Douglas Corley MD, PhD Kaiser Permanente, Northern California · around next expected screen/surveillance exam . Developed with KP an ADR report: a mechanism for providing feedback

Monitoring post-colonoscopy CRC: Outcomes and

lessons learned from the US experience (AKA Right-Sided Lesion and Interval Cancer: Definitions, Measurement & Challenges)

Douglas Corley MD, PhD

Kaiser Permanente, Northern California

Possible conflicts of interest

• None

Name of presenter

Definitions, Measurement & Challenges

• What are key-steps and challenges in performance

– Process/methodology

– How to calculate rates?

– Practices at Kaiser Permanente

• Assigning potential cancer etiology

– Challenges?

– How to resolve?

A key step is defining the question

Translating definitions to measurements and unbiased comparisons requires additional knowledge

• “CRC diagnosed after a CRC screening examination or test in which no cancer is detected and before date of next recommended exam”

– How to classify the first 6-12 months, when highest likelihood of ongoing work-up of prevalent cancer and/or ongoing resection?

Comparisons may be influenced by length of follow-up

• “Screen-detected and non-screen-detected CRC rates should be reported as #/100,000 PY”

The % Interval Cancers varies with definitions & f/up

8%

3% 3%

7%

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%

KP (0.5-10 yrs) Denmark (0-5yrs)

Netherl (<5 yrs) Medicare (0.5-3yrs)

% of CRCs which are Interval Cancers

% of CRCs

The locations & stages may differ with definitions & total follow-up

64%

48%

56%

65% 62%

51% 56%

78%

8% 3% 3%

7%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

KP (0.5-10yrs)

Denmark (0-5yrs)

Netherl (<5yrs)

Medicare(0.5-3 yrs)

% Stage 1/2 v 3/4

% Prox Colon

% of CRCs

CRC incidence comparisons may differ due to variable follow-up independent of P-Yrs

0

10

20

30

40

50

60

70

80

90

Yrs 1-3 Yrs 3.1-5 Yrs. 5.1-7 Yrs. 7.1-9

New CRCs

Missed CRCs

Total CRCs

It is unclear how to classify some exams

• But what if detected one year later, at time of recommended f/up

exam? • What are acceptable windows for f/up exams (e.g. within 6

months, one year of exact time?

Do we think about FIT & Colo differently?

• FIT – Polyp detection + – CRC detection +++ – : early detection>cancer prevention – Possible cancer at next screen/surveillance EXPECTED

• Colonoscopy – Polyp detection +++ – CRC detection ++++ – : early detection ≈ cancer prevention – Possible cancer at next 10 year screen: EXPECTED – Possible cancer at 3 yr surveillance: NOT EXPECTED

CRC causal distribution will change with duration of follow-up

0

10

20

30

40

50

60

70

80

90

Yrs 1-3 Yrs 3.1-5 Yrs. 5.1-7 Yrs. 7.1-9

New CRCs

Missed CRCs

Total CRCs

Missed CRCS: Incomplete exams

(prep & extent)

Incomplete resection/Missed adenomas/new polyps & CAs

We are now looking at detailed characteristics of interval cancers in KPNC

• Study Design: Retrospective Cohort Study • Population: Kaiser Permanente Northern California (KPNC),

integrated healthcare system with a CRC screening program. • Patients: Interval cancer between the years 1998 and 2013

– CRC 1 to 10 years following any colonoscopy negative for CRC (the index exam).

• Exclusions: history of CRC, inflammatory bowel disease, Lynch Syndrome, or familial adenomatous polyposis.

• Data Sources: • Colonoscopies from electronic records • Chart abstractions on all colonoscopies in 10 years before CRC • Cancer stage and diagnosis from KPNC cancer registry.

Factors associated with interval cancer

0

0,5

1

1,5

2

2,5

3

3,5

4

4,5

Poor prep Any prior polyp Polyp index exam Didn’t reach cecum

Series 1

Series 1

OR 0.9 95% CI 0.7-1.23

OR 2.3 95% CI 1.9-2.8

OR 2.4 95% CI 1.9-2.9

OR 4.3 95% CI 2.6-7.1

Name of presenter

Metrics for Interval CRC • Uncommon outcome

– Individual cancer metrics and reporting imprecise • KPNC report ADR individually through clinical operations, but not interval CAs

– May require adjustment for length of f/up or use only cohorts with fixed periods of f/up

• May need different concepts/definitions for colonoscopy vs. FIT based exams • Or change expectations about surveillance expectations

• ? How to define cancers in first year and in “windows” around next expected screen/surveillance exam

Developed with KP an ADR report: a mechanism for providing feedback and measuring change after interventions

Developed with KP and pilot-tested an interactive ADR educational tool using best practices for educational interventions.

It takes a village Thank You!

Analytic Strategies for Prevention of Interval CRC • Knowledge of cancer incidence rates over time

– Different risk groups • Adenoma status, number, histology • Interaction between family history and adenoma status

• Detailed analyses of possible causes over time • Investigations/interventions on those causes

– High quality prep (split prep) – Complete exams – High quality inspection – Complete polyp resection/follow-up – Appropriate follow-up intervals

Interval CRCs were more often in proximal colon