The Appropriateness of More Intensive Colonoscopy Screening than Recommended in Medicare Beneficiaries

Abstract

Importance

Many Medicare beneficiaries undergo more intensive colonoscopy screening than recommended. Whether this is favorable for beneficiaries and efficient from a societal perspective is uncertain.

Objective

To determine whether more intensive colonoscopy screening than recommended is favorable for Medicare beneficiaries (i.e. whether it results in a net health benefit), and if so, whether it is efficient from a societal perspective (i.e. whether the net health benefit justifies the additional resources required).

Design

A microsimulation modeling study.

Setting, Participants

65-year-old Medicare beneficiaries at average risk for colorectal cancer (CRC) and with an average life-expectancy who underwent a negative screening colonoscopy at age 55.

Interventions

Colonoscopy screening as recommended by guidelines (i.e. at ages 65 and 75) versus scenarios of more intensive screening in which either a shorter screening interval was applied (5 or 3 instead of 10 years) or in which screening was continued beyond age 75 (up to age 85 or 95).

Main Outcomes and Measures

Quality-adjusted life-years (QALYs) gained (measure of net health benefit); additional colonoscopies required per additional QALY gained, additional costs per additional QALY gained (measures of efficiency).

Results

Screening previously screened Medicare beneficiaries more intensively than recommended resulted in only small increases in CRC deaths prevented and life-years gained. In comparison, the increases in colonoscopies performed and colonoscopy-related complications experienced were large. As a result, all scenarios of more intensive screening than recommended resulted in a loss of QALYs, rather than a gain (i.e. a net harm). The only exception was shortening the screening interval from 10 to 5 years; however, this scenario was inefficient, as it required no less than 909 additional colonoscopies and an additional $711,000 per additional QALY gained. Results in previously unscreened beneficiaries were slightly less unfavorable, but conclusions were identical.

Conclusion and Relevance

Screening Medicare beneficiaries more intensively than recommended is not only inefficient from a societal perspective; often it is also unfavorable for those being screened. This study provides strong evidence and a clear rationale for clinicians and policy makers to actively discourage this practice.

Introduction

All guidelines for colorectal cancer (CRC) screening recommend a screening interval of 10 years for colonoscopy screening in average risk individuals.^1-4 Moreover, the U.S. Preventive Services Task Force (USPSTF) and the American College of Physicians recommend against routine screening in adults aged over 75 with an adequate screening history.^1,3 While CRC screening is well-known to be underused in many Medicare beneficiaries,^5,6 recent studies have also demonstrated that many beneficiaries undergo more intensive colonoscopy screening than recommended:^7,8 One in five beneficiaries with a negative screening colonoscopy undergo a repeated screening colonoscopy within 5 years' time instead of after 10 years. Furthermore, one in four beneficiaries with a negative screening colonoscopy at age 75 or older receive yet another screening colonoscopy at an even more advanced age. Although the reasons for these practices will vary, in some occasions they are likely to result from the beneficiary's or clinician's perception that screening should occur more frequently than recommended. However, whether such practices are actually favorable for Medicare beneficiaries (i.e. whether they result in a net health benefit) is uncertain: The low risk for CRC after a negative screening colonoscopy limits the life-years (LYs) that can be gained by applying a shorter screening interval than recommended,^9-13 whereas the high risk for other cause mortality at advanced age limits the LYs that can be gained by continuing screening beyond age 75.^13-15 On the other hand, both practices will substantially increase the number of colonoscopies performed, and hence, the number of colonoscopy-related complications experienced.¹⁶ Moreover, continuing screening beyond age 75 might substantially increase over-diagnosis and over-treatment of CRC (i.e. the detection and treatment of cancers that would not have been diagnosed without screening). As a result, more intensive screening than recommended might be associated with a balance among benefits, burden, and harms that is unfavorable for Medicare beneficiaries: it might negatively affect health.

If more intensive screening than recommended is favorable for Medicare beneficiaries, the subsequent question is whether it is efficient from a societal perspective (i.e. whether the net health benefit justifies the additional colonoscopies and financial resources required). This is important, as both colonoscopy capacity and financial resources are constrained.

The objective of this study was to determine whether more intensive colonoscopy screening than recommended is favorable for Medicare beneficiaries, and if so, whether it is efficient from a societal perspective. In a prior analysis,¹³ we already demonstrated that applying a screening interval of 5 instead of 10 years and continuing screening beyond age 75 result in a small increase in LYs gained compared with the increase in colonoscopies performed in those starting screening at age 50. In this study we extend this work by determining the net health benefit and cost-effectiveness of screening. Moreover, in our current analysis, we focus on the Medicare population. Analyses were performed using the microsimulation model MISCAN-Colon.

METHODS

MISCAN-Colon

MISCAN-Colon is a well-established microsimulation model for CRC developed at the Department of Public Health of the Erasmus University Medical Center (Rotterdam, the Netherlands). The model's structure, underlying assumptions, and calibration are described in eAppendix 1. In brief, MISCAN-Colon simulates the life histories of a large population of persons from birth to death. As each simulated person ages, one or more adenomas may develop. These adenomas can progress from small (≤5mm), to medium (6-9mm), to large size (≥10mm). Some adenomas can develop into preclinical cancer, which may progress through stages I to IV. During each stage CRC may be diagnosed because of symptoms. Survival after clinical diagnosis is determined by the stage at diagnosis, the localization of the cancer, and the person's age.¹⁷

Screening will alter some of the simulated life histories: Some cancers will be prevented by the detection and removal of adenomas; other cancers will be detected in an earlier stage with a more favorable survival. However, screening can also result in serious complications and over-diagnosis and over-treatment of CRC. By comparing all life histories with screening with the corresponding life histories without screening, MISCAN-Colon quantifies the effectiveness and efficiency of screening.

MISCAN-Colon was calibrated to the age-, stage-, and localization-specific incidence of CRC as observed before the introduction of screening and the age-specific prevalence and multiplicity distribution of adenomas as observed in autopsy studies.^18-28 The preclinical duration of CRC and the adenoma dwell-time were calibrated to the rates of interval and surveillance detected cancers observed in randomized controlled trials evaluating screening using guaiac fecal occult blood tests and a once-only sigmoidoscopy.^29-33

Model Inputs

Populations Simulated

We simulated two cohorts of 10 million 65-year-old Medicare beneficiaries. In the first cohort, all beneficiaries had received a negative screening colonoscopy at age 55 (i.e. were up-to-date with screening recommendations). In the second cohort, all beneficiaries were previously unscreened. For both cohorts, we assumed the average population risk for CRC and an average life-expectancy.³⁴

Screening Scenarios

In both cohorts we simulated ‘recommended screening’ (i.e. colonoscopy screening at ages 65 and 75) as well as two scenarios in which a shorter screening interval was applied: 1) screening at a 5 year interval (screening at ages 65, 70, and 75) and 2) screening at a 3 year interval (screening at ages 65, 68, 71, and 74). Furthermore, we simulated two scenarios of continued screening beyond age 75: 1) screening up to age 85 (screening at ages 65, 75, and 85) and 2) screening up to age 95 (screening at ages 65, 75, 85, and 95). Beneficiaries in whom adenomas were removed were assumed to undergo colonoscopy surveillance according to current guidelines.³⁵ We assumed that surveillance continued until the diagnosis of CRC or death and that adherence to surveillance was 100%.

Test Characteristics

The sensitivity of colonoscopy for the detection of adenomas and CRC was obtained from a systematic review on miss rates observed in tandem colonoscopy studies and was 75% for small adenomas (≤5mm), 85% for medium-sized adenomas (6-9mm), and 95% for large adenomas (≥10mm) and CRC.³⁶ We assumed that 95% of all colonoscopies reached the cecum; for the remaining 5% the reach of the procedure was assumed to be distributed uniformly over colon and rectum.

Complications of Colonoscopy

Age-specific risks for gastrointestinal and cardiovascular complications associated with colonoscopy were derived by performing additional statistical analyses on Medicare data used in a study by Warren and colleagues (eAppendix 2).¹⁶ According to these analyses colonoscopies with polypectomy were associated with an excess risk for complications, whereas colonoscopies without polypectomy were not. The risks associated with colonoscopies with polypectomy increased exponentially with age (Figure 1). Only complications requiring hospitalization or an emergency department visit were considered. We assumed that one out of every 30,000 colonoscopies involving a polypectomy resulted in death.^16,37

Figure 1.

Model Inputs: Age-Specific Risks for Complications Associated with Colonoscopies with Polypectomy.^a

^aDerived by performing additional statistical analyses on Medicare data used in a study by Warren et al.¹⁶ (eAppendix 2). Only complications requiring hospitalization or an emergency department visit were considered.

^bPerforations, gastrointestinal bleeding or transfusions; risk per colonoscopy = (((EXP(9.27953-0.06105*Age))+1)^-1)-((EXP(10.78719-0.06105*Age))+1)^-1.

^cParalytic ileus, nausea and vomiting, dehydration, abdominal pain; risk per colonoscopy = (((EXP(8.81404-0.05903*Age))+1)^-1)-((EXP(9.61197-0.05903*Age))+1)^-1.

^dMyocardial infarction or angina, arrhythmias, congestive heart failure, cardiac or respiratory arrest, syncope, hypotension, or shock; risk per colonoscopy = (((EXP(9.09053-0.07056*A69))+1)^-1)-((EXP(9.38297-0.07056*A69))+1)^-1.

Utility Losses Associated with Colonoscopy Screening

We assumed a utility loss (i.e. a loss of quality of life) equivalent to two days of life per colonoscopy (0.0055 QALYs) and two weeks of life per complication (0.0384 QALYs) (Table 1). We also assigned a utility loss to each LY with CRC care.³⁸ Assigning utility losses to LYs with CRC care works two ways: On the one hand, screening prevents cancers by the detection and removal of adenomas. This reduces LYs with CRC care, and hence, results in a gain of quality of life. On the other hand, screening results in earlier detection and over-diagnosis and over-treatment of cancers. This adds LYs with CRC care, and hence, results in a loss of quality of life. The resulting net impact on quality of life can be either positive or negative.

Table 1.

Model Inputs: Utility Losses and Costs Associated with Colonoscopy Screening.

UTILITY LOSS (QALYs)a
Per colonoscopy
without polypectomy/ biopsy	0.0055
with polypectomy/ biopsy	0.0055
Per complication of colonoscopy	0.038
Per LY with CRC careb, c	Initial care	Continuing care	Terminal care Death CRC	Terminal care Death other cause
Stage I CRC	0.12	0.05	0.70	0.05
Stage II CRC	0.18	0.05	0.70	0.05
Stage III CRC	0.24	0.24	0.70	0.24
Stage IV CRC	0.70	0.70	0.70	0.70
COSTS (2013 US$)d
Per colonoscopy
without polypectomy/ biopsy	887
with polypectomy/ biopsy	1,096
Per complication of colonoscopy	6,045
Per LY with CRC careb	Initial care	Continuing care	Terminal care Death CRC	Terminal care Death other cause
Stage I CRC	36,683	3,050	63,809	19,176
Stage II CRC	49,234	2,870	63,555	17,279
Stage III CRC	59,759	4,021	67,041	21,457
Stage IV CRC	77,790	12,178	88,368	49,866

QALY = quality-adjusted life-year; LY = life-year; CRC = colorectal cancer

^aThe loss of quality of life associated with a particular event.

^bCare for CRC was divided in three clinically relevant phases: the initial, continuing, and terminal care phase. The initial care phase was defined as the first 12 months after diagnosis; the terminal care phase was defined as the final 12 months of life; the continuing care phase was defined as all months in between. In the terminal care phase, we distinguished between CRC patients dying from CRC and CRC patients dying from another cause. For patients surviving less than 24 months, the final 12 months were allocated to the terminal care phase and the remaining months were allocated to the initial care phase. A patient diagnosed with CRC at age 65, who dies from CRC at age 70, will be in the initial care phase from age 65 up to age 66, in the continuing care phase from age 66 up to age 69, and in the terminal care phase associated with death from CRC from age 69 up to age 70. CRC patients who do not die from CRC will be in the continuing care phase from one year after diagnosis until one year before death from another cause.

^cUtility losses for LYs with initial care were derived from a study by Ness and colleagues.³⁸ For LYs with continuing care for stage I and II CRC, we assumed a utility loss of 0.05 QALYs; for LYs with continuing care for stage III and IV CRC, we assumed the corresponding utility losses for LYs with initial care. For LYs with terminal care for CRC, we assumed the utility loss for LYs with initial care for stage IV CRC. For LYs with terminal care for another cause, we assumed the corresponding utility losses for LYs with continuing care.

^dCosts include copayments and patient time costs (i.e. the opportunity costs of spending time on screening or being treated for a complication or CRC), but do not include travel costs, costs of lost productivity, and unrelated health care and non-health care costs in added years of life. We assumed that the value of patient time was equal to the median wage rate in 2012: $16.71 per hour.⁴¹ We assumed that colonoscopies and complications used up 8 and 16 hours of patient time, respectively. Patient time costs were already included in the estimates for the costs of LYs with CRC care obtained from a study by Yabroff and colleagues.

Costs Associated with Colonoscopy Screening

The cost-effectiveness analyses were conducted from a societal perspective. The costs of colonoscopies were based on 2007 Medicare payment rates and copayments (Table 1).³⁹ The costs of complications were obtained from a cost-analysis of cases of unexpected hospital use after endoscopy in 2007.⁴⁰ We added patient time costs to both.⁴¹ The costs of LYs with CRC care were obtained from an analysis of SEER-Medicare linked data and included patient deductibles, copayments, and patient time costs.⁴² We adjusted all costs to reflect the 2013 level using the U.S. consumer price index.⁴³

Assigning costs to LYs with CRC care also works two ways: On the one hand, screening prevents LYs with CRC care, reducing costs. On the other hand, screening adds LYs with CRC care, increasing costs. The net effect can be either a reduction or an increase in costs.

Outcomes and Analysis

For each scenario of more intensive screening than recommended, we determined the associated increase in CRC cases prevented, CRC deaths prevented, LYs gained, LYs with CRC care prevented, colonoscopies performed, and complications experienced. Subsequently, we determined the resulting increases in QALYs gained: the net health benefit of screening. If more intensive screening resulted in a gain of QALYs, it was considered favorable for Medicare beneficiaries; if it resulted in a loss of QALYs, it was considered unfavorable. In a second step, we determined the increase in costs associated with all scenarios simulated. For those scenarios that were favorable for Medicare beneficiaries, we related the increases in colonoscopies performed and costs to the increase in QALYs gained to determine the efficiency of screening.

We present both undiscounted and discounted results (applying the conventional 3% annual discount rate for both effects and resources required). We based our conclusions on the discounted results. Screening strategies associated with an incremental cost per QALY gained exceeding $100,000 (discounted) were considered inefficient.

Sensitivity Analyses

We repeated our analysis assuming 1) half and twice the base-case utility losses for colonoscopies and colonoscopy-related complications; 2) no utility loss for LYs with continuing care for CRC and a utility loss of 0.12, 0.18, 0.24, and 0.70 QALYs lost for each life-year with continuing care for stage I, II, III, and IV CRC, respectively; 3) twice the base-case costs for LYs with CRC care; 4) twice the base-case miss-rates for proximal adenomas and CRC;^44-47 and 5) twice the base-case miss rates for all adenomas and CRCs. Furthermore, to determine the extent to which our results were driven by more intensive screening, rather than more beneficiaries entering surveillance, we repeated our analysis assuming 0% adherence to surveillance. Finally, we repeated our analysis for beneficiaries at 25% higher and 25% lower risk for CRC and for beneficiaries without comorbidity and with severe comorbidity, using comorbidity status-specific life-tables.⁴⁸

Results

Previously Screened Beneficiaries

Benefits, Burden, and Harms

Screening Medicare beneficiaries with a negative screening colonoscopy at age 55 according to current guideline recommendations (i.e. colonoscopy screening at ages 65 and 75) resulted in 14.1 CRC cases prevented, 7.7 CRC deaths prevented, and 63.1 LYs gained per 1,000 beneficiaries compared with no screening (i.e. an average of 23.0 days per beneficiary) (Figures 2A-C). Moreover, recommended screening prevented 37.5 LYs with CRC care per 1,000 beneficiaries (Figure 2D). To achieve this effect, 2,131 colonoscopies had to be performed causing 8.3 complications (Figures 2E-F).

Figure 2.

The Increases in the Benefits (A-D) and the Burden and Harms (E-F) of Screening Associated with More Intensive Colonoscopy Screening than Recommended in Medicare Beneficiaries with a Negative Screening Colonoscopy at Age 55 (numbers per 1,000 beneficiaries; undiscounted).

CRC = colorectal cancer; LY = life-year

^aFor each scenario, the distribution of additional LYs with CRC care prevented over the different phases of care and stages of CRC is given in eAppendix 3.

^bOnly complications requiring hospitalization or an emergency department visit were considered.

Compared with 10-yearly screening, screening every 5 years resulted in 1.7 additional CRC cases prevented, 0.6 additional CRC deaths prevented, and 5.8 additional LYs gained per 1,000 beneficiaries (i.e. an average of 2.1 additional days of life per beneficiary). Moreover, 5-yearly screening prevented 10.9 additional LYs with CRC care per 1,000 beneficiaries. To achieve this relatively small added benefit, 783 additional colonoscopies had to be performed causing 1.3 additional complications. Continuing screening up to age 85 instead of 75 resulted in even fewer additional CRC cases prevented, CRC deaths prevented, and LYs gained: +0.2, +0.3, and +1.2 per 1,000 beneficiaries (i.e. an average of 0.4 additional days of life per beneficiary), respectively. To achieve this marginal additional benefit 369 additional colonoscopies had to be performed causing 2.4 additional complications. Furthermore, instead of preventing additional LYs with CRC care, screening up to age 85 instead of 75 increased the number of LYs with CRC care. Further intensifying screening by reducing the screening interval to 3 years or by continuing screening up to age 95 resulted in even smaller increases in the benefits of screening, also relative to the corresponding increases in burden and harms (Figures 2A-F).

Net Health Benefit

In previously screened beneficiaries, ‘recommended screening’ resulted in 63.1 LYs gained per 1,000 beneficiaries compared with no screening (Table 2). On top of that, screening resulted in 13.4 QALYs gained per 1,000 beneficiaries through preventing LYs with CRC care. However, to achieve these benefits, colonoscopies had to be performed, resulting in a loss of 11.7 QALYs per 1,000 beneficiaries. Furthermore, these colonoscopies caused complications, resulting in a loss of another 0.3 QALYs per 1,000 beneficiaries. Hence, recommended screening resulted in a net health benefit of 63.1 + 13.4 – 11.7 – 0.3 = 64.5 QALYs gained per 1,000 beneficiaries.

Table 2.

The Net Health Benefit of More Intensive Colonoscopy Screening than Recommended in Medicare Beneficiaries with a Negative Screening Colonoscopy at Age 55 (numbers per 1,000 beneficiaries).

Scenario		LYs gaineda	IMPACT ON QUALITY OF LIFE (QALYs)b			NET HEALTH BENEFITc
			LYs with CRC cared	Colonoscopies	Complications	QALYs gained
		(A)	(B)	(C)	(D)	(A+B+C+D)
UNDISCOUNTED
Recommended screeninge		63.1	13.4	-11.7	-0.3	64.5
Applying shorter screening intervals	5 yearsf, i	5.8	1.8	-4.3	-0.0	3.2
	3 yearsg	1.6	0.7	-4.3	-0.0	-2.1
Continuing screening beyond age 75	up to age 85f	1.2	0.1	-2.0	-0.1	-0.8
	up to age 95h	0.0	-0.0	-0.4	-0.0	-0.4
3% DISCOUNTED
Recommended screeninge		36.0	7.7	-9.9	-0.2	33.6
Applying shorter screening intervals	5 yearsf	3.4	1.1	-3.7	-0.0	0.7
	3 yearsg	1.0	0.5	-3.9	-0.0	-2.5
Continuing screening beyond age 75	up to age 85f	0.5	0.0	-1.1	-0.1	-0.6
	up to age 95h	0.0	-0.0	-0.2	-0.0	-0.2

LY = life-year; CRC = colorectal cancer; QALY = quality-adjusted life-year

^aThe impact of a screening scenario on quantity of life.

^bThe impact of a screening scenario on quality of life.

^cThe impact of a screening scenario on quantity and quality of life incorporated in one measure.

^dScreening results in a gain of quality of life by preventing LYs with CRC care and a loss of quality of life by adding LYs with CRC care. The net effect can be a gain of quality of life (positive values) or a loss of quality of life (negative values).

^eCompared with no screening.

^fCompared with recommended screening.

^gCompared with applying shorter screening intervals - 5 years.

^hCompared with continuing screening beyond age 75 - up to age 85.

ⁱDetailed results for this scenario are given in eAppendix 4.

When applying a screening interval of 5 instead of 10 years, the gain in quality of life by preventing additional LYs with CRC care was exceeded by the loss of quality of life due to additional colonoscopies and additional complications. As a result, applying a screening interval of 5 instead of 10 years resulted in fewer QALYs than LYs gained: 3.2 versus 5.8 per 1,000 beneficiaries. When screening was continued up to age 85 instead of 75, the overall loss of quality of life exceeded the associated increase in LYs gained. Hence, continuing screening up to age 85 instead of 75 resulted in a loss of QALYs, rather than a gain. Both applying a screening interval of 3 instead of 5 years and continuing screening up to age 95 instead of 85 also negatively affected the number of QALYs gained by screening. Discounting did not change the direction of the effect on QALYs gained for any of the scenarios simulated.

Efficiency

Screening previously screened beneficiaries every 5 instead of 10 years was the only scenario of more intensive screening that resulted in QALYs gained: 0.7 per 1,000 beneficiaries (discounted result) (Table 3). To gain these QALYs, 675 additional colonoscopies and an additional $528,000 were required; hence, 909 additional colonoscopies and an additional $711,000 were required per additional QALY gained.

Table 3.

The Efficiency of More Intensive Colonoscopy Screening than Recommended in Medicare Beneficiaries with a Negative Screening Colonoscopy at Age 55 (net health benefit and resources required per 1,000 beneficiaries).

Scenario		NET HEALTH BENEFIT	RESOURCES REQUIRED		EFFICIENCY
		QALYs gained	Colonoscopies	US$ (*1,000)a	ΔColonoscopies/ ΔQALYs gained	ΔUS$/ ΔQALYs gained (*1,000)
UNDISCOUNTED
Recommended screeningb		64.5	2,131	922	33	14
Applying shorter screening intervals	5 yearsc, f	3.2	783	573	245	179
	3 yearsd	-2.1	781	656	unfavorable	unfavorable
Continuing screening beyond age 75	up to age 85c	-0.8	369	349	unfavorable	unfavorable
	up to age 95e	-0.4	69	68	unfavorable	unfavorable
3% DISCOUNTED
Recommended screeningb		33.6	1,815	1,091	54	32
Applying shorter screening intervals	5 yearsc	0.7	675	528	909	711
	3 yearsd	-2.5	711	610	unfavorable	unfavorable
Continuing screening beyond age 75	up to age 85c	-0.6	203	196	unfavorable	unfavorable
	up to age 95e	-0.2	28	28	unfavorable	unfavorable

QALY = quality-adjusted life-year

^aNet costs: the costs of additional screening and surveillance colonoscopies, complications and LYs with CRC care minus the savings by preventing additional LYs with CRC care.

^bCompared with no screening.

^cCompared with recommended screening.

^dCompared with applying shorter screening intervals - 5 years.

^eCompared with continuing screening beyond age 75 - up to age 85.

^fDetailed results for this scenario are given in eAppendix 4.

Previously Unscreened Beneficiaries

In previously unscreened beneficiaries, more intensive screening than recommended was slightly less unfavorable/ inefficient; however, screening every 3 instead of 5 years and continuing screening beyond age 75 were still associated with a loss of QALYs, rather than a gain and screening every 5 instead of 10 years was still inefficient requiring 416 additional colonoscopies and an additional $317,000 per additional QALY gained (discounted results) (eAppendix 5).

Sensitivity Analyses

When the base-case utility losses for colonoscopies and complications were doubled, or when a lower risk for CRC or a worse-than-average life-expectancy was assumed, screening previously screened beneficiaries every 5 instead of 10 years resulted in a loss of QALYs, rather than a gain (Table 4). None of the other sensitivity analyses changed the direction of the effect on QALYs gained for any of the scenarios simulated. Assuming a 25% higher risk for CRC resulted in the least unfavorable efficiency ratios: screening previously screened beneficiaries every 5 instead of 10 years required 249 additional colonoscopies and an additional $181,000 per additional QALY gained (discounted results) (Table 4). Again, results in beneficiaries with prior screening were slightly less unfavorable (eAppendix 5): In previously unscreened beneficiaries at 25% increased risk for CRC, screening every 5 instead of 10 years was associated with 174 additional colonoscopies and an additional $121,000 per additional QALY gained.

Table 4.

The Efficiency of More Intensive Colonoscopy Screening than Recommended in Medicare Beneficiaries with a Negative Screening Colonoscopy at Age 55: Results of Sensitivity Analyses (3% discounted).

Scenario		Outcome	ANALYSIS
			Base case	Utility losses colonoscopies and complications*0.5	Utility losses colonoscopies and complications*2	Utility loss LYs with continuing care = 0	Utility losses LYs with continuing care for stage I and II CRC = 0.12 and 0.18	Costs LYs with CRC care*2	Miss rates colonoscopy - proximal*2a	Miss rates colonoscopy - entire colon and rectum*2a	Adherence to surveillance = 0%	CRCrisk*1.25	CRCrisk*0.75	Life-expectancy individuals with no comorbidityb	Life-expectancy individuals with severe comorbidityb
Applying shorter screening intervals	5 yearsc	ΔColonoscopies/ΔQALYs gained	909	258	unfavorablef	11,787	587	909	536	267	1,488	249	unfavorablef	655	unfavorablef
		ΔUS$/ ΔQALYs gained (*1,000)	711	202		9,224	459	582	413	200	1,167	181		505
	3 yearsd	ΔColonoscopies/ΔQALYs gained	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable
		ΔUS$/ ΔQALYs gained (*1,000)
Continuing screening beyond age 75	up to age 85c	ΔColonoscopies/ΔQALYs gained	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable
		ΔUS$/ ΔQALYs gained (*1,000)
	up to age 95e	ΔColonoscopies/ΔQALYs gained	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable	unfavorable
		ΔUS$/ ΔQALYs gained (*1,000)

QALY = quality-adjusted life-year

^aAssuming twice the base case miss-rates for colonoscopy implies a sensitivity of 50% for small adenomas (≤5mm), 70% for medium-sized adenomas (6-9mm), and 90% for large adenomas (≥10mm) and CRC.

^bIndividuals are classified as having no comorbidity if none of the following conditions is present: an ulcer, a history of acute myocardial infarction, rheumatologic disease, peripheral vascular disease, diabetes, paralysis, cerebrovascular disease, constructive obstructive pulmonary disease, congestive heart failure, moderate or severe liver disease, chronic renal failure, dementia, cirrhosis and chronic hepatitis, or AIDS. Individuals are classified as having severe comorbidity if diagnosed with constructive obstructive pulmonary disease, congestive heart failure, moderate or severe liver disease, chronic renal failure, dementia, cirrhosis and chronic hepatitis, or AIDS.

^cCompared with recommended screening.

^dCompared with applying shorter screening intervals - 5 years.

^eCompared with continuing screening beyond age 75 - up to age 85.

^fThe sensitivity analysis changed the direction of the effect on QALYs gained.

Discussion

Recent studies show that many Medicare beneficiaries undergo more intensive colonoscopy screening than recommended.^7,8 Our study shows that the resulting balance among benefits, burden, and harms is often unfavorable. Screening previously screened Medicare beneficiaries up to age 85 instead of 75, for example, resulted in only 1.2 additional LYs gained per 1,000 beneficiaries, whereas it required 369 additional colonoscopies causing 2.4 additional complications (undiscounted results). As a result, this practice was associated with a loss of QALYs, rather than a gain (i.e. a net harm). The only scenario favorable for beneficiaries: screening every 5 instead of 10 years, required 909 additional colonoscopies and an additional $711,000 per additional QALY gained (discounted results). This well exceeds the conventional thresholds for the willingness-to-pay per QALY gained of $50,000 and $100,000; and although some researchers regard these thresholds as being too low,⁴⁹ even these researchers suggest a threshold well below $700,000. Results in previously unscreened beneficiaries were slightly less unfavorable. However, screening every 3 instead of 5 years and continuing screening beyond age 75 were still associated with a loss of QALYs, rather than a gain; and screening every 5 instead of 10 years still required 416 additional colonoscopies and an additional $317,000 per additional QALY gained.

The small increase in LYs gained by applying a shorter screening interval than recommended is explained by a combination of two factors: 1) the high sensitivity of colonoscopy for the detection of adenomas and CRC³⁶ and 2) the low progression rate of adenomas into CRC.²⁹ As a result of the former, adenomas and CRC are unlikely to be prevalent in individuals who just underwent a negative screening colonoscopy. As a result of the latter, adenomas that remain undetected during the first colonoscopy at age 65 or newly developed adenomas after this colonoscopy are unlikely to develop into CRC before age 75, at which the next recommended screening colonoscopy would take place. Hence, an additional screening colonoscopy at age 70 is unlikely to add much benefit. The small increase in LYs gained by continuing screening beyond age 75 is explained by the same two factors and by the high risk for other cause mortality at advanced age, which reduces both the probability that screening will prevent CRC mortality and the number of LYs gained if CRC mortality is prevented. Moreover, the risks for colonoscopy-related complications and over-diagnosis and over-treatment of CRC increase with age, negatively affecting the net health benefit of screening.

In the analysis underlying the USPSTF recommendation statement on CRC screening,¹³ we already demonstrated that both applying shorter screening intervals than 10 years and continuing screening beyond age 75 result in a small increase in LYs gained and a large increase in colonoscopies performed in those starting screening at age 50. However, in that analysis we did not quantify the possible harms of screening (i.e. non-lethal colonoscopy-related complications and over-diagnosis and over-treatment of CRC), nor did we explicitly weigh the benefits against the burden and harms. This is one of the strengths of our current study, as it allows us to draw conclusions about the net health benefit of screening. Whereas clinicians and patients might be reluctant to apply a relatively long screening interval or to discontinue screening after age 75 based on a certain balance between colonoscopies and LYs gained, they are more likely to respond to evidence demonstrating that more intensive screening than recommended negatively affects health. Furthermore, in our earlier study we did not consider costs, which is necessary to evaluate the appropriateness of screening from a societal perspective.

Our study has three main limitations. First of all, although CRC screening is recommended from age 50 onwards, we focused our analysis on the Medicare population, i.e. we addressed the appropriateness of more intensive screening than recommended from age 65 onwards. We chose to do so, as patterns of more intensive screening than recommended have mainly been documented in the Medicare population. Nevertheless, an additional analysis shows that more intensive screening than recommended is also inefficient when started at age 50 (eAppendix 6). Secondly, as we aimed to illustrate the impact of more intensive screening than recommended, we assumed 100% attendance to all screening scenarios. However, in reality, a beneficiary with negative screening colonoscopies at ages 65 and 68 might be unlikely to receive another screening colonoscopy at age 71. If a lower attendance rate would be assumed, the scenarios of more intensive screening than recommended would be more similar to recommended screening. Thirdly, although we did perform a sensitivity analysis on the sensitivity of colonoscopy for adenomas and CRC, we did not perform an analysis assuming low quality colonoscopies. If a proper colonoscopy cannot be performed due to a bad bowel preparation, for example, an early repeat screening colonoscopy is of course justified.

Our analysis highlights some critical future research directions. First of all, it shows that continuing screening up to very advanced age can be inefficient/ harmful. This is also likely to be true for surveillance in adenoma patients; particularly in those at relatively low risk for CRC. Investigating the appropriateness of surveillance at advanced age is particularly important as a substantial proportion of those being screened eventually enter surveillance. Furthermore, our sensitivity analyses demonstrate that the effectiveness and cost-effectiveness of screening depend on an individual's life expectancy and, more importantly, risk for CRC, which reinforces the need for research on personalizing CRC screening recommendations. Finally, our study demonstrates the importance of considering effects on quality of life when evaluating screening. However, data regarding the utility losses associated with CRC screening are sparse or even absent. More research is needed in this area.

In conclusion, screening Medicare beneficiaries intensively than recommended is not only inefficient from a societal perspective; often it is also unfavorable for those being screened. This study provides strong evidence and a clear rationale for clinicians and policy makers to actively discourage this practice.