SEARCH

SEARCH BY CITATION

Keywords:

  • bladder cancer;
  • deprivation;
  • delay;
  • survival

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

OBJECTIVES

To investigate the relationship between deprivation, delay and survival from bladder cancer in the West Midlands, as socio-economic deprivation is associated with worse survival in many malignancies, and it has been suggested that treatment differences and delay in seeking care are major contributing causes.

PATIENTS AND METHODS

Data were prospectively collected on 1537 newly diagnosed cases of urothelial cancer presenting in the West Midlands between January 1991 and June 1992. Survival was censored at 31 July 2000, when 785 (51%) patients had died. The influence of deprivation on survival was explored using cause-specific and all-cause mortality.

RESULTS

Patients in less affluent groups had significantly worse survival than patients in more affluent groups when considering deaths from all causes (P = 0.02), which held true when adjusting for independent prognostic factors (age, smoking history, and tumour grade, stage, type and size). Bladder cancer-specific mortality showed no significant difference between socio-economic groups (P = 0.30).

CONCLUSION

Socio-economic deprivation is a significant predictor of survival when death from all causes is considered. However, this does not hold true for bladder cancer-specific death. The perceived differences in treatment and delay between socio-economic groups do not seem to occur for bladder cancer in the West Midlands.


Abbreviations
ED

enumeration district

IQR

interquartile range;

INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Several biological tumour characteristics have been identified as prognostic factors for urothelial cancer of the bladder. These include tumour stage, grade, size, and multiplicity, in addition to many molecular markers (reviewed in [1]). Several nonbiological factors also appear to play a role in cancer prognosis and these include delay in diagnosis and treatment, and socio-economic deprivation. Large studies of several malignancies have provided conflicting evidence of the impact of delay on survival [2–8]. However, in bladder cancer a longer delay between the onset of symptoms and referral to a urologist results in higher stage tumours and worse patient survival, and a shorter hospital delay (the delay from referral by the family practitioner to treatment by a urologist) also results in worse patient survival [9,10]. In addition, for many malignancies patients in lower socio-economic groups have worse survival than have more affluent patients [11–21], and it has been proposed that differences in treatment and delay in seeking care are major contributing causes [12]. However, there have been no detailed studies specifically of bladder cancer. Several different indicators of socio-economic status have been used in previous studies, including housing tenure [12], occupation [13], the Carstairs Index [14,15,19], and the Townsend score [17]. However, the variations in cancer survival in different socio-economic groups occur regardless of sex, country and the source of the study population, and do not depend on the socio-economic indicator used [18]. These previous studies show clearly that there is large potential for reducing cancer mortality in lower socio-economic groups if these differences can be understood. The aim of this research was to study the factors involved with socio-economic deprivation, delay and survival, specifically for bladder cancer.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

In 1991 data were collected on all patients newly presenting with urothelial cancers in the West Midlands Region, as part of a study investigating occupational histories. Additional data were collected on delay times, as previously published [10]. Delay times were calculated in days from four time points: delay 1, the date of onset of symptoms to date of first referral by a family practitioner; delay 2, the date of family practitioner referral to the date of first attendance at hospital; delay 3, the date of first hospital attendance to the date of first treatment; total delay, the date of onset of symptoms to date of first treatment.

Patients were asked to provide their full postcode; the measure of deprivation for each patient was based on the postcode from their usual residence at time of diagnosis by linking their postcode to the corresponding census enumeration district (ED) (http://census.ac.uk/cdu/Datasets/Lookup_tables/Postal/Postcode_Enumeration_District_Directory.htm). Within the UK each ED contains on average 400 households. For each of the 10 774 EDs in the West Midlands, data from the 1991 Census were obtained through Manchester Information & Associated Services and the West Midlands Cancer Intelligence Unit for four variables, i.e. unemployment, car and home ownership, and overcrowding. The Townsend Material Deprivation score is calculated by combining these four variables to provide a single score, representing material deprivation, which is then standardized for the West Midlands population. The Townsend score was ranked for all EDs from low scores (affluent patients) to high scores (deprived patients) and categorized into five deprivation groups using thresholds devised from the West Midlands population available on the Multi-Agency Internet Geographic Information Service website (http://www.maigis.wmpho.org.uk/), group 1 being the most affluent and group 5 being the most deprived. Patients were flagged with the West Midlands Cancer Registry and censored to 31 July 2000. Although we aimed to recruit all incident patients, we may have missed up to 14%, as calculated from final cancer registry values for this period. Private patients may be under-represented, along with patients presenting with advanced disease who may not have been diagnosed by a urological unit.

STATISTICS

The data were analysed using commercial software; differences in patient and tumour characteristics and Townsend score were investigated using the Mantel-Haenszel chi-squared test. The primary endpoint for analysis was death, where both cause-specific and all-cause mortality were investigated. Survival curves were constructed using the method of Kaplan and Meier [22] and the log-rank test [23] used to assess differences between groups. Survival was defined as date of first treatment to the date of death or until the censor date of 31 July 2000. Survival was compared in terms of deprivation and known prognostic factors; age, smoking history, tumour type, size, grade and stage. Cox regression [24] was used to determine independent predictors of survival. All influential factors were included in a stepwise analysis, then to minimize the effect of missing data the analysis was re-run including only the independent factors identified from the final model.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Data were collected on 1537 patients across 10 774 EDs within the West Midlands; comparing the percentage of cases in each deprivation group in this study (groups 1–5, respectively 20%, 13%, 17%, 28% and 22%) with the percentage of EDs within the West Midlands (24%, 14%, 17%, 23% and 22%) verified that the study population was a representative sample from the whole West Midlands population, with half of the patients in the two most deprived groups and only a fifth in the most affluent group. At presentation the median age of patients was 69 years, interquartile range (IQR) 62–78 years; 75% were male, 87% had macroscopic haematuria, 77% were current or previous smokers, 65% had papillary tumours, 75% had a single tumour, 73% of tumours were stage Ta/T1 and 27% were WHO grade 3.

SOCIO-ECONOMIC FACTORS, CHARACTERISTICS, DELAY AND SURVIVAL

Patient and tumour characteristics were analysed by deprivation group, as shown in Table 1. Less affluent patients tended to be female (P = 0.05), be current or ex-smokers (P = 0.02) and presented with macroscopic haematuria (P = 0.04). There were no significant differences for age, tumour type, tumour stage, tumour size, tumour multiplicity and tumour grade among socio-economic groups.

Table 1.  Patient and tumour characteristics by socio-economic deprivation group
FactorGroupDeprivation group, n (%)Chi squareP
Affluent234Deprived
  1. Delay 1, onset of symptoms to first referral by GP; Delay 2, first referral to first attendance at hospital; Delay 3, first attendance at hospital to first treatment; Total delay, onset of symptoms to first treatment.

GenderFemale 62 (21) 54 (27) 64 (25)108 (27) 90 (28)3.820.05
Male240 (79)144 (73)188 (75)298 (73)231 (72)  
Age, years≤60 70 (23) 54 (27) 38 (15) 81 (20) 70 (22)0.130.72
61–70 95 (31) 64 (32) 82 (33)126 (31)115 (36)  
71–80104 (34) 51 (26) 88 (35)150 (37)102 (32)  
≥80 33 (11) 29 (15) 44 (17) 49 (12) 34 (11)  
SmokedNever 73 (27) 53 (29) 64 (29) 69 (19) 59 (22)5.100.02
Ever201 (73)129 (71)155 (71)285 (81)207 (78)  
TumourPapillary197 (69)131 (70)154 (65)231 (59)197 (64)2.110.15
Mixed 41 (14) 29 (16) 39 (16) 79 (20) 69 (22)  
Solid 47 (16) 27 (14) 44 (19) 81 (21) 41 (13)  
StagePTa144 (55) 96 (55)115 (53)162 (45)143 (50)2.840.09
pT1 52 (20) 39 (22) 38 (17) 92 (26) 69 (24)  
T2–4 66 (25) 39 (22) 65 (30)103 (29) 75 (26)  
Size, cm≤2120 (43) 78 (42) 91 (41)151 (39)110 (37)2.900.09
>2160 (57)106 (58)129 (59)234 (61)191 (63)  
Single225 (79)133 (71)172 (76)282 (72)227 (75)1.020.31
Multiple 61 (21) 55 (29) 55 (24)107 (28) 77 (25)  
GradeWell 84 (31) 72 (39) 85 (38)129 (35)105 (35)0.260.61
Moderate126 (46) 64 (35) 81 (36)137 (37)110 (37)  
Poor 61 (23) 48 (26) 60 (27)105 (28) 82 (28)  
HaematuriaNone 18 (8) 21 (14) 14 (8) 15 (5) 16 (6)4.420.04
Micro 13 (5) 13 (9) 10 (5) 17 (5) 14 (5)  
Macro206 (87)115 (77)161 (87)281 (90)231 (89)  
Delay 1, days≤14152 (51) 98 (50)133 (55)197 (49)155 (49)2.610.62
>14145 (49) 97 (50)107 (45)201 (51)159 (51)  
Delay 2≤28147 (49) 94 (48)128 (53)186 (46)173 (55)6.800.15
>28150 (51)101 (52)112 (47)216 (54)141 (45)  
Delay 3≤20155 (52) 99 (51)120 (50)196 (49)151 (48)1.300.86
>20142 (48) 96 (49)120 (50)206 (51)163 (52)  
Total delay≤110156 (52) 86 (44)134 (55)194 (48)164 (52)6.650.16
>110143 (48)109 (56)109 (45)207 (52)151 (48)  

There were no significant differences in delay times among socio-economic groups for all delay categories (Table 1). At the censor date of 31 July 2000, 752 patients (49%) remained alive and the median (IQR) follow-up was 8.4 (7.8–9.0) years. There was a significant difference in survival among the five socio-economic groups when considering all causes of deaths (log-rank P = 0.02; Fig. 1). Survival was highest in the two most affluent groups (1 and 2). Comparing the two most affluent groups with the two most deprived groups exaggerated this survival difference (log-rank P = 0.004). The 5- and 8-year survival rates (56% and 45%, respectively) were identical for groups 4 and 5, and substantially lower than the 5- and 8-year survival for groups 1 and 2 (63% and 67%, and 53% and 57%, respectively; Table 2). However, cause-specific analysis showed no difference in survival among deprivation groups (log-rank P= 0.30).

image

Figure 1. Survival by socio-economic deprivation group.

Download figure to PowerPoint

Table 2.  Survival rates at 5 and 8 years by socio-economic group
Deprivation groupSurvival, % (95% CI)
5-year8-year
1 Affluent63 (56, 69)53 (44, 62)
267 (58, 75)57 (46, 67)
358 (50, 66)48 (37, 59)
456 (50, 63)45 (36, 54)
5 Deprived56 (49, 64)45 (36, 55)
All combined59 (56, 62)49 (44, 53)

CAUSES OF DEATH

In all, 785 patients (51%) had died at the censor date, of whom 720 (92%) had a known cause of death; 350 died from bladder cancer (49% of deaths), 135 from other cancers (19%), and there were 235 deaths from other than cancer (32%). In 708 cases both the deprivation group and cause of death were known; there was no relationship between certified cause of death and deprivation group (chi square 0.05, Ptrend= 0.82; Table 3).

Table 3.  Certified causes of death by deprivation group, as n (%), in the 708 cases where both deprivation group and cause of death were known
Cause of deathAffluent234DeprivedTotal
Bladder cancer 67 (52)45 (52) 59 (48)114 (54) 70 (45)355
Other cancer 38 (29)18 (21) 38 (31) 59 (28) 53 (34)206
Other causes 25 (19)24 (28) 26 (21) 40 (19) 32 (21)147
Total13087123213155708

MULTIVARIATE ANALYSIS

To assess the independent effect of deprivation on survival a Cox regression model was applied to the data. A base model of independent prognostic factors was formed consisting of age (continuous), smoking history (continuous), and tumour type (papillary, solid, mixed), stage (Ta, T1, T2–T4), grade (well, moderate, poor) and size (continuous). Applying this model confirmed deprivation as an independent prognostic factor (P= 0.04; Table 4).

Table 4.  Cox regression table to predict survival for patients with bladder cancer
Factor, groupingCoefficientChi square to remove/enter*PRisk ratio95% CI
  • *

    If a variable does not enter the model, the chi-square to remove is the chi-square to enter.

Age (continuous)0.05264104.74<0.0011.051.04–1.07
Tumour type (papillary, mixed, solid)0.271 12.60<0.0011.311.13–1.52
Tumour stage (pTa, pT1, T2-T4)0.27086 11.09<0.0011.311.12–1.54
Smoking history (continuous)0.00688 11.08<0.0011.011.00–1.01
Grade (well, moderate, poor)0.18880  7.560.0061.211.06–1.38
Tumour size (continuous)0.08205  7.120.0081.091.02–1.15
Deprivation (1–5)0.06815  4.390.041.071.00–1.14

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

For a wide range of malignancies, patients in lower socio-economic groups have poorer cancer survival than more affluent patients [11–21]. Proposed explanations for these differences include delays in seeking care (resulting in higher-stage tumours at presentation) and treatment differences [12]. However, the present study does not confirm these findings for bladder cancer in the West Midlands. In these patients there were no differences in delays in diagnosis and treatment, stage at diagnosis, or tumour characteristics (type, stage, size, number or grade) among socio-economic groups. There were some differences in patient characteristics between socio-economic groups, including type of haematuria (P = 0.04), smoking history (P = 0.02) and gender (P = 0.05). The incidence of macroscopic haematuria at presentation was higher in the more deprived groups of patients. These findings may reflect a difference in the availability and use of corporate health schemes, private health insurance examinations, and family practitioner screening clinics (e.g. ‘well-woman’ clinics) where urine dipstick testing is common. More deprived groups tended to be previous or current smokers. Smoking is considered to be the single most important cause of urothelial cancer [25], but this had no effect on bladder cancer-specific survival among different socio-economic groups in this study. However, patients in more deprived groups had a significantly worse all-cause survival than patients in more affluent groups (P = 0.02), and smoking may have influenced this through cardiorespiratory disease and other smoking-related illnesses. It therefore appears that comorbid factors have a significant influence on the bladder cancer population, as previously reported [10].

The National Health Service Cancer Plan introduced in September 2000 (http://www.doh.gov.uk/cancer/pdfs/cancerplan.pdf) set out the first ever comprehensive strategy to tackle the disease, linking prevention, diagnosis, treatment, care and research. One of the four main aims was ‘to tackle the inequalities in health that mean unskilled workers are twice as likely to die from cancer as professionals’. One of the proposed methods was to reduce waiting times, with the delay from urgent referral to treatment to be no longer than 1 month. In addition, this Cancer Plan also recognized that the public could be more aware of important symptoms, thus prompting patients to visit their family practitioners earlier. However, we detected no such ‘inequalities in health’ for the diagnosis and treatment of patients with bladder cancer in the West Midlands.

In addition, the two factors the government have identified to reduce cancer risk are diet and smoking. They quote that ‘Children in disadvantaged families are 50% less likely to eat fruit and vegetables than those in a high income families’. Although diet was not investigated in this study, previous studies have shown that dietary factors can have an effect in both the primary prevention of malignancies (including bladder cancer) and in improving survival in patients with established malignancies [26–30]. Specifically, the consumption of green-yellow vegetables and fruit appears to be protective against bladder cancer [27], and reducing dietary fat and increasing fruit and vegetable intake can improve survival in patients diagnosed with cancer [28–30]. Smoking was included in the multivariate analysis, although we did not collect detailed data about the level of smoking or smoking cessation after a diagnosis of cancer. However, diet and smoking had no influence on bladder cancer-specific survival but, along with comorbid factors, may contribute to the significant differences in all-cause mortality.

Clearly the present study only considered patients who have already presented with bladder cancer. It is impossible to know whether there are significant differences in the number of patients with undiagnosed bladder cancer among the different socio-economic groups.

In conclusion, there were no significant differences in bladder cancer-specific mortality among five socio-economic groups, showing that there are no ‘inequalities in health’ for the diagnosis and treatment of bladder cancer in the West Midlands. However, there were significant differences in all-cause mortality, and these may reflect differences in smoking habits, diet and comorbidity in these groups. Addressing these factors may improve the overall survival of the bladder cancer population.

ACKNOWLEDGEMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

We thank all participating clinicians and their staff from within the West Midlands Region and the West Midlands Cancer Intelligence Unit for their support. We also thank Ralph Smith for providing the Townsend scores. The work was funded by a grant from the Health and Safety Executive and supported by the Cancer Research UK Clinical Trials Unit and the Institute of Occupational Health, at The University of Birmingham. R.T. Bryan is funded by a Medical Research Council Research Training Fellowship.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES