Open Access
Open Peer Review

This article has Open Peer Review reports available.

How does Open Peer Review work?

Oesophagectomy rates and post-resection outcomes in patients with cancer of the oesophagus and gastro-oesophageal junction: a population-based study using linked health administrative linked data

BMC Health Services Research201212:384

DOI: 10.1186/1472-6963-12-384

Received: 12 June 2012

Accepted: 31 October 2012

Published: 8 November 2012

Abstract

Background

Hospital performance is being benchmarked increasingly against surgical indicators such as 30-day mortality, length-of-stay, survival and post-surgery complication rates. The aim of this paper was to examine oesophagectomy rates and post-surgical outcomes in cancers of the oesophagus and gastro-oesophageal junction and to determine how the addition of gastro-oesophageal cancer to oesophageal cancer impacts on these outcomes.

Methods

Our study population consisted of patients with a primary invasive oesophageal or gastro-oesophageal cancer identified from the NSW Cancer Registry from July 2000-Dec 2007. Their records were linked to the hospital separation data for determination of resection rates and post-resection outcomes. We used multivariate logistic regression analyses to examine factors associated with oesophagectomy and post-resection outcomes. Cox-proportional hazard regression analysis was used to examine one-year cancer survival following oesophagectomy.

Results

We observed some changes in resection rates and surgical outcomes with the addition of gastro-oesophageal cancer patients to the oesophageal cancer cohort. 14.6% of oesophageal cancer patients and 26.4% of gastro-oesophageal cancer patients had an oesophagectomy; an overall oesophagectomy rate of 18.2% in the combined cohort. In the combined cohort, oesophagectomy was associated with younger age, being male and Australian-born, having non-metastatic disease or adenocarcinoma and being admitted in a co-located hospital. Rates of length-of-stay >28 days (20.9% vs 19.7%), 30-day mortality (3.8% vs 2.7%) and one-year survival post-surgery (24.5% vs 23.1%) were similar between oesophageal cancer alone and the combined cohort; whilst 30-day complication rates were 21.5% versus 17.0% respectively. Some factors statistically associated with post-resection complication in oesophageal cancer alone were not significant in the overall cohort. Poorer post-resection outcomes were associated with some patient (older age, birthplace) and hospital-related characteristics (fiscal sector, area health service).

Conclusion

Outcomes following oesophagectomy in oesophageal and gastro-oesophageal cancer patients in NSW are within world benchmarks. Our study demonstrates that the inclusion of gastro-oesophageal cancer did alter some outcomes compared to analysis based solely on oesophageal cancer. As such, care must be taken with analyses based on administrative health data to capture all populations eligible for treatment and to understand the contribution of these subpopulations to overall outcomes.

Keywords

Hospital outcomes Oesophageal cancer Administrative health data, Data linkage

Background

Accountability and transparency in health care, at individual and institutional levels, has gained increasing importance in recent years [13]. This growing focus on quality of practise and safety of services has implications for patients, clinicians, administrators and policy makers. Quality assessments of hospital performance are now conducted routinely throughout the world, the results of which are made available to all stakeholders. Information of this kind has been used for the purposes of benchmarking, so as to improve performance across all institutions [4] and to empower the general public to make informed decisions about where they choose to receive care.

Surgical resection for curable oesophageal and gastro-oesophageal cancer is the mainstay of treatment for all patients who are fit for major surgery. However, despite being the only curative treatment for oesophageal/gastro-oesophageal cancer patients, oesophagectomy is associated with significant operative morbidity and mortality [57]. Key performance measures of oesophagectomy include 30-day mortality, hospital length-of-stay and post-surgery complication rate [817]. Factors influencing outcomes following oesophagectomy include hospital/surgeon factors (peer group/volume, surgical experience), tumour stage, histology and location, surgery type and patient comorbidity; with better outcomes reported in high volume hospitals, patients with non-metastatic disease (approximately 60% of patients), adenocarcinoma (the incidence of which is increasing), and when transhiatal oesophagectomy is performed [1825].

In Australia, there has been keen interest in using administrative datasets to examine hospital performance [2628] with surgeons, administrators and patients having endorsed independent review of surgical care [29]. We previously examined patient outcomes for oesophageal cancer following oesophagectomy in one Australian state using data linkage of administrative datasets [9]. We found oesophagectomy was performed with satisfactory outcomes however resection rates were lower than reported elsewhere, as other studies included cancer of the gastro-oesophageal junction in their analyses since gastro-oesophageal cancer is treated in a similar manner to oesophageal cancer [30, 31].

Therefore, our aim was to examine rates of oesophogectomy and post-resection outcomes for patients diagnosed with oesophageal cancer or gastro-oesophageal cancer in the state of New South Wales (NSW) between 2000 and 2007. Specifically, we examined the way in which the inclusion of gastro-oesophageal cancer to the oesophageal cancer cohort impacted on resection rates and post-resection outcomes.

Methods

Setting

Australia has a publicly-funded universal health care system and health care data is collected by Commonwealth and state-based jurisdictions. The Australian Government Department of Human Services administer the Medicare Benefits Scheme (MBS; which provides free or subsidised treatment by health professionals and fully subsidised treatment in public hospitals) and the Pharmaceutical Benefits Scheme (PBS; which provides subsidised access to prescription drugs). The individual state and territory health authorities are responsible for the collection of public and private hospital separation data within their jurisdiction. However, cross-jurisdictional data sets have not been linked routinely due to legislative and privacy issues.

Our current study uses routinely collected hospital separation data from New South Wales (NSW). NSW is the largest jurisdiction in Australia with eight area health districts (at the time this research commenced) and consists of more than 200 public and 80 private hospitals, some of which are in proximity to each other (co-located). As such medical specialists practise at both the public and affiliated private hospital.

Study population

Our study population consisted of patients with a primary invasive oesophageal and gastro-oesophageal cancer (International Classification of Diseases v3 [ICD-O-3] codes C15.0-C15.9 and C16.0).

Data sources and linkage

We accessed and linked two population-based data sets in this study:

NSW Central Cancer Registry (1 July 2000-31Dec2007) - (CCR) receives notifications of cancer in NSW and maintains a record of all cancer cases diagnosed in NSW residents since 1972. The registry is managed according to the International Association of Cancer Registries (IACR) rules [32] and is one of the few Australian Cancer Registries to record degree-of-spread at first diagnosis for all solid malignant tumours [33]. Degree-of-spread is assigned by the NSW CCR into one of four summary stages (localised, regional, distant or unknown) [25]. Histology type was based on Berg groupings of ICD-O-3 classifications and grouped as adenocarcinomas, squamous cell carcinoma (SCC) or other.

Admitted Patient Data Collection (1 July 2000 – 30 June 2008) - is a census of all inpatient separations from all public, private and repatriation hospitals, private day procedures centres and public nursing homes in NSW. We identified surgical resections in the separations using the International Classification of Diseases (ICD_10_AM) procedural block codes for oesophagectomy by abdominal and thoracic/cervical/transthoracic mobilisation and endoscopic mucosal resection 0858–0860 (if recorded) [34].

Data linkage was undertaken by a third-party, the NSW Centre for Health Record Linkage (CHeReL) using best practice privacy preserving protocols. Using probabilistic linkage, CHeReL identified all appropriate records from the APDC and CCR. They then assigned unique identifiers (or Project Person Number: PPN) to the APDC and CCR files. We received two individual data files and linked the unit-record data using the unique PPNs.

Outcomes measures and statistical analyses

We reported oesophagectomy rates for oesophageal cancer, gastro-oesophageal cancer and the combined cohort. As oesophagectomy is most beneficial for, and is associated with, non-metastatic disease and patients with adenocarcinoma, we also examined rates of surgical resection and factors associated with surgical resection and post-surgical outcomes for this sub-group of patients. Multivariate logistic regression was used to determine which factors were associated with (i) receiving an oesophagectomy and (ii) the post-resection outcomes of: longer length of hospital stay (>28 days); 30-day complication and 30-day mortality. Post resection complications arising within 30 days were derived from ICD-10-AM diagnostic codes and included any post-procedural disorders or complications, haemorrhages, pulmonary and cardiac complications.

Cox proportional hazard regression survival analysis was also undertaken to examine one-year survival following oesophagectomy. We followed patients diagnosed with oesophageal cancer or cancer of the gastro-oesophageal junction in the NSW CCR to 31st December, 2007 for death from the cancer.

We did not identify individual hospitals in this study, rather we grouped hospitals by health service (identified as Area 1 – Area 8) and according to fiscal sector (public co-located; private co-located; public non co-located; private non co-located). We used the Charlson Comorbidity Index (CCI) to account for comorbidity before cancer diagnosis [35]. This index is commonly used to adjust for comorbidity and was derived from ICD-10-AM diagnostic codes as recorded in any hospital separation prior to and including the date of cancer diagnosis [36, 37]. Other covariates of interest included patient factors of age at diagnosis, gender, Australian or overseas born, tumour location and surgical procedure (where appropriate).

Statistical significance was taken at the p<0.05 level. All statistical analysis was performed using SAS version 11.

Ethical approval

This research was approved by the NSW Population and Health Service Research Ethics Committee (2010/05/235).

Results

Cohort Characteristics (Table 1)

Our cohort comprised 5,024 patients, 3,456 (69%) of whom were diagnosed with oesophageal cancer and 1,568 with gastro-oesophageal cancer.

More than half of oesophageal cancer patients were ≥70 years, 67% were male, 45% were diagnosed with adenocarcinoma, 20% with regional spread and 18% with distant spread-of-disease. The vast majority (90%) of oesophageal cancer patients were defined by the CCI as having no pre-existing comorbidity. Similarly, most gastro-oesophageal patients were ≥70 years and male. Owing to the nature of gastro-oesophageal cancer, 89% of this group were diagnosed with adenocarcinoma, 35% with regional and 25% with distant disease spread. The addition of gastro-oesophageal patients to the cohort increased the proportion of patients in the overall cohort with adenocarcinoma to approximately 60%, and patients with regional or distant spread increased from 38% to 45%.
Table 1

Characteristics of patients diagnosed with oesophageal (C15) and gastro-oesophageal cancer (C160)

Variable

C15

C160

C15 & C160

N=3,456

N=1,568

N=5,024

n

%

n

%

n

%

Age at diagnosis group (years)

      

<60

657

19.0

402

25.6

1,059

21.1

60–69

840

24.3

399

25.4

1,239

24.7

70–79

1,048

30.3

486

31.0

1,534

30.5

80+

911

26.4

281

17.9

1,192

23.7

Gender

      

Female

1,159

33.5

362

23.1

1,521

30.3

Male

2,297

66.5

1,206

76.9

3,503

69.7

Country of Birth

      

Australia

2,497

72.2

1,063

67.8

3,560

70.9

Other

959

37.8

505

32.2

1,464

29.1

Histology group

      

SCC

1,478

42.8

70

4.5

1,548

30.8

Adenocarcinomas

1,549

44.8

1,398

89.2

2,947

58.7

Other

293

8.5

79

5.0

372

7.4

Degree of Spread

      

Localised

1,290

37.3

435

27.7

1,725

34.3

Regional

692

20.0

543

34.6

1,235

24.6

Distant

622

18.0

387

24.7

1,009

20.1

Unknown

721

20.9

203

12.9

924

18.4

Tumour location

      

Upper

297

8.6

  

297

5.9

Middle

470

13.8

  

470

9.4

Lower

1,528

45.0

  

1,528

30.9

NOS*

1,113

32.7

  

1,113

22.2

GOJ*

  

1,568

 

1,568

31.2

Charlson comorbidity

      

0

3,098

89.6

1,431

91.2

4,529

90.1

1,2

331

9.6

124

7.9

455

9.1

3+

27

0.8

13

0.8

40

0.8

*NOS = not otherwise specified; GOJ = gastro-oesophageal junction.

Rates of surgical resection and factors associated with receiving surgical resection (Tables 2,3)

The rate of resection in oesophageal cancer patients was 14.6%, whereas 26.4% of gastro-oesophageal patients underwent surgical resection. The addition of gastro-oesophageal cancer patients to the cohort increased the overall surgery rates to 18.2%.

The majority of oesophageal cancer patients who underwent resection were <70 years (71%), male (76%), born in Australia (68%), had no pre-existing comorbidity (70%) and were diagnosed with adenocarcinoma (62%) and with loco-regional disease at the time of notification (85%). Gastro-oesophageal cancer patients had even higher proportions of males (87%) and adenocarcinoma patients (93%) who underwent resection. After adjustment for covariates, surgical resection for oesophageal cancer patients was associated with younger age (p<0.001), adenocarcinoma (p<0.001), non-metastatic disease (p<0.001), no pre-existing comorbidity (p=0.006) and the resection being conducted in a public or private co-located hospital (p<0.001) (Table 2). Not surprisingly the addition of gastro-oesophageal patients to the cohort increased patient numbers and resulted in some additional significant predictors of surgical resection (male gender [aOR=1.31 95%CI 1.05-1.64], born in Australia [aOR=1.32, 95%CI 1.09-1.61] and undergoing surgery at specific area health services).
Table 2

Patient characteristics, location and surgery type of cohort who received resection

Variable

C15

C160

C15 & C160

N=3,456

N=1,568

N=5,024

n

%

n

%

n

%

Oesophagectomy Yes

503

14.6

412

26.3

915

18.2

PATIENT CHARACTERISTICS

Age at diagnosis (years)

      

<60

164

4.7

137

8.7

301

6.0

60–69

195

5.6

137

8.7

332

6.6

70–79

116

3.4

120

7.7

236

4.7

80+

28

0.8

18

1.1

46

0.9

Gender

      

Female

118

3.4

54

3.4

172

3.4

Male

385

11.1

358

22.8

743

14.8

Country of Birth

      

Australia

345

10.0

291

18.6

636

12.7

Other

158

4.6

121

7.7

279

5.6

Histology group

      

SCC

169

4.9

12

0.8

181

3.6

Adenocarcinomas

312

9.0

385

24.6

697

13.9

Other

11

0.3

11

0.7

22

0.4

Degree of Spread

      

Localised

197

5.7

123

7.8

320

6.4

Regional

232

6.7

239

15.2

471

9.4

Distant

38

1.1

35

2.2

73

1.5

Unknown

36

1.0

15

1.0

51

1.0

Tumour location

      

Upper

17

0.5

  

17

0.3

Middle

55

1.6

  

55

1.1

Lower

331

9.6

  

331

6.6

NOS*

99

2.9

  

99

2.0

GOJ*

  

412

26.3

412

8.3

Charlson comorbidity

      

0

353

10.2

284

18.1

637

12.7

1,2

32

0.9

22

1.4

54

1.1

3+

0

0.0

2

0.1

2

0.0

HOSPITAL CHARACTERISTICS

Surgery type

      

Abdothoracic

79

2.3

65

4.1

144

2.9

Abdocervical

64

1.9

45

2.9

109

2.2

Transhiatal

360

10.4

301

19.2

661

13.2

Hospital admission

      

Private co-located

97

2.8

75

4.8

172

3.4

Private non co-located

90

2.6

87

5.5

177

3.5

Public co-located

182

5.3

131

8.4

313

6.2

Public non co-located

134

3.9

119

7.6

253

5.0

Hospital area service

      

Area 1

99

2.9

90

5.7

189

3.8

Area 2

107

3.1

85

5.4

192

3.8

Area 3

93

2.7

70

4.5

163

3.2

Area 4

126

3.6

90

5.7

216

4.3

Area 5

53

1.5

51

3.3

104

2.1

Area 6

11

0.3

9

0.6

20

0.4

Area 7

14

0.4

17

1.1

31

0.6

Area 8

0

0.0

0

0.0

0

0.0

* NOS = not otherwise specified; GOJ = gastro-oesophageal junction.

Table 3

Surgery rates and factors associated with receiving surgery by topography

Variable

C15 Surgical Rate (per 100)

aOR#(95%CI)

C15&C160 Surgical rate (per 100)

aOR (95% CI)#

Age at diagnosis (years)

    

<60

25.0

8.96 (5.63-14.3)*

28.4

8.12 (5.64-11.7)*

60–69

23.2

9.85 (6.25-15.5)*

26.8

8.48 (5.94-12.1)*

70–79

11.1

3.99 (2.52-6.32)*

15.4

4.20 (2.94-6.00)*

80+

3.1

Referent

3.9

Referent

Gender

    

Female

10.2

Referent

11.3

Referent

Male

16.8

1.04 (0.78-1.38)

21.2

1.31 (1.05-1.64)*

Country of Birth

    

Australia

13.8

Referent

17.9

Referent

Other

16.5

0.90 (0.70-1.16)

19.1

0.76 (0.62-0.92)*

Histology group

    

SCC

11.4

0.40 (0.31-0.52)*

11.7

0.41 (0.32-0.51)*

Adenocarcinomas

20.1

Referent

23.6

Referent

Other

3.8

0.28 (0.17-0.48)*

5.9

0.36 (0.24-0.53)*

Degree of Spread

    

Localised

15.3

6.06 (4.06-9.06)*

18.6

5.50 (4.08-7.41)*

Regional

33.5

12.34 (8.22-18.52)*

38.1

10.29 (7.67-13.79)*

Distant

6.1

Referent

6.4

Referent

Unknown

5.0

2.38 (1.42-3.98)*

5.5

1.71 (1.14-2.56)*

Tumour location

    

Upper

5.8

0.22 (0.12-0.39)

5.8

0.22 (0.13-0.40)

Mid

10.9

0.51 (0.34-0.76)

10.9

0.52 (0.35-0.77)

Distal

21.5

Referent

21.5 26.4

Referent 1.10 (0.88-1.36)

Charlson comorbidity

    

0

11.4

Referent

14.1

Referent

1,2

9.7

0.52 (0.32-0.83)*

11.9

0.56 (0.39-0.80)*

3+

0

-

5.0

0.25 (0.06-1.14)

Hospital admission

    

Private co-located

43.1

Referent

46.5

Referent

Private non co-located

4.4

0.06 (0.04-0.10)*

6.2

0.07 (0.05-0.10)*

Public co-located

34.0

0.78 (0.53-1.16)

37.8

0.80 (0.59-1.08)

Public non co-located

20.6

0.26 (0.16-0.42)*

26.5

0.26 (0.18-0.38)*

Hospital area service

    

Area 1

18.1

2.30 (1.38-3.84)*

23.6

2.99 (2.03-4.39)*

Area 2

16.2

0.73 (0.48-1.09)

19.4

0.80 (0.59-1.09)

Area 3

25.8

Referent

28.6

Referent

Area 4

19.6

1.29 (0.86-1.93)

23.8

1.58 (1.16-2.17)*

Area 5

12.2

1.71 (0.99-2.97)

16.5

1.92 (1.27-2.90)*

Area 6

3.5

0.82 (0.38-1.77)

4.6

0.88 (0.49-1.56)

Area 7

5.1

0.71 (0.34-1.47)

8.2

0.98 (0.59-1.65)

Area 8

0.0

-

0.0

-

# Adjusted for age-group, gender, histology, country of birth, histology, degree-of-spread, tumour location, hospital admitted, area of service, Charlson comorbidity index.

* Significant at P<0.05 (two-tailed) level.

Sub-analyses of patients with non-metastatic adenocarcinoma

When restricted to the 2,155 non-metastatic adenocarcinoma patients, the addition of gastro-oesophageal cancer patients increased the overall surgery rate to 29.6% (from 26.1% for oesophageal cancer alone). Patient and hospital factors associated for surgery in this sub-cohort were the same as the entire cohort. The addition of the gastro-oesophageal patients again added the significant predictors of surgical resection (male gender [aOR=2.03 95%CI 1.44-2.85], being born in Australia [aOR=1.41, 95%CI 1.35-3.44] and undergoing surgery at specific area health services).

Outcomes following surgery (Table 4)

Table 4

Patient and hospital factors associated with outcomes following surgical resection

 

C15

C15 & C16.0

Outcome

N (Rate per 100)

aOR (95%CI)#

N (Rate per 100)

aOR (95%CI)#

(i) Length-of-stay >28 days

105 (20.9)

 

186 (19.7)

 

<60 years

23 (14.0)

0.20 (0.08-0.53)*

43 (14.3)

0.24 (0.12-0.51)*

60–64 years

43 (22.1)

0.41 (0.17-1.00)

67 (20.2)

0.43 (0.22-0.87)*

70–74 years

28 (24.1)

0.44 (0.17-1.13)

60 (25.4)

0.57 (0.28-1.17)

80+ years

11 (39.3)

Referent

16 (34.8)

Referent

Australian born

75 (21.7)

Referent

133 (20.9)

Referent

Overseas born

30 (19.0)

0.62 (0.38-1.05)

53 (19.0)

0.67 (0.46-0.99)*

Private co-located

14 (14.4)

Referent

22 (12.8)

Referent

Private non co-located

11 (12.2)

1.06 (0.43-2.63)

23 (13.0)

1.92 (0.90-4.13)

Public co-located

43 (23.6)

1.61 (0.74-3.49)

71 (22.7)

2.10 (1.17-3.78)*

Public non co-located

37 (27.6)

3.16 (0.98-10.8)

70 (27.7)

7.89 (3.16-19.5)*

Area 1

29 (29.3)

0.91 (0.26-3.18)

52 (27.5)

0.44 (0.17-1.10)

Area 2

22(20.6)

0.91 (0.43-1.96)

43 (22.4)

1.24 (0.70-2.19)

Area 3

22 (23.7)

Referent

37 (22.7)

Referent

Area 4

21 (16.7)

0.74 (0.33-1.66)

33 (15.3)

0.64 (0.34-1.20)

Area 5

8 (15.1)

0.30 (0.07-1.26)

14 (13.5)

0.15 (0.05-0.43)*

Area 6

2 (18.2)

0.82 (0.12-5.59)

3 (15.0)

0.30 (0.07-1.39)

Area 7

1 (7.1)

0.18 (0.02-1.92)

4 (12.9)

0.22 (0.06-0.86)*

(ii) 30 day complication

108 (21.5)

 

168 (17.0)

 

<60 years

35 (21.3)

0.40 (0.17-0.98)*

53 (17.6)

0.53 (0.26-1.10)

60–64 years

36 (18.5)

0.33 (0.14-0.80)*

59 (17.8)

0.55 (0.27-1.13)

70–74 years

25 (21.6)

0.39 (0.16-0.99)*

52 (22.0)

0.70(0.34-1.46)

80+ years

12 (42.9)

Referent

14 (30.4)

Referent

Australian born

71 (20.6)

Referent

129 (20.3)

Referent

Overseas born

37 (23.4)

1.28 (0.79-2.08)

49 (17.6)

0.87 (0.92-1.06)

Private co-located

16 (16.5)

Referent

28 (16.3)

Referent

Private non co-located

22 (24.4)

1.13 (0.43-2.98)

32 (18.1)

1.07 (0.51-2.23)

Public co-located

40 (22.0)

1.82 (0.87-3.78)

61 (19.5)

1.37 (0.79-2.40)

Public non co-located

30 (22.4)

0.74 (0.23-2.37)

57 (22.5)

1.47 (0.62-3.49)

Area 1

25 (25.3)

3.88 (1.12-13.5)*

40 (21.2)

1.36 (0.54-3.47)

Area 2

21 (19.6)

1.91 (0.85-4.28)

37 (19.3)

1.40 (0.76-2.57)

Area 3

15 (16.1)

Referent

28 (17.2)

Referent

Area 4

29 (23.0)

2.24 (0.99-5.09)

40 (18.5)

1.29 (0.68-2.42)

Area 5

13 (24.5)

4.01 (1.04-15.4)*

23 (22.1)

1.34 (0.49-3.65)

Area 6

9 (18.2)

2.86 (0.43-19.1)

2 (10.0)

0.74 (0.14-3.97)

Area 7

11 (21.4)

3.87 (0.68-22.1)

8 (25.8)

2.14 (0.66-6.91)

 

Mortality N (rate per 100)

aHR (95%CI)

Mortality N (rate per 100)

aHR (95%CI)

(iv) One year survival

123 (24.5)

 

218 (23.1)

 

<60 years

36 (22.0)

0.20 (0.10-0.41)*

64 (21.3)

0.41 (0.23-0.74)*

60–64 years

38 (30.9)

0.19 (0.09-0.38)*

65 (19.6)

0.36 (0.20-0.65)*

70–74 years

32 (27.6)

0.22 (0.10-0.47)*

68 (28.8)

0.44 (0.24-0.81)*

80+ years

17 (60.7)

Referent

21 (45.6)

Referent

Australian born

90 (26.1)

Referent

155 (24.4)

Referent

Overseas born

33 (20.9)

0.64 (0.39-1.06)

63 (22.6)

0.74 (0.51-1.08)

Private co-located

10 (10.3)

Referent

20 (11.6)

Referent

Private non co-located

22 (24.4)

1.58 (0.57-4.42)

40 (22.6)

2.20 (1.03-4.70)*

Public co-located

48 (26.4)

1.25 (0.55-2.83)

80 (25.6)

1.80 (0.97-3.32)

Public non co-located

43 (32.1)

1.34 (0.45-4.02)

78 (30.8)

1.66 (0.71-3.89)

Area 1

34 (34.3)

1.06 (0.36-3.16)

61 (32.3)

1.28 (0.54-3.03)

Area 2

23 (21.5)

0.93 (0.42-2.04)

41 (21.4)

1.08 (0.60-1.96)

Area 3

25 (26.9)

Referent

38 (23.3)

Referent

Area 4

19 (15.1)

0.73 (0.32-1.68)

38 (17.6)

0.93 (0.49-1.76)

Area 5

15 (28.3)

0.85 (0.26-2.82)

30 (28.9)

1.67 (0.67-4.11)

Area 6

2 (18.2)

1.52 (0.27-8.52)

5 (25.0)

1.92 (0.59-6.22)

Area 7

5 (35.7)

1.45 (0.36-5.87)

5 (16.1)

0.76 (0.22-2.66)

# Adjusted for age-group, gender, histology, country of birth, degree-of-spread, tumour location, hospital admitted, area of service, Charlson comorbidity index, surgery type.

* Significant at the P<0.05 (two-tailed) level.

Rates of hospital length-of-stay >28 days (20.9% and 19.7%), 30-day mortality (3.8% and 2.7%) and one-year cancer survival (75.5% and 76.9%) were similar for oesophageal cancer and the combined cohort respectively, but differed by 4.5% for 30-day complication rates (21.5% and 17.0%). Mean one year survival time for the combined cohort was 328 days (SD 48 days).

When investigating the factors associated with longer length of hospital stay and one-year survival, the inclusion of cancer of the gastro-oesophageal junction did not alter the trends established with that of oesophageal cancer alone. However, due to increased precision, some of the associations that were not statistically significant in oesophageal cancer alone became significant in the overall cohort. For longer length of stay these factors were: being Australian born (aOR=1.49, 95%CI 1.01-2.17) and having surgery in a public co-located hospital (aOR=2.10, 95%CI 1.17-3.78). Other factors associated with length of stay >28 days included older age (p<0.001), having an admission in a public hospital (p<0.001) and not being admitted in certain area health services. Poorer survival in the overall cohort was associated with older age (p<0.001) and private non-co-located hospitals compared with private co-located hospitals (aHR=2.20, 95%CI 1.03-4.70).

We did not observe any significant associations for 30 day post-resection complication in the overall cohort. However, in the oesophageal cohort we found older age and some area health services were associated with increased complication rates. The loss of statistical significance in the overall cohort may be attributed to the lower complication rate with gastro-oesophageal cancer.

The small number of patients dying within 30-days of surgery did not permit us to perform adjusted analysis for this outcome measure.

Outcomes following surgery in patients with non-metastatic adenocarcinoma

Rates of hospital length-of-stay >28 days (17.3% and 19.6%), 30-day mortality (2.4% and 2.8%), 30-day complication rates (19.0% and 19.7%) and one-year cancer survival (80.6% and 79.7%) were similar for patients with non-metastatic adenocarcinoma oesophageal cancer and the combined cohort respectively.

Factors associated with poorer post-resection outcomes in the sub-cohort were similar to those reported in all patients. However, in this group of patients with non-metastatic adenocarcinoma, we did not observe any significant associations for 30-day post-resection complication in the combined cohort or in oesophageal cancer alone.

Discussion

There is an increasing use of health administrative data for research and in health policy and planning. We examined rates of oesophagectomy on oesophageal cancer patients and whether the inclusion of patients with gastro-oesophageal cancer affected oesophagectomy rates and post-resection outcomes in NSW. Unsurprisingly, the addition of gastro-oesophageal cancer patients to our cohort increased the oesophagectomy rates by 3.5%. Clearly if hospital funding is based partly on patient admissions and the volume of procedures undertaken, care must be taken to ensure that all potential ‘qualifying’ patients are included in any performance assessment. Our reported rate of oesophagectomy in patients is comparable to previously reported European and UK studies [30, 31] However, the UK rate has been decreasing [38] and may be due to the highly specialised pre- and peri-operative assessment and management required for conducting an oesophagectomy [39, 40], and the support for centralization of oesophagectomies being performed in dedicated multidisciplinary centres [41, 42].

Due to increased precision, the inclusion of gastro-oesophageal cancer resulted in an increase in the number of predictor variables associated with rates of resection and post-resection outcomes that were not seen with oesophageal cancer alone (with the exception of 30-day complication rates). For our overall cohort, factors associated with receiving an oesophagectomy included patient characteristics (age, gender, birthplace) cancer characteristics (histology, disease spread) and hospital related factors (location and funding sector); similar to previously published work [31]. 30-day mortality in this study was low, in-line with previous studies and international benchmarks [9, 38, 43]. We found that patients undergoing surgical resection in public hospitals were more likely to have a length of stay >28 days. This may reflect the epidemiology of oesophageal and gastro-oesophageal cancer and that more complex cases are treated in public hospitals (teaching hospitals linked to academic centres in Australia). Thus, the addition of gastro-oesophageal cancer did impact on factors associated with receiving an oesophagectomy and on some post-surgery outcomes.

One-year cancer survival was worse in those who underwent surgery in private hospitals that were not co-located with public hospitals, potentially indicating the benefits of surgery being conducted in higher volume or teaching hospitals [7, 9, 18, 21, 30, 41, 44, 45].

Our current study demonstrated a reduced 30-day complication rate in gastro-oesophageal cancer compared with oesophageal cancer, hence statistical associations tended towards the null in the combined cohort. This may have been due to the slightly higher rate of transhiatal oesophagectomies undertaken for adenocarcinoma, compared with that in oesophageal cancer and hence lower peri-operative morbidity [46], however the only factors which were significantly associated with higher complication rate in patients with oesophageal cancer were older age and resection being conducted in some area health services.

Although we attempted to account for comorbidity with the use of the CCI, the proportion of patients with comorbidity preceding cancer diagnosis was low and as aforementioned, due to how CCI is calculated, comorbidity was most likely under-ascertained. The under-reporting of comorbidities in administrative data may be due to incomplete data transposition from medical records in individual hospitals to administrative databases [4749]. Further, non-surgical treatments such as chemotherapy and radiotherapy are funded by the PBS and MBS respectively and this data is not linked routinely to hospital morbidity data thus it was not available for this population linkage study. Further, the only measure of disease severity available was degree-of-spread at diagnosis; there were no indicators of performance status available.

Conclusion

Oesophagectomy in NSW is performed with good outcomes. Our own experience highlights the need for identifying the appropriate patient populations under study and their unique contributions to the overall cohort, given the change in surgical resection rate and longer length-of-stay post-resection with the inclusion of gastro-oesophageal cancer. The measurement of hospital performance is a powerful tool to stimulate quality improvement, transparency and accountability and health administrative data sets have become pivotal to this activity.

Declarations

Acknowledgements

The authors wish to acknowledge NSW Health for use of the data and the CHeReL for undertaking the data linkage. This study was partially funded by the Cancer Services Cancer Institute NSW (block funding agreement with South East Sydney & Illawarra Area Health Service). The funders had no role in the study design, data collection, data analysis, data interpretation, or writing of the manuscript. The views expressed are not necessarily those of the funders.

Authors’ Affiliations

(1)
Lowy Cancer Research Centre, Prince of Wales Clinical School, University of New South Wales
(2)
Faculty of Pharmacy, University of Sydney
(3)
Kolling Institute of Medical Research, University of Sydney

References

  1. Berwick DM, Leape LL: Reducing errors in medicine. BMJ. 1999, 319: 136-137.View ArticlePubMedPubMed CentralGoogle Scholar
  2. Leathley CM, Gilbert R, Kennedy PJ, Hughes CF: Measuring hospital performance - 2008 forum summary. Med J Aust. 2010, 193: S95-S96.PubMedGoogle Scholar
  3. Clinical Excellence Commission (CEC) and NSW Department of Health: Clinical Incident Management in the NSW Public Health System 2009. 2010, Sydney: NSW Department of HealthGoogle Scholar
  4. Department of Health and Human Services Agency for Healthcare Research and Quality: Guide to Inpatient Quality Indicators: Quality of care in hospitals - volume, mortality, and utilization. 2007, Rockville: Department of Health and Human ServicesGoogle Scholar
  5. Ra J, Paulson EC, Kucharczuk J, Armstrong K, Wirtalla C, Rapaport-Kelz R, Kaiser LR, Spitz FR: Postoperative mortality after esophagectomy for cancer: Development of a preoperative risk prediction model. Ann Surg Oncol. 2008, 15 (6): 1577-1584.View ArticlePubMedGoogle Scholar
  6. Reavis KM, Smith BR, Hinojosa MW, Nguyen NT: Outcomes of Esophagectomy at Academic Centers: An Association between Volume and Outcome. Am Surg. 2008, 74 (10): 939-943.PubMedGoogle Scholar
  7. Verhoef C, van de Weyer R, Schaapveld M, Bastiaannet E, Plukker JTM: Better survival in patients with oesophageal cancer after surgical treatment in University hospitals: A plea for performance by surgical oncologists. Ann Surg Oncol. 2007, 14: 1678-1687.View ArticlePubMedPubMed CentralGoogle Scholar
  8. Spilsbury K, Semmens JB, Saunders CM, Holman CDJ: Long-term survival outcomes following breast cancer surgery in Western Australia. ANZ J Surg. 2005, 75 (8): 625-U624.View ArticlePubMedGoogle Scholar
  9. Stavrou EP, Smith GS, Baker DF: Surgical outcomes associated with oesophagectomy in New South Wales; an investigation of hospital volume. J Gastrointest Surg. 2010, 14: 951-957.View ArticleGoogle Scholar
  10. Ben-Tovim DI, Pointer SC, Woodman R, Hakendorf PH, Harrison JE: Routine use of administrative data for safety and quality purposes - hospital mortality. Med J Aust. 2010, 193: S100-S103.PubMedGoogle Scholar
  11. Chowdhury MM, Dagash H, Pierro A: A systematic review of the impact of volume of surgery and specialization on patient outcome. Br J Surg. 2007, 94 (2): 145-161.View ArticlePubMedGoogle Scholar
  12. Finlayson EVA, Goodney PP, Birkmeyer JD: Hospital volume and operative mortality in cancer surgery - A national study. Archives Surg. 2003, 138 (7): 721-725.View ArticleGoogle Scholar
  13. Holman CDJ, Wisniewski ZS, Semmens JB, Rouse IL, Bass AJ: Mortality and prostate cancer risk in 19 598 men after surgery for benign prostatic hyperplasia. BJU Int. 1999, 84 (1): 37-42.View ArticlePubMedGoogle Scholar
  14. Karanicolas PJ, Dubois L, Colquhoun PHD, Swallow CJ, Walter SD, Guyatt GH: The more the better? The impact of surgeon and hospital volume on in-hospital mortality following colorectal resection. Ann Surg. 2009, 249 (6): 954-959.View ArticlePubMedGoogle Scholar
  15. Reavis KM, Hinojosa MW, Smith BR, Wooldridge JB, Krishnan S, Nguyen NT: Hospital volume is not a predictor of outcomes after gastrectomy for neoplasm. Am Surg. 2009, 75 (10): 932-936.PubMedGoogle Scholar
  16. Shahian DM, O'Brien SM, Normand SLT, Peterson ED, Edwards FH: Association of hospital coronary artery bypass volume with processes of care, mortality, morbidity, and the Society of Thoracic Surgeons composite quality score. J Thoracic Cardiovasc Surg. 2010, 139 (2): 273-282.View ArticleGoogle Scholar
  17. Smith FJ, Holman CDJ, Moorin RE, Fletcher DR: Incidence of bariatric surgery and postoperative outcomes: a population-based analysis in Western Australia. Med J Aust. 2008, 189 (4): 198-202.PubMedGoogle Scholar
  18. Bachmann MO, Alderson N, Edwards D, Wotton S, Bedford C, Peters TJ, et al: Cohort study in South and West England of the influence of specialization on the management and outcome of patients with oesophageal and gastric cancers. Br J Surg. 2002, 89: 914-922.View ArticlePubMedGoogle Scholar
  19. Pye JK, Crumplin MKH, Charles J, Kerwat R, Foster ME, Biffin A: One-year survey of carcinoma of the oesophagus and stomach in Wales. Br J Surg. 2001, 88: 278-285.View ArticlePubMedGoogle Scholar
  20. Swisher SG, Deford L, Merriman KW, Walsh GL, Smythe R, Vaporicyan A, al E: Effect of operative volume on morbidity, mortality and hospital use after esophagectomy for cancer. J Thorac Cardivasc Surg. 2000, 119: 1126-32.View ArticleGoogle Scholar
  21. van Lanschot JJB, Hulscher JB, Buskens CJ, Tilanus HW, ten Kate FJW, Obertop H: Hospital volume and hospital mortality for esophagectomy. Cancer. 2001, 91: 1574-1578.View ArticlePubMedGoogle Scholar
  22. Wouters M, Krijnen P, Le Cessie S, Gooiker GA, Guicherit OR, Marinelli A, Kievit J, Tollenaar R: Volume- or Outcome-Based Referral to Improve Quality of Care for Esophageal Cancer Surgery in The Netherlands. J Surg Oncol. 2009, 99 (8): 481-487.View ArticlePubMedGoogle Scholar
  23. Ajani J, Barthel JS, Bekaii-Saab T, Bentrem D, D'Amico TA, Das P, et al: Gastric Cancer, NCCN Practice Guidelines in Oncology v22010. 2010, Fort Washington: National Comprehensive Cancer NetworkGoogle Scholar
  24. Ajani J, Barthel JS, Bekaii-Saab T, Bentrem DJ, D'Amico TA, Das P, et al: Esophageal Cancer, NCCN Clinical Practice Guidelines in Oncology v12010. 2010, Fort Washington: National Comprehensive Cancer NetworkGoogle Scholar
  25. Stavrou EP, McElroy HJ, Baker DF, Bishop JF: Adenocarcinoma of the oesophagus: incidence and survival rates in New South Wales, 1972–2005. Med J Aust. 2009, 191: 310-314.PubMedGoogle Scholar
  26. Semmens JB, Lawrence-Brown MMD, Fletcher DR, Rouse IL, Holman CDJ: The quality of surgical care project: A model to evaluate surgical outcomes in Western Australia using population-based record linkage. ANZ J Surg. 1998, 68 (6): 397-403.View ArticleGoogle Scholar
  27. Semmens JB, Aitken RJ, Sanfilippo FM, Mukhtar SA, Haynes NS, Mountain JA: The Western Australian audit of surgical mortality: advancing surgical accountability. Med J Aust. 2005, 183: 504-508.PubMedGoogle Scholar
  28. Board N, Watson DE: Using what we gather - harnessing information for improved care. Med J Aust. 2010, 193: S93-S94.PubMedGoogle Scholar
  29. Semmens JB, Mountain JA, Sanfilippo FM, Barraclough JY, McKenzie A, Mukhtar SA, Haynes NS, Aitken RJ: Providers and consumers support the Western Australian Audit of surgical mortality. ANZ J Surg. 2006, 76: 442-447.View ArticlePubMedGoogle Scholar
  30. Gillison EW, Powell J, McConkey CC, Spychal RT: Surgical workload and outcome after resection for carcinoma of the oesophagus and cardia. Br J Surg. 2002, 89: 344-348.View ArticlePubMedGoogle Scholar
  31. Verhorf C, van de Weyer R, Schaapveld M, Bastiaannet E, Plukker JTM: Better survival in patients with oesophageal cancer after surgical treatment in University hospitals: A plea for performance by surgical oncologists. Ann Surg Oncol. 2007, 14: 1678-1687.View ArticleGoogle Scholar
  32. International Agency for Research on Cancer: Manual for Cancer Registry Personnel: IARC Technical Report No 10. Lyon: IARC. 1995, 81: 1-4.Google Scholar
  33. Barraclough H, Morrell S, Arcorace M, McElroy H, Baker DF: Degree-of-spread artefact in the New South Wales Central Cancer Registry. Aust NZ J Public Health. 2008, 32: 414-416.View ArticleGoogle Scholar
  34. National Centre for Classification in Health: ICD-10-AM/ACHI/ACS. 2008, Sydney: University of Sydney, SixthGoogle Scholar
  35. Charlson ME, Pompei P, Ales KL, McKenzie CR: A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987, 40: 373-383.View ArticlePubMedGoogle Scholar
  36. Deyo RA, Cherkin DC, Ciol MA: Adapting a clinical co-morbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992, 45: 613-619.View ArticlePubMedGoogle Scholar
  37. Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi JC, et al: Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. 2005, 43: 1130-1139.View ArticlePubMedGoogle Scholar
  38. Allum WH, Blazeby JM, Griffin SM, Cunningham D, Jankowski JA, Wong R: On behalf of the Association of Upper Gastrointestinal Surgeons of Great Britain and Ireland, the British Society of Gastroenterology and the British Association of Surgical Oncology: Guidelines for the management of oesophageal and gastric cancer. Gut. 2011, 60: 1449-1472.View ArticlePubMedGoogle Scholar
  39. Morita M, Nakanoko T, Fujinaka Y, Kubo N, Yamashita Y, Yoshinaga K, et al: In-hospital mortality after a surgical resection for esophageal cancer: Analyses of the associated factors and historical changes. Ann Surg Oncol. 2011, 18: 1757-1765.View ArticlePubMedGoogle Scholar
  40. Fujita H, Ozawa S, Kuwano H, Ueda Y, Hattori S, Yanagawa T: Japanese Assoc Thoracic S: Esophagectomy for cancer: clinical concerns support centralizing operations within the larger hospitals. Dis Esophagus. 2010, 23 (2): 145-152.View ArticlePubMedGoogle Scholar
  41. Rouvelas I, Lagergren J: The impact of volume on outcomes after oeseophageal cancer surgery. ANZ J Surg. 2010, 80: 634-641.View ArticlePubMedGoogle Scholar
  42. Anderson O, Ni Z, Moller H, Coupland VH, Davies EA, Allum WH, Hanna GB: Hospital volume and survival in oesophagectomy and gastrectomy for cancer. Eur J Cancer. 2011, 47: 2408-2414.View ArticlePubMedGoogle Scholar
  43. Royal College of Surgeons of England: National Oesophago-Gastric Cancer Audit. Third Annual Report. 2010, London: The NHS Information Centre, 1-69.Google Scholar
  44. Wouters MW, Wijnhoven BP, Karim-Kos HE, Blaauwgeers HG, Stassen LP, Steup W-H, et al: High-volume versus low-volume for esophageal resections for cancer: The essential role of case-mix adjustments based on clinical data. Ann Surg Oncol. 2008, 15: 80-87.View ArticlePubMedGoogle Scholar
  45. Fujita H, Kuwano H, Ueda Y, Hattori S, Yanagawa T: Esophagectomy for cancer: clinical concerns support centralizing operations within the larger hospitals. Dis Esophagus. 2010, 23: 145-152.View ArticlePubMedGoogle Scholar
  46. Hulscher JB, van Sandick JW, de Boer AG, Wijnhoven BP, Tijssen JG, Fockens P, et al: Extended transthoracic resection compared with limited transhiatal resection for adenocarcinoma of the esophagus. N Engl J Med. 2002, 347: 1662-1669.View ArticlePubMedGoogle Scholar
  47. Preen DB, Holman CDJ, Lawrence DM, Baynham NJ, Semmens JB: Hospital chart review provided more accurate comorbidity information than data from a general practitioner survey or an administrative database. J Clin Epidemiol. 2004, 57: 1295-1304.View ArticlePubMedGoogle Scholar
  48. Powell H, Lim LL-Y, Heller RF: Accuracy of administrative data to assess comorbidity in patients with heart disease: an Australian perspective. J Clin Epidemiol. 2001, 54: 687-693.View ArticlePubMedGoogle Scholar
  49. Hawker GA, Coyte PC, Wright JG, Paul JE, Bombardier C: Accuracy of administrative data for assessing outcomes after knee replacement surgery. J Clin Epidemiol. 1997, 50: 265-273.View ArticlePubMedGoogle Scholar
  50. Pre-publication history

    1. The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1472-6963/12/384/prepub

Copyright

© Stavrou et al.; licensee BioMed Central Ltd. 2012

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.