Abstract
Objective
We describe factors, in the context of the Social Determinants of Health model, associated with receiving Pap smears within risk-appropriate guidelines (i.e., guidelines that specify screening intervals based upon a woman’s individual risk of developing cervical cancer).
Methods
Completed in June 2006, we conducted a cross-sectional survey of women from 14 health clinics in Ohio Appalachia pertaining to psychosocial, demographic, biological, and health-related factors. A logistic regression model was constructed to predict whether or not a woman was within risk-appropriate cervical cancer screening guidelines.
Results
Of 562 women with a date of last Pap smear, 380 (68%) were within risk-appropriate guidelines. Logistic regression showed that, compared to women with low-level SES, women with middle- and high-level SES had 3.39 [1.85, 6.21] and 3.86 [2.03, 7.34] times the odds, respectively, of being within risk-appropriate guidelines. Odds of being within guidelines increased 1.09 [1.04, 1.15] fold for each decrease of one major life event. Additionally, women that were financially better off or financially worse off than their parents at the same age had lower odds (0.41 [0.23, 0.73] and 0.49 [0.24, 0.98], respectively) of being within guidelines than women who reported their finances were the same as their parents. Results also showed an interaction between marital status and age at first intercourse (p=0.001).
Conclusion
The results suggest an impact of psychosocial factors on Pap smear testing behaviors, and illustrate the need to examine risk-appropriate interventions to improve screening.
Keywords: cervical cancer, Pap smear, Pap test, risk-appropriate, screening, guidelines, women, barriers, underserved populations, health disparities
INTRODUCTION
Since its introduction in the 1950’s, the Pap smear has been associated with sharp declines in cervical cancer incidence and mortality. While Pap smear testing has been recommended and well-established for many years, not all women report receiving screening at appropriate intervals. Data from the Ohio Behavioral Risk Factor Surveillance System (BRFSS) in 2006 show that 83.4% of women aged 18 and older reported having had a Pap smear within the past 3 years (Ohio Behavioral Risk Factor Surveillance System, 2006), yet this rate of Pap smear utilization is considerably lower than the Healthy People 2010 target of 90% (U.S. Department of Health and Human Services, November 2000). Moreover, this rate does not consider a woman’s risk status for cervical cancer in determining if Pap smears were received at optimum intervals. For example, women with any risk factors for cervical cancer (i.e., smoking, early age at first intercourse, five or more sexual partners in a lifetime, or having a personal history or partner with a history of human papillomavirus (HPV) or a sexually transmitted infection (STI)) should have a Pap smears annually, while those women with no identified risk factors should have a Pap smears at least every 3 years (American College of Obstetricians and Gynecologists, 1995; Cyrus-David, Michielutte et al., 2002). Additionally, past studies have documented that certain groups of women (i.e., poorer income, lower educated, elderly, minority, geographically isolated) are less likely to report having had a Pap smear within recommended guidelines (Akers, Newmann et al., 2007).
The Community Awareness, Resources, and Education (CARE) project is a P50 Center grant funded through the National Institutes of Health Centers for Population Health and Health Disparities (CPHHD) initiative. This initiative has proposed a novel framework for addressing health disparities that integrates both upstream (i.e., individual-level demographics, behaviors, and biology) and downstream (i.e., community characteristics, policy and institutional contexts, and societal relationships and conditions) multilevel factors (Warnecke, Oh et al., 2008). Using this framework and consisting of both observational and intervention-based projects, CARE uses a transdisciplinary approach, focused on the Social Determinants of Health (SDH) model (Marmont, 1999), to address the problem of high cervical cancer incidence and mortality rates in Ohio Appalachia, a region, at the time the study was conducted, comprised of 29 counties in the state.
The first phase of CARE, completed in June 2006, involved collecting cross-sectional in-person survey data on women from participating clinics in order to characterize social, demographic, biological, and health-related factors associated with obtaining regular Pap smears. The focus of this paper is to describe factors that were associated with receiving Pap smears within risk-appropriate guidelines among women who completed the survey, using the CPHHD overall multilevel framework and the SDH model. To our knowledge, this is the first time that several of these factors have been examined in relation to risk-appropriate guidelines for Pap smear utilization.
METHODS
Study Design
The CARE project divided Ohio Appalachia into four regions (Hade, 2009 submitted to Prev Med). The northeast, central, southern, and southwest parts of Ohio Appalachia correspond to Regions 1, 2, 3, and 4, respectively. Twenty-two health clinics in Ohio Appalachia were approached to participate in CARE, of which 14 (63.6%) agreed. Monthly random samples of women from these clinics were selected, and medical records were reviewed to determine eligibility. To be eligible, a woman had to be at least 18 years of age, a resident of an Ohio Appalachian county, not pregnant, seen in a participating clinic within the last two years, and have no history of hysterectomy or invasive cervical cancer. If a woman was eligible and agreed to participate, a baseline survey was conducted, usually at the participant’s home. Of the 801 women who were contacted and determined to be eligible, 571 completed the interview yielding a cooperation rate of 71% (Hade, 2009 submitted to Prev Med). Women who refused to participate in the survey were demographically similar (age, race, and ethnicity) to those who did participate, except more women from Region 4 (21% vs. 13%, p=0.029), and more women who reported being current smokers (41% vs. 29%, p=0.004) refused to participate. The study was approved by the Institutional Review Boards at Ohio State University, the University of Michigan, and the Centers for Disease Control and Prevention (CDC).
Measures
Outcome Variable
The outcome of interest was whether or not a woman was within risk-appropriate cervical cancer screening guidelines, which was dependent on both screening history and cervical cancer risk level. A woman was classified as high risk for cervical cancer if she reported any of the following behaviors: 1) five or more sexual partners during lifetime; 2) sexual intercourse before age 18; 3) being a current smoker; 4) ever being diagnosed with an STI (HPV, genital warts, genital herpes, syphilis, gonorrhea, human immunodeficiency virus (HIV)); or 5) having a sexual partner with history of an STI (American College of Obstetricians and Gynecologists, 1995; Cyrus-David, Michielutte et al., 2002). Conversely, a woman was classified as low risk if she did not report any of the previous behaviors or conditions. High risk women were considered to be within guidelines if their most recent Pap smear was in the 13 months prior to their cross-sectional interview, while low risk women were classified as being within screening guidelines if their most recent Pap smear was in the 37 months prior to their interview. These definitions of compliance with risk-appropriate screening guidelines were drawn from the consensus recommendation of the National Institutes of Health with additional guidelines provided by the American College of Obstetricians and Gynecologists, and have been established for over a decade (American College of Obstetricians and Gynecologists, 1995; Cyrus-David, Michielutte et al., 2002). To determine the date of a woman’s most recent Pap smear, information from medical records review was used, if available (N=428, 76.2%). Otherwise, self-reported dates were used (N=134, 23.8%). Of the 421 women who had both medical record information and self-reported information available, 340 (80.7%) had agreement between the two. Among women who self-reported being within guidelines (N=342), 19% (N=65) were not based on medical record review. In contrast, of those who self-reported not being within cervical cancer screening guidelines (N=79), 20% (N=16) actually did receive a Pap smear based on medical record review, thus discrepancies were evenly balanced.
Independent Variables
Independent variables were selected based on their relationships to Pap smear screening behaviors and were categorized using the SDH model. The categories included social and cultural factors, material factors, health and health behaviors, psychological factors, and environmental factors (Table 1).
Table 1.
Descriptive statistics and univariate model results by risk-appropriate cervical cancer screening status among participating women in Ohio Appalachia, 2006 (n=562)
| Within Risk-Appropriate Cervical Cancer Screening Guidelines, n (%) | |||||
|---|---|---|---|---|---|
| Predictor Variable | No | Yes | Totala | ORcrude | 95% CI |
| Demographics | |||||
| Age (in years) | |||||
| Older than 50 | 49 (27%) | 71 (19%) | 120 (21%) | 1.00 | |
| 31–50 | 82 (45%) | 177 (47%) | 259 (46%) | 1.38 | [0.87, 2.20] |
| 18–30 | 51 (28%) | 132 (35%) | 183 (33%) | 1.61 | [0.95, 2.73] |
| Race | |||||
| White | 173 (95%) | 359 (94%) | 532 (95%) | 1.00 | |
| Other | 9 (5%) | 21 (6%) | 30 (5%) | 1.31 | [0.57, 3.02] |
| Socioeconomic Status | |||||
| Education | |||||
| Less than high school | 25 (14%) | 20 (5%) | 45 (8%) | 1.00 | |
| High school graduate/GED | 72 (40%) | 130 (34%) | 202 (36%) | 2.07 | [1.05, 4.06] |
| Some college/college graduate | 85 (47%) | 230 (61%) | 315 (56%) | 2.80 | [1.44, 5.45] |
| Employment Status | |||||
| Other | 48 (26%) | 75 (20%) | 123 (22%) | 1.00 | |
| Unemployed/Disabled | 27 (15%) | 50 (13%) | 78 (14%) | 1.17 | [0.64, 2.16] |
| Full-time/Part-time | 107 (59%) | 254 (67%) | 360 (64%) | 1.41 | [0.91, 2.18] |
| Income | |||||
| Less than $20,000 | 74 (43%) | 109 (30%) | 183 (34%) | 1.00 | |
| $20,000–$50,000 | 64 (37%) | 143 (40%) | 207 (39%) | 1.48 | [0.96, 2.30] |
| More than $50,000 | 36 (21%) | 109 (30%) | 145 (27%) | 1.97 | [1.18, 3.28] |
| Poverty Income Ratio (PIR, by score)b | |||||
| Less than 1.0 | 54 (31%) | 83 (23%) | 137 (26%) | 1.00 | |
| 1.0–1.9 | 61 (35%) | 102 (28%) | 163 (30%) | 1.13 | [0.70, 1.83] |
| 2.0–2.9 | 19 (11%) | 65 (18%) | 84 (16%) | 2.40 | [1.26, 4.57] |
| 3.0 or above | 40 (23%) | 111 (31%) | 151 (28%) | 1.84 | [1.08, 3.13] |
| Socioeconomic Status (SES, by category) | |||||
| 0–1 | 53 (29%) | 48 (13%) | 101 (19%) | 1.00 | |
| 2–3 | 63 (35%) | 151 (40%) | 214 (40%) | 2.43 | [1.43, 4.03] |
| 4–6 | 58 (32%) | 162 (43%) | 220 (41%) | 2.76 | [1.63, 4.66] |
| Housing & Household Characteristics | |||||
| Marital Status | |||||
| Divorced/widowed/separated | 58 (32%) | 60 (16%) | 118 (21%) | 1.00 | |
| Married/member of a couple | 102 (56%) | 243 (64%) | 345 (62%) | 2.00 | [1.28, 3.12] |
| Never married | 21 (12%) | 77 (20%) | 98 (17%) | 3.58 | [1.89, 6.81] |
| Number in Householdc | 3.1 (1–9) | 3.3 (1–17) | 560 (100%) | 1.04d | [0.92, 1.18] |
| Number of years lived in countyc | 28.9 (0–82) | 26.0 (0–73) | 562 (100%) | 1.05e | [0.99, 1.11] |
| Health Care Access | |||||
| Health Insurance | |||||
| No | 45 (25%) | 88 (23%) | 133 (24%) | 1.00 | |
| Yes | 137 (75%) | 290 (77%) | 427 (76%) | 0.93 | [0.60, 1.45] |
| Type of Insurance | |||||
| Private | 108 (60%) | 235 (62%) | 343 (61%) | 1.00 | |
| Other | 73 (40%) | 142 (38%) | 215 (39%) | 1.01 | [0.68, 1.51] |
| Diagnosed with Non-STD Comorbidityf | |||||
| No | 141 (77%) | 322 (85%) | 463 (82%) | 1.00 | |
| Yes | 41 (23%) | 58 (15%) | 99 (18%) | 0.65 | [0.41, 1.03] |
| Psychological Factors | |||||
| Center for Epidemiological Studies Depression Scale (CES-D) | |||||
| 16 or above | 62 (34%) | 113 (30%) | 175 (31%) | 1.00 | |
| Less than 16 | 120 (66%) | 266 (70%) | 386 (69%) | 1.12 | [0.75, 1.65] |
| Perceived Stress Scale (PSS)c | 18.3 (3–40) | 17.3 (0–39) | 561 (100%) | 1.06g | [0.94, 1.20] |
| Life Stressor Checklist—Revised (LSC-R)c | 36.4 (0–102) | 27.1 (0–99) | 512 (100%) | 1.15g | [1.09, 1.21] |
| Number of Major Life Events (from LSC-R)c | 8.4 (0–22) | 6.3 (0–21) | 512 (100%) | 1.85g | [1.45, 2.36] |
| Beck Anxiety Inventory (BAI)c | 12.6 (0–45) | 10.9 (0–53) | 474 (100%) | 1.05g | [0.96, 1.16] |
| Trust in Physician Scale (TPS)c | 40.5 (20–55) | 41.0 (13–55) | 437 (100%) | 0.95g | [0.82, 1.09] |
| Spirituality (God Locus of Health Control, GLHC)c | 19.0 (6–36) | 19.4 (6–36) | 435 (100%) | 1.03h | [0.90, 1.18] |
| Discrimination | |||||
| Discrimination feelings scale (from DAS-DQ)c | 10.3 (0–40) | 8.5 (0–35) | 561 (100%) | 1.17g | [1.03, 1.33] |
| Lifetime discrimination (from DAS-DQ) | |||||
| 3 or above | 14 (8%) | 23 (6%) | 37 (7%) | 1.00 | |
| 2 | 31 (17%) | 38 (10%) | 69 (12%) | 0.74 | [0.32, 1.72] |
| 1 | 46 (25%) | 116 (31%) | 162 (29%) | 1.49 | [0.69, 3.23] |
| 0 | 90 (50%) | 197 (53%) | 287 (52%) | 1.38 | [0.66, 2.91] |
| Culture | |||||
| Self-identification as Appalachian | |||||
| No/Don’t know | 117 (64%) | 248 (65%) | 365 (65%) | 1.00 | |
| Yes | 65 (36%) | 132 (35%) | 197 (35%) | 0.92 | [0.62, 1.36] |
| Neighborhood has a name | |||||
| Yes | 84 (46%) | 140 (37%) | 224 (40%) | 1.00 | |
| No | 97 (54%) | 235 (63%) | 332 (60%) | 1.36 | [0.93, 1.97] |
| Social Factors | |||||
| Neighborhood (social cohesion level)c | 12.4 (5–25) | 12.2 (5–25) | 561 (100%) | 1.06g | [0.83, 1.35] |
| Collective efficacyc | 12.6 (0–20) | 13.0 (0–20) | 460 (100%) | 1.05h | [0.83, 1.33] |
| Number of close contacts | |||||
| 0–5 | 24 (13%) | 52 (14%) | 76 (13%) | 1.00 | |
| 6–10 | 26 (14%) | 52 (14%) | 78 (14%) | 0.83 | [0.41, 1.67] |
| 11–20 | 34 (19%) | 70 (19%) | 104 (19%) | 0.79 | [0.41, 1.53] |
| 20 or more | 98 (54%) | 204 (54%) | 302 (54%) | 0.81 | [0.45, 1.44] |
| Perceived relative disadvantage | |||||
| Satisfaction with current financial situation | |||||
| Not at all satisfied | 63 (35%) | 123 (32%) | 186 (33%) | 1.00 | |
| More or less satisfied | 63 (35%) | 135 (36%) | 198 (35%) | 1.14 | [0.73, 1.78] |
| Pretty well satisfied | 56 (31%) | 121 (32%) | 177 (32%) | 1.11 | [0.70, 1.77] |
| Financial situation during past few years | |||||
| Getting worse | 50 (27%) | 98 (26%) | 148 (26%) | 1.00 | |
| Stayed the same | 61 (34%) | 128 (34%) | 189 (34%) | 1.23 | [0.76, 1.97] |
| Getting better | 71 (39%) | 152 (40%) | 223 (40%) | 1.12 | [0.71, 1.76] |
| Finances compared to parents at same age | |||||
| Worse off | 37 (21%) | 64 (17%) | 101 (19%) | 1.00 | |
| About the same | 35 (19%) | 92 (25%) | 127 (23%) | 1.45 | [0.81, 2.59] |
| Better off | 108 (60%) | 210 (57%) | 318 (58%) | 1.06 | [0.66, 1.72] |
| Family income compared to other families | |||||
| Far below average or below average | 86 (48%) | 135 (36%) | 221 (40%) | 1.00 | |
| Average | 71 (39%) | 183 (48%) | 254 (45%) | 1.57 | [1.05, 2.35] |
| Above average or far above average | 24 (13%) | 60 (16%) | 84 (15%) | 1.54 | [0.87, 2.74] |
| Environmental Factorsi | |||||
| Region | |||||
| 4: Northeast | 46 (25%) | 62 (16%) | 108 (19%) | 1.00 | |
| 3: Central | 48 (26%) | 113 (30%) | 161 (29%) | 1.68 | [0.63, 4.50] |
| 2: Southern | 31 (17%) | 114 (30%) | 145 (26%) | 2.67 | [0.97, 7.34] |
| 1: Southwest | 57 (31%) | 91 (24%) | 148 (26%) | 1.05 | [0.39, 2.81] |
| County of Residence Type | |||||
| Urban | 64 (35%) | 155 (41%) | 270 (48%) | 1.00 | |
| Rural | 118 (65%) | 225 (59%) | 292 (52%) | 0.88 | [0.48, 1.60] |
| Per capita incomec | $20,337 | $20,144 | 552 (100%) | 0.98j | [0.92, 1.04] |
| median 1999 earnings for adults 16 years and older | ($13,217–$27,529) | ($3,778–$30,859) | |||
| Poverty Income Ratio (census-tract level)b,c | 14.0% (1.5%–33.0%) | 14.3% (1.5%–84.3%) | 552 (100%) | 1.08k | [0.81, 1.45] |
| High school graduation percentagec | 77.5% (66.7%–89.7%) | 78.6% (57.3%–100.0%) | 552 (100%) | 1.16l | [0.99, 1.37] |
| Percentage of households with female as headc,m | 6.7% (2.8%–13.2%) | 6.5% (1.0%–20.8%) | 552 (100%) | 0.97n | [0.89, 1.05] |
ORcrude is the crude (unadjusted) odds ratio of being within risk-appropriate cervical cancer screening guidelines versus not, including clinic location as a random effect.
95% CI is the corresponding 95% confidence interval for the crude odds ratio.
GED is General Equivalency Degree.
DAS-DQ is the Detroit Area Study Discrimination Questionnaire.
May not sum to 562 due to missing data
Annual household income divided by the federal poverty level for the relevant household size
Continuous variables show mean value with range in parentheses. Per capita income shows median value with range in parenthesis.
Odds ratio for a 1-member increase in the household
Odds ratio for a 5-year decrease
Includes lupus, colitis, Crohn’s disease, chronic hepatitis, thyroiditis, HIV, rheumatoid arthritis, tuberculosis, or diabetes
Odds ratio for a 5-unit decrease
Odds ratio for a 5-unit increase
Geocoded to census tract or county-level
Odds ratio for a $1,000 increase in median tract income
Odds ratio for a 10% increase in poverty ratio
Odds ratio for a 5% increase in high school graduation rate
With own children less than 18 years of age
Odds ratio for a 1% increase in female head of household rate
Instruments used for the independent variables included social cohesion level (Sampson, Raudenbush et al., 1997), Detroit Area Study Discrimination Questionnaire (DAS-DQ) (Taylor, Kamarck et al., 2004), Trust in Physician Scale (TPS) (Anderson and Dedrick, 1990), Center for Epidemiologic Studies Depression (CES-D) scale (Radloff, 1977), Beck Anxiety Inventory (BAI) (Beck, 1990), Life Stressor Checklist-Revised (LSC-R) scale (Wolfe, 1997), Perceived Stress Scale (Cohen, Kamarck et al., 1983), and God Locus of Health Control (GLHC) measure (Wallston, 1999). Additionally, the collective efficacy scale utilized was based on a similar concept used by Sampson et al (Sampson, Raudenbush et al., 1997).
The measure of socioeconomic status (SES) used in this analysis was loosely based on the Hollingshead scale (Hollingshead, 1975), and was derived by combining information on occupation, education, and income. Scores ranged from 0 to 6, with a higher score suggesting better SES. All three variables were categorized into three groups and summed as follows: occupation (professional (+2), skilled laborer (+1), unskilled laborer (0)); income (over $50K (+2), $25K–$50K (+1), under $25K (0)); education (more than high school (+2), high school or GED (+1), less than high school (0)). For analysis purposes, SES was grouped into three levels: SES scores of 0–1, 2–3, and 4–6 represent low-, middle-, and high-level SES, respectively.
The barriers to cervical cancer screening score, the knowledge about cervical cancer and Pap smear testing scale, and the beliefs about Pap smear testing scale were all based on previously used measures (Paskett, Rushing et al., 1997; Paskett, Tatum et al., 2006). For the barriers score, each item was worth one point with possible scores ranging from 0 to 10. Higher scores reflected more barriers to getting a Pap smear. Scoring methods for the knowledge and beliefs scales were similar (Paskett, Rushing et al., 1997). Environmental factors were based on Census 2000 data at either the census tract or county level.
Statistical Analysis
Descriptive statistics were used to provide overall sample characteristics. Univariate logistic regression models incorporating clinic as a random effect were utilized to assess the association of each independent variable with whether a woman was within or outside risk-appropriate cervical cancer screening guidelines. Available case analysis was used for instances of item non-response at the univariate level.
Next, a multivariable random effects model was built using a forward stepwise selection procedure. After the initial main effects were determined, the presence of significant two-way interactions was assessed. Odds ratios and 95% confidence intervals were calculated from the final multivariable model. Complete case analysis was used for logistic regression modeling. All analyses were performed using SAS v9.1.3 (SAS Institute, Inc., Cary, NC) and Stata SE 9.2 (StataCorp, College Station, Texas).
RESULTS
A total of 571 women in the CARE study completed the cross sectional baseline survey. Of those, 562 (98%) had either a self-reported last Pap date or last Pap date confirmed by medical record review that was used to determine their screening status. The nine (2%) women who did not have either were excluded from all subsequent analysis.
Of the 562 participants with a date of last Pap smear, 380 (68%) were within risk-appropriate screening guidelines and 182 (32%) were not. Summary statistics for participant characteristics by cervical cancer screening status are listed in Table 1. At the univariate level, women who were outside guidelines tended to be less well-educated, had lower measures of SES, and were more likely to be divorced, widowed, or separated. Additionally, women who reported having had a greater number of major life events on the LSC-R, who had a higher score on the life stressor sum section of the LSC-R, and who reported greater feelings of discrimination on the DAS-DQ had higher odds of being outside of guidelines.
Table 2 displays summary statistics for variables related to cervical cancer risk status, as well as other variables involving Pap smear testing. Cervical cancer risk status, number of lifetime sexual partners, sexual intercourse before the age of 18, and smoking status were all highly associated with whether a woman was within risk-appropriate cervical cancer screening guidelines. Because cervical cancer risk status is a combination of components, including the number of lifetime sexual partners, age at first intercourse, history of STI, and smoking status, the overall risk covariate was not included in the selection procedure. Instead, the separate components of risk status were modeled as individual covariates in the forward selection process.
Table 2.
Relationship between variables pertaining to cervical cancer risk and risk-appropriate cervical cancer screening status in Ohio Appalachia, 2006 (n=562)
| Within Risk-Appropriate Cervical Cancer Screening Guidelines, n (%) | |||||
|---|---|---|---|---|---|
| Predictor Variable | No | Yes | Totala | ORcrude | 95% CI |
| Cervical Cancer Risk Status | |||||
| Low | 12 (7%) | 105 (28%) | 117 (21%) | 1.00 | |
| High | 170 (93%) | 268 (72%) | 438 (79%) | 0.16 | [0.08, 0.30] |
| Individual Risk Status Components | |||||
| 5 or more lifetime sexual partners | |||||
| No | 87 (48%) | 230 (61%) | 317 (56%) | 1.00 | |
| Yes | 95 (52%) | 150 (39%) | 245 (44%) | 0.59 | [0.41, 0.85] |
| Had sex before age 18 | |||||
| No | 48 (27%) | 161 (44%) | 209 (38%) | 1.00 | |
| Yes | 133 (73%) | 209 (56%) | 342 (62%) | 0.43 | [0.29, 0.65] |
| Current smoker | |||||
| No | 119 (65%) | 288 (76%) | 407 (72%) | 1.00 | |
| Yes | 63 (35%) | 92 (24%) | 155 (28%) | 0.62 | [0.42, 0.93] |
| Previous sexual partners with sexually transmitted infection | |||||
| No | 148 (88%) | 300 (88%) | 448 (88%) | 1.00 | |
| Yes | 21 (12%) | 41 (12%) | 62 (12%) | 0.98 | [0.55, 1.77] |
| Ever diagnosed with venereal warts, condylomas, or HPV | |||||
| No | 165 (91%) | 346 (91%) | 511 (91%) | 1.00 | |
| Yes | 16 (9%) | 33 (9%) | 49 (9%) | 0.90 | [0.47, 1.72] |
| Ever diagnosed with genital herpes, syphilis, or gonorrhea | |||||
| No | 165 (91%) | 361 (95%) | 526 (94%) | 1.00 | |
| Yes | 16 (9%) | 18 (5%) | 34 (6%) | 0.52 | [0.25, 1.07] |
| Belief Scale Score | |||||
| Negative (−5 to 1) | 71 (39%) | 149 (39%) | 220 (39%) | 1.00 | |
| Positive (2 to 5) | 111 (61%) | 230 (61%) | 341 (61%) | 0.95 | [0.65, 1.38] |
| Knowledge Scale Score | |||||
| Poor to Fair (−14 to 4) | 51 (28%) | 87 (23%) | 138 (25%) | 1.00 | |
| Good (5 to 6) | 31 (17%) | 95 (25%) | 126 (22%) | 1.62 | [0.94, 2.82] |
| Very Good (7 to 8) | 40 (22%) | 95 (25%) | 135 (24%) | 1.25 | [0.74, 2.10] |
| Excellent (9 to 14) | 60 (33%) | 102 (27%) | 162 (29%) | 0.88 | [0.53, 1.46] |
| Barriers Scale Score (0 to 10) | |||||
| Mean (Range) | 1.93 (0–7) | 1.67 (0–6) | 438 (100%) | 0.90b | [0.77, 1.05] |
| Previous Abnormal Pap Smear | |||||
| No | 125 (70%) | 239 (63%) | 364 (65%) | 1.00 | |
| Yes | 54 (30%) | 140 (37%) | 194 (35%) | 1.25 | [0.84, 1.86] |
ORcrude is the crude (unadjusted) odds ratio of being within risk-appropriate cervical cancer screening guidelines versus not, including clinic location as a random effect.
95% CI is the corresponding 95% confidence interval for the crude odds ratio.
HPV is human papillomavirus
May not sum to 562 due to missing data
Odds ratio for a 1-point increase in the barriers scale score
Odds ratios and 95% confidence intervals from the final multivariable model (n=479, after exclusion of incomplete cases) are shown in Table 3. Socioeconomic status, number of major life events, and perceived financial status compared to the participant’s parents were all significantly associated with risk-appropriate screening status. In addition, there was a significant interaction between marital status and whether the participant first had sexual intercourse before age 18 (p = 0.009).
Table 3.
Final model for variables associated with being within risk-appropriate cervical cancer screening guidelines in Ohio Appalachia, 2006 (n=479)
| Predictor Variable | ORadj | 95% CI | p-value |
|---|---|---|---|
| Socioeconomic Status (SES, by category) | |||
| 0–1 | 1.00 | ||
| 2–3 | 3.39 | [1.85, 6.21] | <0.001 |
| 4–6 | 3.86 | [2.03, 7.34] | <0.001 |
| Decrease of one major life event on Life Stressor Checklist—Revised (LSC-R) | 1.09 | [1.04, 1.15] | 0.001 |
| Finances compared to parents at the same age | |||
| About the same | 1.00 | ||
| Better off | 0.41 | [0.23, 0.73] | 0.002 |
| Worse off | 0.49 | [0.24, 0.98] | 0.042 |
| Interaction: Marital Status × Sexual Intercourse Before Age 18 | 0.009a | ||
| Among women who reported having had sex before age 18 | |||
| Never married | 1.00 | ||
| Divorced/widowed/separated | 0.19 | [0.08, 0.46] | <0.001 |
| Married/couple | 0.24 | [0.11, 0.53] | <0.001 |
| Among women who reported not having had sex before age 18 | |||
| Never married | 1.00 | ||
| Divorced/widowed/separated | 1.12 | [0.33, 3.87] | 0.852 |
| Married/couple | 1.97 | [0.64, 6.06] | 0.235 |
| Among women who have never been married | |||
| Reported having had sex before age 18 | 1.00 | ||
| Reported not having had sex before age 18 | 0.32 | [0.09, 1.08] | 0.067 |
| Among women who have been divorced/widowed/separated | |||
| Reported having had sex before age 18 | 1.00 | ||
| Reported not having had sex before age 18 | 1.92 | [0.76, 4.82] | 0.165 |
| Among women who are married/couple | |||
| Reported having had sex before age 18 | 1.00 | ||
| Reported not having had sex before age 18 | 2.63 | [1.47, 4.71] | 0.001 |
ORadj is the adjusted odds ratio (for all variables presented in the final model) of being within risk-appropriate cervical cancer screening guidelines versus not including clinic location as a random effect. All other variables were considered in the modeling process but did not exert independent, confounding, or modifying effects.
95% CI is the corresponding 95% confidence interval for the adjusted odds ratio.
p-value is from a likelihood ratio Chi-square test for the interaction effect
The final model showed good fit (Hosmer-Lemeshow goodness-of-fit test p=0.795) (Hosmer and Lemeshow, 2000) and acceptable discrimination (area under the ROC curve = 0.767). Fractional polynomials suggested appropriate modeling of continuous variables.
DISCUSSION
Recent cervical cancer incidence rates among counties in Ohio Appalachia ranged from 11.4 to 20.3 per 100,000 women, which was noticeably higher than the national rate of 9.6 per 100,000 women (Ohio Cancer Incidence Surveillance System, 2003). One of the possible reasons for the elevated cervical cancer rates observed in this region is the underutilization of the Pap smear. The main focus of this research was to identify factors associated with being within risk appropriate cervical cancer screening guidelines among women in Ohio Appalachia.
Results indicate that multiple factors were significantly associated with Pap smear screening status. Socioeconomic status and the number of major life events experienced had the strongest associations with whether or not a woman was within risk-appropriate screening guidelines. A woman’s financial status compared to her parents at the same age was also found to be associated with the outcome. Lastly, there was an interesting interaction between marital status and if a woman had sexual intercourse before the age of 18. Among women who reported never being married, those who did not have sexual intercourse before age 18 had significantly lower odds of being within screening guidelines compared to those that did have sex before age 18. Additionally, among women who reported having had sex before age 18, those who were married and those who had been divorced, widowed, or separated had lower odds of being within risk appropriate guidelines than women who were never married.
Previous studies have examined predictors of Pap smear testing using a variety of data sources and a wide range of individual- and community-level variables. These earlier studies found that a woman’s marital status, socioeconomic status, education, insurance status, smoking status, physical and mental heath, and family income were all significantly related to being within recommended cervical cancer screening guidelines (Snider, Beauvais et al., 1996; Lee, Parsons et al., 1998; Snider and Beauvais, 1998; Simoes, Newschaffer et al., 1999; Fontaine, Heo et al., 2001; Maxwell, Bancej et al., 2001; Cyrus-David, Michielutte et al., 2002; Ostbye, Greenberg et al., 2003; Sambamoorthi and McAlpine, 2003; Carrasquillo and Pati, 2004; Coughlin, Uhler et al., 2004; Diab and Johnston, 2004; Lockwood-Rayermann, 2004; Pirraglia, Sanyal et al., 2004; Ferrante, Chen et al., 2006; Akers, Newmann et al., 2007; Blackwell, Martinez et al., 2008). This study highlights some of these same findings in a unique population of Appalachian women (i.e., that poorer, less educated, and non-married women were generally more likely to be outside of cervical cancer screening guidelines). In addition, this is the first study to our knowledge to highlight the potential association between major life events and adherence to cervical cancer screening guidelines. Furthermore, this study highlights a potential modifying effect involving age at first intercourse and marital status regarding risk-appropriate screening guidelines that has not been otherwise reported. While previous research has found age to be a significant predictor of cervical cancer screening compliance (Lockwood-Rayermann, 2004), after controlling for other multilevel factors, this study did not.
The finding that generational variation in financial status was associated with being within risk-appropriate cervical cancer screening guidelines is also unique to this study. Thus, not only is a woman’s own SES an important correlate of Pap smear compliance, but a woman’s financial status relative to her parents was also associated with risk-appropriate screening compliance, perhaps suggesting that early life influences exert an effect on cancer screening behavior. Lower odds of being within risk appropriate screening guidelines for women who reported their finances were worse off than their parents at the same age (compared to approximately the same) may be the result of having financial barriers to access, lack of knowledge about the need for regular screening, or facing added access barriers related to low SES. Contrastingly, lower odds of being within guidelines for women who reported being financially better off than their parents (versus approximately the same) may be due to financial disqualification from government sponsored screening programs like the Breast and Cervical Cancer Early Detection Program (BCCEDP) or ineligibility for Medicaid. These differences in screening adherence based on intergenerational financial differences should be examined in future studies.
Lastly, this study contributes to existing literature because variables considered in this analysis were selected from a multilevel framework (Warnecke, Oh et al., 2008) encompassing a robust range of social and cultural factors, material factors, health and health behaviors, psychological factors, and environmental factors. In determining compliance with cervical cancer screening guidelines, it is imperative to evaluate compliance in accordance with a woman’s risk for developing cervical cancer. Notably, this study evaluated risk status simultaneously in investigating correlates of compliance with Pap smear testing guidelines. Aside from Cyrus-David et al. (Cyrus-David, Michielutte et al., 2002), this is the only study to evaluate a woman’s Pap smear compliance in a risk-appropriate fashion. Our results are similar to those of Cyrus-David et al. in terms of the proportion of women determined to be at high risk for cervical cancer, but this study had higher risk-appropriate screening rates. This may be attributable to the multiracial composition of the Cyrus-David et al. study, and their use of only low-income clinic populations. Additionally, this study refrained from modeling risk status as a predictor variable to prevent high correlation with other independent variables from which a woman’s risk status was determined.
Strengths and Limitations
This study has several strengths. Women were randomly selected from clinics to participate, and eligibility was confirmed using medical record review. Additionally, this study identified whether women were within screening guidelines based on their cervical cancer risk status. Furthermore, we captured information that has not been traditionally measured in previous studies focusing on the receipt of Pap smears. These factors, like major life events, extend our current knowledge about women who comply or do not comply with cervical cancer screening guidelines.
Limitations of this report include that it is cross-sectional, and was limited to women who had a medical visit in the past two years at one of the study clinics. In addition, the participation rate was 71.3%. Thus, the women in this study may not represent all women living in the Appalachia Ohio region. Another limitation stems from the reliance on self-reported date of Pap smear for 23.8% of the women, which may vary when compared to medical record. The validation rate of 80% for self-reports, however, is encouraging.
CONCLUSIONS
Among a population of women who reside in a region with high cervical cancer rates, an observational study found a relationship between social and psychological factors and being within risk-appropriate guidelines for Pap smear utilization. Risk-appropriate screening intervals could actually better reflect clinical practice recommendations (American College of Obstetricians and Gynecologists, January 2008) and may offer a way to reduce cervical cancer incidence and mortality in high risk populations, like women in Appalachia. While it is certainly important to promote cervical cancer screening adherence among all women, it is equally important that providers have heightened vigilance among women identified to be at high risk of this disease. This information can be used to develop and test risk-appropriate interventions to improve cervical cancer screening with the overall goal of lowering cervical cancer incidence and mortality rates in Appalachian Ohio.
Acknowledgments
The authors would like to acknowledge support from the following grants: P50 CA105632 (EDP); P30 CA016058 (Behavioral Measurement Shared Resource at The Ohio State University Comprehensive Cancer Center); and K07 CA107079 (MLK)
Footnotes
Conflicts of Interest: None
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References Cited
- Akers AY, Newmann SJ, et al. Factors underlying disparities in cervical cancer incidence, screening, and treatment in the United States. Curr Probl Cancer. 2007;31(3):157–81. doi: 10.1016/j.currproblcancer.2007.01.001. [DOI] [PubMed] [Google Scholar]
- American College of Obstetricians and Gynecologists. ACOG committee opinion. Recommendations on frequency of Pap test screening. Number 152--March 1995. Committee on Gynecologic Practice. American College of Obstetricians and Gynecologists. Int J Gynaecol Obstet. 1995;49(2):210–1. [PubMed] [Google Scholar]
- American College of Obstetricians and Gynecologists. New Evidence-Based Screening Guidelines for Pap Tests. Washington, D.C: The American College of Obstetricians and Gynecologists; Jan, 2008. [Google Scholar]
- Anderson LA, Dedrick RF. Development of the Trust in Physician scale: a measure to assess interpersonal trust in patient-physician relationships. Psychol Rep. 1990;67(3 Pt 2):1091–100. doi: 10.2466/pr0.1990.67.3f.1091. [DOI] [PubMed] [Google Scholar]
- Beck AT, Steer RA. Beck anxiety inventory manual. San Antonio, TX: Psychological Corporation; 1990. [Google Scholar]
- Blackwell DL, Martinez ME, et al. Women’s compliance with public health guidelines for mammograms and pap tests in Canada and the United States: an analysis of data from the Joint Canada/United States Survey Of Health. Womens Health Issues. 2008;18(2):85–99. doi: 10.1016/j.whi.2007.10.006. [DOI] [PubMed] [Google Scholar]
- Carrasquillo O, Pati S. The role of health insurance on Pap smear and mammography utilization by immigrants living in the United States. Prev Med. 2004;39(5):943–50. doi: 10.1016/j.ypmed.2004.03.033. [DOI] [PubMed] [Google Scholar]
- Cohen S, Kamarck T, et al. A global measure of perceived stress. J Health Soc Behav. 1983;24(4):385–96. [PubMed] [Google Scholar]
- Coughlin SS, Uhler RJ, et al. Nonadherence to breast and cervical cancer screening: what are the linkages to chronic disease risk? Prev Chronic Dis. 2004;1(1):A04. [PMC free article] [PubMed] [Google Scholar]
- Cyrus-David MS, Michielutte R, et al. Cervical cancer risk as a predictor of Pap smear use in rural North Carolina. J Rural Health. 2002;18(1):67–76. doi: 10.1111/j.1748-0361.2002.tb00878.x. [DOI] [PubMed] [Google Scholar]
- Diab ME, Johnston MV. Relationships between level of disability and receipt of preventive health services. Arch Phys Med Rehabil. 2004;85(5):749–57. doi: 10.1016/j.apmr.2003.06.028. [DOI] [PubMed] [Google Scholar]
- Ferrante JM, Chen PH, et al. Breast and cervical cancer screening in obese minority women. J Womens Health (Larchmt) 2006;15(5):531–41. doi: 10.1089/jwh.2006.15.531. [DOI] [PubMed] [Google Scholar]
- Fontaine KR, Heo M, et al. Body weight and cancer screening among women. J Womens Health Gend Based Med. 2001;10(5):463–70. doi: 10.1089/152460901300233939. [DOI] [PubMed] [Google Scholar]
- Hollingshead A. A four-factor index of social status. New Haven, CT: Department of Sociology, Yale University; 1975. [Google Scholar]
- Hosmer DW, Lemeshow S. Applied Logistic Regression. Hoboken, NJ: Wiley & Sons, Inc; 2000. [Google Scholar]
- Lee J, Parsons GF, et al. Falling short of Pap test guidelines. Health Rep. 1998;10(1):9–19(ENG). 9–21(FRE). [PubMed] [Google Scholar]
- Lockwood-Rayermann S. Characteristics of participation in cervical cancer screening. Cancer Nurs. 2004;27(5):353–63. doi: 10.1097/00002820-200409000-00004. [DOI] [PubMed] [Google Scholar]
- Marmont M, Wilkinson R. Social determinants of health. Oxford: Oxford Press; 1999. [Google Scholar]
- Maxwell CJ, Bancej CM, et al. Factors important in promoting cervical cancer screening among Canadian women: findings from the 1996–97 National Population Health Survey (NPHS) Can J Public Health. 2001;92(2):127–33. doi: 10.1007/BF03404946. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ohio Behavioral Risk Factor Surveillance System. Prevalence Data: Women aged 18+ who have had a pap test within the past three years. Centers for Disease Control; 2006. [Google Scholar]
- Ohio Cancer Incidence Surveillance System. Cervical cancer in Ohio, 1996–2000. Columbus, OH: Ohio Department of Health; 2003. [Google Scholar]
- Ostbye T, Greenberg GN, et al. Screening mammography and Pap tests among older American women 1996–2000: results from the Health and Retirement Study (HRS) and Asset and Health Dynamics Among the Oldest Old (AHEAD) Ann Fam Med. 2003;1(4):209–17. doi: 10.1370/afm.54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Paskett E, Tatum C, et al. Randomized trial of an intervention to improve mammography utilization among a triracial rural population of women. J Natl Cancer Inst. 2006;98(17):1226–37. doi: 10.1093/jnci/djj333. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Paskett ED, Rushing J, et al. Cancer screening behaviors of low-income women: the impact of race. Womens Health. 1997;3(3–4):203–26. [PubMed] [Google Scholar]
- Pirraglia PA, Sanyal P, et al. Depressive symptom burden as a barrier to screening for breast and cervical cancers. J Womens Health (Larchmt) 2004;13(6):731–8. doi: 10.1089/jwh.2004.13.731. [DOI] [PubMed] [Google Scholar]
- Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. Applied Psychological Measurement. 1977;1:385–401. [Google Scholar]
- Sambamoorthi U, McAlpine DD. Racial, ethnic, socioeconomic, and access disparities in the use of preventive services among women. Prev Med. 2003;37(5):475–84. doi: 10.1016/s0091-7435(03)00172-5. [DOI] [PubMed] [Google Scholar]
- Sampson RJ, Raudenbush SW, et al. Neighborhoods and violent crime: a multilevel study of collective efficacy. Science. 1997;277(5328):918–24. doi: 10.1126/science.277.5328.918. [DOI] [PubMed] [Google Scholar]
- Simoes EJ, Newschaffer CJ, et al. Predictors of compliance with recommended cervical cancer screening schedule: a population-based study. J Community Health. 1999;24(2):115–30. doi: 10.1023/a:1018754307718. [DOI] [PubMed] [Google Scholar]
- Snider J, Beauvais J, et al. Trends in mammography and Pap smear utilization in Canada. Chronic Dis Can. 1996;17(3–4):108–17. [PubMed] [Google Scholar]
- Snider JA, Beauvais JE. Pap smear utilization in Canada: estimates after adjusting the eligible population for hysterectomy status. Chronic Dis Can. 1998;19(1):19–24. [PubMed] [Google Scholar]
- Taylor TR, Kamarck TW, et al. Validation of the Detroit Area Study Discrimination Scale in a community sample of older African American adults: the Pittsburgh healthy heart project. Int J Behav Med. 2004;11(2):88–94. doi: 10.1207/s15327558ijbm1102_4. [DOI] [PubMed] [Google Scholar]
- U.S. Department of Health and Human Services. Healthy People 2010: Understanding and Improving Health. 2. Washington, DC: Government Printing Office; Nov, 2000. [Google Scholar]
- Wallston K, Malcarne VL, Flores L, Hansdottir I, Smith CA, Stein MJ, et al. Does God determine your health? The God locus of health control scale. Cognitive Therapy and Research. 1999;23(2):131–142. [Google Scholar]
- Warnecke RB, Oh A, et al. Approaching Health Disparities From a Population Perspective: The National Institutes of Health Centers for Population Health and Health Disparities. Am J Public Health. 2008 doi: 10.2105/AJPH.2006.102525. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wolfe J, Kimerling R. Assessing psychological trauma and PTSD. New York: Guilford Press; 1997. Gender issues in the assessment of posttraumatic stress disorder. [Google Scholar]