (n = 2337) and the Thames Cancer Registry (n = 3772) were retrieved and analysed by sex, age, period, histology, region and extent of surgery. Multivariable logistic regression analysis was used to determine prognostic factors, to calculate odds ratios (OR) and to develop a case-mix model.

Results

POM was 4.2% (n = 257) on average and increased with age from 1.7% for patients younger than 60 years up to 9.4% for octogenarians. After lobectomy, POM was 2.9% against 6.0% and 9.5% after pneumonectomy left and right, respectively. Multivariable analysis showed higher risk for men (OR = 1.4) and lower risk for adenocarcinoma (OR = 0.6).

Conclusions

The final prediction model supports comparison and monitoring of POM rates for lung cancer. Only a limited number of risk factors need to be registered to allow adjustment for case-mix.

Introduction

Despite the progress in radiotherapy and medical oncology, surgical resection is still the cornerstone for curative treatment in patients with lung cancer. The pursued long-term survival benefit can only be obtained at the cost of the morbidity and mortality related to the resection. Particularly in the elderly, this should be weighed against the increase in life expectancy.

The risk of dying within 30 days of the operation, irrespective of cause, is generally referred to as postoperative mortality (POM) and has been proposed as a performance indicator. Some studies suggest that POM is lower in high volume hospitals [1], [2] and for operations performed by specialist surgeons [3]. To enable meaningful comparison between studies, regions, hospitals or surgeons, a case-mix model is needed to adjust for variation between patient series.

Clinicians also need contemporary information about the size of the operative risk and its prognostic factors so as to enable proper consultation and decision making with their patients. Accurate reference figures are difficult to find because information is mainly derived from specialist centres and subject to publication bias. Moreover, results obtained in other health care systems or patient populations may not be representative for the situation in other regions. To provide current, population-based estimates for POM and to develop a case-mix model for comparative studies, we combined information from the Thames and Rotterdam Cancer Registries.

Section snippets

Patients and methods

Electronic records for patients with lung cancer (ICD C34) who underwent lobectomy, bilobectomy (OPCS4 E54.2/3) or pneumonectomy (OPCS4 E54.1) in the period 1993–2002 were derived from the Rotterdam and Thames Cancer Registries (N = 6110).

The Rotterdam Cancer Registry covers the southwestern part of the Netherlands, a region with 16 hospitals and 2.3 million inhabitants. Specially trained registration clerks collect medical information on patient characteristics, tumour type (ICD-O topography and

Results

This study comprises a total number of 6110 patients with a mean POM of 4.2% [Table 1]. Thirty-two percent of patients were females and 35% were 70 years or older. Due to decreasing incidence of lung cancer in men, the absolute number of operated patients was lower in the period 1998–2002 than in 1993–1997. Adenocarcinoma and squamous carcinoma were diagnosed in 37% and 49% of patients, respectively. Treatment consisted of left pneumonectomy in 16%, right pneumonectomy in 11% and the side of

Discussion

This population-based study demonstrates a negligible variation in POM between the Rotterdam (4.3%) and Thames region (4.1%). The findings are similar to those of the Flemish Lung Cancer Registry (4.6%, n = 1614) [10] but clearly better than those of the Norwegian Cancer Registry (5.6%, n = 2354) [11]. For the USA, Harpole et al. [12] reported a POM rate of 5.2% (n = 3516) but in his series only 16% of patients received a pneumonectomy. The Japanese repeatedly report better results (1.2%, n = 6195) [13]

References (22)

G.A. Silvestri et al.
Specialists achieve better outcomes than generalists for lung cancer surgery
Chest
(1998)
D.H. Harpole et al.
Prognostic models of thirty-day mortality and morbidity after major pulmonary resection
J Thorac Cardiovasc Surg
(1999)
H. Wada et al.
Thirty-day operative mortality for thoracotomy in lung cancer
J Thorac Cardiovasc Surg
(1998)
S. Cykert
Risk acceptance and risk aversion: patients’ perspectives on lung surgery
Thorac Surg Clin
(2004)
S.A. Smulders et al.
Actual and predicted postoperative changes in lung function after pneumonectomy
Chest
(2004)
J.D. Birkmeyer et al.
Hospital volume and surgical mortality in the United States
N Engl J Med
(2002)
P.B. Bach et al.
The influence of hospital volume on survival after resection for lung cancer
N Engl J Med
(2001)
AN.J. Graham
Cardiothoracic surgeons do a good job
BMJ
(2002)
D.W. Hosmer et al.
Applied logistic regression
(2000)
R.G. Newcombe
Two-sided confidence intervals for the single proportion: comparison of seven methods
Statist Med
(1998)

J. Lovegrove et al.

Monitoring the performance of cardiac surgeons

J Oper Res Soc

(1998)

Cited by (28)

Stereotactic body radiotherapy to treat small lung lesions clinically diagnosed as primary lung cancer by radiological examination: A prospective observational study
2018, Lung Cancer
Citation Excerpt :
Recently, Detillon et al. showed that the mortality rate is 6.0% for patients 80 years or older who received surgical lung cancer resection [1]. Damhuis et al. reported that postoperative mortality increased with age: 1.7% for patients <60 years old versus 9.4% for patients of 80 years or older using data from the Rotterdam and Thames Cancer registries and a prediction model [2]. Occasionally physicians recommend watchful waiting until the nodule grows.
Even with advanced image guidance, biopsies occasionally fail to diagnose small lung lesions, which are highly suggestive of primary lung cancer by radiological examination. The aim of this study was to evaluate the outcome of stereotactic body radiotherapy (SBRT) to treat small lung lesions clinically diagnosed as primary lung cancer.
This is a prospective, multi-institutional observation study. Strict inclusion and exclusion criteria were determined in a nation-wide consensus meeting and used to include patients who were clinically diagnosed with primary lung cancer using precise imaging modalities, for whom further surgical intervention was not feasible, who refused watchful waiting, and who were highly tolerable of SBRT with informed consent. SBRT was performed with 48 Gy in 4 fractions at the tumor isocenter.
From August 2009 to August 2014, 62 patients from 11 institutions were enrolled. Their median age was 80 years. The tumors ranged in size from 9 to 30 mm in diameter (median, 18 mm). The median follow-up interval was 55 months. The 3-year overall survival rate was 83.3% (95% confidence interval (CI) 71.1–90.7%) for all the patients and 94.7% (95% CI 68.1–99.2%) for the patients younger than 75 years. Local failure, regional lymph node metastases and distant metastases occurred in 4 (6.4%), 3 (4.8%) and 11 (17.7%) patients, respectively. Grades 3 and 4 toxicities were observed in 8 (12.9%) patients and 1 (1.6%) patient, respectively. No grade 5 toxicities were observed.
SBRT is safe and effective for patients with small lung lesions clinically diagnosed as primary lung cancer that satisfied the proposed strict indication criteria as previously reported. A prospective interventional study is required to ascertain if SBRT is an alternative strategy for these patients.
Early stage non-small-cell lung cancer in octogenarian and older patients: A SEER database analysis
2016, Clinical Lung Cancer
Citation Excerpt :
These outcomes are reported by large centers, and may not reflect the experience of smaller, low-volume centers. A recent prediction model using the data from the Rotterdam and Thames Cancer registries found that the postoperative mortality increased with age; 1.7% for patients < 60 years versus 9.4% for patients ≥ 80 years.16 A comparison of the outcomes of octogenarian patients with early stage NSCLC after surgery and non-surgical management in a well-defined US population, included in the SEER database, was conducted to help define the best treatment options for this cohort.
The median age at diagnosis of lung cancer is 70 years. However, the evidence guiding the management of octogenarians and older patients with non–small-cell lung cancer (NSCLC), is based on data derived from younger patients and may not be appropriate.
Patients ≥ 80 years diagnosed with clinical stages I and II NSCLC, between 1988 and 2007, were identified from the SEER database. Patients were classified according to treatments received: no treatment, surgery only, radiation only, and surgery + radiation. Factors associated with survival were assessed using the Cox proportional hazards model.
There were 1338 cases of early stage NSCLC in octogenarians. Surgery was the most common treatment modality. The median overall survival was 3.8 years for patients who had surgery, compared with 1.6 years, 1.6 years, and 0.9 years for those who received surgery + radiation, radiation alone, and no treatment, respectively (P < .0001). Factors significantly associated with worse overall survival following surgery included increasing age (hazard ratio [HR], 1.08; P = .0005), male gender (HR, 1.33; P = .01), stage II (HR, 2.21; P < .0001), and squamous histology (HR, 1.36; P = .01).
Surgical resection is associated with long-term survival outcomes in a substantial proportion of octogenarian and older patients with early stage lung cancer and should not be withheld on the basis of age alone.
Nationwide quality improvement in lung cancer care: The role of the Danish Lung Cancer Group and Registry
2013, Journal of Thoracic Oncology
To improve prognosis and quality of lung cancer care the Danish Lung Cancer Group has developed a strategy consisting of national clinical guidelines and a clinical quality and research database. The first edition of our guidelines was published in 1998 and our national lung cancer registry was opened for registrations in 2000. This article describes methods and results obtained by multidisciplinary collaboration and illustrates how quality of lung cancer care can be improved by establishing and monitoring result and process indicators.
A wide range of indicators was established, validated, and monitored. By registration of all lung cancer patients since the year 2000, data on more than 40,000 patients have been included in the database. Results are reported periodically/quarterly and submitted to formal auditing on an annual basis.
Improvements in all outcome indicators are documented and statistically significant. Thus the 1-year overall survival rate has increased between 2003 and 2011 from 36.6% to 42.7%, the 2-year survival rate from 19.8% to 24.3%, and the 5-year survival rate from 9.8% to 12.1%. Five-year survival after surgical resection has increased from 39.5% to 48.1%. Improvements of waiting times, accordance between cTNM and pTNM, and resection rates are documented.
The Danish experience shows that a national quality management system including national guidelines, a database with high data quality, frequent reports, audit and commitment from all stakeholders can contribute to improve clinical practice, improve core results, and reduce regional differences.
Outcomes of stereotactic ablative radiotherapy in patients with potentially operable stage i non-small cell lung cancer
2012, International Journal of Radiation Oncology Biology Physics
Citation Excerpt :
However, our findings of no 30-day mortality following SABR contrasts with the predicted 30-day surgical mortality of 2.6% determined by using the Thoracoscore (16). This mortality rate of 2.6% is based on the assumption that all patients would have undergone a lobectomy, which probably underestimates the actual risks, as 10% or more of Dutch patients have undergone a pneumonectomy for a stage I NSCLC in recent years (26–28). A reliable comparison between the OS of our SABR patients and that of patients reported in the surgical literature is difficult, despite our efforts to select patients whose conditions were potentially operable.
Approximately two-thirds of patients with early-stage non-small-cell lung cancer (NSCLC) in The Netherlands currently undergo surgical resection. As an increasing number of fit patients have elected to undergo stereotactic ablative radiotherapy (SABR) in recent years, we studied outcomes after SABR in patients with potentially operable stage I NSCLC.
In an institutional prospective database collected since 2003, 25% of lung SABR cases (n = 177 patients) were found to be potentially operable when the following patients were excluded: those with (1) synchronous lung tumors or other malignancy, (2) prior high-dose radiotherapy/pneumonectomy, (3) chronic obstructive pulmonary disease with a severity score of 3-4 according to the Global initiative for Obstructive Lung Disease classification. (4) a performance score of ≥3, and (5) other comorbidity precluding surgery. Study patients included 101 males and 76 females, with a median age of 76 years old, 60% of whom were staged as T1 and 40% of whom were T2. Median Charlson comorbidity score was 2 (range, 0–5). A SABR dose of 60 Gy was delivered using a risk-adapted scheme in 3, 5, or 8 fractions, depending on tumor size and location. Follow-up chest computed tomography scans were obtained at 3, 6, and 12 months and yearly thereafter.
Median follow-up was 31.5 months; and median overall survival (OS) was 61.5 months, with 1- and 3-year survival rates of 94.7% and 84.7%, respectively. OS rates at 3 years in patients with (n = 59) and without (n = 118) histological diagnosis did not differ significantly (96% versus 81%, respectively, p = 0.39). Post-SABR 30-day mortality was 0%, while predicted 30-day mortality for a lobectomy, derived using the Thoracoscore predictive model (Falcoz PE et al. J Thorac Cardiovasc Surg 2007;133:325–332), would have been 2.6%. Local control rates at 1 and 3 years were 98% and 93%, respectively. Regional and distant failure rates at 3 years were each 9.7%. Toxicity was mild, with grade ≥3 radiation pneumonitis and rib fractures in 2% and 3%, respectively.
Patients with potentially operable disease who underwent primary SABR had a median OS that exceeded 5 years. This finding supports ongoing randomized clinical trials comparing surgery and SABR in cases of operable stage I NSCLC.
Outcomes of stereotactic ablative radiotherapy for centrally located early-stage lung cancer
2011, Journal of Thoracic Oncology
The use of stereotactic ablative radiotherapy (SABR) in centrally located early-stage lung tumors has been associated with increased toxicity. We studied outcomes after delivery of risk-adapted SABR of central tumors.
SABR was delivered in eight fractions of 7.5 Gy to 63 such patients between 2003 and 2009. Of these, 37 patients had a tumor at a central hilar location, whereas 26 patients had tumors abutting the pericardium or mediastinal structures. Survival outcomes were compared with patients with peripheral tumors treated during the same time period using fewer fractions of SABR.
Median follow-up was 35 months. Late grade III toxicity was limited to chest wall pain (n = 2) and increased dyspnoea (n = 2). No grade IV/V toxicity was observed, but grade V toxicity could not be excluded with certainty in nine patients who died of cardiopulmonary causes. Distant metastases were the predominant cause of death; cardiovascular deaths were not associated with a paracardial tumor location. No significant differences in outcomes were observed between these 63 patients and 445 other SABR patients treated for peripheral early-stage lung tumors. Three-year local control rates were 92.6% and 90.2% (p = 0.9). Three-year overall survival rates were 64.3% and 51.1% with median survival rates of 47 and 36 months, in favor of the group of patients with central tumors (p = 0.09).
Use of risk-adapted SABR delivered in eight fractions of 7.5 Gy did not result in excess toxicity for centrally located early-stage lung tumors, and clinical outcomes were comparable with those seen for peripheral lesions.
Locoregional and distant failure following image-guided stereotactic body radiation for early-stage primary lung cancer
2011, Radiotherapy and Oncology
To report our institutional experience using image-guided stereotactic body radiation therapy (SBRT) for early stage lung cancer, including an analysis into factors associated with nodal and distant failures (NF, DF).
Forty-eight patients with early-stage primary lung cancer were treated with image-guided SBRT between 2007 and 2009. Median prescription dose was 48 Gy in 4 fractions. Toxicity was graded according to the NCI CTCAE v3.0 scale.
Local failure was detected in two lesions and actuarial 24-month local control was 95%. At 24 months, the cumulative incidence of NF was 6%, and DF was 29%. Larger lesions (>3 cm) and younger age (<70 years) were the only factors found to be significantly correlated with increased DF (p = 0.005 and p = 0.015, respectively). A single grade ⩾3 toxicity was observed. After adjusting for age and lesion size, distant failure was significantly associated with a poorer OS (Cox regression, p = 0.0059).
Image-guided SBRT can produce excellent LC rates with minimal toxicity. Distant failure was a major determinant of OS and the most common pattern of failure, indicating a potential role for systemic therapy in younger patients with large lesions.

View all citing articles on Scopus

View full text

A case-mix model for monitoring of postoperative mortality after surgery for lung cancer

Summary

Background

Patients and methods

Results

Conclusions

Introduction

Section snippets

Patients and methods

Results

Discussion

Chest

J Thorac Cardiovasc Surg

J Thorac Cardiovasc Surg

Thorac Surg Clin

Chest

Hospital volume and surgical mortality in the United States

N Engl J Med

The influence of hospital volume on survival after resection for lung cancer

N Engl J Med

Cardiothoracic surgeons do a good job

BMJ

Applied logistic regression

Two-sided confidence intervals for the single proportion: comparison of seven methods

Statist Med

Monitoring the performance of cardiac surgeons

J Oper Res Soc