Djaa 048

JNCI J Natl Cancer Inst (2020) 112(7): djaa048
doi: 10.1093/jnci/djaa048
First published online April 27, 2020
Review
Health Insurance Coverage Disruptions and Cancer Care and

Outcomes: Systematic Review of Published Research
K. Robin Yabroff, PhD ,1,* Katherine Reeder-Hayes, MD,2 Jingxuan Zhao, MPH ,1 Michael T. Halpern, MD,
Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

PhD ,3 Ana Maria Lopez, MD,4 Leon Bernal-Mizrachi, MD,5 Anderson B. Collier, MD,6 Joan Neuner, MD,7
Jonathan Phillips, MPH ,8 William Blackstock, MD,9 Manali Patel, MD10
1
Surveillance and Health Services Research, American Cancer Society, Atlanta, GA, USA; 2Lineberger Comprehensive Cancer Center, University of North Carolina,
Chapel Hill, NC, USA; 3Division of Cancer Control and Population Sciences, National Cancer Institute, Rockville, MD, USA; 4Sidney Kimmel Cancer Center, Thomas
Jefferson University, Philadelphia, PA, USA; 5Winship Cancer Institute, Emory University School of Medicine, Atlanta, GA, USA; 6Children’s Cancer Center, The
University of Mississippi Medical Center, Jackson, MS, USA; 7Medical College of Wisconsin, Milwaukee, WI, USA; 8American Society of Clinical Oncology, Alexandria,
VA, USA; 9Wake Forest School of Medicine, Winston-Salem, NC, USA; and 10Stanford University School of Medicine, Stanford, CA, USA
*Correspondence to: K. Robin Yabroff, PhD, Surveillance and Health Services Research Program, American Cancer Society, 250 Williams St, Atlanta, GA 30303, USA (e-
mail: [email protected]).
Abstract
Background: Lack of health insurance coverage is associated with poor access and receipt of cancer care and survival in the
United States. Disruptions in coverage are common among low-income populations, but little is known about associations of
disruptions with cancer care, including prevention, screening, and treatment, as well as outcomes of stage at diagnosis and
survival. Methods: We conducted a systematic review of studies of health insurance coverage disruptions and cancer care
REVIEW
and outcomes published between 1980 and 2019. We used the PubMed, EMBASE, Scopus, and CINAHL databases and
identified 29 observational studies. Study characteristics and key findings were abstracted and synthesized qualitatively.
Results: Studies evaluated associations between coverage disruptions and prevention or screening (31.0%), treatment (13.8%),
end-of-life care (10.3%), stage at diagnosis (44.8%), and survival (20.7%). Coverage disruptions ranged from 4.3% to 32.8% of
patients age-eligible for breast, cervical, or colorectal cancer screening. Between 22.1% and 59.5% of patients with Medicaid
gained coverage only at or after cancer diagnosis. Coverage disruptions were consistently statistically significantly associated
with lower receipt of prevention, screening, and treatment. Among patients with cancer, those with Medicaid disruptions
were statistically significantly more likely to have advanced stage (odds ratios ¼ 1.2-3.8) and worse survival (hazard ratios ¼
1.28-2.43) than patients without disruptions. Conclusions: Health insurance coverage disruptions are common and adversely
associated with receipt of cancer care and survival. Improved data infrastructure and quasi-experimental study designs will
be important for evaluating the associations of federal and state policies on coverage disruptions and care and outcomes.
Lack of health insurance coverage is one of the strongest predic- Following the implementation of the Affordable Care Act
tors of poor cancer outcomes in the United States (1–4). The unin- (ACA), there were historic increases in the number of working-
sured are less likely to receive evidence-based care throughout age Americans with health insurance coverage (14). Even so,
the cancer control continuum, including prevention and screen- some adults experience insurance coverage losses and/or gains
ing, diagnosis, treatment (ie, surgery, radiation therapy, and sys- within a single year (15,16). Coverage disruptions and health in-
temic therapies) and symptom management, survivorship, and surance coverage churn are especially common among the poor
end-of-life care than their counterparts with health insurance and those with Medicaid coverage (17). Research conducted in
coverage (2,5). The uninsured are also more likely to have later pediatric populations consistently shows that health insurance
stage of disease at diagnosis (6,7) and poorer survival (4,7). coverage disruptions are associated with reduced access to care
Because health insurance coverage can reduce health disparities (18). In the few studies of coverage disruptions conducted in
in populations defined by race or ethnicity, poverty, and geogra- adults, even a single loss of coverage of at least 1 month is asso-
phy (2,8–13), expanding public and private health insurance cov- ciated with worse access to care (19), delaying or forgoing care
erage options has been the focus of many policy efforts. (20), more emergency department use (15,21,22), and declines in
Received: December 27, 2019; Revised: February 12, 2020; Accepted: March 27, 2020
Published by Oxford University Press 2020. This work is written by US Government employees and is in the public domain in the US.
671
672 | JNCI J Natl Cancer Inst, 2020, Vol. 112, No. 7
overall health (15). Transitions between health plans can also literature review, and Supplemental Figure 1 (available online)
disrupt access to usual source of care and provider networks. illustrates the search process.
However, to date, most research on the health effects of insur-
ance coverage has measured and evaluated coverage at only a
single point in time (eg, at cancer diagnosis) (3,4,23–26), and
Data Abstraction
synthesis of research addressing the effects of disruptions in
insurance coverage on cancer care and outcomes across the Data were abstracted on study characteristics, including year of
cancer control continuum is needed. Understanding the asso- publication, data source(s), geographic setting (national, multi-
ciations of coverage disruptions with care is especially rele- ple states/cities, single state), and study design (cross-sectional,
vant given the rapidly changing health insurance landscape cohort, intervention). Component(s) of the cancer control con-
in the United States. In this study, we conducted a systematic tinuum (prevention, screening, treatment, survivorship/sur-
review of published peer-reviewed research to assess the vival, and end-of-life care) and outcome measures (receipt of
associations of health insurance coverage disruptions with care, stage at diagnosis, survival and/or mortality, spending)
care and outcomes, including cancer prevention, screening, were also recorded. Although earlier stage of disease at diagno-
and stage of disease at diagnosis, treatment, survival, and sis can reflect receipt of regular screening for breast, cervical,

end-of-life care. and colorectal cancers, screening tests are not recommended
for other cancers. Additionally, early evaluation of signs and
symptoms can play an important role in diagnosis. For these
reasons, we abstracted and reported findings for stage at diag-
nosis separately from cancer screening.
Methods
Patient characteristics included sample size, age range or
age distribution, and cancer diagnosis if the study was con-
Literature Review
ducted in patients with cancer. The source of health insurance
Because a standard definition of coverage disruptions does not coverage measures included registry, self-report, and Medicaid
exist, we defined health insurance coverage disruptions as gaps enrollment. The type of insurance coverage disruption was ab-
in coverage, losses or gains of public or private coverage, timing stracted as reported in underlying articles and later classified as
of coverage (eg, pre-, peri-, or postcancer diagnosis), or transi- a coverage gap, the measured duration of coverage or coverage
tions between types of coverage (eg, public and private) or spe- gap, and timing of coverage (eg, pre-, peri-, or postcancer diag-
cific health insurance plans. We used the PubMed, EMBASE, nosis). Similarly, data on comparison groups were classified by
Scopus, and CINAHL databases to identify studies assessing insurance type, continuously insured or previously insured, du-
health insurance coverage disruptions and cancer care or out- ration of coverage (eg, <12 months vs 12 months), and timing
comes in the United States and published in English between of coverage. Key findings abstracted from each article noted the
REVIEW
January 1, 1980, and July 31, 2019. We started the literature prevalence and type of coverage disruption and the associations
search in 1980 and ended with the most recent year to ensure between coverage disruption and care and/or outcomes. A sin-
we identified as many relevant studies as possible. In the gle author (J.Z.) abstracted data from the included studies, and
PubMed database, our search strategy combined Medical another author (K.R.Y.) reviewed these data. Any inconsisten-
Subject Heading and title and abstract terms for neoplasms, cies were resolved by consensus. The heterogeneity of underly-
health insurance coverage, and health insurance enrollment ing study populations, coverage disruption measures, time
(see Supplemental Methods, available online). This search strat- periods, and component of the cancer control continuum pre-
egy was replicated in the EMBASE, Scopus, and CINAHL data- cluded a quantitative data synthesis. We performed a qualita-
bases, and the combined searches yielded 1523 unique articles. tive synthesis of study findings by outcome examined.
A single reviewer (J.Z.) assessed the abstracts of these articles
for eligibility. Included studies were required to quantitatively
assess insurance coverage at more than 1 time point to allow
Results
measurement of coverage disruptions (a period with insurance
coverage and a period either without coverage or with a change
Study, Patient, and Insurance Coverage Characteristics
in coverage) and examine the association between coverage dis-
ruption and cancer-related care (ie, prevention, screening, Most studies were conducted and published before 2014, con-
follow-up of abnormal findings, treatment, survivorship, or ducted in a single state, and used cancer registry data linked to
end-of-life care) or outcomes (ie, stage of disease at diagnosis, Medicaid enrollment data (Table 1). Studies were published af-
survival). We excluded studies conducted outside the United ter the passage of the ACA in 2010, but none evaluated the
States or whose only source of health insurance coverage infor- effects of the ACA on coverage disruptions or on the association
mation was from cancer registry data, because registries only of disruptions with care receipt or outcomes. Studies also used
report coverage at a single time point consolidated after cancer survey data or Medicaid enrollment and claims data only with-
diagnosis. We also excluded editorials, commentaries, and re- out registry linkage. Outcomes from claims, such as screening
view articles. Questions about article eligibility were resolved by or treatment, are reported only during periods of continuous
consensus. Following abstract review, 66 full articles were coverage. Studies evaluated associations between coverage dis-
reviewed, and of these, 24 met the inclusion criteria. Reference ruptions and prevention or screening (31.0%), stage of disease at
lists of included articles were hand-searched, and an additional diagnosis (44.8%), treatment (13.8%), survivorship or survival
5 articles were identified for a total of 29 articles (27–55). This (20.7%), and end-of-life care (10.3%). A single study evaluated
systematic review was conducted in accordance with Preferred health-care spending. Most were conducted in samples of at
Reporting Items for Systematic Reviews and Meta-Analyses least 1000 patients, and among studies that evaluated patients
guidelines (http://prisma-statement.org/). Supplemental Table 1 after a cancer diagnosis, the most common types were breast,
(available online) describes the search terms used in the cervical, colorectal, and lung.
K. R. Yabroff et al. | 673
Table 1. Study, patient, and insurance coverage characteristics Table 1. (continued)
No. of studies (%) No. of studies (%)

Characteristics (N ¼ 29) Characteristics (N ¼ 29)
Study characteristics Duration of coverage 12 (41.4)

Publication year Timing of coverage (eg, pre- or 18 (62.1)
2000-2004 10 (34.5) postdiagnosis)
2005-2009 5 (17.2) Comparison groupa
2010-2014 7 (24.1) Continuously insured or continuously 11 (37.9)
2015-2019 7 (24.1) uninsured
Geographic setting Continuously insured and continuously 5 (17.2)
National 4 (13.8) uninsured
Multiple cities and/or states 2 (6.9) Prediagnosis coverage 9 (31.0)
Single state 23 (79.3) Duration of coverage 4 (13.8)
Data source Other 2 (6.9)

Cancer registry - Medicaid enrollment, 6 19 (65.5) a
Categories are not mutually exclusive, and studies were included in multiple
claims
categories. AYA ¼ Adolescents and young adult.
Medicaid enrollment, claims 4 (13.8)
Survey 6 (20.7)
Study design
Cross-sectional 4 (13.8) Nearly three-quarters of studies used Medicaid enrollment
Cohort 25 (86.2) and claims data as the source of health insurance coverage;
Component of cancer control continuuma self-report was less common. There was substantial heteroge-
Prevention 1 (3.4) neity across studies in the types of coverage disruptions evalu-
Screening 8 (27.6) ated. Studies evaluated coverage gaps (27.6%), timing of
Stage of diagnosis 13 (44.8) coverage, either pre- or postdiagnosis (62.1%), as well as the du-
Treatment 4 (13.8) ration of coverage (41.4%) or coverage gap (6.9%). Many studies
Survivorship/survival 6 (20.7) evaluated multiple types of disruptions. The type of coverage
End-of-life care 3 (10.3) disruption varied by data source and outcome measured. For
Study outcome(s)a example, most studies evaluating stage of disease at diagnosis
Access to care or receipt of care 15 (51.7) used measures about the timing of coverage pre-, peri-, and/or
Stage of disease at diagnosis 13 (44.8)
postdiagnosis. The number of months used to measure the tim-
Survival or mortality 7 (24.1)
REVIEW
ing of coverage varied widely.
Spending 1 (3.1)
All studies included a comparison group, although compar-
Patient characteristics
ison groups varied, and included the continuously insured,
No. of patients
without Medicaid coverage (both uninsured and insured), in-
<999 3 (10.3)
1000-9999 11 (37.9)
sured before diagnosis, and a combination of continuously in-
10 000þ 15 (51.7) sured and continuously uninsured. Although some studies
Age groupa included populations with all types of coverage, including pri-
<18 13 (44.8) vate coverage, none of the studies evaluated coverage disrup-
18-39 21 (72.4) tions exclusively among those with private coverage or
40-64 25 (82.8) transitions between different private insurance plans.
Cancer site(s)a Heterogeneity of geographic region, patient population, meas-
Breast 8 (27.6) ures, and outcomes in the underlying studies precluded the
Cervical 6 (20.7) measurement of trends over time.
Colorectal 6 (20.7)
Lung 6 (20.7)
Other cancer sites 4 (13.8) Cancer Prevention and Screening
Multiple cancer types (eg, gynecologic, 3 (10.3)
AYA) For the studies evaluating cancer prevention and screening,
All cancer sites 2 (6.9) details of study populations, settings, measures of coverage dis-
Without cancer 9 (31.0) ruptions and outcomes, comparison groups, and key findings
Insurance coverage are listed in Table 2. One study evaluated associations between
Coverage disruption evaluated coverage disruptions and HPV vaccination in adolescent girls
Medicaid only 22 (75.9) (39), and 8 studies examined associations between disruptions
Multiple types of coverage 7 (24.1) and cancer screening in age-eligible adults (27–32,37,38).
Source of coverage measure Prevalence of coverage disruptions ranged from 4.3% to 32.8% in
Registry 1 (3.4) samples of adults without a cancer history and age-eligible for
Self-report 6 (20.7)
cancer screening in studies that reported this information
Enrollment and claims 21 (72.4)
(27,30). Coverage disruptions were statistically significantly as-
Other 1 (3.4)
sociated with less frequent receipt of prevention or screening in
Coverage disruption measure(s)a
7 of 9 studies; associations were null in the remaining 2 studies.
Coverage gap 8 (27.6)
Coverage disruptions were statistically significantly associated
Duration of coverage gap 2 (6.9)
with less use of mammography (27–29,32,37) and Pap testing
(continued)
(27–29) compared with continuously insured women or women
REVIEW
Table 2. Health insurance coverage disruptions and cancer screeninga
Sample size, Insurance coverage

Reference No. Setting, data source, and year measures Outcome measures Key findings
Ayanian et al., 2000 223 128 adults National; BRFSS, 1997-1998 Self-reported short-term Self-reported Pap test 3 y, 4.3 % short-term uninsured
(27) aged 18-64 uninsured (uninsured mammography 2 y, fe- Compared with currently insured, short-term uninsured
y <1 y) with coverage gap cal occult blood test 2 y, were less likely to receive mammography (78.7% vs
<1 year vs currently in- and sigmoidoscopy 5 y 89.0%, P < .05) or Pap test (89.5% vs 93.7%, P < .05)
sured (private and/or among eligible adults No differences observed for receipt of CRC screenings
public)
Bednarek and 11 755 women National; MEPS, 1996 Self-reported duration of Self-reported use of mam- Percentage short-term insured 1-6 and 7-11 mo not
Schone, 2003 (28) aged 21-64 private or public cover- mography and Pap test reported
y age, measured as 1-6 mo, 2 y among eligible Compared with continuously insured for 12 mo, insured
7-11 mo vs all 12 mo women for 1-6 mo less likely to have Pap smears (79.9% vs 70.7%,
P < .05) or mammograms (66.7% vs 53.6%, P < .05)
Compared with insured 7-12 mo, insured 1-6 mo were
less likely to have Pap smears (81.8% vs 70.7%, P < .05)
or mammograms (67.0% vs 53.6%, P < .05).
Continuously insured and insured 7-12 mo did not sta-
tistically significant differ

Broyles et al., 2002 1512 women Oklahoma’s BRFSS, 1993 Self-reported temporarily Self-reported use of mam- Percentage temporarily uninsured in previous year not
(29) aged 18 y uninsured with coverage mography (in the past reported
gap previous year vs con- 2 y) and Pap test screen- Use of mammograms and Pap smears similar in tempo-
tinuously insured ing (in the past 3 y) rarily uninsured and continuously insured (OR ¼ 0.8
among eligible women and OR ¼ 0.9, respectively; both P > .05)
Freund et al., 2019 333 adults Minority race/ethnicity or low SES Self-reported insurance in- Self-reported use of screen- 32.8% reported insurance instability
(30) aged 40-74 participants recruited across stability in past 12 mo de- ing and date of last test. No statistically significant differences in BC (72.6% vs
y 4 sites in 3 states: Chinese fined as uninsured, losing Up to date for BC, cervical 80.6%, P ¼ .23), cervical cancer (67.2% vs 73.4%, P ¼ .48),
Americans in Boston, MA; coverage, or changing in- cancer, or CRC screening and CRC (61.4% vs 70.1%, P ¼ .19) screening between
Hispanic in Columbus, OH; surance vs stable insur- status calculated per adults with insurance coverage instability compared
Appalachian populations in ance status (insured, USPSTF with those with stable coverage
OH’s Appalachian region; and uninsured)
African American and Black
populations in Philadelphia, PA.
Patients recruited from commu-
nity-based organizations, faith-
based organizations, public
housing, screening events,
health fairs, and from existing
research studies. Years of data
collection not stated
Jerant et al., 2013 (31) 92 809 adults National; MEPS, 2000-2008; 2-y Self-reported insurance loss Self-reported use of CRC 3193 adults lost insurance. Those with insurance loss
aged 18 y panels during 2-y follow-up pe- screening (fecal occult less likely to receive Pap (OR ¼ 0.6, 95% CI ¼ 0.5 to 0.8)
riod vs no private and/or blood testing 2 y and/or and mammography (OR ¼ 0.6, 95% CI ¼ 0.4 to 0.8)
public coverage change endoscopy 5 y), Pap test than those without insurance change. Also less likely
(continuously insured 3 y, and mammography to receive CRC screenings, but association not statisti-
cally significant (OR ¼ 0.7, 95% CI ¼ 0.4 to 1.0)
(continued)

Table 2. (continued)
Sample size, Insurance coverage

Reference No. Setting, data source, and year measures Outcome measures Key findings
and continuously 2 y among eligible

uninsured) adults
Koroukian, 2004 (32) 140 592 Ohio Medicaid claims and enroll- Duration of coverage ¼ Receipt of annual and regu- 40.7% enrolled 12 mo, 17.6% enrolled 13-24 mo, 41.7%
women ment files, 1992-1999 length of enrollment in lar annual screening enrolled >24 mo
aged 40-64 Medicaid (12, 13-24, 25- mammography from Proportion of women receiving screening mammogra-
y 36 mo, etc [up to 8 y]) claims phy increased statistically significantly each addi-
tional year of Medicaid enrollment (AOR ¼ 1.6, 95% CI
¼ 1.6 to 1.6)
Mean annual mammograms increased from 0.08 in
women with enrollment 12 mo to 0.26 in women
with enrollment 7 y
McWilliams et al., 2036 adults National; Health and Retirement Self-reported coverage gap Self-reported mammogra- 216 were intermittently insured
2003 (37) aged 60-64 Study, 1994, 1996 measured as intermit- phy 2 y Compared with continuously insured, intermittently in-
y in 1996; tently insured (insured in sured less likely to receive mammography (76.0% vs
No. of either 1994 or 1996) vs 57.7%, P < .05)
women not privately and/or publicly
stated insured in both 1994 and
1996 (continuously
insured)
O’Leary et al., 2019 10 831 adults Oregon Medicaid claims data, New enrollment in CRC screening with colo- Percentage newly enrolled not reported
(38) who turned 2010-2014 Medicaid at 50 y vs prior noscopy, sigmoidoscopy, No differences in CRC screening newly enrolled in
50 y during enrollment before 50 y or stool testing from Medicaid and prior enrollment (RR ¼ 1.0, 95% CI ¼ 0.8
2010-2013 claims within 12 mo of to 1.3, P ¼ .87)
age 50 y
Staras et al., 2010 237 015 girls Florida Medicaid enrollment and Coverage duration mea- At least 1 HPV vaccine Percentages by duration of months enrolled in Medicaid
(39) aged 9-17 y claims, June 2006 to Dec 2008 sured as no. of months claim not reported
enrolled in Medicaid Longer length of Medicaid enrollment positively associ-
(1–31) ated with receipt of at least 1 HPV vaccination. HPV
vaccination rates were 3.5%, 12.8%, 19.0%, 22.2%, 28.4%
for girls enrolled in Medicaid for 1-7, 8-13, 14-19, 20-25,
26-31 mo, respectively
a
AOR ¼ adjusted odds ratio; BC ¼ breast cancer; BRFSS ¼ Behavioral Risk Factor Surveillance Survey; CI ¼ confidence interval; CRC ¼ colorectal cancer; MEPS ¼ Medical Expenditure Panel Survey; OR ¼ odds ratio; RR ¼ risk ratio; SES
¼ socioeconomic status; USPSTF ¼ United States Preventive Services Task Force.

REVIEW
with longer durations of insurance coverage. Shorter durations disruptions and outcomes, comparison groups, and key findings
of insurance coverage were statistically significantly associated are listed in Table 4. Studies were conducted in California,
with less frequent receipt of HPV vaccination among adolescent Georgia, Michigan, North Carolina, New Jersey, and Ohio.
girls (39). Several studies reported a dose-response relationship Among newly diagnosed patients with Medicaid coverage, be-
between coverage duration and greater receipt of prevention tween 22.1% and 59.5% enrolled in Medicaid at or after diagno-
and screening (28,32,39). sis. In 3 of 4 studies evaluating treatment, coverage disruptions
Coverage disruption was not statistically significantly asso- were associated with treatment delay (51) and lower likelihood
ciated with receipt of screening in 1 study with a relatively of receiving treatment (44,52). Treatment delay and lack of de-
small sample (n ¼ 333) that included adults older than age finitive surgery were reported for multiple cancer sites, includ-
65 years (who are generally age-eligible for continuous Medicare ing breast, colon, lung, and gastric cancers (44,51,52). Two
coverage) (30). Associations between coverage disruptions and relatively small studies reported null findings for the associa-
colorectal cancer screening (eg, fecal occult blood test, flexible tion between coverage disruption and receipt of treatment
sigmoidoscopy, or colonoscopy) in men and women were null (52,53); however, they measured both coverage disruptions and
(27,30,38). Two of the 4 studies did not include colonoscopy treatment with Medicaid claims after diagnosis, limiting inter-
(27,38), currently the most common colorectal cancer screening pretation of findings because treatment could not be measured

modality. among those without Medicaid coverage.
All 7 studies reported that survival following diagnosis was
statistically significantly worse for patients who gained
Stage of Disease at Diagnosis Medicaid coverage at or after diagnosis. Worse survival was
reported for multiple cancer types, including, breast, cervical,
Among the 13 studies that examined coverage disruptions and
colon, lung, and gastric cancers, with odds ratios or hazard ra-
stage of disease at diagnosis, details for study populations, set-
tios ranging from 1.28 to 2.43 (40,42,51). Several studies
tings, measures of coverage disruptions and outcomes, compar-
reported that disparities in survival for those with coverage
ison groups, and key findings are listed in Table 3
disruptions were mediated by differences in stage at diagnosis
(33,34,40,41,43–51). Each of these studies examined timing of
(40,45).
Medicaid enrollment in a single state, including California,
Although studies did not explicitly compare the magnitude
Michigan, New Jersey, North Carolina, Ohio, or Washington.
of associations between coverage disruptions for cancers with
Coverage disruption was measured as Medicaid enrollment tim-
and without effective screening tests, 1 study of patients with
ing in relation to diagnosis or as continuity of Medicaid cover-
breast cancer reported better cancer-specific and overall sur-
age. The number of months used to define timing of coverage or
vival among previously uninsured one-time and repeat BCCEDP
coverage continuity varied widely. Information about insurance
users compared with other Medicaid enrollees (34).
coverage before Medicaid enrollment was not available, and
REVIEW
patients could have been previously enrolled in private health

insurance plans or uninsured.
End-of-Life Care and Health-Care Spending
Among newly diagnosed patients with Medicaid coverage,
between 22.1% and 59.5% enrolled in Medicaid at or after diag- Two studies evaluated coverage disruptions and end-of-life care
nosis in studies that reported this information. Coverage dis- (54,55) and 1 evaluated spending among Medicaid enrollees
ruptions were associated with advanced stage in all 13 studies, who died of cancer. Details for each study are listed in Table 5
with odds ratios ranging from 1.2 to 3.8. In studies that evalu- (36). Studies were conducted in California, New York, and Ohio.
ated the effects of coverage disruptions by cancer site, disrup- Disruptions were associated with statistically significantly less
tions were statistically significantly associated with advanced use of hospice before death among patients diagnosed with
stage for breast, cervical, colorectal, thyroid, lung, melanoma, stage IV lung cancer compared with those with continuous
non-Hodgkin and Hodgkin lymphomas, and ovarian cancers. Medicaid coverage in 2 states (54). Similarly, disruptions were
None of the studies explicitly compared the magnitude of associated with statistically significantly less hospice use
associations between coverage disruptions for those cancers among those diagnosed with any cancer as adolescent and
with effective screening tests (ie, breast, cervical, colorectal) and young adults (55). Both studies reported a dose-response associ-
those cancers without effective screening tests that are typically ation, with lower odds of hospice use among those with greater
diagnosed at a later stage of disease. One study linked cancer coverage discontinuity between diagnosis and death (54,55).
registry, Medicaid enrollment, and the Breast and Cervical Discontinuous Medicaid coverage was also associated with less
Cancer Early Detection Program (BCCEDP) data to assess one- use of other types of care, including aggressive treatment, mea-
time and repeat BCCEDP use and stage at breast cancer diagno- sured as chemotherapy within 14 days of death, emergency
sis (34) (previously uninsured women diagnosed with cancer as room visit, hospitalization, or intensive care unit stay within
a result of screening through BCCEDP are automatically enrolled 30 days of death (55). Longer Medicaid enrollment was associ-
in Medicaid). Compared with prediagnosis Medicaid enrollees, ated with higher per-person, per-month medical care spending
one-time BCCEDP users were more likely to be diagnosed with among patient with any cancer indicated as the underlying
advanced-stage breast cancer, but there was no statistically sig- cause of death (36), although no age restriction on the date of
nificant association with stage for repeat BCCEDP users. death was used and 61% were aged 65 years and older and age-
eligible for Medicare coverage.
Cancer Treatment and Survival

Discussion
Four studies evaluated the effects of Medicaid coverage disrup-
tions on receipt of treatment (44,51–53), and 7 studies evaluated In this study, we conducted a systematic review of the pub-
the effects of disruptions on survival (34,35,40,42,44,45,51). lished literature to assess the role of health insurance coverage
Details of study populations, settings, measures of coverage disruptions on receipt of cancer care and outcomes in the
Table 3. Health insurance coverage disruptions and stage of disease at cancer diagnosisa
Setting, data source, and Insurance coverage

Reference Sample size, No. year measures Outcome measures Key findings
Bradley et al., 2003a (41) 5852 women and men aged Michigan Cancer Registry - Timing of coverage mea- Early stage (in situ or local- 36% enrolled in Medicaid after diagnosis
25-64 y newly diagnosed Medicaid enrollment, sured as Medicaid-en- ized) vs late stage (re- Postdiagnosis Medicaid enrolled more likely
with BC, CRC, or cervical 1996-1997 rolled after diagnosis vs gional, distant, or to have late-stage diagnosis compared with
or lung cancer Medicaid-enrolled for >1 invasive/unknown) from previously enrolled for BC (1.328, 95% CI ¼
mo at diagnosis registry 0.95 to 1.67), cervical cancer (2.96, 95% CI ¼
1.85 to 4.75), CRC (2.08, 95% CI ¼ 1.30 to 3.33,
and lung cancer (3.40, 95% CI ¼ 2.13 to 5.43)
Bradley et al., 2003b (40) 598 women aged 29-64 y Michigan Cancer Registry - Timing of coverage mea- Early stage (in situ or local- 22.1% enrolled in Medicaid after diagnosis
with newly diagnosed BC Medicaid enrollment, sured as Medicaid-en- ized) vs late stage (re- Lack of Medicaid coverage before diagnosis
1996-1997 rolled after diagnosis vs gional, distant, or increased odds of late-stage diagnosis
Medicaid-enrolled for >1 invasive/unknown) from (HR ¼ 1.71, 95% CI ¼ 1.13 to 2.58, P < .05)
mo at diagnosis registry
Bradley et al., 2004 (43) 1063 women aged <65 y Michigan Cancer Registry - Timing of coverage mea- Early-stage (in situ or local)Percentage enrolled in Medicaid after diagno-
with newly diagnosed Medicaid enrollment, sured as Medicaid-en- vs late stage (regional, sis not reported. Compared with those en-
cervical cancer 1996-1997 rolled after diagnosis vs distant, or invasive/un- rolled in Medicaid >1 mo before diagnosis,
Medicaid-enrolled for >1 known) from registry those enrolled after diagnosis more likely
mo at diagnosis diagnosed at late stage (4% vs 8%, P < .05)
Dawes et al., 2014 (44) 96 220 women and men California Cancer Registry, Discontinuous Medicaid en- Early stage (localized and 59.5% of patients with Medicaid discontinu-
aged <65 y newly diag- California’s Patient rollment (did not have regional) vs advanced ously enrolled
nosed with colon, esoph- Discharge Database, and consecutive 6 mo cover- stage (remote) from Compared with those continuously enrolled
ageal, lung, ovarian, Medicaid enrollment age prediagnosis) vs con- registry in Medicaid, those discontinuously enrolled
pancreatic, or gastric files, 2002-2008 tinuous Medicaid more likely to have advanced stage of diag-
cancers enrollment (>6 mo nosis for colon (29.8% vs 41.7%), esophageal
prediagnosis) (51.6% vs 58.5%), lung (67.8% vs 78.5%), ovar-
ian (59.8% vs 69.4%), pancreatic (66.1% vs
69.5%), and gastric (52.6% vs 61.6%) cancers
in unadjusted analyses
Doll et al., 2016 (45) 782 women aged <65 y North Carolina Central Timing of coverage mea- Early stage (local) vs ad- 38.6% enrolled in Medicaid after diagnosis
newly diagnosed with Cancer Registry - sured as Medicaid enroll- vanced stage (regional After propensity matching, postdiagnosis en-
cancers of cervix, uterus, Medicaid enrollment, ment after diagnosis vs and distant) from registry rollment group more likely to have ad-
ovary, and vulva or 2003-2008. Follow-up Medicaid enrollment pre- vanced stage (OR ¼ 1.46, 95% CI ¼ 1.03 to
vagina through 2010 diagnosis (1-6 mo be- 2.05). When stratified by cancer site, effect
fore diagnosis) greatest in uterine cancer (OR ¼ 1.74, 95%
CI ¼ 0.87 to 3.47) and cervix (OR ¼ 1.50, 95%
CI ¼ 0.91 to 2.49), but not statistically
significant
Keegan et al., 2019 (46) 52 774 AYAs aged 15-39 y California Cancer Registry Timing of coverage and AJCC stage I vs II-IV or AJCC Of Medicaid patients, 34.2% peridiagnosis en-
newly diagnosed with 9 data (2005–2014), linked coverage gap measured stage I-II vs III-IV from rolled <1 mo and 15.7% discontinuously
common cancers, includ- to Medicaid enrollment as Medicaid peridiagnosis registry enrolled
ing BC, CRC, and thyroid, (2004-2014) enrolled <1 mo, or dis- Compared with AYAs with private insurance,
melanoma, testicular, continuous Medicaid vs AYAs who gained Medicaid coverage at di-
non-Hodgkin lymphoma, continuous Medicaid (en- agnosis 2.2-2.5 times more likely later stage
rolled 5 mo before (stage II-IV vs I: OR ¼ 2.46, 95% CI ¼ 2.26 to
(continued)

REVIEW
REVIEW

Hodgkin lymphoma, cer- diagnosis) and private or 2.69; III-IV vs I-II: OR ¼ 2.16, 95% CI ¼ 2.00 to
vical, and ovarian military coverage 2.33) and AYAs with discontinuous
Medicaid 1.7-1.9 times more likely later
stage (stage II-IV vs I: OR ¼ 1.93, 95% CI ¼
1.70 to 2.18; III-IV vs I-II: OR ¼ 1.74, 95% CI ¼
1.56 to 1.95) in adjusted analyses. Findings
statistically significant for all 9 cancers. In
analyses limited to AYAs with Medicaid,
continuous coverage improved odds of ear-
lier stage diagnosis compared with both
peridiagnosis enrollment or discontinuous
enrollment
Koroukian, 2003 (33) 2576 women aged 15 y, Ohio Cancer Registry; Medicaid vs non-Medicaid Early stage (in situ or local- Of patients with Medicaid, 70.8% enrolled
newly diagnosed with BC Medicaid Enrollment coverage, and timing of ized) vs advanced stage prediagnosis, 25.2% peridiagnosis, and 4.0%
or cervical cancer data, 1996-1998 Medicaid enrollment: (regional or distant) from postdiagnosis
prediagnosis (enrolled >3 registry Compared with prediagnosis enrollment,
mo preceding diagnosis), peridiagnosis group had increased risk of
peridiagnosis (2 mo pre- advanced cancer overall (AOR ¼ 3.8, 95%
ceding diagnosis), and CI ¼ 2.8 to 5.0) and for BC (AOR ¼ 3.8, 95%
postdiagnosis (enrolled CI ¼ 2.8 to 5.2) and cervical cancers (AOR ¼
3 mo after diagnosis) 3.6, 95% CI ¼ 1.8 to 7.3)
Postdiagnosis enrollment increased risk of
advanced cancer compared with prediagno-
sis group (AOR ¼ 2.2, 95% CI ¼ 1.2 to 4.0). By
cancer site, advanced disease higher for BC
(AOR ¼ 3.8, 95% CI ¼ 2.8 to 5.2; AOR ¼ 2.1,
95% CI ¼ 1.1 to 4.1) and cervical cancers
(AOR ¼ 3.6, 95% CI ¼ 1.8 to 7.3; AOR ¼ 2.8,
95% CI ¼ 0.5 to 14.2) for peri- and postdiag-
nosis groups compared with prediagnosis
group, respectively
Koroukian et al., 2017 (34) 26 426 women aged 40-64 y Ohio Cancer Registry, Timing of Medicaid cover- Localized vs advanced- 31.6% of patients in peri- or postdiagnosis
newly diagnosed with in- Medicaid enrollment age peridiagnosis, (en- stage (regional or distant) group
vasive BC data, and BCCEDP data- rolled at diagnosis or <3 from registry Peridiagnosis group more likely to be diag-
base (diagnoses 2002- mo of diagnosis), BCCEDP nosed with advanced-stage disease (AOR ¼
2008, deaths 2002-2010) repeat user vs enrolled 2.20; 95% CI ¼ 1.83 to 2.66)
3 mo before diagnosis Compared with Medicaid prediagnosis,
BCCEDP 1-time users more likely to be diag-
nosed late stage (AOR ¼ 1.48, 95% CI ¼ 1.17
to 1.87) but repeat users similar stage at di-
agnosis (AOR ¼ 0.83, 95% CI ¼ 0.59 to 1.18)
(continued)


O’Malley et al., 2006 (47) 4682 women aged <65 y California Cancer Registry- Timing of Medicaid cover- Early stage (localized) vs 16.0% enrolled in Medicaid during month of
newly diagnosed with in- Medicaid enrollment; age measured as prediag- late stage (regional and diagnosis, 23.6% 1-11 mo before diagnosis,
vasive cervical cancer 1996-1999 nosis: 1) first enrolled in remote) from registry and 60.4% continuously enrolled 1 y be-
month of diagnosis; 2) fore diagnosis
enrolled at time of diag- Compared with non-Medicaid coverage
nosis and for 1-11 mo in (uninsured and insured), adjusted odds ra-
year before diagnosis (in- tios for late- or unknown-stage diagnosis
termittently enrolled); 3) were: 2.8 (95% CI ¼ 1.9 to 4.2) for those en-
enrolled 12 mo before di- rolled at diagnosis, 1.34 (95% CI ¼ 1.00 to
agnosis, including at di- 1.80) for those enrolled 1-11 mo before diag-
agnosis; and 4) not nosis, and 1.08 (95% CI ¼ 0.89 to 1.33) for
enrolled at diagnosis those continuously enrolled in Medicaid for
1 y
Perkins et al., 2000 (48) 10 016 women aged 30-64 y, California Cancer Registry, Timing of prediagnosis Early stage (in situ and lo- 18.0% enrolled in Medicaid at time of
newly diagnosed with BC and linked Medi-Cal en- Medicaid enrollment: 1) calized) vs late stage (any diagnosis
rollment files, 1992-1993 entire 12 mo before diag- extension beyond the Compared with non-Medicaid (uninsured/pri-
nosis; 2) part of 12 mo be- breast, including regional vately insured combined), odds ratio for
fore diagnosis; and 3) not lymph nodes) from late-stage disease among all women on
covered by Medicaid in registry Medi-Cal was 1.67 (95% CI ¼ 1.41 to 1.97) but
any of 12 mo before was reduced by 42% to 1.39 (95% CI ¼ 1.15 to
diagnosis 1.67) when women without benefits before
diagnosis excluded
Pollitt et al., 2008 (49) 4558 women and men aged California Cancer Registry Medicaid enrollment at di- Localized and advanced- Among Medicaid-insured, 13.7% first enrolled
15-64 y, newly diagnosed and linked Medicaid en- agnosis (yes/no); timing stage (regional or distant) month of diagnosis; 31.6% enrolled for 1 -11
with melanoma rollment files, 1998-1999 of enrollment: 1) first en- from registry mo (continuously or noncontinuously) be-
rolled at month of diag- fore diagnosis; and 54.7% enrolled entire
nosis, 2) enrolled during past year.
month of diagnosis and Compared with non-Medicaid coverage
1-11 mo before diagnosis, (uninsured and insured combined), ad-
3) enrolled during month justed odds ratios for late-stage diagnosis
of diagnosis and 12 mo were: 13.64 (95% CI ¼ 4.43 to 41.98) for those
before diagnosis, and 4) enrolled at diagnosis, 2.77 (95% CI ¼ 1.28 to
not enrolled at diagnosis 5.99) for those enrolled 1-11 mo before diag-
nosis, and 1.30 (95% CI ¼ 0.64 to 2.64) for
those continuously enrolled in Medicaid 1
Ramsey et al., 2008 (50) 5009 women and men aged Washington State Cancer Timing of Medicaid enroll- In situ, localized, regional, 57.2% of Medicaid patients enrolled <3 mo
<65 y with newly diag- Registry and Medicaid ment measured as previ- and distant from registry before diagnosis
nosed BC, CRC, or cervi- enrollment 1997-2002 ously enrolled (3 mo Those enrolled in Medicaid at diagnosis more
cal, lung, or prostate before diagnosis) vs en- likely to have regional and distant stage dis-
cancer rolled at diagnosis (<3 ease at diagnosis than previously enrolled
mo before diagnosis to 6 (P < .001) and more likely to disenroll at 12
mo after diagnosis) mo (76.4% vs 23.6%)
(continued)

REVIEW
REVIEW

Tsui et al., 2018 (51) 19 209 women aged 21-64 y New Jersey Cancer Registry Timing of coverage mea- Early stage (in situ or local- 25.3% with Medicaid newly enrolled at
newly diagnosed with BC, and New Jersey Medicaid sured as longer term/ ized) or late (regional or diagnosis
CRC, or cervical cancer Management established Medicaid (en- distant) from registry For all cancer sites, statistically significantly
Information System, rolled 6 mo at diagnosis) higher proportions of Medicaid patients
2012-2014 or newly enrolled and especially newly enrolled Medicaid
Length of Medicaid enroll- patients diagnosed with late-stage cancer
ment (11, 6 to <11, 1 to compared with non-Medicaid patients (BC,
<6, <1 mo) 20% and 23% vs 11%, P < .001; CRC, 46% and
Continuous enrollment sta- 56% vs 42%, P ¼ .025; ICC, 41% and 38% vs
tus (defined as no gaps 30%, P < .001)
>30 d in prior year) For all cancers combined, shorter enrollment
length was associated with higher likeli-
hood of late-stage diagnosis (57.5%, 48.1%,
39.9%, and 41.5 for <1, 1 to <6, 6 to <11, and
11 mo, respectively. P < .001)
No statistically significant differences in

stage at diagnosis observed comparing
those continuously covered and with cover-
age disruptions (46.3% vs 45.2%, P ¼ .86)
a
AJCC ¼ American Joint Committee on Cancer; AOR ¼ Adjusted odds ratio; AYA ¼ Adolescents and young adult; BC ¼ breast cancer; BCCEDP ¼ Breast and Cervical Cancer Early Detection Program; CI ¼ confidence interval; CRC ¼ co-
lorectal cancer; HR ¼ hazard ratio; ICC ¼ invasive cervical cancer; OR ¼ odds ratio; RR ¼ risk ratio.

Table 4. Health insurance coverage disruptions and treatment and mortalitya

Adams et al., 2012 2048 women aged Georgia Cancer Registry Timing of coverage Receipt of lumpectomy, 51.7% enrolled after diagnosis and 48.9% continuously
(52) 19-63 y with newly -Medicaid enrollment measured as mastectomy, any enrolled
diagnosed BC and BCCEDP data Medicaid-enrolled be- drug regimen (hor- Compared with women enrolled in Medicaid after cancer
2002–2004 with 2-y fore diagnosis vs monal or chemother- diagnosis, those previously enrolled more likely to receive
follow-up Medicaid-enrolled af- apy), radiation, and any treatment (OR ¼ 2.41, 95% CI ¼ 1.28 to 4.56) or any de-
ter diagnosis and any treatment from finitive surgery (OR ¼ 7.66, 95% CI ¼ 5.06 to 11.59) in ad-
Medicaid-enrolled as diagnosis to end of justed analyses
part of BCCEDP follow-up. Receipt of Compared with “other” and disabled Medicaid enrollment,
Prevention and treatment measured BCCEDP enrolled more likely to receive any treatment (OR
Treatment Act, dis- by claims ¼ 4.71, 95% CI ¼ 2.48 to 8.96), any drug regimen (OR ¼ 3.58,
abled, and other 95% CI ¼ 2.32 to 5.51), and any definitive surgery (OR ¼
Medicaid 2.52, 95% CI ¼ 1.74 to 3.66) in adjusted analyses
Bradley et al., 2003b 598 women aged 29- Michigan Cancer Timing of coverage All-cause mortality 22.1% enrolled in Medicaid after diagnosis
(40) 64 y with newly Registry - Medicaid measured as Vital status from Among those aged <65 y, mortality risk was higher for
diagnosed BC enrollment, 1996- Medicaid-enrolled for registry those without Medicaid coverage at diagnosis (HR ¼ 1.67,
1997. Follow-up until >1 mo at diagnosis, 95% CI ¼ 1.09 to 2.56, P < .05). Overall, late stage was
1998 Medicaid-enrolled af- strongest predictor of increased risk of mortality (HR ¼
ter diagnosis 4.40, 95% CI ¼ 2.8 to 6.9, P < .05).
When late-stage disease added to model, effects of
Medicaid insurance no longer statistically significant
Bradley et al., 2005 13 740 women and Michigan Cancer Timing of coverage All-cause mortality 42% enrolled in Medicaid after diagnosis
(42) men aged <65 y Registry - Medicaid measured as Vital status from Median survival in postdiagnosis enrollment group was 19
newly diagnosed enrollment, 1996-1997 Medicaid-enrolled for registry mo (95% CI ¼ 17 to 22 mo) compared with 38 mo (95% CI ¼
with BC, CRC, or and follow-up >1 mo at diagnosis, 32 to 44 mo) in prediagnosis enrollment group. When
lung cancer through 2003 Medicaid-enrolled af- assessed by cancer site and stage, HR for Medicaid en-
ter diagnosis, or non- rolled after diagnosis vs non-Medicaid generally higher
Medicaid (including for women with early-stage BC (HR ¼ 3.10, 95% CI ¼ 2.35
uninsured or privately to 4.10), CRC (HR ¼ 2.78, 95% CI ¼ 1.87 to 4.14), and lung
insured) cancer (HR ¼ 1.64, 95% CI ¼ 1.19 to 2.26) than late-stage BC
(HR ¼ 2.43, 95% CI ¼ 1.94 to 3.04), CRC (HR ¼ 2.18, 95% CI ¼
1.59 to 2.98), and lung cancer (HR ¼ 1.28, 95% CI ¼ 1.08 to
1.52)
(continued)

REVIEW
REVIEW

Dawes et al., 2014 96 220 women and California Cancer Continuous Medicaid Receipt of definitive op- 59.5% of patients with Medicaid discontinuously enrolled
(44) men aged <65 y Registry, California’s enrollment (>6 mo eration status from Compared with those continuously enrolled in Medicaid,
newly diagnosed Patient Discharge prediagnosis) vs dis- hospital discharge, discontinuously enrolled patients less likely to receive de-
with colon, esoph- Database, and continuous Medicaid death within 1 y of finitive surgery for colon (70.5% vs 61.9%, P < .001), lung
ageal, lung, ovar- Medicaid enrollment enrollment (did not diagnosis (17.7% vs 15.3%, P ¼ .012), and gastric (37.2% vs 31.0%, P ¼
ian, pancreatic, or 2002-2008 have 6 consecutive Vital status from .015) cancers. Adjusted models not reported for esopha-
gastric cancers mo coverage registry gus, ovary, and pancreas cancers. Statistically significant
prediagnosis) 1-y mortality benefit in patients with continuous (vs dis-
continuous) Medicaid coverage in 3 cancer types: colon
(23.0% vs 19.1%, P ¼ .001), lung (66.7 vs 62.4%, P ¼ .002),
and gastric (57.8 vs 49.0%, P ¼ .001) cancers. No statisti-

cally significant mortality differences were observed for
esophagus (66.6% vs 70.4%, P ¼ .33), ovary (21.5% vs 22.1%
P ¼ .80), and pancreas (76.1% vs 75.7%, P ¼ .87) cancers in
adjusted analyses
Doll et al., 2016 (45) 782 women aged <65 North Carolina Central Timing of coverage All-cause mortality (me- 38.6% enrolled in Medicaid after diagnosis
y newly diagnosed Cancer Registry - measured as dian follow-up ¼ 22 Lack of prediagnosis Medicaid coverage had mortality HR of
with cancers of Medicaid enrollment, Medicaid enrollment mo) 1.28 (95% CI ¼ 0.99 to 1.65) when stage not included and
cervix, uterus, 2003-2008. Follow-up before diagnosis (1 Vital status from mortality HR of 1.19 (95% CI ¼ 0.92 to 1.53) when adjusted
ovary, and vulva/ through 2010 mo of coverage in 6 registry for stage
vagina mo before diagnosis)
vs Medicaid enroll-
ment after diagnosis
Koroukian et al., 12 703 women and Ohio Cancer Registry; Timing of coverage Survival and 5-y 43.9% of Medicaid enrollees peri-/postdiagnosis.
2012 (35) men aged 15-54 y Ohio Medicaid measured as mortality Adjusted AORs for 5-y disease-specific mortality were 1.58
newly diagnosed Enrollment data, Medicaid prediagno- Vital status from for prediagnosis Medicaid enrollees (95% CI ¼ 1.25 to 1.99)
with bladder, co- 1996-2002. Follow-up sis (3 mo coverage registry and 2.43 for peri-/postdiagnosis Medicaid enrollees (95%
lon, lung, and tes- through 2007 before diagnosis) vs CI ¼ 1.94 to 3.04) compared with non-Medicaid population
ticular cancers, Medicaid peri-/post- with either private coverage or uninsured (for both, P <
melanoma, pedi- diagnosis (enrollment .001). Effects of coverage disruptions not reported by can-
atric malignan- during <3-mo win- cer site
cies, or Hodgkin dow before or after
diagnosis)
(continued)


and non-Hodgkin
lymphoma
Koroukian et al., 26 426 women aged Ohio Cancer Registry, Timing of coverage Overall and cancer-spe- 31.6% enrolled peri-/postdiagnosis.
2017 (34) 40-64 y newly di- Medicaid enrollment measured as cific survival Peridiagnosis Medicaid worse overall and cancer-specific
agnosed with in- data, and BCCEDP Medicaid status (en- Vital status from survival compared with non-Medicaid (P < .05).
vasive BC data. Patients diag- rolled 3 mo before registry Peridiagnosis Medicaid similar overall (HR ¼ 0.87, 95% CI
nosed 2002-2008; diagnosis or peridiag- ¼ 0.74 to 1.02) and cancer-specific (HR ¼ 0.92, 95% CI ¼
deaths through 2010 nosis, [enrolled at di- 0.76 to 1.11) survival compared with prediagnosis
agnosis or <3 mo of Medicaid.
diagnosis]), BCCEDP Compared with Medicaid prediagnosis, BCCEDP 1 time and
repeat user repeat users better overall (AHR ¼ 0.60, 95% CI ¼ 0.45 to
0.80 and AHR ¼ 0.27, 95% CI ¼ 0.13 to 0.54, respectively)
and cancer-specific survival (AHR ¼ 0.77, 95% CI ¼ 0.57 to
1.06 and AHR ¼ 0.36, 95% CI ¼ 0.16 to 0.80, respectively)
Subramanian and Adults aged 18-64 y California and Georgia Continuous Medicaid Receipt of first course of 9.3% and 31.5% patients alive 12 mo after diagnosis not con-
Chen, 2013 (53) from California (N State Cancer Registry from diagnosis to treatment (surgery, tinuously enrolled in Medicaid after diagnosis in
¼ 691) and Georgia - California and 12 mo postdiagnosis: radiation, and chemo- California and Georgia, respectively
(N ¼ 225) newly di- Georgia Medicaid vs not continuously therapy) from cancer Compared with continuously enrolled, no statistically sig-
agnosed with Claims, 2002-2006. enrolled (gaps in cov- diagnosis to end of nificant differences for receipt of surgery (OR ¼ 1.05, 95%
head and neck Follow-up time not erage of 2 mo) vs follow-up from claims CI ¼ 0.72 to 1.56), radiation (OR ¼ 0.93, 95% CI ¼ 0.63 to
cancer and alive stated continuously 1.37), and chemotherapy (OR ¼ 1.49, 95% CI ¼ 1.00 to 2.27)
12 mo after Medicaid enrolled (no with those not continuously insured
diagnosis gaps in coverage or
gaps in coverage for
<2 mo)
Tsui et al., 2018 (51) 19 209 women newly New Jersey Cancer Timing of coverage Treatment delay (>90 d 25.3% with Medicaid newly enrolled at diagnosis
diagnosed with Registry and New measured as longer after diagnosis) from Newly enrolled patients higher likelihood of treatment de-
BC, CRC, or cervi- Jersey Medicaid term/established claims, and 2-y lay for BC (OR ¼ 8.79, 95% CI ¼ 5.91 to 13.10), cervical can-
cal cancer Management Medicaid patients (en- survival cer (OR ¼ 2.47, 95% CI ¼ 1.00 to 6.15), and CRC (OR ¼ 3.02,
Information System, rolled 6 mo at diag- Vital status from 95% CI ¼ 1.94 to 4.71) cancers
2012-2014; survival nosis), newly enrolled registry Shorter enrollment time associated with treatment delay
through 2016 Medicaid patients (<6 (76.3%, 41.6%, 46.5%, and 51.9 for <1 mo, 1 to <6 mo, 6 to
mo). Duration of cov- <11 mo, and 11 mo, respectively; P < .001)
erage measured as No statistically significant differences in treatment delay
length of Medicaid between continuously covered and coverage disruptions
enrollment (11, 6 to (46.9% vs 51.3%, P ¼ .45)
<11, 1 to <6, <1 mo) Newly enrolled in Medicaid had lowest 2-y survival com-
vs continuous enroll- pared with established Medicaid and patients without
ment status (no gaps Medicaid coverage (private insurance and uninsured;
>30 d in prior year). P < .001)
a
BC ¼ breast cancer; BCCEDP ¼ Breast and Cervical Cancer Early Detection Program; CI ¼ confidence interval; CRC ¼ colorectal cancer; HR ¼ hazard ratio; OR ¼ odds ratio; AOR ¼ adjusted odds ratio.

REVIEW
REVIEW
Table 5. Health insurance coverage disruptions and other health services use and health-care spendinga

Koroukian et al., 44 509 decedents with Ohio Medicaid Duration of coverage mea- Per person per month 2.5%, 12.7%, 7.8%, 5.3%, 4.7%, and 67.0% of decedents en-
2006 (36) cancer as underlying Enrollment and sured as Medicaid enroll- enrolled total medical rolled in Medicaid at month of death, 1-3, 4-6, 7-9, 10-12,
cause of death (no age claims data: death ment months prior death expenditures from or >12 mo prior death, respectively.
restriction) certificate, 1992-2002 (enrolled at month of claims Overall, longer time of Medicaid enrollment was associated
death, or 1-3, 4-6, 7-9, 10- with higher monthly total expenditures. Monthly expen-
12, or >12 mo prior ditures were $770, $1105, $1674, $1941, $1987, and $1905
death) for those enrolled at month of death, 1-3, 4-6, 7-9, 10-12,
or >12 mo before death, respectively. Year of dollars not
stated
Mack et al., 2013 (54) 4797 California patients California and New York Medicaid enrollment be- Hospice use from claims 69%, 21%, and 10% patients continuously enrolled in
and 4001 New York State Cancer Registry tween month of diagno- Medicaid, enrolled >50% of time but not continuously, en-
patients aged 21-64 y - California and New sis and month of death or rolled <50% of time in California, respectively; 64%, 24%,
with newly diagnosed York Medicaid censoring (continuous, and 12% patients continuously enrolled in Medicaid, en-
stage IV lung cancer Enrollment, 2002- enrolled more than 50% rolled >50% of time but not continuously, enrolled <50%
2006. Follow-up of the time but not con- of time in New York, respectively.
through 2017 tinuously, enrolled <50% In both states, compared with continuously enrolled,
of time) patients enrolled >50% of time but not continuously (OR
¼ 0.82, 95% CI ¼ 0.69 to 0.98) or enrolled <50% of time (OR
¼ 0.45, 95% CI ¼ 0.35 to 0.57) had lower hospice use
Mack et al., 2015 (55) 705 decedents previ- New York State Cancer Timing of coverage mea- 1) Hospice use; 2) EOL 15.4% of patients enrolled in Medicaid at or after cancer di-
ously diagnosed with Registry - Medicaid sured as 1) enrolled in intensity measured agnosis. 65.5%, 28.5%, and 6.0% patients continuously en-
cancer between ages Enrollment, 2004- Medicaid before diagno- by chemotherapy use rolled in Medicaid, enrolled >50% of time but not
of 15 and 29 y 2011. Deaths by Dec. sis or around time of di- within 14 d of death, continuously, enrolled <50% of time, respectively.
31, 2011 agnosis (duration of care in ICU within 30 Compared with patients enrolled before cancer diagnosis,
enrollment before diag- d of death, more than those enrolled at or after diagnosis had lower hospice use
nosis not stated); 2) 1 ER visit within 30 d (OR ¼ 0.26, 95% CI ¼ 0.08 to 0.83). Compared with continu-
Medicaid enrollment be- of death, hospitaliza- ously enrolled, patients who enrolled >50% of time but
tween month of diagno- tion within 30 d of not continuously (OR ¼ 0.36, 95% CI ¼ 0.23 to 0.56) or en-
sis and month of death death from claims rolled <50% of time (OR ¼ 0.22, 95% CI ¼ 0.10 to 0.51) were
(continuous, enrolled less likely to have intensive EOL care
>50% of the time but not
continuously, enrolled
<50% of time)
a
CI ¼ confidence interval; EOL ¼ end-of-life; ER ¼ emergency room; ICU ¼ intensive care unit; OR ¼ odds ratio.

United States. We identified 29 observational published studies We did not identify any published studies of effective inter-
and found that coverage disruptions were common and despite ventions at the patient, provider, employer, health system, pol-
heterogeneity in populations and measures, disruptions were icy, or regulatory levels that help patients maintain continuous
consistently statistically significantly associated with less fre- health insurance coverage throughout the cancer control con-
quent receipt of cancer care and poorer cancer outcomes. tinuum. Several studies testing interventions to improve cover-
Specifically, those with coverage disruptions were less likely to age continuity are in progress (60,61), however. For example,
receive cancer prevention or screening (27–29,32,37,39), and if within federally qualified health centers in multiple states, the
diagnosed with cancer, they were more likely to have advanced introduction of electronic health record tools for identifying
disease (33,34,40,41,43–51), be less likely to receive treatment patients in advance of their Medicaid recertification to ensure
(44,52,54,55), and have worse survival (34,35,40,42,44,45,51) than continuity of coverage is being evaluated for potential improve-
their counterparts without coverage disruptions. Findings were ments in receipt of cancer prevention and screening (60).
consistent across multiple cancer sites. Additionally, several Another ongoing intervention study involves a multi-employer
studies reported a dose-response relationship between cover- Taft-Hartley Trust Fund that provides health benefits to hourly,
age duration and receipt of prevention (39), screening (28,32), low-wage employees who would otherwise not have health in-
earlier stage (51), and timeliness of treatment (51). The consis- surance coverage. This ongoing study is testing provision of con-

tency of these findings across the cancer control continuum tinuous health benefits coverage for low-wage employees after a
highlight the importance of minimizing health insurance cover- cancer diagnosis and incentives for the use of high-quality cancer
age disruptions in addressing cancer disparities and promoting care providers through travel benefits and narrow networks (61).
health equity. Further development and testing of interventions to improve cov-
Policies and trends that exacerbate coverage disruptions may erage continuity in public and private settings using experimental
increase disparities. The recent emergence of work requirements designs, especially for patients at risk of coverage disruptions,
for some state Medicaid programs might increase the prevalence will be important.
of coverage disruptions. Research conducted in Arkansas sug- Most studies in this review were conducted in single states
gests that Medicaid work requirements are associated with disen- and used cancer registry–Medicaid enrollment and claims link-
rollment of eligible residents without increases in employment ages to evaluate the effects of coverage disruptions. Although
(56). Broader employment trends, such as the increased preva- the prevalence of coverage disruptions varied, among those
lence of “gig” workers (eg, Uber, Lyft, TaskRabbit) (57) and associ- with Medicaid coverage, up to 59.5% of newly diagnosed
ated income fluctuations, may increase disruptions in coverage patients enrolled only at or after their cancer diagnosis (42,50).
among the nongroup self-insured facing frequent changes in eli- Findings that peri- and postdiagnosis enrollment were consis-
gibility for subsidies and coverage affordability. Despite the im- tently associated with worse outcomes suggest that at least
portance of private health insurance coverage for the working some of the worse survival associated with Medicaid compared
age population in the United States, we did not identify any stud- with private insurance coverage observed elsewhere (4,7) may
REVIEW
ies specifically addressing the effects of private health insurance be attributable to the impact of coverage disruptions.
coverage disruptions on cancer care and outcomes. Preliminary Coverage disruptions were consistently associated with ad-
research suggests that for cancer survivors, the magnitude of as- vanced stage and worse survival for cancers with effective
sociation between disruptions in private health insurance and screening tests—namely, breast, cervical, and colorectal cancers
worse access to and receipt of care is similar to that for disrup- (33,34,40–43,46–48,50,51). Two studies included in this review
tions in public insurance (58); additional research is warranted in evaluated the effects of breast and cervical cancer screening
both public and private coverage settings. through BCCEDP before Medicaid enrollment in 2 states (34,52),
Conversely, policies that facilitate health insurance coverage but they could not fully disentangle the potentially positive
continuity may minimize disparities. Increased availability of effects of screen detection vs the potentially negative effects of
health insurance coverage options through the ACA, including limited access to usual source of care or symptom evaluation
individual purchase through the ACA Marketplace, availability associated with lack of insurance coverage before diagnosis. A
of subsidies to reduce premium costs, expansion of dependent large body of research has consistently found that in addition to
coverage on parents’ private plans for young adults until age 26 health insurance coverage, having a usual source of health care
years, and Medicaid eligibility in some states, might minimize is strongly associated with receipt of breast and cervical cancer
disruptions and facilitate continuous health insurance cover- screening (5), but neither study reported this information.
age. State policies, such as the proposed New York Medicaid Medicaid patients with coverage disruptions have also been
waiver to allow prisoners with health conditions to receive reported to be more likely to disenroll in the year after their can-
Medicaid coverage before and following their release from jail, cer diagnosis (50), which may adversely affect completion of
may also help minimize disruptions and maintain care and pro- recommended treatment(s) and access to and receipt of high-
vider network continuity. To date, most research evaluating the quality survivorship care. Care received after Medicaid disen-
effects of the ACA has focused on coverage gains, and none rollment, including assessment for recurrence, surveillance for
assessed the effects of the ACA on reducing the prevalence or new cancers, symptom management, or end-of-life care, is un-
frequency of coverage disruptions and effects on cancer care known. Among adults with private insurance coverage, a cancer
and outcomes. A recent study found Medicaid expansion was diagnosis and its treatment(s) can lead to time away from work,
associated with reductions in coverage disruptions in the low- job loss, and loss of employer-sponsored health insurance cov-
income general population living in expansion compared with erage. The longitudinal effects of coverage disruptions across
nonexpansion states (59). Given the rapidly changing health in- the cancer control continuum are not easily addressed with cur-
surance landscape in the United States and the maturation of rent data infrastructure, however.
data post-ACA, evaluating the effects of specific provisions of Improvements in cancer registry and health insurance data
the ACA, especially Medicaid expansions, on coverage disrup- infrastructure resulting in comprehensive longitudinal data can
tions and health outcomes will be important for future research help to quantify the effects of coverage disruptions and differ-
using quasi-experimental designs. ences in state-level policies, such as Medicaid generosity,
physician reimbursement, managed care penetration, and tim- Nonetheless, we were able to qualitatively synthesize a large
ing of Medicaid eligibility recertification, on receipt of cancer body of research and identify research gaps and opportunities
care and outcomes. Current individual state-level linkages, for data infrastructure improvements.
such as the ones identified in this review, cannot be used to In summary, we found that health insurance coverage dis-
evaluate the effects of differences between states in state-level ruptions were consistently adversely associated with receipt of
policies. Centralized data linkages across multiple states, such cancer prevention and screening and among those diagnosed
as SEER-Medicaid, could transform the ability to evaluate with cancer, later stage of disease, delayed treatment if any,
Medicaid policies, especially as the updated version of the na- and poorer survival. Future research identifying modifiable fac-
tional Medicaid data, Transformed Medicaid Statistical tors at the patient, employer, state, and federal policy levels to
Information System, becomes available (62). Currently, more minimize coverage disruptions may also reduce cancer dispar-
than 15 states have legislation requiring aggregation of all- ities. Improved data infrastructure and quasi-experimental and
payer claims data (APCD); as these data become increasingly experimental study designs will be important for evaluating the
available and research-ready, state-level cancer registry–APCD associations of federal and state policies on coverage disrup-
linkages may be especially useful for evaluating the effects of tions and care and outcomes.
private coverage disruptions and cancer outcomes and poten-

tially, differences in state policies. As APCD data become stan-
dardized across states, centralized cancer registry data linkages Funding
across multiple states with APCD could further these efforts.
No funding was provided for this research.
We identified many limitations in existing studies, including
inconsistency of coverage disruption measures and lack of in-
formation about prior coverage (private or uninsured) or rea-
sons for prior uninsurance or coverage change among those
Notes
newly enrolled in Medicaid at or after cancer diagnosis. Because Disclosures: The authors do not report any conflicts of interest.
previous private insurance coverage would be expected to con-
vey better access to care, any misclassification of previously Acknowledgments: The authors acknowledge feedback on early
uninsured patients suggests that the effect of gaining coverage versions of this work from the Health Equity Committee of the
only after diagnosis is understated. Similarly, some studies American Society of Clinical Oncology.
compared patients gaining Medicaid coverage at diagnosis to all
non-Medicaid-covered patients, including privately insured and References
uninsured, which would underestimate the magnitude of asso-
1. Institute of Medicine. Delivering High-Quality Cancer Care: Charting a New Course
ciations between disruptions and more advanced stage or sur- for a System in Crisis. Washington, DC: The National Academies Press; 2013.
vival following diagnosis. Generalizability of findings of the 2. Institute of Medicine. America’s Uninsured Crisis: Consequences for Health and
REVIEW
adverse effects of coverage disruptions in populations of Health Care. Washington, DC: National Academies of Medicine; 2009.
3. Ellis L, Canchola AJ, Spiegel D, Ladabaum U, Haile R, Gomez SL. Trends in can-
Medicaid enrollees to populations with private health insurance cer survival by health insurance status in California From 1997 to 2014. JAMA
coverage may be limited. Oncol. 2018;4(3):317–323.
All studies included in this systematic review were observa- 4. Walker GV, Grant SR, Guadagnolo BA, et al. Disparities in stage at diagnosis,
treatment, and survival in nonelderly adult patients with cancer according to
tional, and we cannot infer causality between health insurance insurance status. J Clin Oncol. 2014;32(28):3118–3125.
coverage disruptions and worse care receipt and outcomes. 5. Mandelblatt JS, Yabroff KR, Kerner JF. Equitable access to cancer services: a
However, all studies included comparison groups and despite review of barriers to quality care. Cancer. 1999;86(11):2378–2390.
6. Halpern MT, Ward EM, Pavluck AL, Schrag NM, Bian J, Chen AY. Association
the heterogeneity of measures of coverage disruptions, findings
of insurance status and ethnicity with cancer stage at diagnosis for 12 cancer
were consistent across studies and the magnitude of many sites: a retrospective analysis. Lancet Oncol. 2008;9(3):222–231.
associations was large. Nonetheless, studies with experimental 7. Ward E, Halpern M, Schrag N, et al. Association of insurance with cancer care
utilization and outcomes. CA Cancer J Clin. 2008;58(1):9–31.
designs are needed to infer causality of coverage disruptions
8. Sommers BD, Long SK, Baicker K. Changes in mortality after Massachusetts
and intervention effects to minimize disruptions and receipt of health care reform. Ann Intern Med. 2015;162(9):668–669.
cancer care and health outcomes. In addition, evaluation of 9. Sommers BD, Gunja MZ, Finegold K, Musco T. Changes in self-reported insur-
trends and comparisons of findings from specific states at dif- ance coverage, access to care, and health under the Affordable Care Act.
JAMA. 2015;314(4):366–374.
ferent time points are limited by differences in state-level poli- 10. Sommers BD, Gawande AA, Baicker K. Health insurance coverage and health.
cies (eg, Medicaid eligibility threshold, recertification timing), N Engl J Med. 2017;377(20):2000–2001.
socioeconomic characteristics, cancer sites, and heterogeneity 11. Sommers BD, Baicker K, Epstein AM. Mortality and access to care among
adults after state Medicaid expansions. N Engl J Med. 2012;367(11):1025–1034.
of definitions of coverage disruptions. 12. Goldman AL, McCormick D, Haas JS, Sommers BD. Effects of the ACA’s health
An important strength of this systematic review is that we insurance marketplaces on the previously uninsured: a quasi-experimental
used multiple scientific publication databases to identify analysis. Health Aff. 2018;37(4):591–599.
13. Wallace J, Sommers BD. Health insurance effects on preventive care and
articles in our efforts to comprehensively assess the effects of health: a methodologic review. Am J Prev Med. 2016;50(5):S27–S33.
health insurance coverage disruptions on care and cancer out- 14. Uberoi NF, Gee E. Health Insurance coverage and the Affordable Care Act
comes. We included many years of published studies in our 2010-2016. ASPE Issue Brief. https://collections.nlm.nih.gov/master/borndig/
101701796/ACA2010-2016.pdf Accessed October 11, 2018.
search, starting in 1980. We also hand-searched the reference
15. Sommers BD, Gourevitch R, Maylone B, Blendon RJ, Epstein AM. Insurance
lists of each included article for any additional published stud- churning rates for low-income adults under health reform: lower than
ies that were not initially captured in the electronic search pro- expected but still harmful for many. Health Aff. 2016;35(10):1816–1824.
16. Vistnes JP, Cohen JW. Duration of uninsured spells for nonelderly adults de-
cess. It is possible we missed some relevant studies, however.
clined after 2014. Health Aff. 2018;37(6):951–955.
In addition, although we used standardized data abstraction 17. Short PF, Graefe DR, Swartz K, Uberoi N. New estimates of gaps and transi-
measures, there still may be some unavoidable subjectivity in tions in health insurance. Med Care Res Rev. 2012;69(6):721–736.
some of the measures we abstracted and reported. Finally, be- 18. Guevara JP, Moon J, Hines EM, et al. Continuity of public insurance coverage:
a systematic review of the literature. Med Care Res Rev. 2014;71(2):115–137.
cause of the heterogeneity of measures, study populations, and 19. Yabroff KR, Kirby J, Zodet M. Association of insurance gains and losses with
cancer sites, we could not conduct quantitative meta-analyses. access to prescription drugs. JAMA Intern Med. 2017;177(10):1531–1532.
20. Villarroel MA, Cohen RA. Health insurance continuity and health care access 43. Bradley CJ, Given CW, Roberts C. Health care disparities and cervical cancer.
and utilization. NCHS Data Brief. 2014;2016(249):1–8. Am J Public Health. 2004;94(12):2098–2103.
21. Ginde AA, Lowe RA, Wiler JL. Health insurance status change and emergency 44. Dawes AJ, Louie R, Nguyen DK, et al. The impact of continuous Medicaid en-
department use among US adults. Arch Intern Med. 2012;172(8):642–647. rollment on diagnosis, treatment, and survival in six surgical cancers. Health
22. Roberts ET, Pollack CE. Does churning in Medicaid affect health care use? Med Serv Res. 2014;49(6):1787–1811.
Care. 2016;54(5):483–489. 45. Doll KM, Basch EM, Meng K, et al. Clinical benefits associated with Medicaid
23. Davidoff AJ, Guy GP Jr, Hu X, et al. Changes in health insurance coverage associ- coverage before diagnosis of gynecologic cancers. J Oncol Pract. 2016;12(6):
ated with the Affordable Care Act among adults with and without a cancer his- e724–733.
tory: population-based national estimates. Med Care. 2018;56(3):220–227. 46. Keegan THM, Parsons HM, Chen Y, et al. Impact of health insurance on stage
24. Jemal A, Lin CC, Davidoff AJ, Han X. Changes in insurance coverage and stage at cancer diagnosis among adolescents and young adults. J Natl Cancer Inst.
at diagnosis among nonelderly patients with cancer after the Affordable Care 2019;111(11):1152–1160.
Act. J Clin Oncol. 2017;35(35):3906–3915. 47. O’Malley CD, Shema SJ, Clarke LS, Clarke CA, Perkins CI. Medicaid status and
25. Robin Yabroff K, Short PF, Machlin S, et al. Access to preventive health care stage at diagnosis of cervical cancer. Am J Public Health. 2006;96(12):2179–2185.
for cancer survivors. Am J Prev Med. 2013;45(3):304–312. 48. Perkins CI, Wright WE, Allen M, Samuels SJ, Romano PS. Breast cancer stage
26. Han X, Yabroff KR, Ward E, Brawley OW, Jemal A. Changes in percent unin- at diagnosis in relation to duration of Medicaid enrollment. Med Care. 2001;
sured among cancer patients by state and sociodemographic factors follow- 39(11):1224–1233.
ing the Affordable Care Act. JAMA Oncol. 2018;4(12):1713. 49. Pollitt RA, Clarke CA, Shema SJ, Swetter SM. California Medicaid enrollment
27. Ayanian JZ, Weissman JS, Schneider EC, Ginsburg JA, Zaslavsky AM. Unmet and melanoma stage at diagnosis: a population-based study. Am J Prev Med.
health needs of uninsured adults in the United States. JAMA. 2000;284(16): 2008;35(1):7–13.
2061–2069. 50. Ramsey SD, Zeliadt SB, Richardson LC, et al. Disenrollment from Medicaid af-

28. Bednarek HL, Schone BS. Variation in preventive service use among the inter recent cancer diagnosis. Med Care. 2008;46(1):49–57.
sured and uninsured: does length of time without coverage matter? J Health 51. Tsui J, DeLia D, Stroup AM, et al. Association of Medicaid enrollee charac-
Care Poor Underserved. 2003;14(3):403–419. teristics and primary care utilization with cancer outcomes for the period
29. Broyles RW, Narine L, Brandt EN Jr. The temporarily and chronically unin- spanning Medicaid expansion in New Jersey. Cancer. 2019;125(8):
sured: does their use of primary care differ. J Health Care Poor Underserved. 1330–1340.
2002;13(1):95–111. 52. Adams EK, Chien LN, Gabram-Mendola SG. Treatment patterns among
30. Freund KM, Reisinger SA, LeClair AM, et al. Insurance stability and cancer Medicaid-eligible women with breast cancer in Georgia: are patterns differ-
screening behaviors. Health Equity. 2019;3(1):177–182. ent under the breast and cervical cancer prevention and treatment act?
31. Jerant A, Fiscella K, Tancredi DJ, Franks P. Health insurance is associated J Oncol Pract. 2012;8(1):46–52.
with preventive care but not personal health behaviors. J Am Board Fam Med. 53. Subramanian S, Chen A. Treatment patterns and survival among low-
2013;26(6):759–767. income Medicaid patients with head and neck cancer. JAMA Otolaryngol Head
32. Koroukian SM. Screening mammography was used more, and more fre- Neck Surg. 2013;139(5):489–495.
quently, by longer than shorter term Medicaid enrollees. J Clin Epidemiol. 54. Mack JW, Chen K, Boscoe FP, et al. Underuse of hospice care by Medicaid-
2004;57(8):824–831. insured patients with stage IV lung cancer in New York and California. J Clin
33. Koroukian SM. Assessing the effectiveness of Medicaid in breast and cervical Oncol. 2013;31(20):2569–2579.
cancer prevention. J Public Health Manag Pract. 2003;9(4):306–314. 55. Mack JW, Chen K, Boscoe FP, et al. High intensity of end-of-life care among
34. Koroukian SM, Bakaki PM, Htoo PT, et al. The Breast and Cervical Cancer adolescent and young adult cancer patients in the New York State Medicaid
Early Detection Program, Medicaid, and breast cancer outcomes among Program. Med Care. 2015;53(12):1018–1026.
Ohio’s underserved women. Cancer. 2017;123(16):3097–3106. 56. Sommers BD, Goldman AL, Blendon RJ, Orav EJ, Epstein AM. Medicaid work
35. Koroukian SM, Bakaki PM, Raghavan D. Survival disparities by Medicaid sta- requirements - results from the first year in Arkansas. N Engl J Med. 2019;
tus: an analysis of 8 cancers. Cancer. 2012;118(17):4271–4279. 381(11):1073–1082.
36. Koroukian SM, Beaird H, Madigan E, Diaz M. End-of-life expenditures by Ohio 57. Katz LK. The Rise and Nature of Alternative Work Arrangements in the United
REVIEW
Medicaid beneficiaries dying of cancer. Health Care Financ Rev. 2006;28(2): States, 1995-2015. National Bureau of Economic Research - NBER Working
65–80. Paper. 2016; 22667.
37. McWilliams JM, Zaslavsky AM, Meara E, Ayanian JZ. Impact of Medicare cov- 58. Zhao J HX, Nogueira L, Jemal A, Yabroff KR. The association between health
erage on basic clinical services for previously uninsured adults. JAMA. 2003; insurance coverage disruptions and access to care and affordability among
290(6):757–764. cancer survivors in the United States. http://abstractsascoorg/244/AbstView_
38. O’Leary MC, Lich KH, Gu Y, et al. Colorectal cancer screening in newly insured 244_274497html. Accessed September 28, 2019.
Medicaid members: a review of concurrent federal and state policies. BMC 59. Goldman AL, Sommers BD. Among low-income adults enrolled in Medicaid,
Health Serv Res. 2019;19(1):298. churning decreased after The Affordable Care Act. Health Aff. 2020;39(1):
39. Staras SA, Vadaparampil ST, Haderxhanaj LT, Shenkman EA. Disparities in 85–93.
human papillomavirus vaccine series initiation among adolescent girls en- 60. Angier H, Huguet N, Marino M, et al. Assessing Community Cancer care after
rolled in Florida Medicaid programs, 2006-2008. J Adolesc Health. 2010;47(4): insurance ExpanSionS (ACCESS) study protocol. Contem Clin Trials Commun.
381–388. 2017;7:136–140.
40. Bradley CJ, Given CW, Roberts C. Correlates of late stage breast cancer and 61. Patel M, Andrea N, Jay B, Coker TR. A community-partnered, evidence-based
death in a Medicaid-insured population. J Health Care Poor Underserved. 2003; approach to improving cancer care delivery for low-income and minority
14(4):503–515. patients with cancer. J Community Health. 2019;44(5):912–920.
41. Bradley CJ, Given CW, Roberts C. Late stage cancers in a Medicaid-insured 62. Centers for Medicare and Medicaid Services. Transformed Medicaid
population. Med Care. 2003;41(6):722–728. Statistical Information System (T-MSIS). https://wwwmedicaidgov/medic-
42. Bradley CJ, Gardiner J, Given CW, Roberts C. Cancer, Medicaid enrollment, aid/data-and-systems/macbis/tmsis/indexhtml. Accessed September 28,
and survival disparities. Cancer. 2005;103(8):1712–1718. 2019.

Djaa 048

Uploaded by

Copyright:

Available Formats

Djaa 048

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Djaa 048

Uploaded by

Copyright:

Available Formats

JNCI J Natl Cancer Inst (2020) 112(7): djaa048

Health Insurance Coverage Disruptions and Cancer Care and

Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

Table 1. Study, patient, and insurance coverage characteristics Table 1. (continued)

No. of studies (%) No. of studies (%)

Study characteristics Duration of coverage 12 (41.4)

Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

Table 2. Health insurance coverage disruptions and cancer screeninga

Sample size, Insurance coverage

tistically significant differ

Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

Sample size, Insurance coverage

and continuously 2 y among eligible

Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

patients could have been previously enrolled in private health

Cancer Treatment and Survival

Setting, data source, and Insurance coverage

Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

Setting, data source, and Insurance coverage

Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

Setting, data source, and Insurance coverage

Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

Setting, data source, and Insurance coverage

No statistically significant differences in

Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

Setting, data source, and Insurance coverage

Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

Setting, data source, and Insurance coverage

and gastric (57.8 vs 49.0%, P ¼ .001) cancers. No statisti-

Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

Setting, data source, and Insurance coverage

Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

Setting, data source, and Insurance coverage

Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

Downloaded from https://academic.oup.com/jnci/article/112/7/671/5821430 by guest on 12 April 2022

You might also like