STUDY: The use of screening
colonoscopy for patients cared for by the Department of Veterans
Affairs | Authors, ref ID:
El-Serag, Peterson, Hampel, Richardson, Cooper 100
Year of
publication: 2006
Dates of data
collection: 1998 – 2003
Trial
name: NA |
|
OBJECTIVE OR AIM:
| To investigate whether
colonoscopy use increased disproportionately in the VA system and
changes in rates of FS, DCBE, and FOBT use between October 1, 1998, and
September 30, 2003. |
|
DESIGN:
| Setting: United
States, VA medical centers
Study design:
cross-sectional
Duration (mean followup):
NA
Overall study size (N enrolled/N
analyzed): |
|
Sample size:
| NA |
Describe
intervention: NA |
|
RECRUITMENT: (population-based, clinic-based, volunteer,
other)
| Population based- from VA
databases |
|
INCLUSION CRITERIA:
| Procedures for VA users aged 49
to 75 years. |
|
EXCLUSION CRITERIA:
| NR |
|
POPULATION CHARACTERISTICS:
| Feature | Screening c-scope
(n=178,853) | FOBT
(n=1,635,364) | DCBE (n=78,830) | FS (n=217,327) |
Mean age & range (years):
|
Colonoscopy: 62.3 (7.6) years | Race |
FOBT: 63.9 (7.8) | White | 105,746 (59.1) | 815,582 (49.9) | 45,894 (58.2) | 118,789 (54.7) |
DCBE: 62.6 (7.8) | Black | 17,934 (10.0) | 133,822 (8.2) | 10,458 (13.3) | 20,358 (9.4) |
FS: 61/8 (7.7) | Other | 55,173 (30.8) | 685,960 (41.9) | 22,478 (28.5) | 78,180 (36.0) |
Sex (% female):
| Sex |
| Male | 174,356 (97.5( | 1,583,775 (96.8) | 76,427 (97.0) | 211,824 (97.5) |
| Female | 4,497 (2.5) | 51,589 (3.2) | 2,403 (3.0) | 5,503 (2.5) |
Race:
| |
Other:
| The mean (SD) age of those
undergoing screening colonoscopy was 62.3 (7.6) years; FOBT, 63.9 (7.8)
years; DCBE, 62.6 (7.8) years; and FS, 61.8 (7.7) years. |
|
Attrition/Drop-out (not available for endpoint
measurement):
| NR |
Adherence:
| |
Contamination:
| |
Response Rates (e.g. for surveys):
| |
|
STATISTICAL ANALYSES:
| The frequencies and proportions
of the 4 types of CRC screening procedures, and of unique individuals
undergoing these procedures, were calculated for each fiscal year. The
temporal changes and potential determinants (age, sex, and race) of
screening colonoscopy (vs other CRC screening tests) were examined in
unadjusted and adjusted logistic regression analyses.
Similar
calculations were performed for screening colonoscopy (vs other
colonoscopy). Statistical comparisons of these proportions were not
performed because of overlapping groups.
The authors used
the predictive values to perform a sensitivity analysis of the
calculated proportions for screening colonoscopy. |
|
ASSESSMENT OF EXPOSURES AND POTENTIAL CONFOUNDERS:
| National inpatient and outpatient
VA databases used to assess potential confounders/covariates |
|
OUTCOME ASSESSMENT:
| National inpatient and outpatient
VA databases were searched for codes indicative of colonoscopy, FS,
FOBT, and DCBE recorded during fiscal years 1998 to 2003. The authors
also used the VA Patient Treatment File, which contains hospital
discharge records and up to 10 diagnostic codes, 5 operating room
procedures, and 32 nonoperating room procedures coded according to the
International Classification of Diseases, Ninth Revision,
Clinical Modification.
The indications for
CRC screening tests were classified as screening, diagnosis, or
surveillance based on the predefined Algorithm. All FOBT procedures were
designated as screening procedures. Flexible sigmoidoscopy and DCBE were
considered diagnostic in the presence of specific conditions recorded
within the year before the date of the procedure. They were considered
to be surveillance procedures in the presence of a second set of
prespecified conditions (coded as 17–28). The remaining
procedures were considered to be screening procedures Colonoscopy was
considered for CRC screening in the absence of conditions associated
with diagnostic or surveillance indications and if no colonoscopy had
been performed within the past 4 years.
Because of
concerns of the accuracy of diagnosis and procedure codes for specifying
procedural indications, the authors also conducted a medical record
review study in a subset of colonoscopic procedures nested within the
main study cohort to validate and refine the algorithm that was used. A
review of procedure, pathology, and progress notes was performed by 2
boardcertified gastroenterologist investigators who were blinded to the
designated status based on the VA administrative data sets. They
categorized indications for procedures as screening, surveillance, or
diagnostic.
A total of 303 medical records of unique patients
with colonoscopy performed at the Michael E. DeBakey VA Medical Center,
between October 6, 1999, and September 30, 2003, were identified at
random using a computer generated algorithm, and reviewed from the
national databases (ie, a subset from the main study cohort). Agreement
between the 2 reviewers was achieved in 92.0% of cases, and differences
were resolved by discussion. The predictive values of the database
algorithm for identifying screening colonoscopy (compared with the
medical record as a gold standard) were calculated. |
|
RESULTS:
|
|
KQ2 - What factors influence the use of colorectal cancer
screening?
| NA |
|
KQ3 - Which strategies are effective in increasing the appropriate
use of colorectal cancer screening and followup?
| NA |
|
KQ4 - What are the current and projected capacities to deliver
colorectal cancer screening and surveillance at the population
level?
| NA |
|
KQ5 - What are the effective approaches for monitoring the use and
quality of colorectal cancer screening?
| The authors calculated that the
algorithm has approximately 70.1% sensitivity and 71.6% specificity to
define screening colonoscopy.
The findings of medical record
review were then applied in a sensitivity analysis to recalculate the
estimated annual frequency of screening colonoscopy. Apart from reducing
the total number of screening colonoscopies by up to 25%, changing the
definition of screening colonoscopy had little effect on the observed
trends.
Concordance or kappa not calculated |
|
QUALITY RATING:
| Fair |
|
Quality Assessment-Internal Validity for Observational
Studies
|
|
|
Yes
|
No
|
Other (CD, NR, NA)
|
|
Were the groups similar at
baseline regarding the most important prognostic
indicators? | | | NA |
|
Were the drop-out or response
rates acceptable (≤ 20%)? [If
between 20% and 60%, check other and explain.] | X | | |
|
Were the differential drop-out or
response rates acceptable (≤ 15%)? | X | | |
|
Were intervention/exposure
measures valid, reliable, and equally applied? | X | | |
|
Were the outcome assessors blinded
to the intervention or exposure status of subjects? | | | NA |
|
Were outcome measures valid,
reliable, and equally applied? | x | | Algorithm used to differentiate
screening from diagnostic procedures, verified through dual review (see
statistical methods section of abstraction)
Concordance/kappa not
calculated |
|
Does the analysis control for
baseline differences? | x | | Logistic regression models
used |
|
Were important potential
confounding and modifying variables taken into account in the design and
analysis (e.g., through matching, stratification, or statistical
adjustment)? | x | | |
|
Were the statistical methods used
to assess the abstracted outcomes appropriate? | x | | |
|
Quality Rating (Good, Fair,
or Poor): Fair |
|
STUDY:
| Authors, ref ID:
Fiscella101
Year of publication:
2006
Dates of data collection:
1998–2002
Trial name: NR |
|
OBJECTIVE OR AIM:
| To determine whether estimates of
racial and racial disparities in receipt of six different types of
largely preventive procedures differ between self-report and Medicare
claims data |
|
DESIGN:
| Setting: Medicare
claims data
Study design:
observational
Duration (mean followup): 2
to 4 years
Overall study size (N enrolled/N
analyzed): 88509 # observations (n =
1474 for colorectal testing) |
|
Sample size:
| 1474 |
Describe intervention:
| Includes fecal occult blood
testing, sigmoidoscopy, or colonoscopy |
|
RECRUITMENT: (population-based, clinic-based, volunteer,
other)
| Population |
|
INCLUSION CRITERIA:
| Medicare Beneficiaries 65 and
older who participated in the Medicare Current Beneficiary Survey,
1999–2002. |
|
EXCLUSION CRITERIA:
| Participated in facility
interviews (i.e., resided in long-term care facilities), were less than
65 years of age (i.e., were Medicare recipients due to having a
qualifying disability), reported race/ethnicity other
than Hispanic, non- Hispanic African American, or
non-Hispanic White, i.e. majority, were enrolled in a Medicare HMO, or
were not eligible for Medicare B (or Medicare A and B) coverage dropped
due to incomplete claims. |
|
POPULATION CHARACTERISTICS:
|
Overall
|
Mean age & range (years):
| 65–69
16.7%
70–74 35.1%
75–79
26.7%
80–84 14.3%
85+ 7.2% |
Sex (% female):
| 56% female |
Race:
| Hispanic, African American
8.8%
Majority 91.2% |
Other:
| |
|
Attrition/Drop-out (not available for endpoint
measurement):
| NA |
Adherence:
| |
Contamination:
| |
Response Rates (e.g. for surveys):
| |
|
STATISTICAL ANALYSES:
| Describe: used SAS Survey and
logistic regression to assess the relationship, adjusted, between
minority status and self report |
|
ASSESSMENT OF EXPOSURES AND POTENTIAL CONFOUNDERS:
| Predictors of self report |
|
OUTCOME ASSESSMENT:
| Outcome
Measures:
• Prevalence of receipt
of FOBT, FS, or colonoscopy as measured
by:
• Self-report in the MCBS of having
any of the tests in the last year (MCBS) (indication was not
specified)
• Medicare claims, including
both screening and diagnostic codes (administrative data) |
|
RESULTS:
|
|
KQ2 - What factors influence the use of colorectal cancer
screening?
| NA |
|
KQ3 - Which strategies are effective in increasing the appropriate
use of colorectal cancer screening and followup?
| NA |
|
KQ4 - What are the current and projected capacities to deliver
colorectal cancer screening and surveillance at the population
level?
| NA |
|
KQ5 - What are the effective approaches for monitoring the use and
quality of colorectal cancer screening?
| Outcomes:
Unadjusted prevalence of CRC
screening
White:
Survey 38.0
Administrative
30.1
Minority:
Survey 34.8
Administrative
20.4
Concordance between self-report and administrative
data (measured by kappa score) for CRC screening
White
0.37
Minority 0.19
The authors also calculated an
odds ratios for reporting a procedure in the absence of a claim, or vice
versa. Minorities were more likely to report receipt of CRC screening in
the absence of a claim (OR=1.92, 95% CI, 1.32‐
2.79), with little change after adjustment for age, gender, income,
educational level, health status, proxy response, and supplemental
insurance. Having a claim for CRC testing in the absence of self-report
did not differ by race or ethnicity. |
|
QUALITY RATING:
| Good |
|
Quality Assessment-Internal Validity for Observational
Studies
|
|
|
Yes
|
No
|
Other (CD, NR, NA)
|
|
Were the groups similar at
baseline regarding the most important prognostic
indicators? | | | NA |
|
Were the drop-out or response
rates acceptable (≤ 20%)? [If
between 20% and 60%, check other and explain.] | | | Response rate for MCBS not
reported.
Possible that non-respondents differ from respondents
in terms of associations between self report and administrative
claims. |
|
Were the differential drop-out or
response rates acceptable (≤ 15%)? | | | |
|
Were intervention/exposure
measures valid, reliable, and equally applied? | | | NA |
|
Were the outcome assessors blinded
to the intervention or exposure status of subjects? | | | NA |
|
Were outcome measures valid,
reliable, and equally applied? | x | | |
|
Does the analysis control for
baseline differences? | x | | |
|
Were important potential
confounding and modifying variables taken into account in the design and
analysis (e.g., through matching, stratification, or statistical
adjustment)? | x | | |
|
Were the statistical methods used
to assess the abstracted outcomes appropriate? | x | | |
|
Quality Rating (Good, Fair, or
Poor): Good |
|
STUDY:
| Authors, ref ID:
Hall et al.102
Year of publication:
2004
Dates of data collection: December 1999
– June 2001
Trial name: NA |
|
OBJECTIVE OR AIM:
| To examine the accuracy of
self-reports of prostate and colorectal cancer testing and reasons for
testing using questions included in national surveys, the authors
conducted a survey and reviewed the medical records for procedures
performed among members of three health maintenance organizations
(HMOs) |
|
DESIGN:
| Setting: member
lists of Kaiser Permanente-Northern California (NC), Kaiser Permanente-
Georgia (GA), and HealthPartners (HP), Minnesota
Study
design: cross-sectional
Duration (mean
follow-up): NA
Overall study size (N enrolled/N
analyzed): 3546 |
|
| Black men | White and other men | Women |
Sample size:
| 363 | 847 | 920 |
Describe intervention
| | | |
|
RECRUITMENT: (population-based, clinic-based, volunteer,
other)
| H.M.O. population based |
|
INCLUSION CRITERIA:
| Men aged 45 years and older and
women aged 55 years and older as of September 1, 1999 who had been
enrolled in the plan for at least 5 years were eligible for
selection.
Simple random sample for women. For men, an SRS for
men at one site. At 2 sites, separate sample of black and white men
recruited until all strata were filled. Also at one site, sample was
drawn from members who had participated in a past survey. |
|
EXCLUSION CRITERIA:
| Screened out (statrum filled);
language barrier, dead, out of network, phone disconnected, other |
|
POPULATION CHARACTERISTICS:
| Characteristics of study sample
from 3 HMO’s |
Mean age & range (years):
Sex (% female):
| | Black men (n=363) | White and other men (n
= 847) | Women (n = 920) |
Age (y) | % | (n) | % | (n) | % | (n) |
45 – 49 | 19.8 | (72) | 14.3 | (121) | 0.0 | 0 |
50 – 59 | 41.0 | 149 | 38.6 | 327 | 23.3 | (214) |
60 – 69 | 19.3 | (70) | 23.4 | (198) | 33.8 | (311) |
70+ | 19.8 | (72) | 23.7 | (201) | 42.9 | (394) |
Race:
| Ethnicity |
Hispanic | 0.6 | (2) | 3.5 | (30) | 3.2 | (29) |
Non-Hispanic | 99.4 | (361) | 96.5 | (816) | 96.8 | (888) |
Other:
| Education |
< HS | 10.2 | (37) | 6.3 | (53) | 9.9 | (91) |
HS/GED | 22.6 | (82) | 17.9 | (152) | 27.1 | (249) |
Some college/tech school | 38.8 | (141) | 31.6 | (268) | 35.7 | (328) |
College grad | 28.4 | (103) | 44.2 | (374) | 27.2 | (250) |
Marital status |
Married | 78.0 | (283) | 82.4 | (698) | 55.8 | (512) |
Unmarried | 22.0 | (80) | 17.6 | (149) | 44.2 | (406) |
Employment |
Employed | 62.9 | (227) | 59.2 | (499) | 31.5 | (288) |
Unemployed | 33.8 | (122) | 39.3 | (331) | 62.0 | (566) |
Retired/other | 3.3 | (12) | 1.5 | (13) | 6.5 | (59) |
Income |
<=
$20,000 | 13.6 | (46) | 9.4 | (72) | 29.5 | (236) |
$20,000
– $40,000 | 24.2 | (82) | 16.9 | (130) | 30.0 | (240) |
$40,001
– $60,000 | 23.9 | (81) | 25.8 | (199) | 21.4 | (171) |
>
$60,000 | 38.3 | (130) | 47.9 | (369) | 19.0 | (152) |
Health status |
Excellent/good | 82.3 | (298) | 87.6 | (741) | 82.1 | (754) |
Fair/poor | 17.7 | (64) | 12.4 | (105) | 17.9 | (164) |
Smoking status |
Current | 17.9 | (65) | 10.0 | (85) | 7.5 | (69) |
Former | 46.6 | (169) | 49.9 | (423) | 36.2 | (333) |
Never | 35.5 | (129) | 40.0 | (339) | 56.3 | (518) |
|
Attrition/Drop-out (not available for Endpoint
Measurement):
| Response Rates (for
survey):
4,833 members contacted
3,546
eligible
1,248 refused
2,298
completed |
Adherence:
| Response rate = 64.8%
among those contacted and eligible |
Contamination:
| 1181 not contacted (no answer,
busy, answering service, exceeded call limit) |
|
STATISTICAL ANALYSES:
| The authors analyzed the data
according to our sampling frame, that is, for all women, black men, and
white and other men separately. Few participants were of races other
than white or black (67 [3.1%] were of other
races and 83 [3.9%] were multiracial).
Therefore, non- Black men of other races and multiracial men were
grouped together with white men. At HP, few men were black, and no
statistics are presented for them. Participants with a history of
prostate or colorectal cancer were excluded from the analyses. The final
sample sizes were 363 black men, 847 white and other men, and 920
women.
Respondent characteristics were summarized as frequencies
and percentages. The authors calculated the percentages of respondents
who reported having had their most recent DRE, PSA, or FOBT within 2
years before interview and sigmoidoscopy or colonoscopy within 5 years
before interview. The percentages of respondents who had tests as
determined by the medical record audit were similarly calculated; for
FOBT, the authors counted only home kits to match with the survey
question. On the basis of these time frames, the authors calculated the
concordance (agreement) between the self-report and medical record audit
for the tests. The authors also calculated agreement for endoscopy by
combining sigmoidoscopy and colonoscopy.
The authors calculated
the kappa statistic by the method of Landis and Koch, which accounts for
agreement expected by chance. Excellent agreement is defined as a kappa
statistic greater than 0.75, fair to good agreement as 0.40 to 0.75, and
poor agreement as less than 0.40. The authors also calculated the
sensitivity and specificity of self-reports using the information from
the medical records as the standard. Differences in sensitivity and
specificity between health plans were determined with a pairwise test of
proportions. The authors assessed the relationships between agreement
and demographic characteristics (age, sex-race groups, and education)
with polytomous regression models. The reference was agreement between
self-reports and medical records and was compared to over- and
underreporting. The authors also assessed the relationship with income,
but because of the relatively large number of missing observations and
no significant associations in any of the models, it was not included in
the final models. All analyses were adjusted for study
site.
Further, the authors determined the reasons for testing and
the concordance between the reasons reported by participants and the
reasons recorded in the medical records. Tests were defined as screening
tests when the reason for testing was screening, family history, or
“part of exam/ doctor just did it.” Tests were
defined as being performed for “other reasons”
when the reasons were symptoms, problems, follow-up of an abnormal test,
or other reasons. Finally, the authors evaluated whether and how
frequently participants confused sigmoidoscopy and colonoscopy. |
|
ASSESSMENT OF EXPOSURES AND POTENTIAL CONFOUNDERS:
| Computer assited telephone
interviews (CATI): The survey instrument elicited information
on demographic and health characteristics, including a personal or
family history of prostate or colorectal cancer, whether they had ever
been tested for prostate or colorectal cancer and, if so, when they had
the most recent tests (see Appendix for the questions asked). In
addition, participants were asked the reasons for testing and the test
results. The initial questions on prostate and colorectal cancer testing
were based on the questions proposed for the Year 2000 National Health
Interview Survey Cancer Control Supplement and Behavioral Risk Factor
Surveillance System. The questionnaire is available from the authors
upon request.
Medical record reviews:
Medical records were reviewed to determine whether any of the cancer
tests included in the survey had been recorded within 5 years before the
interview date. In addition, the dates and results of the tests were
obtained. For each test recorded in the medical records, the authors
ascertained the reason for or symptoms associated with the test by
reviewing the records for up to 6 months before the test date but no
more than 5 years before the survey (index) date. At each site, each
page of all relevant medical records in the study time period was
physically reviewed by a trained medical record analyst. Each provider
note, all laboratory, radiology, endoscopy, and pharmacy information
were abstracted using a standardized medical record abstraction form.
Quality control procedures included a review of all abstraction forms
for missing data and ambiguous responses and duplicate abstraction for
10% of the records. Data entry for all forms was conducted at Kaiser
Permanente- NC with double key entry. Programs for internal consistency
and validity were used to identify and correct errors of coding and data
entry. |
|
OUTCOME ASSESSMENT:
| (see above) |
|
RESULTS:
|
|
KQ2 - What factors influence the use of colorectal cancer
screening?
| NA |
|
KQ3 - Which strategies are effective in increasing the appropriate
use of colorectal cancer screening and follow-up?
| NA |
|
KQ4 - What are the current and projected capacities to deliver
colorectal cancer screening and surveillance at the population
level?
| NA |
|
KQ5 - What are the effective approaches for monitoring the use and
quality of colorectal cancer screening?
| Percentage of respondents who
received prostate or colorectal cancer tests according to self- report
and medical record audits, 1999 - 2001 |
| Black men (n=363) | White + other men
(n=847) | Women (n = 920) |
| % | (n) | % | (n) | % | (n) |
FOBT* |
Survey | 22.2 | (79) | 20.3 | (169) | 25.9 | (234) |
Med recs | 11.6 | (42) | 9.5 | (80) | 14.1 | (129) |
Sig’oscpy** |
Survey | 38.4 | (138) | 42.0 | (352) | 50.0 | (447) |
Med rec | 29.6 | (107) | 30.6 | (258) | 34.1 | (313) |
C-scope** |
Survey | 13.7 | (49) | 14.6 | (121) | 15.7 | (140) |
Med rec | 8.1 | (29) | 11.1 | (94) | 9.6 | (88) |
End’scopy** |
Survey | 44.4 | (159) | 49.9 | (415) | 58.6 | (523) |
Med rec | 34.4 | (124) | 37.8 | (319) | 39.8 | (365) |
Note: endoscopy indicates
sigmoidoscopy OR colonoscopy
* test within past 2
years
** test within past 5
years
CRC screening information: agreement between
self-reports and medical record audits, 1999 - 2001 |
procedure | Black men | White and other men | Women |
| | sens | spec | conc | K | n | sens | spec | conc | K | n | sens | spec | conc | K |
Kaiser Permanente-Georgia
|
Fobt | 187 | 0.89 | 0.86 | 0.86 | 0.57 | 293 | 0.90 | 0.85 | 0.86 | 0.56 | 310 | 0.89 | 0.82 | 0.84 | 0.62 |
Sig | 188 | 0.84 | 0.81 | 0.82 | 0.56 | 294 | 0.95 | 0.86 | 0.89 | 0.77 | 310 | 0.87 | 0.77 | 0.81 | 0.61 |
Col | 185 | 0.56 | 0.93 | 0.90 | 0.43 | 292 | 0.73 | 0.95 | 0.91 | 0.69 | 309 | 0.81 | 0.93 | 0.91 | 0.68 |
End | 186 | 0.87 | 0.83 | 0.84 | 0.64 | 294 | 0.96 | 0.88 | 0.92 | 0.83 | 311 | 0.95 | 0.73 | 0.83 | 0.67 |
Kaiser Permanente – northern California
|
Fobt | 165 | 0.73 | 0.86 | 0.85 | 0.29 | 271 | 0.72 | 0.86 | 0.85 | 0.32 | 304 | 0.79 | 0.86 | 0.86 | 0.35 |
Sig | 166 | 0.83 | 0.77 | 0.80 | 0.58 | 271 | 0.80 | 0.77 | 0.78 | 0.53 | 301 | 0.83 | 0.76 | 0.78 | 0.55 |
Col | 165 | 0.67 | 0.88 | 0.87 | 0.36 | 270 | 0.80 | 0.93 | 0.92 | 0.60 | 300 | 0.95 | 0.94 | 0.94 | 0.65 |
end | 165 | 0.93 | 0.76 | 0.82 | 0.65 | 269 | 0.91 | 0.75 | 0.81 | 0.62 | 299 | 0.91 | 0.76 | 0.82 | 0.63 |
HealthPartners, Minnesota |
Fobt | 4 | NA | NA | NA | NA | 269 | 0.59 | 0.87 | 0.84 | 0.31 | 287 | 0.55 | 0.81 | 0.78 | 0.23 |
Sig | 4 | NA | NA | NA | NA | 270 | 0.77 | 0.67 | 0.70 | 0.38 | 280 | 0.86 | 0.53 | 0.63 | 0.31 |
Col | 4 | NA | NA | NA | NA | 267 | 0.69 | 0.90 | 0.89 | 0.38 | 280 | 0.73 | 0.87 | 0.86 | 0/30 |
end | 4 | NA | NA | NA | NA | 267 | 0.85 | 0.64 | 0.71 | 0.43 | 279 | 0.93 | 0.45 | 0.61 | 0.30 |
Fobt = fecal occult
blood test
Sig = flexible sigmoidoscopy
Col
= colonoscopy
End =
endoscopy
Predictors of agreement between self-reports
and medical record audits of crc screening tests, generalized logits
models, 1999 – 2001 |
Predictor | Outcome | FOBT | Sig | Col |
| | OR | 95% CI | OR | 95% CI | OR | 95% CI |
Age (y) |
50 –
59 | Over | 1.69 | 0.91, 3.13 | 2.13 | 1.20, 3.80 | 6.81 | 1.62, 28.58 |
Under | 1.22 | 0.34, 4.41 | 3.98 | 1.20, 13.19 | 6.86 | 0.90, 52.15 |
60 –
69 | Over | 2.8 | 1.50, 5.25 | 3.07 | 1.70, 5.54 | 9.29 | 2.19, 39.42 |
Under | 1.9 | 0.51, 7.05 | 3.41 | 0.99, 11.72 | 9.95 | 1.27, 77.94 |
70+ | Over | 2.78 | 1.46, 5.29 | 2.85 | 1.56, 5.19 | 14.36 | 3.38, 61.06 |
Under | 2.47 | 0.66, 9.31 | 4.06 | 1.18, 13.93 | 13.91 | 1.72, 111.45 |
Sex/race |
White/other men | Over | 0.92 | 0.62, 1.37 | 0.75 | 0.52, 1.10 | 0.51 | 0.31, 0.83 |
Under | 0.74 | 0.29, 1.90 | 0.82 | 0.45, 1.5 | 0.86 | 0.40, 1.85 |
women | Over | 0.9 | 0.60, 1.34 | 0.91 | 0.62, 1.32 | 0.44 | 0.27, 0.73 |
Under | 0.88 | 0.34, 2.24 | 0.79 | 0.43, 1.45 | 0.34 | 0.14, 0.81 |
Education |
HS/GED | Over | 0.87 | 0.54, 1.4 | 1.48 | 0.91, 2.39 | 1.17 | 0.64, 2.16 |
Under | 1.74 | 0.49, 6.19 | 1.42 | 0.63, 3.24 | 0.93 | 0.32, 2.73 |
Some college / tech
school | Over | 0.73 | 0.46, 1.16 | 1.15 | 0.72, 1.84 | 0.84 | 0.46, 1.55 |
Under | 1.55 | 0.45, 5.40 | 1.07 | 0.48, 2.40 | 0.92 | 0.33, 2.56 |
College grad | Over | 0.85 | 0.54, 1.36 | 1.24 | 0.77, 2.00 | 0.77 | 0.41, 1.45 |
Under | 1.6 | 0.45, 5.70 | 1.09 | 0.48, 2.48 | 0.73 | 0.25, 2.10 |
aReference levels for
predictors are: age = 40– 49; sex/race group
= black men; education = < high school;
all analyses are adjusted for site.
bReference level
for outcome variable is “agreement” between the
survey and medical record.
Over =
overreport
Under = underreport |
|
QUALITY RATING:
| Good |
|
Quality Assessment-Internal Validity for Observational
Studies
|
|
|
Yes
|
No
|
Other (CD, NR, NA)
|
|
Were the groups similar at
baseline regarding the most important prognostic
indicators? | | | NA |
|
Were the drop-out or response
rates acceptable (≤ 20%)? [If
between 20% and 60%, check other and explain.] | X | | Cooperation rate 68% Some concern
about exclusion criteria (language barrier) |
|
Were the differential drop-out or
response rates acceptable (≤ 15%)? | | | NA |
|
Were intervention/exposure
measures valid, reliable, and equally applied? | X | | |
|
Were the outcome assessors blinded
to the intervention or exposure status of subjects? | X | | |
|
Were outcome measures valid,
reliable, and equally applied? | X | | |
|
Does the analysis control for
baseline differences? | X | | |
|
Were important potential
confounding and modifying variables taken into account in the design and
analysis (e.g., through matching, stratification, or statistical
adjustment)? | X | | |
|
Were the statistical methods used
to assess the abstracted outcomes appropriate? | X | | |
|
Quality Rating (Good, Fair,
or Poor): Good |
|
STUDY:
| Authors, ref ID:
Haque et al.103
Year of publication:
2005
Dates of data
collection:
Trial name: NA |
|
OBJECTIVE OR AIM:
| The goal of this study was to
develop an automated data algorithm designed to distinguish screening
and diagnostic endoscopy (sigmoidoscopy and colonoscopy) exams. The
authors assessed the algorithm’s ability to correctly
classify the exams using paper medical records as the “gold
standard.” |
|
DESIGN:
| Setting: Kaiser
Permanente Southern California (KPSC) cares for approximately 3 million
members, of whom 13% are older than 50 years and targeted for colorectal
cancer screening. Automated data tracks outpatient and inpatient care
received.
Study design:
cross-sectional
Duration (mean
follow-up):
Overall study size (N enrolled/N
analyzed): Stratified random sample of 220 medical records
reviewed. |
|
|
Had Colonoscopies
| Had
Sigmoidoscopies_110 |
Sample size:
| 110 | |
Describe intervention:
| | |
|
RECRUITMENT: (population-based, clinic-based, volunteer,
other)
| HMO based (Kaiser
Permanente) |
|
INCLUSION CRITERIA:
| Participants included all health
plan members between the ages 50 and 70 years, who were continuously
enrolled from 1998 to 2002, and completed an endoscopy during those
years.
Stratified random sample based on the
algorithm’s classification. 110 FS, 30 classified as
diagnostic and 80 as screening, and 110 COLON, 30 diagnostic and 80
screening. |
|
EXCLUSION CRITERIA:
| Participants with a history of
colorectal cancer were excluded (N =
1972).
Of the 220, 32 excluded due to mismatches in
participants endoscopy dates. |
|
POPULATION CHARACTERISTICS:
| NA |
Mean age & range (years):
| |
Sex (% female):
| |
Race:
| |
Other:
| |
|
Attrition/Drop-out (not available for endpoint
measurement):
| NA |
Adherence:
| |
Contamination:
| |
Response Rates (e.g. for surveys):
| |
|
STATISTICAL ANALYSES:
| The authors conducted
cross-tabulations between the algorithm and medical review
classification to examine the sensitivity, specificity, and
κ. The classification after medical record review was
considered the gold standard. Sensitivity indicates the probability that
a diagnostic endoscopy was classified as such by medical records review.
Specificity indicates the probability that a nondiagnostic endoscopy was
classified as screening. The κ indicates the overall
agreement between the two sources. |
|
ASSESSMENT OF EXPOSURES AND POTENTIAL CONFOUNDERS:
| NA |
|
OUTCOME ASSESSMENT:
| Endoscopies were identified using
International Classification of Disease (ICD 9 CM) and Current
Procedural Technology-4 codes. In instances in which a participant
completed multiple endoscopies in the 5-year period, the authors
retrieved data for the first endoscopy. FOBT (due to poor sensitivity)
and BE (due to infrequent use) were not included in the
study.
The algorithm used automated data to presumptively
classify the endoscopies as diagnostic or screening. Endoscopies were
classified as diagnostic if automated data included certain
gastrointestinal conditions in the year prior to the exam, or signs or
symptoms or a FOBT in the 45 days prior. The study gastroenterologist
(KRM) identified the conditions and signs and symptoms likely to result
in diagnostic endoscopies. All other endoscopies were classified into
the screening group.
Two trained abstractors reviewed medical
records from 1997 to 2002 to confirm endoscopy use. The abstractors also
assessed screening or diagnostic indications for the endoscopies,
including the presence of gastrointestinal conditions or signs and
symptoms. To minimize interrater variability, one abstractor reviewed
all participants’ medical records classified as a diagnostic
exam while the second reviewed all participants’ medical
records classified as a screening exam. Abstractors classified the
endoscopies as diagnostic if the exam was a follow-up to a previous
abnormality or when clear-cut conditions or signs and symptoms were
present, using the same list and time frames as the algorithm. All other
endoscopies were classified as screening. |
|
RESULTS:
|
|
KQ2 - What factors influence the use of colorectal cancer
screening?
| NA |
|
KQ3 - Which strategies are effective in increasing the appropriate
use of colorectal cancer screening and follow-
up?
| NA |
|
KQ4 - What are the current and projected capacities to deliver
colorectal cancer screening and surveillance at the population
level?
| NA |
|
KQ5 - What are the effective approaches for monitoring the use and
quality of colorectal cancer screening?
| The sensitivities for identifying
diagnostic sigmoidoscopy and colonoscopy were 48.1% and 23.8%,
respectively.
The algorithm missed most of the diagnostic
endoscopies. Conversely, the sensitivities for identifying screening
sigmoidoscopy and colonoscopy were high (87.9% and 84.4%, respectively)
but were associated with low specificities.
Comparison of
classification by automated algorithm versus medical record review |
| Classification by medical record
review |
Classification by automated
algorithm | Diagnostic (n =
90) | Screening (n=98) | % sensitivity | % specificity | kappa |
Sigmoidoscopy |
Diagnostic | 13 | 8 | 48.1 (13/27) | 12.1 (8/66) | 76.3 |
Screening | 14 | 58 | 87.9 (58/66) | 51.9 (14/27) | |
Total | 27 | 66 | | | |
Colonoscopy |
Diagnostic | 15 | 5 | 23.8 (15/63) | 15.6 (5/32) | 44.2 |
Screening | 48 | 27 | 84.4 (27/32) | 76.2 (48/63) | |
Total | 63 | 32 | | | |
The authors conclude that studies
relying solely on automated data overestimate screening rates if
indication is not considered. They also conclude that the automated
algorithm needs further improvements to better differentiate screening
from diagnostic exams. |
|
QUALITY RATING:
| Fair |
|
Quality Assessment-Internal Validity for Observational
Studies
|
|
|
Yes
|
No
|
Other (CD, NR, NA)
|
|
Were the groups similar at
baseline regarding the most important prognostic
indicators? | | | |
|
Were the drop-out or response
rates acceptable (≤ 20%)? [If
between 20% and 60%, check other and explain.] | | | |
|
Were the differential drop-out or
response rates acceptable (≤ 15%)? | | | |
|
Were intervention/exposure
measures valid, reliable, and equally applied? | X | | Algorithm developed by
investigator |
|
Were the outcome assessors blinded
to the intervention or exposure status of subjects? | | X | |
|
Were outcome measures valid,
reliable, and equally applied? | | X | Medical chart
review—single person review. Little detail on training
provided, no IRR assessment done |
|
Does the analysis control for
baseline differences? | | | |
|
Were important potential
confounding and modifying variables taken into account in the design and
analysis (e.g., through matching, stratification, or statistical
adjustment)? | | | |
|
Were the statistical methods used
to assess the abstracted outcomes appropriate? | | | |
|
Were data inputs
valid? | X | | |
|
Was an appropriate search strategy
used to find data inputs? | | X | Does not say how charts used were
randomly selected |
|
Were the calculations and
statistical analyses adequate? | X | | |
|
Were appropriate sensitivity
analyses conducted (especially for any variables that were not based on
data from literature)? | | | |
|
Other considerations: | X | | |
|
Quality Rating (Good, Fair,
or Poor): Fair |
|
STUDY: Yield of claims data and surveys for determining cancer
screening among health plan members
| Authors, ref ID:
Pignone et al 104
Year of publication:
2009
Dates of data
collection:
Trial name: CHOICE |
|
OBJECTIVE OR AIM:
| To evaluate the independent and
combined yield of claims and direct survey for identifying colorectal
cancer screening among average risk health plan beneficiaries. |
|
DESIGN:
| Setting: 32 Primary
care practices in Georgia, Florida taking part in a randomized trial of
a CRC decision aid and practice-level academic
detailing
Study design:
Observational
Duration (mean followup):
2005-2007
Overall study size (N enrolled/N
analyzed): 2558/1595 (responded to survey) |
|
|
Group 1
|
Sample size:
| 1595 |
Describe intervention:
| none |
|
RECRUITMENT: (population-based, clinic-based, volunteer,
other)
| Claims data from Aetna |
|
INCLUSION CRITERIA:
| Members with ages between 52 and
80 years whose primary care physicians had agreed to participate in the
CHOICE study |
|
EXCLUSION CRITERIA:
| Individuals at increased risk for
colorectal cancer (because the intervention was designed for
average-risk patients) or with medical conditions that would limit their
ability to participate in the study or who might not be considered
reasonable candidates for screening. Above average-risk persons were
defined as adults with a personal history of colorectal cancer or
polyps, a known history of colorectal cancer or polyps in a first-degree
relative, or a known history of inflammatory bowel disease also,
dementia, chronic obstructive pulmonary disease, heart failure, coronary
artery disease, current treatment for cancer or history of metastatic
cancer, cirrhosis, upper or lower gastrointestinal bleeding,
unintentional weight loss of >10% within 6 months, blindness, or
uncorrectable hearing impairment. To be excluded, the individual had to
have at least two claims with either a diagnosis or procedure code
indicating that they had one of these additional conditions. |
|
POPULATION CHARACTERISTICS:
|
Group 1
|
Mean age & range (years):
| Age group
52-59 887
(69.90)
60-64 283 (22.30)
65-69 81 (6.38)
70-82 18
(1.42) |
Sex (% female):
| % female 60% |
Race:
| 75% white/ 19% black/ 6%
other |
Other:
| |
|
Attrition/Drop-out (not available for endpoint
measurement):
| Overall |
Adherence:
| |
Contamination:
| |
Response Rates (e.g. for surveys):
| Response rate was 62% |
|
STATISTICAL ANALYSES:
|
Describe: Descriptive analyses, logistic regression and
examination of confounders
|
|
ASSESSMENT OF EXPOSURES AND POTENTIAL CONFOUNDERS:
| Yes |
|
OUTCOME ASSESSMENT:
| Outcome Measures:
primary outcome was proportion of persons up to date with CRC screening
Survey asked about individuals completion of CRC screening, whether even
had and when tests were conducted Claims data on CRC screening,
including FOBT (within 1 year), FS (5 years), colonoscopy (10 years) or
BE (5 years), including most recent data for each type. Indication for
the procedure was not specified. |
|
RESULTS:
|
|
KQ2 - What factors influence the use of colorectal cancer
screening?
| NA |
|
KQ3 - Which strategies are effective in increasing the appropriate
use of colorectal cancer screening and followup?
| NA |
|
KQ4 - What are the current and projected capacities to deliver
colorectal cancer screening and surveillance at the population
level?
| NA |
|
KQ5 - What are the effective approaches for monitoring the use and
quality of colorectal cancer screening?
| Outcomes:
• Of 4,020
average-risk members identified, claims data indicated that 1,066 (27%)
had recent colorectal cancer screening. Among the 1,269 average risk
members with no evidence of screening by claims data who returned
surveys, 498 (39%) reported being up-to-date with
screening.
• Combining claims data and
survey data:
• Prevalence of CRC
screening, combining claims data plus self-reported data (and not
including nonresponders to the survey):
47%
• Prevalence of CRC screening,
combining claims data plus self-reported data (assuming nonresponders
were screened at the same rate as average-risk responders): 59% |
|
QUALITY RATING:
| Fair |
|
Quality Assessment-Internal Validity for Observational
Studies
|
|
|
Yes
|
No
|
Other (CD, NR, NA)
|
|
Were the groups similar at
baseline regarding the most important prognostic
indicators? | | | NA |
|
Were the drop-out or response
rates acceptable (≤20%)? [If between
20% and 60%, check other and explain.] | x | | 62% response rate |
|
Were the differential drop-out or
response rates acceptable (≤15%)? | | | Not known |
|
Were intervention/exposure
measures valid, reliable, and equally applied? | | | NA |
|
Were the outcome assessors blinded
to the intervention or exposure status of subjects? | | | NA |
|
Were outcome measures valid,
reliable, and equally applied? | x | | However, screening vs. diagnostic
procedures were not distinguished |
|
Does the analysis control for
baseline differences? | | | NA |
|
Were important potential
confounding and modifying variables taken into account in the design and
analysis (e.g., through matching, stratification, or statistical
adjustment)? | | | NA |
|
Were the statistical methods used
to assess the abstracted outcomes appropriate? | x | | |
|
Quality Rating (Good, Fair,
or Poor): Fair |
|
STUDY:
Data sources for
measuring colorectal endoscopy use among Medicare enrollees Evaluation
of claims, medical records, and self-report for measuring fecal occult
blood testing among Medicare enrollees in fee for service | Authors, ref ID:
Schenck 105–106
Year of publication:
2007 and 2008
Dates of data collection:
1998–2002
Trial name: |
|
OBJECTIVE OR AIM:
| Comparison of different data
sources used for measuring CRC testing |
|
DESIGN:
| Setting:
Study design:
Observational
Duration (mean followup):
Brief
Overall study size (N enrolled/N
analyzed): 936 eligible 561 analyzed |
|
|
Overall
|
Sample size:
| 561 |
Describe intervention: NA
| endoscopy and FOBT |
|
RECRUITMENT: (population-based, clinic-based, volunteer,
other)
| Population – NC
Medicare patients from 10 different counties |
|
INCLUSION CRITERIA:
| North Carolina Medicare
enrollees, African American or white, in fee for service Medicare,
between ages 55 and 80 without history of CRC, who had responded to a
2002 survey. |
|
EXCLUSION CRITERIA:
| HMO coverage or gap in
coverage |
|
POPULATION CHARACTERISTICS:
|
Overall
|
Mean age & range (years):
| 50–64
10.5%
65–74 62.7%
75–80 26.8% |
Sex (% female):
| 61% female |
Race:
| African American 23.5% White
76.5% |
|
Attrition/Drop-out (not available for endpoint
measurement):
| Overall – NA |
Adherence:
| |
Contamination:
| |
Response Rates (e.g. for surveys):
| Response rate to survey not
reported. Sample consisted of 1001 persons who had responded to the
survey. |
|
STATISTICAL ANALYSES:
| Describe: descriptive statistics,
report to record ratio to detect bias, concordance amongst sources |
|
ASSESSMENT OF EXPOSURES AND POTENTIAL CONFOUNDERS:
| No |
|
OUTCOME ASSESSMENT:
| Outcome
Measures:
• Sigmoidoscoy,
colonoscopy and FOBT (both at home and in
office)
• Ever use of an endoscopic
procedure measured by survey, and date of most recent (outcome of FS in
last 4 years and colonscopy in last 5
years)
• FOBT, distinguishing in office
from home, and whether test was part of a check up or because of a
problem, from survey
• Medical record
review (using a hierarchical algorithm to link the patient to a
provider) to record date of endoscopic procedure, and whether screening
or diagnostic; and dates of four most recent
FOBTs
• Claims
(1/1998–12/2002) data using Medicare inpatient, outpatient,
and physician claims. Screening and diagnostic codes were
available. |
|
RESULTS:
|
|
KQ2 - What factors influence the use of colorectal cancer
screening?
| Outcomes:
NA |
|
KQ3 - Which strategies are effective in increasing the appropriate
use of colorectal cancer screening and followup?
| Outcomes:
NA |
|
KQ4 - What are the current and projected capacities to deliver
colorectal cancer screening and surveillance at the population
level?
| Outcomes:
NA |
|
KQ5 - What are the effective approaches for monitoring the use and
quality of colorectal cancer screening?
|
Outcomes: %
|
| Self report | Claim | Medical record |
Sigmoidoscopy | 22.8 | 21.6 | 15.2 |
Colonoscopy | 38.5 | 35.1 | 34.1 |
Endoscopy | 50.1 | 44.9 | 42.3 |
In last 5 yrs | NR | NR | |
FOBT- home | | | 30.1 |
FOBT office | | | 31.4 |
FOBT in last
year | 28.7 | 21.2 | 19.4 |
FOBT in last 2
years | 44.0 | 34.2 | 29.2 |
Prevalence of endoscopy in
the past year
Overall:
Survey
50.1
Administrative 44.9
Medical record review
42.3
By sociodemographic characteristics:
Age
55–64; 65–74; 65–80
Survey
50.8; 52.4; 44.0
Administrative 35.6; 43.9; 50.7
Medical
record review 32.2; 40.7; 50.0
All African Americans; all
whites; all women, all men:
Survey 40.9; 52.9; 46.8;
55.3
Administrative 41.7; 45.9; 43.6; 47.0
Medical record
review 42.4; 42.2; 42.7; 41.6
Less than high school; high
school diploma; more than high school:
Survey 28.7; 46.9;
59.8
Administrative 39.4; 45.9; 45.8
Medical record review
38.3; 41.8; 43.6
Prevalence of FOBT in the past
year
Overall:
Survey
28.7
Administrative 21.2
Medical record review:
19.4
By sociodemographic
characteristics:
Age 55–64;
65–74; 65–80
Survey 35.2; 27.9;
28.4
Administrative 19.3; 21.0; 23.6
Medical record
review: 19.3; 19.8; 19.6
All African Americans; all
whites; all women, all men
Survey 32.0; 27.8; 30.6;
25.9
Administrative 18.8; 22.4; 25.5; 15.3
Medical record
review: 12.5; 21.9; 21.7; 16.7
Less than high school;
high school diploma; more than high school
Survey 26.6; 26.0;
31.6
Administrative 20.2; 20.4; 22.8
Medical record
review: 19.1; 16.3; 22.4
Measures of concordance
for endoscopy use
Administrative to medical
record review
Agreement: 95 (93–97)
Kappa:
0.89 (0.81–0.98)
Self-report to medical
record review
Agreement: 70 (66–73)
Kappa:
0.39 (0.31–0.47)
Self-report to
administrative
Agreement: 70 (66–74)
Kappa:
0.40 (0.32–0.49)
Agreement regarding test
type (FS or colonoscopy)
Claims to medical record review:
93 (88–97)
Self-report to medical record review: 82
(75–89)
Self-report to claims: 77
(70–85)
Agreement regarding test purpose
(screening or diagnostic):
Administrative to medical
record review: 52 (43–61)
Self-report to medical
record review: 65 (55–74)
Self-report to
administrative: 29 (20–36)
Measures of
concordance for FOBT
Administrative to medical
record review
Agreement: 82
(79–85)
Self-report to medical record
review
Agreement: 70 (66–74)
Self-report to
administrative
Agreement: 67
(63–71)
Sensitivity analyses included:
excluding claims of FOBT on day of medical visit; including all medical
record review of FOBTs (likely including in-office, single card FOBTs
with digital rectal exams); did not appreciably change the measures |
|
QUALITY RATING:
| Good |
|
Quality Assessment-Internal Validity for Observational
Studies
|
|
|
Yes
|
No
|
Other (CD, NR, NA)
|
|
Were the groups similar at
baseline regarding the most important prognostic
indicators? | | | NA |
|
Were the drop-out or response
rates acceptable (≤20%)? [If between
20% and 60%, check other and explain.] | | | Survey response not recorded;
possibility that non respondents to survey would be different in some
way to alter agreement between a survey response and medical
records/chart review |
|
Were the differential drop-out or
response rates acceptable (≤15%)? | | | NA |
|
Were intervention/exposure
measures valid, reliable, and equally applied? | | | NA |
|
Were the outcome assessors blinded
to the intervention or exposure status of subjects? | | | NA |
|
Were outcome measures valid,
reliable, and equally applied? | x | | |
|
Does the analysis control for
baseline differences? | | | NA |
|
Were important potential
confounding and modifying variables taken into account in the design and
analysis (e.g., through matching, stratification, or statistical
adjustment)? | | | NA |
|
Were the statistical methods used
to assess the abstracted outcomes appropriate? | x | | |
|
Quality Rating (Good, Fair,
or Poor): Good |
|
STUDY:
| Authors, ref ID:
Schneider EC, et al.107
Year: 2008
Trial
Name: |
|
RESEARCH OBJECTIVE OR AIM:
| Evaluate quality measures by
describing a field test of the colorectal cancer screening measure
included in the Health Plan Employer Data and Information Set of the
National Committee for Quality Assurance – |
|
DESIGN:
| Setting: 5
geographically dispersed healthcare plans that enrolled 189,193
considered eligible for CRC screening
Study design:
Observational
Duration (mean
follow-up):
Overall study size (N enrolled/N
analyzed): 1000 plus an additional random sample of 400
enrollees from the original health care plan cohort (3000
enrollees) |
|
INTERVENTIONS:
| NA |
Sample size:
| |
Describe intervention:
| |
|
RECRUITMENT: (population-based, clinic-based, volunteer,
other)
| Each of the 5 enrolled health
plans identified all enrollees > 51 y/o, sent this to RAND Corp
for further analysis. Research staff at RAND Corp. randomly selected
1000 enrollees (200 per plan) who lacked evidence of CRC screening
(based on administrative data). Then, medical record abstractors looked
for evidence of CRC screening in these pts. Health plan also surveyed
patients via mail. During a 6 week period, nonrespondents received 3
survey mailings + a reminder postcard and a final overnight
mailing. Response rate was 48.1% (range 37.8% – 57.5% |
|
INCLUSION CRITERIA:
| > or = 51 y/o
continuously enrolled for at least 2 years in one of the 5 health
plans |
|
EXCLUSION CRITERIA:
| NR |
|
POPULATION CHARACTERISTICS:
| Survey respondents older, longer
enrollment
Survey respondents more likely than non respondents to
have evidence of CRC screening |
Mean age & range (years):
| |
Sex (% female):
| |
Ethnicity:
| |
Other:
| |
|
OUTCOME ASSESSMENT:
| | Plan A | Plan B | Plan C | Plan D | Plan E |
Age of plan, y | 53 | 19 | 17 | 17 | 17 |
Submitted cohort aged
51–80 and enrolled continuously for 2 years | 65,241 | 68,659 | 28,564 | 3,561 | 23,168 |
Age, mean, y | 64.2 | 60.9 | 59.7 | 58.8 | 57.4 |
Female, % | 54.1% | 53.0 | 51.4 | 52.8 | 51.4 |
Enrollment, mo | 73.3 | NA | 69.1 | 83.1 | 46.8 |
Survey rspndts | N=1250 | | | | |
Age, mean, y | 65.9 | 60.4 | 60.9 | 59.4 | 56.8 |
Female, % | 58.9 | 51.7 | 50.2 | 54.7 | 51.3 |
Enrollment, mo | 76.5 | NA | 75.5 | 86.4 | 48.3 |
Survey nonrespondents | N=1349 | | | | |
Age, mean, y | 64.6 | 60.3 | 59.9 | 58.8 | 56.7 |
Female, % | 54.2 | 50.7 | 52.0 | 51.8 | 49.3 |
Enrollment, | 81.8 | NA | 72.0 | 82.6 | 48.5 |
The authors defined, for each
data source (administrative, medical record, and survey results), the
data elements needed to implement all of the measure specifications. The
authors developed a list of outpatient diagnosis codes that represent a
prior diagnosis of colorectal cancer, CPT codes related
to an acceptable screening procedure, and historical
CPT codes used within the previous 10 years. For
the medical record method, the authors designed an abstraction tool to
collect data on screening tests, clinical risk factors for colorectal
cancer, and evidence of limited life expectancy and trained experienced
nurses from each health care plan during a pilot test of the abstraction
protocol using a common set of records. For the enrollee survey, the
authors developed a sequence of questions (based on the Behavioral Risk
Factor Surveillance System) that addressed each of the screening tests,
the time frames in which they occurred, and a measure of risk status
based on report of a family history of colorectal
cancer.
Data from all sources (administrative, survey,
and medical record) were linked at the enrollee level to create a single
analytic file. For each of the 4 colorectal screening tests, the authors
compared across health care plans the percentage of sampled enrollees
identified as screened based on administrative data and the percentage
based on survey data. The authors could not determine the percentage
screened by each test under the hybrid method because medical record
abstractors were instructed to stop after identifying the occurrence of
1 of the tests. For the single measure that combined all tests, the
authors calculated colorectal cancer screening performance scores based
on administrative data only, survey data only, and combined
administrative and medical record data (the hybrid method), weighing the
medical record sample to represent the population from which they were
drawn (unscreened based on administrative data).. Among the survey
respondents, the authors assessed agreement of the sampled
enrollees’ screening status according to the survey data and
the hybrid data. The authors compared the rate of screening among survey
respondents and nonrespondents using the hybrid estimation
procedure. |
|
RESULTS:
|
|
KQ4 - What are the current and projected capacities to deliver
colorectal cancer screening and surveillance at the population
level?
| NA |
|
KQ5 - What are the effective approaches for monitoring the use and
quality of colorectal cancer screening?
| Compared with survey
non-respondents, respondents were older (60.4 vs. 59.4 years;
P_.001) and had longer enrollment (73.3 vs. 67.6
months; P = .001), but the 2 groups had
similar percentages of female participants (53.3% vs. 51.1%;
P = .28).
Survey
respondents were more likely than non-respondents to have evidence of
CRC screening (62.7% vs. 46.5%; p<0.001)
The
percentage of enrollees having specific tests varied between the
administrative and survey data methods. In health care plans A and E,
the percentages of enrollees screened by FOBT according to the 2 methods
were similar; in health care plan B, the percentage based on survey data
was nearly twice that based on administrative data; and in health care
plan C, the percentage based on administrative data exceeded that based
on survey data. For the procedural tests (flexible sigmoidoscopy,
doublecontrast barium enema, and colonoscopy), the rates based on survey
data were 2 to 3 times higher than the rates based on administrative
data. |
Method | Plan A | Plan B | Plan C | Plan D | Plan E |
FOBT |
Administrative | 23.6 | 15.0 | 31.1 | NA | 24.7 |
Survey | 25.4 | 26.3 | 20.5 | 21.8 | 25.1 |
Flex Sig |
Administrative | 14.2 | 17.9 | 18.4 | 15.3 | 15.4 |
Survey | 29.7 | 39.6 | 33.9 | 33.6 | 30.6 |
COLON |
Administrative | 12.8 | 12.1 | 9.4 | 10.5 | 14.2 |
Survey | 19.9 | 39.0 | 33.6 | 33.7 | 40.7 |
Using the single measure of
screening (based on any 1 of the 4 tests), the percentage of enrollees
with evidence of colorectal cancer screening varied substantially
depending on the data sources used. For 4 of the 5 health care plans,
the hybrid method produced a higher screening rate than the
administrative data method. The difference ranged from 0 (health care
plan A) to 14.9 (health care plan B) percentage points. For all 5 health
care plans, the survey data method produced a higher calculated
screening rate than the other 2 methods. |
Method | Plan A | Plan B | Plan C | Plan D | Plan E |
Administrative | 41.5 (41.1–41.9) | 38.6 (38.2–38.9) | 47.1 (46.5–47.6) | 27.3 (25.8–28.7) | 44.4 (43.8–45.1) |
Survey | 53.2 (42.1–64.4) | 69.7 (60.3–79.2) | 55.0 (41.1–68.8) | 62.1 (53.8–70.4) | 66.2 (57.1–75.2) |
Hybrid | 41.5 (41.1–41.9) | 53.5 (48.5–56.8) | 52.6 (48.3–56.8) | 38.8 (34.3–43.4) | 45.6 (44.0–47.2) |
|
ANALYSIS:
| |
|
ARE GROUPS COMPARABLE AT BASELINE:
| |
|
ASCERTAINMENT METHODS ADEQUATE AND EQUALLY APPLIED:
| |
|
STATISTICAL ANALYIS ADEQUATE:
| |
|
ATTRITION:
| NA |
Dropout (not available for endpoint measurement):
| |
Adherence in control group:
| |
Contamination in control group:
| |
Differential dropouts:
| |
|
QUALITY RATING:
| Fair |
|
Quality Assessment-Internal Validity for Observational
Studies
|
|
|
Yes
|
No
|
Other (CD, NR, NA)
|
|
Were the groups similar at
baseline regarding the most important prognostic
indicators? | | | |
|
Were the drop-out or response
rates acceptable (≤ 20%)? [If
between 20% and 60%, check other and explain.] | | | RR varied among plans from 38% to
58% |
|
Were the differential drop-out or
response rates acceptable (≤ 15%)? | | | |
|
Were intervention/exposure
measures valid, reliable, and equally applied? | X | | |
|
Were the outcome assessors blinded
to the intervention or exposure status of subjects? | | | |
|
Were outcome measures valid,
reliable, and equally applied? | X | | Questions Based on BRFSS |
|
Does the analysis control for
baseline differences? | | | |
|
Were important potential
confounding and modifying variables taken into account in the design and
analysis (e.g., through matching, stratification, or statistical
adjustment)? | | | |
|
Were the statistical methods used
to assess the abstracted outcomes appropriate? | | | Very specific purpose to this
study to compare results of 3 different methods of estimating overall
screening rates. Hence outcomes not directly comparable to other
studies. |
|
Quality Rating (Good, Fair,
or Poor): Fair |