Cost-effectiveness of an early awareness campaign for colorectal cancer Sophie Whyte 1, Sue Harnan 1, Paul Tappenden 1, Mark Sculpher 2, Seb Hinde 2, Claire Mckenna 2 Policy Research Unit in Economic Evaluation of Health and Care Intervention (EEPRU) 1 School of Health and Related Research (ScHARR), University of Sheffield 2 Centre for Health Economics, University of York Contact: [email protected]
18
Embed
Cost-effectiveness of an early awareness campaign for colorectal cancer
Cost-effectiveness of an early awareness campaign for colorectal cancer
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Cost-effectiveness of an early awareness campaign for colorectal cancer
Sophie Whyte 1, Sue Harnan 1, Paul Tappenden 1, Mark Sculpher 2, Seb Hinde 2, Claire Mckenna 2
Policy Research Unit in Economic Evaluation of Health and Care Intervention (EEPRU)
1 School of Health and Related Research (ScHARR), University of Sheffield
• The primary aim of cancer awareness campaigns is earlier presentation of symptomatic cancers through improved public knowledge of the symptoms
Data from early awareness campaigns
Data from the pilot early awareness campaign for colorectal cancer showed an increase in GP referrals for a short period following the campaign.
Awareness campaign efficacy
Key questions:• Does this data mean the campaign
‘worked’?• Should it be repeated?
To understand benefit of campaign need information on longer-term outcomes such as ‘change in cancer incidence or mortality’
• A campaign may also lead to increased numbers of GP attendances by the ‘worried well’
• Some critics assert that the campaigns will ‘undo years of work persuading patients with minor ailments to stay at home’.
Project Aim
• Estimate cost effectiveness of an early awareness campaign using data from the colorectal cancer early awareness campaign piloted in the East of England and south west regions in January 2011.
• The estimates provide an improved understanding of the benefits of such a campaign so can be used to inform policy decisions.
Potential effects of an early awareness campaign for colorectal cancer
Public awareness of signs and symptoms Increase in awareness?
Early Awareness campaign TV, radio, press, online, etc.
GP consultations Increase?
GP referrals for suspected CRC Increase?
Cancer diagnoses Change in incidence/stage distribution?
Diagnosis of other lower GI conditions Increase?
Cancer mortality Decrease?
CRC screening Increase in uptake?
lives saved
treatment costs
consultation costs
campaign costs
referral costs
screening costs
Scope of analysis
The analysis captures:• the direct costs of the campaign, • the costs any additional GP
consultations/appointments in secondary care resulting from the campaign
• benefits of the campaign due to earlier diagnosis and any change in screening uptake.
Data from the pilot campaign used in the modelling
Data observed from pilot campaign
Base case assumption in model
Scenario analyses
GP attendances 700 increase over 3 month period (532 increase if diarrhoea included as a symptom) Equivalent to 60,000-80,000 nationally.
70,000 more attendances nationally over 3 month period
Assumed 50% ‘additional’& 50% ‘earlier’
Assumed 90% ‘additional’ & 10% ‘earlier’
GP referrals 1,956 increase in referrals over 5 month period (+28%)
17,519 additional referrals nationally
Assumed 50% ‘additional’& 50% ‘earlier’
Assumed 90% ‘additional’ & 10% ‘earlier’
CRC incidence 7-11% increase in incidence for 1 month
10% increase in presentation rates for 1 month
5-20% magnitude 1-6 month duration
CRC incidence stage distribution
Numbers too small to draw any conclusions
Campaign assumed to have the same proportional effect on presentation rates for each CRC stage.
Short term increase in incidence only consists of Dukes stages C & D
CRC screening uptake
No significant change which could be attributed to the campaign
Assume screening uptake unaffected by campaign
Exploratory analysis undertaken
Cost of running campaign
£5 million £5 million -
Methods• Pilot data demonstrates short term
impacts of the awareness campaign.• A mathematical model was used in
combination with the pilot data to predict long term impacts of the campaign on cancer incidence, mortality and costs.
• An existing mathematical model [1] was adapted (representing the CRC disease natural history, symptomatic presentation and the bowel cancer screening programme).
[1] Re-appraisal of the options for colorectal cancer screening in England; Whyte S, Chilcott J, Halloran S, (Colorectal Disease, March 2012)
CRC=colorectal cancer, LR=low risk, HR=high risk
Normal Epithelium
LR adenomas
HR adenomas
Dukes’ A CRC
Dukes’ B CRC
Dukes’ C CRC
Stage D CRC
Dead (CRC)
Dukes’ A CRC clinical
Dukes’ C CRC clinical
Stage D CRC clinical
Dukes’ B CRC clinical
Dead (non-CRC)
Transition estimated within model calibration
Transition estimated directly from mortality data
CRC natural history model CRC screening pathways
Invited to screening
Screening test completed
Do not attend screening
Do not attend follow up
No adenomas
Positive screening result – refer to follow up (colonoscopy)
Negative screening test result / LR adenomas found
Return to general screening population
LR adenomas
HR adenomas
CRC CRC treatment
Attend follow up
Surveillance (annual/ 3-yearly colonoscopy)
Model structure
Modelling methodology
• Four rates relating to symptomatic or chance presentation with Dukes’ A-D CRC. Baseline presentation rates reflect the England population from years 2004 to 2006 i.e. before screening commenced.
• The four transition probabilities are increased to result in an increase in incidence which matches the observed increase seen in the pilot campaign.
• Assumption: campaign causes a temporary change in the transition probabilities and that subsequently these probabilities will return to their pre-campaign values.
• No data available on stage distribution of incidence. -> Assume that the extra incidence due to the awareness campaign has the usual CRC stage distribution.
Results-Effectiveness
The campaign causes: Dukes’ stage A-C CRC presenting
Overall the campaign lead to increase in NHS costs
Campaign running cost¯ Screening costs (caused by a decrease in
positives at screening since more CRC presents symptomatically)
CRC treatment costs (1) CRC is presenting at younger ages which are
associated with higher treatment costs.
(2) A shift of cases from stage D to Dukes’ C and Dukes’ C CRC is associated with higher treatment costs than Dukes’ D.
Costs associated with increased GP consultations and referrals (account for only a small proportion of total costs and are considerably less than the cost of the campaign itself)
CRC incidence - symptomatic presentation TOTAL 20CRC incidence screen/surveillance detected Dukes Stage A -0
B -1 C -2 D -2
CRC incidence - screening/surveillance detected TOTAL -5 CRC-specific deaths -66 Deaths with undiagnosed CRC -14 Total costs related to screening (discounted) 3,407-£ Cancer management (inc. pathology) costs (discounted) 94,443£ Cost of additional GP consultations/referrals (discounted) 855,716£ Cost of awareness campaign (discounted) 4,499,995£ Total cost (discounted) 5,446,745£ Total life years gained (discounted) 622 Total QALYs gained (discounted) 404 ICER 13,496£ NMB 2,624,770£
Model predictions for the current population of England evaluated over a lifetime: Change compared to 'No awareness campaign'
For a CRC awareness campaign resulting in a 10% increase in presentation rates for a period of one month
Modelling uncertainty in change in presentation rates due to campaign: duration and magnitude
0%
5%
10%
15%
20%
0 1 2 3 4 5 6
Increase in symptomatic
presentation rate (%)
Duration of increase in symptomatic presentation rate (months)
Number of CRC deaths prevented
800-900
700-800
600-700
500-600
400-500
300-400
200-300
100-200
0-100
base case= 66 deaths prevented
0.05
0.1
0.15
0.2
1 2 3 4 5 6
Increase in symptomatic presentation rate (%)
Duration of increase in symptomatic presentation rate (months)
ICER
25000-30000
20000-25000
15000-20000
10000-15000
5000-10000
0-5000
base case= ICER=£13K per QALY
Modelling uncertainty in change in presentation rates due to campaign: duration and magnitude
Priorities for future research
Co-ordinate and maximise the evaluation and dissemination of efforts that have already been made to increase cancer awareness.• comparison with non-intervention regions• clear reporting of completeness of data and potential
data limitations
Information of importance for future modelling studies:• duration of effect of campaign• effect of campaign on CRC incidence• effect of campaign on emergency presentation rates• effect of campaign by age• differential diagnoses costs associated with emergency
presentation versus two-week wait referrals • rates of diagnosis of other lower GI conditions with