Disparities in Breast and Cervical CancerScreening in Women with Mental Illness

A Systematic Literature Review

Arpita Aggarwal, MD, MSc, Ananda Pandurangi, MD, Wally Smith, MD

Context: Breast and cervical cancer screening rates have improved substantially in the U.S. duringthe past decade. Cancer screening and other health outcomes in patients with mental illnesses, suchas major depression and schizophrenia, remain suboptimal. Understanding the prevalence and rootcauses of these disparities is an essential fırst step toward developing effective interventions. Thispaper presents a systematic literature review of current evidence on breast and cervical cancerscreening disparities in women with mental illness.

Evidence acquisition: A systematic PubMed/MEDLINE and PsycINFO search completed inMay2012 retrieved articles pertaining to cancer screening andmentally ill patients using pertinent searchterms. Articles that met the inclusion criteria were appraised critically for evidence quality related toscreening disparities using defıned criteria. Articles that reported cancer screening rates in patientswith mental illness were reviewed to determine whether any barriers to screening or factors thatpromote screening were identifıed.

Evidence synthesis: Nineteen studies met the inclusion criteria. Many articles contributed tomore than one of the identifıed areas of interest (i.e., screening utilization, barriers to screening, andfactors that encourage screening).

Conclusions: Substantial evidence in the current literature confırms disparities in breast andcervical cancer screening rates among women with mental illness. However, the mentally ill popu-lation is more complex and diverse thanmany studies imply. Using a global functional indicator thatmeasures the overall impact of mental illness may yield amore useful categorization of influences oncancer screening.(Am J Prev Med 2013;44(4):392–398) © 2013 Published by Elsevier Inc. on behalf of American Journal ofPreventive Medicine

Context

In 2012, there will be an estimated 241,230 new casesand 44,140 deaths attributed to breast and cervicalcancer.1 To combat this morbidity andmortality, the

current national guidelines for breast and cervical cancerscreening in women aim to detect cancer at an earlystage.2–4 In 2010, the screening mammography rateswithin the prior 2 years for women aged �40 yearsranged from 49.1% to 78.2%. In that same period, 54.1%–77.8% of women aged �18 years reported having a Papfor cervical cancer screening in the prior 3 years.5,6

From the Division of General Internal Medicine (Aggarwal, Smith), theDepartment of Psychiatry (Pandurangi), Virginia Commonwealth Univer-sity, Richmond, Virginia

Address correspondence to: Arpita Aggarwal, MD, MSc, Division ofGeneral Internal Medicine, Virginia Commonwealth University, 730 EastBroad Street, Room 435, PO Box 980102, Richmond VA 23298. E-mail:aaggarwal@mcvh-vcu.edu.

p0749-3797/$36.00http://dx.doi.org/10.1016/j.amepre.2012.12.006

392 Am J PrevMed 2013;44(4):392–398 ©2013Publishe

Screening rates have improved substantially in the U.S.during the past decade.5,6 However, cancer screening inmedically underserved, minority, and lower-SES popula-tions remains suboptimal.7–9 A number of specifıc inter-ventions have therefore targeted these women.10–12

When cancer screening is considered, people withmental illnesses occasionally are overlooked.Majormen-tal illnesses include medical conditions such as majordepression, schizophrenia, bipolar disorder, obsessivecompulsive disorder, panic disorder, and post-traumaticstress disorder. Because of increasing evidence of dispar-ities in disease-specifıc morbidity and mortality in men-tally ill patients, there is a growing need to understand theimpact of mental illness as a barrier to cancer screen-ing.13–15 The burden of mental illness in the U.S. andother developed countries is high, and an estimated26.2% of individuals aged �18 years suffer from a diag-nosable mental disorder.16 The annual prevalence of bi-

olar disorder, post-traumatic stress disorder, and

dbyElsevier Inc. on behalf ofAmerican Journal of PreventiveMedicine

intbm

wmpmmchasmms

mcooFptstp

ctrss

illne

Aggarwal et al / Am J Prev Med 2013;44(4):392–398 393

A

schizophrenia are estimated at 2.6%, 3.5%, and 1.1%,respectively.16

Individuals with mental illness have a higher medi-cal disease burden, higher use of emergency services,longer hospital stays, and shorter life spans than indi-viduals with no mental illness.17–23 Studies in mentallyll individuals with diabetes, hypertension, and coro-ary artery disease reveal poor adherence to medica-ion regimens, greater complications, increased num-ers of emergency room visits, and poorer physical andental functioning.14,24–26

The current authors developed a conceptual frame-work (Figure 1) that incorporates social determinants ofhealth disparities, as theorized by Freeman and Chu,59 asell as Katon’s15 theory about the relationship betweenental illness and health behaviors. The framework hy-othesizes a complex, bidirectional relationship betweenental illness and medical outcomes, suggesting that de-ographic factors (gender, race/ethnicity, SES, and edu-ation); social determinants (social support, neighbor-ood and household stability, and childhood exposure tobuse and neglect); and genetic factors independently orynergistically may increase susceptibility to not onlyental illness but also poor health behaviors. In turn,ental illness or poor health behaviors independently orynergistically may lead to worse cancer outcomes.Factors outside the patient that complicate the assess-ent of patient factors related to screening are not in-luded in this model. There is variation in the prevalencef mental illness, with overlapping diagnostic categoriesf mental illness and poor or under-reported diagnoses.urther, health disparities in the mentally ill may be ex-lained by disparities in care for the mentally ill, such ashe under-recommendation and underuse of appropriatecreening, which may be explained by provider distrac-ion and time away from prevention activities due to therimacy of patients’ mental illnesses.However, despite the importance, prevalence, and

omplexity of physical illnesses, including cancer, amonghe mentally ill, the authors could not fınd a systematiceview of health disparities in breast and cervical cancercreening among women with mental illness. However,ome primary evidence was found: a mixture of studies

ure 1. Conceptual framework to understand mental

ss and cancer prevention

pril 2013

have examined cancer screening disparities in popula-tions withmental illnesses. Therefore, the authors under-took a systematic review of the current literature onbreast and cervical cancer screening inwomenwithmen-tal illnesses.

Evidence AcquisitionA systematic, limited review of the literature was conducted and acritical appraisal was made of selected articles believed to containmaterial that would provide evidence of cancer screening ratesamong the mentally ill. The consecutive steps of the methods usedduring this review process are summarized in Figure 2 and brieflydescribed below.

Step 1: Literature Search

APubMed/MEDLINEandPsycINFO search, performed from Jan-uary 2010 to March 2012, retrieved 3720 articles pertaining tomentally ill patients and cancer screening. PubMed/MEDLINE is aprimary database of information in the biomedical and healthsciences fıelds. PsycINFO is an American Psychological Associa-tion database of abstracts for psychological literature, includingjournals, dissertations, reports, book chapters, and other scholarlydocuments from 1887 to the present.The search strategy involved a list of keywords that progressed

systematically as follows:

● Mental illness and each of the following secondary terms indi-vidually: cancer screening, breast cancer screening,mammogram,self-breast exam, cervical cancer screening, Pap smear.

● Retaining the secondary terms,mental illnesswas replaced by thefollowing topics individually: chronic mental illness, psychiatricdisorder, depression, anxiety, mood disorders, bipolar disorder,post-traumatic stress disorder, schizophrenia, psychosis.Comorbidity,mental illness, and cancer screening also were used

as search topics. To obtain all relevant literature, gender was not acriterion. These search terms and the procedure, generated by thereviewers, were deemed adequate, as they (1) included differentterminology for the same procedures (i.e., breast cancer screeningversusmammogram); (2) included both broad (i.e., chronicmentalillness) and more specifıc (e.g., depression, anxiety, bipolar disor-der) terminology; and (3) utilized different combinations of thesekey terms to obtain as comprehensive a search as possible.

Step 2: Abstract Review

After each of the keyword searches, the title, abstracts, and key-

3720 articles identified in PubMed/Medline and PsychINFO

databases

1180 abstracts reviewed

35 full papers reviewed

19 articles given critical appraisal

2540 articles excluded because they did not meet inclusion criteria

1145 articles excluded as they did not include evidence of cancer screening rates in the mentally ill

Figure 2. Article retrieval methods

Fig

words of the articles were reviewed for the inclusion criteria. A

tiw

pmqphop

wrt

394 Aggarwal et al / Am J Prev Med 2013;44(4):392–398

record of articles in each keyword search was maintained to iden-tify duplicates. The exclusion criteria were as follows:

● Basic science: articles reporting basic science only;● Irrelevant cancer type: articles pertaining to cancer types otherthan breast and cervical cancers;

● Irrelevant mental diagnosis: Given the exhaustive nature of var-ious mental illnesses, the researchers based this review on themore common mental illness disorders listed in the keywordsearch. There was concern that including too many mental ill-nesses would diminish the focus of this review.

● Mental illness resulting from a cancer diagnosis: any articlesassessing the causal relationship between cancer diagnosis andsubsequent mental illness;

● Case studies: Studies with intensive analyses of individual unitswere excluded because they did not address disparities.Further, this search excluded editorials, letters, and disserta-

ions. Articles that pertained to incidence rates were reviewed todentify any reference to breast and cervical cancer screening inomen with mental illness.

Step 3: Reviews of Full Articles

Thirty-fıve articles were identifıed as likely to contain originalevidence of cancer screening rates in the mentally ill from theabstract review, and after critical review of these articles in theirentirety, 19 were found to be contributory and relevant. Thesearticles contained one or all of three topics: (1) barriers to screen-ing; (2) screening utilization; and (3) factors that encourage screen-ing. These areas of interest were chosen because they were deter-mined to provide a comprehensive approach to understandingscreening in this population and also focused on the research topic.

Step 4: Critical Appraisal

The critical appraisal guidelines for an article proposed by the U.S.Preventive Health Services Task Force were adapted for this re-view.27–29 Two examiners independently reviewed all the acceptedpapers, evaluating them in six different categories: research design,research quality, generalizability, existence of additional studiesaddressing the same key question, consistency/coherence of re-sults, and any other additional factors affecting the certainty of theresults.For each article, the strength of evidence was judged for each of

the six categories as (1) convincing (a well-designed, well-conducted study in representative populations that directly as-sessed effects on cancer screening); (2) adequate (evidence suffı-cient to determine effects on cancer screening but limited bynumber, quality, or consistency of the study, generalizability toroutine practice, or indirect nature of the evidence); or (3) inade-quate (insuffıcient evidence to determine the effect on cancerscreening because of limited study power, important flaws in thestudy design or conduct, gaps in the chain of evidence, or lack ofinformation on cancer screening).This was followed by an overall certainty ranking of each paper

as (1) high (the conclusion is unlikely to be strongly affected by theresults of future studies); (2) moderate (as information becomesavailable, the magnitude or direction of the observed effect couldchange, and this change may be large enough to alter the conclu-sion); or (3) low (the available evidence is insuffıcient to assesseffects of cancer screening). In the case of disagreement regarding

the article evaluation, differences were discussed between the re-

viewers. If there was still disagreement, a third party reviewed thearticle in question.

Step 5: Combining Findings

A summary table of selected article descriptions, key areas of inves-tigation, and critical appraisal results was developed based on theresults of the previous steps. For each paper, the following infor-mation was documented: lead author; cohort description; researchsetting; fındings as they relate to screening utilization, barriers toscreening, and factors that encourage screening; and overall cer-tainty of the results (see Appendix A, available online at www.ajpmonline.org).

Evidence SynthesisAmong the articles selected, 15 studies were from theU.S., and four were from Europe, Canada, or Australia.Two articles addressed cervical cancer screening, eightaddressed breast cancer screening, and nine addressedboth cancers. The sample size of these studies varied from26 participants to a larger national data set of 335,294participants. The studies included were prospective(n�9) and retrospective studies (n�10); small clinicalpractice (n�7) and larger national database analyses(n�10); and cross-sectional (n�4) and qualitative stud-ies (n�4). Many articles contributed to more than one ofthe identifıed areas of interest (i.e., screening utilization,barriers to screening, and factors that encouragescreening).

Screening UtilizationFifteen studies addressed cancer-screening utilization(see Appendix A, available online at www.ajpmonline.org) in patients with mental illness, including seven thatassociated the presence of mental illness with lower can-cer screening utilization.30–35 In a cohort of femalepatients with schizophrenia (n�46), Lindamer30 re-orted that adherence to preventive screening recom-endations was less likely, even though more than threeuarters of the subjects had insurance, a primary carehysician, and annual physical examinations. Less thanalf of the study population reported having received oner more gender-specifıc preventive services during therevious year.30

Martens found that women with schizophrenia(n�3220) were less likely to have a Pap compared to allother women.31 Being aged �50 years and living in alow-income area further decreased the likelihood of re-ceiving a Pap.31 Similarly, Tilbrook et al. found thatomen with psychosis (n�51) were fıvefold less likely toeceive adequate cervical screening when compared withhe general population without mental illness (n�118).32

Interestingly, these women also exhibited higher smok-ing rates and higher healthcare utilization measured as

the number of primary care visits per year.32

www.ajpmonline.org

(wa

a(gtwswaiasic

s

ts

crprdbfr

Aggarwal et al / Am J Prev Med 2013;44(4):392–398 395

A

Pirraglia33 studied cancer screening in women withdepressive symptoms and reported that women with ahigher depressive burden (n�492) had signifıcantlylower rates of mammography utilization; this factor didnot affect cervical cancer screening. Salsberry35 foundthat mental illness negatively affected mammogramscreening among Medicaid patients (n�669), and Mas-terson36 found that poor mental health, depression, andanxiety (n�2781) might be associated with non-receiptof mammograms. Finally, Yee37 found that womenn�606) with a mental health diagnosis were as likely asomen without (53% and 47%, respectively) to receiveny breast or cervical cancer screening.Werneke and colleagues38 found that, overall, psychi-

tric patients were as likely as the reference groupn�53,340women aged 50–64 years residing in the threeeographic areas) to receive breast screening. However,hose with a history ofmultiple detentions as an inpatientere signifıcantly less likely to receive screening. Theubgroups least likely to receive breast cancer screeningere patients with a psychotic disorder diagnosis. Olderge, social deprivation, and a history of detention in annpatient unit remained independent predictors for non-ttendance.38 As an example of exposure bias in Long’study (n�105), not being psychiatrically hospitalizedwasdentifıed as an independent predictor for not having aurrent clinical breast exam.39

Compared to women without mental illness, Carney40

found that women with low-severity mood disorders(n�18,958) were only 7% less likely to have ever had amammogram. Women with high-severity mood disor-ders (n�540) were 66% less likely to have done so. Pa-tients with psychotic and substance-use disorders hadlow screening rates overall, regardless of severity, sug-gesting that the nature and type of mental illness alsocontributed to screening.40

A few studies have shown no major association ofmental illness with breast and cervical cancer screening.Kahn41 studied women with a history of any mental dis-order or substance-use disorder using focus groups. Thisqualitative study with a purposeful sample was not de-signed to assess screening utilization in this population.Nonetheless, Kahn reported that 25 of the 26 participantsreported mammography utilization within the previous18months despite evident educational defıciencies.41 In amall cross-sectional study, Carney42 found high rates ofmammography utilization over a 1-year period amongpatients (n�164) with a dual mental illness andsubstance-use disorder diagnosis. These patients wererecruited from inpatient and outpatient psychiatricunits and an outpatient drug rehabilitation unit.Owen identifıed no difference in screening utilization

in a cohort of women with mental health problems

pril 2013

(n�100) recruited from an outpatient mental healthclinic in Australia.43 The screening rates were comparedo the general Australian population. A slightly largertudy by Lasser44 found no difference in mammographyutilization among new female, uninsured, immigrantoutpatients (n�526) with or without mental illness. Thestudy failed to address other known predictors of cancerscreening such as low health literacy, having had a pri-mary care physician, and insurance status in this immi-grant population.

Barriers to ScreeningMultiple barriers to breast and cervical cancer screeningin thementally ill were discussed in seven articles selected(see Appendix A, available online at www.ajpmonline.org), with agreement and disagreement between authors.Some authors, including Carney40 and Werneke,38 dis-ussed the importance of mental illness severity as a bar-ier to screening. They identifıed mood disorders andsychosis as specifıc barriers to screening utilization. Pir-aglia and colleagues33 identifıed a high depressive bur-en as an independent barrier to breast cancer screeningut not cervical cancer screening. Several articles identi-ıed low SES as an independent predictor of low screeningates. Friedman45 found that low-income female psychi-atric patients were as likely as low-income women with-out psychiatric disorders to be screened at inappropriateintervals for breast and colorectal cancer.Owen43 found that consumers and healthcare provid-

ers listed transportation diffıculties, embarrassment, ad-verse experiences, lack of reminders, and familiar health-care providers as barriers to screening. For 51% of thecohort, the gender of the primary care physician also wasnamed as a barrier; female patients were less likely to acton Pap screening recommendations made by male prac-titioners.43 Kahn et al.41 similarly found that femalemammography technologists predicted higher rates ofmammogram screening. Fear of pain or discomfort due tocoldness of the machine and compression of the breast alsowere reported as barriers formentally ill participants. Otherconcerns included delays in appointment scheduling, inphysicians’ waiting rooms, and while awaiting results.41

Factors That Encourage ScreeningSix studies focusedon factors thatmight influencepositivelycancer screening among women with mental illnesses (seeAppendix A, available online at www.ajpmonline.org).Four of these concluded that primary care played a criti-cal role in the rates of cancer screening. Using a nationaldata set of a veteran cohort of patients (n�113,505) withsome mental illness and/or substance-use disorders,Druss46 found a decreased likelihood of receipt of all

preventive services regardless of demographics, health

K

iwcvttpb

p

cp7

alitid

pmitsc

daiwpwi

uiiirGsch

ubeltane

ss

396 Aggarwal et al / Am J Prev Med 2013;44(4):392–398

status, or facility-level characteristics. The majority ofpreventive services were cancer screenings, specifıcallyfor colorectal, breast, cervical, and prostate cancer. Drussoutlined the importance of healthcare providers’ referralsfor preventive services and their recognition of psychiat-ric and somatic issues when recommending individualsfor screening.46

Miller47 found that all of thewomenwithmental illnesseswho completed age-appropriate breast- and cervical-cancerscreening in his study were connected to primary care pro-viders. He concluded that better communication betweenprimary care and mental health care providers would en-hance cancer screening participation.47 Friedman45 andahn41 supported this conclusion. Friedman48 demon-

strated the importance and effect of physician recommen-dations amongpatientswithmental illness.Kahnnoted thatphysician recommendations and positive family historywere motivators for cancer screening.41

Long et al.39 studied homeless women with mentalllness and found that higher rates of cancer screeningere associatedwithAfrican-American ethnicity, health-are facility and provider continuity, frequent medicalisits in the previous 12 months, provider recommenda-ion, and having insurance. Having no psychiatric hospi-alization and higher levels of education were the mostredictive for having, respectively, a current clinicalreast exam and Pap.39

DiscussionSubstantial primary evidence was found that suggests thatimportant disparities persist in breast and cervical cancerscreening utilization among women with chronic mentalillness. Cancer screening prevalence for these women lagsbehind that for women without mental illness, which hasimproved over the past decade, and that barriers to screen-ing specifıc to thementally illmay exist. A few studies inves-tigated specifıc barriers to or factors that promote screeningamong womenwithmental illness. These studies suggestedthatwomenwithmental illness, especiallymoreseveremen-tal illness, have unique barriers to breast and cervical cancerscreening and might require innovative approaches to in-crease screening utilization.The review also found that studies examining screen-

ing utilization have produced contrasting results con-cerning whether mental illness is a barrier to screening.Although some studies30,33,34,40 found that mental illnesslayed a negative role in screening utilization, Owen43

found no difference in screening utilization among thementally ill. Conversely, Kahn41 found higher breast can-er screening rates (�90% women were screened) com-ared to the general U.S. population (ranges from 29% to

0%). These contradictory resultsmay be due to the vary- s

ing cohort size, study settings, and defınitions of mentalillness used in the studies.Setting and cohort size affected study fındings, and

smaller studies showed fewer disparities. Lindamer’s30

and Owen and Kahn’s41 sample sizes were small anddmittedly unrepresentative. Only studies with relativelyarger samples33,36,38,40,46,49 showed marginally to signif-cantly lower cancer screening rates in women withmen-al illness. Additionally, cohort composition varied, mak-ng generalization of the results diffıcult. For example, theiverse psychiatric patient cohort in the Halbreich49

study is not comparable to the small Jewish schizophreniccohort of Grinshpoon.50

Other major variations were patient population andrecruitment settings. The small study cohorts used byKahn41 and Owen43 were Medicaid and managed-careopulations with lower SES. Moreover, both studies’ pri-ary focus was understanding barriers to cancer screen-

ng. Retrospective analyses of larger primary care prac-ices or healthcare systems yielded more-representativecreening utilization results. These showed lower rates ofancer screening in women with mental illness.Some studies of more severe mental illness showmoreisparities, but lower cancer screening utilization persistscross the spectrumofmental illness diagnosis and sever-ty. The current fındings suggest that grouping patientsith mental illness by only diagnostic disorder for theurpose of studying cancer screening is a process fraughtith misclassifıcation errors. The mentally ill populations more complex and diverse than many studies imply.For example, anxiety disorders, depression, and substance-se disorders often co-occur, and the impact of mentalllness may vary with disease severity and treatment. Us-ng a global functional indicator thatmeasures the overallmpact of mental illness may yield a more useful catego-ization of influences on cancer screening.51,52 Thelobal Assessment of Function (GAF) assesses bothymptoms of mental illness and disability related to psy-hological functioning and is a reliable measure in a co-ort of long-term mentally ill patients.There is an increased burden of cancer with lowerptake of cancer screening in certain populations, andarriers exist that are associated with age, race/ethnicity,ducation/health literacy, the medically underserved,ack of a primary care provider, and medical comorbidi-ies. Interestingly, these factors also may affect the prev-lence, accurate diagnosis, and treatment of mental ill-ess. The recent literature shows poor assessment ofducation/health literacy,44 having a primary care pro-vider,39,44 and medical comorbidities.33,41 A careful as-essment of the role of known disparities in cancercreening and its interaction withmental disorders in the

tudy sample may yield less-divergent results.

www.ajpmonline.org

aetswih

ms

a

btppdpa

1

1

1

1

1

1

1

1

1

1

Aggarwal et al / Am J Prev Med 2013;44(4):392–398 397

A

Even assuming that a burden of illness scale can beconstructed and applied, common mechanisms explain-ing the lower prevalence of cancer screening among themore severely ill did not emerge from the current review.For instance, specifıc factors, such as a specifıc diagnosis,provider gender, or having a primary care provider, wereidentifıed as barriers to screening in some studies33,43 buts promoters of screening in other studies.39,45,48 How-ver, under-reporting, denial, symptom minimization andolerance, and poor insight are common in patients withchizophrenia and may play a role.53 The mechanisms atorkamongpatientswithhigherdepressive symptomsmaynclude neglect of screening as part of overall self-neglect,elplessness, and sense of lack of control.33

Additionally, it is important to note that many of thebarriers and promoters related to screening identifıed inthese studies are not unique to thementally ill, as barriersand fears are evident in any population.41 Having a pri-ary care provider generally is associated with bettercreening in all populations,54,55 but trust, compliance,and social support might be more important in patientswith mental illness. Having a regular primary care pro-vider and their engagement in patient care may play aneven more important role in cancer screening for thispopulation. The gender of providers and techniciansmayinfluence screening behaviors in all populations.56,57 Inthe current review, studies were found that suggest thatthese factors are associated with better screening amongthe mentally ill,41,43,54 but unique barriers or promoterslso may exist for the mentally ill.Providers may fail to screen the mentally ill for cancerecause of confusion about symptom attributes, distrac-ion by comorbidities, and multiple symptoms. One pa-er stated that “providers tend to ascribe many of theseatients’ somatic complaints to their psychiatric disor-ers, which may result in an underestimate of the pretestrobability of other medical conditions.”36 Redelmeier etl.58 suggested that comorbidity decreases the time andattention primary care physicians can devote to any givenmedical problem because they must balance many com-peting demands.

ConclusionFactors that inhibit or encourage cancer screening amongthementally ill demand further attention to eliminate thescreening rate disparities between these populations andthose without mental illness. This is the fırst review todiscuss assessments of cancer screening utilizationamong women with chronic mental illness and to surveyfactors identifıed as promoting or interfering withscreening in this complex and variable subpopulation,which has low screening rates. Mechanisms for screening

failures are just now emerging, suggesting targets for

pril 2013

intervention. Although the current review suggests thatseveral targets for intervention are shared by thementallyill and non–mentally ill, a few studies have identifıedfactors unique to thementally ill. These studies raise hopefor unique interventions to erase cervical and breast can-cer screening disparities in this subpopulation.

No fınancial disclosures were reported by the authors of thispaper.

References1. American Cancer Society. Cancer facts and fıgures 2012. Atlanta GA:

American Cancer Society, 2012.2. CDC. FastStats: AIDS and HIV. www.cdc.gov/nchs/fastats/aids-

hiv.htm.3. American Cancer Society. Cancer screening guidelines. www.cancer.

org/Healthy/FindCancerEarly/CancerScreeningGuidelines/index.4. U.S. Preventive Services Task Force (USPSTF). Prevention & chronic

care program: improving primary care. www.ahrq.gov/clinic/uspstfıx.htm.

5. American Cancer Society (ACS). Cancer prevention & early detectionfacts and fıgures 2011. Atlanta GA: ACS, 2011.

6. Anderson L, May D. Has the use of cervical, breast, and colorectalcancer screening increased in the U.S.? Am J Public Health1995;85(6):840–2.

7. Howlader N, Noone A, Krapcho M, et al., eds. SEER cancer statisticsreview, 1975-2008. Bethesda MD: National Cancer Institute, 2011.seer.cancer.gov/csr/1975_2008/.

8. Eley J, Hill H, ChenV. Racial differences in survival frombreast cancer.JAMA 1994;272:947–54.

9. Hiatt R, Pasick R. Unsolved problems in early breast cancer detection:focus on the underserved. Breast Cancer Res Treat 1996;40:37–51.

0. Hiatt R, Pasick R, Stewart S, et al. Community-based cancer screeningfor underserved women: design and baseline fındings from the Breastand Cervical Cancer Intervention Study. Prev Med 2001;33:190–203.

1. Valanis B, Whitlock E, Mullooly J, et al. Screening rarely screenedwomen: time-to-service and 24-month outcomes of tailored interven-tions. Prev Med 2003;37:442–50.

2. Whitman S, Lacey L, Ansell D, et al. An intervention to increase breastand cervical cancer screening in low-income African-Americanwomen. Fam Community Health 1994;17(1):56–63.

3. Felker B, Yazell J, Short D. Mortality and medical comorbidity amongpsychiatric patients: a review. Psychiatr Serv 1996;47:1356–63.

4. Wassertheil-Smoller S, Shumaker S, et al. Depression and cardiovascu-lar sequelae in postmenopausal women. Arch Intern Med 2004;164:289–98.

5. Katon W, Rutter C, Simon G, et al. The association of comorbiddepression with mortality in patients with type 2 diabetes. DiabetesCare 2005;28:2668–72.

6. NIMH. NIMH: The number count: mental disorders in America.www.nimh.nih.gov/health/publications/the-numbers-count-mental-disorders-in-america/index.shtml.

7. CDC.Nationaldiabetes fact sheet:national estimatesandgeneral informa-tion on diabetes and prediabetes in the U.S. Atlanta GA: DHHS, CDC,2011.

8. Insel T. Assessing the economic costs of serious mental illness. Am JPsychiatry 2008;165:663–5.

9. Mark T, Levit K, Buck J, et al. Mental health treatment expenditure

trends, 1986-2003. Psychiatr Serv 2007;58:1041–8.

398 Aggarwal et al / Am J Prev Med 2013;44(4):392–398

20. Kessler R, Nelson C, McGonagle K, et al. The epidemiology of co-occurring addictive and mental disorders: implications for preventionand service utilization. Am J Orthopsychiatry 1996;66:17–31.

21. Jones D, Macias C, Barreira P, et al. Prevalence, severity, and co-occurrence of chronic physical health problems of persons with seriousmental illness. Psychiatr Serv 2004;55:1250–7.

22. Druss BG,Walker ER.Mental disorders andmedical comorbidity. RobertWood Johnson Foundation Research Synthesis Project, 2011;21:1–26.

23. Owens P, Myers M, Elixhauser A, Brach C. Care of adults with mentalhealth and substance abuse disorders in U.S. community hospitals,2004. Agency for Healthcare Research and Quality, 2007. HCUP FactBook No. 10. AHRQ Publication No. 07-0008. ISBN 1-58763-229-2.

24. Frayne S,Halanych J,MillerD, et al. Disparities in diabetes care: impactof mental illness. Arch Intern Med 2005;165:2631–8.

25. Sederer L, Silver L,McVeighK, Levy J. Integrating care formedical andmental illnesses. Prev Chronic Dis 2006;3:A33.

26. Carney R, Freeland K, Eisen S, Rich M, Jaffe A. Major depression andmedication adherence in elderly patients with coronary artery disease.Health Psychol 1995;14:88–90.

27. U.S. Preventive Services Task Force. Slide presentation from theAHRQ 2008 Annual Conference (Text Version). January 2009. Rock-villeMD:Agency forHealthcare Research andQuality. www.ahrq.gov/about/annualmtg08/090808slides/Lin2.htm.

28. Fleming C,Whitlock EP, Beil TL, Lederle FA. Screening for abdominalaortic aneurism: a best-evidence systematic review for the U.S. Preven-tive Services Task Force. Ann Intern Med 2005;142:203–11.

29. Buckley DI, Rongwei F, FreemanM, Rogers K, HelfandM. C-Reactiveprotein as a risk factor for coronary heart disease: a systematic reviewand meta-analysis for the U.S. Preventive Services Task Force. AnnIntern Med 2009;151:483–95.

30. Lindamer LA, Wear E, Sadler GR. Mammography stages of change inmiddle-agedwomenwith schizophrenia: an exploratory analysis. BMCPsychiatry 2006;6:49.

31. Martens PJ, ChochinovHM, Prior HJ, et al. Are cervical cancer screen-ing rates different in women with schizophrenia? AManitoba popula-tion-based study. Schizophr Res 2009;113:101–6.

32. Tilbrook D, Polsky J, Lofters A. Are women with psychosis receivingadequate cervical cancer screening? Can Fam Physician 2010;56:358–63.

33. Pirraglia PA, Sanyal P, Singer DE, Ferris TG. Depressive symptomburden as a barrier to screening for breast and cervical cancers. JWomens Health (Larchmt) 2004;13:731–8.

34. Xiong G, Bermudes R, Torres S, Hales R. Use of cancer-screeningservices among persons with serious mental illness in SacramentoCounty. Psychiatr Serv 2008;59(8):929–32.

35. Salsberry P, Chipps E, Kennedy C. Use of general medical servicesamong Medicaid patients with severe and persistent mental illness.Psychiatr Serv 2005;56(4):458–62.

36. Masterson EA, Hopenhayn C, Christian WJ. Self-reported mentalhealth status and recent mammography screening. J Women’s Health2010;19:1569–76.

37. Yee EFT, White R, Lee SJ, et al. Mental illness: is there an associationwith cancer screening among women veterans? Women’s Health Is-sues 2011;21-4S:S195–S202.

38. Werneke U, HornO,Maryon-Davis A,Wessely S, Donnan S,McPher-son K. Uptake of screening for breast cancer in patients with mentalhealth problems. J Epidemiol Community Health 2006;60:600–5.

39. LongHL, Tulsky JP, Chambers DB, et al. Cancer screening in homelesswomen: attitudes and behaviors. J Health Care Poor Underserved1998;9:276–92.

40. Carney CP, Jones LE. The influence of type and severity of mentalillness on receipt of screening mammography. J Gen Intern Med

2006;21:1097–104.

41. Kahn LS, Fox CH, Krause-Kelly J, BerdineDE, CadzowRB. Identifyingbarriers and facilitating factors to improve screening mammographyrates in women diagnosed with mental illness and substance use disor-ders. Women Health 2005;42:111–26.

42. Carney C, Allen J, Doebbeling B. Receipt of clinical preventive medicalservices among psychiatric patients. Psychiatr Serv 2002;53(8):1028–30.

43. Owen C, Jessie D, De Vries Robbe M. Barriers to cancer screeningamongst women with mental health problems. Health Care WomenInt 2002;23:561–6.

44. Lasser KE, Zeytinoglu H, Miller E, Becker AE, Hermann RC, Bor DH.Do women who screen positive for mental disorders in primary carehave lowermammography rates?GenHospPsychiatry 2003;25:214–6.

45. Friedman LC, Moore A, Webb JA, Puryear LJ. Breast cancer screeningamong ethnically diverse low-income women in a general hospitalpsychiatry clinic. Gen Hosp Psychiatry 1999;21:374–81.

46. Druss B, Rosenheck R, DesaiM, Perlin J. Quality of preventivemedicalcare for patients with mental disorders. Med Care 2002;40:129–36.

47. Miller E, Lasser K, Becker A. Breast and cervical cancer screening forwomen with mental illness: patient and provider perspectives on im-proving linkages between primary care and mental health. Arch Wo-mens Ment Health 2007;10(5):189–97.

48. FriedmanL, Puryear L,MooreA,GreenC. Breast and colorectal cancerscreening among low-income women with psychiatric disorders. Psy-chooncology 2005;14:786–91.

49. Halbreich U, Shen J, Panaro V. Are chronic psychiatric patients atincreased risk for developing breast cancer? Am J Psychiatry1996;153:559–60.

50. GrinshpoonA, BarchanaM, PonizovskyA. Cancer in schizophrenia: isthe risk higher or lower. Schizophr Res 2005;73:333–41.

51. Hermann RC, Mattke S, Somekh D, et al. Quality indicators for inter-national benchmarking of mental health care. Int J Qual Health Care2006;18(1):31–8.

52. Jones SH, Thornicroft G, Coffey M, Dunn G. A brief mental healthoutcome scale—reliability and validity of the Global Assessment ofFunctioning (GAF). Br J Psychiatry 1995;166:654–9.

53. Barak Y, Achiron A, Mandel M, Mirecki I, Aizenberg D. Reducedcancer incidence among patients with schizophrenia. Cancer 2005;104:2817–21.

54. Starfıeld B, Shi L, Macinko J. Contribution of primary care to healthsystems and health. Milbank Q 2005;83:457–502.

55. Starfıeld B. Is primary care essential? Lancet 1994;344:1129–33.56. Lurie N, Margolis KL, McGovern PG, Mink PJ, Slater JS. Why do

patients of female physicians have higher rates of breast and cervicalcancer screening? J Gen Intern Med 1997;12:34–43.

57. Fang MC, McCarthy EP, Singer DE. Are patients more likely to seephysicians of the same sex? Recent national trends in primary caremedicine. Am J Med 2004;117:575–81.

58. Redelmeier D, Tan S, Booth G. The treatment of unrelated disorders inpatients with chronic medical diseases. N Engl J Med 1998;338:1516–20.

59. Freeman HP, Chu KC. Determinants of cancer disparities: barriers tocancer screening, diagnosis, and treatment. Surg Oncol Clin N Am2005;14:655–69.

Appendix

Supplementary data

Supplementary data associatedwith this article can be found, in the

online version, at http://dx.doi.org/10.1016/j.amepre.2012.12.006.

www.ajpmonline.org