Stage 3 – Quantitative Data Collection and Analyses

Correlates of Cervical Cancer Screening among Women in Taiwan

---

Su-I Hou, DrPH, CPH, MCHES, RN

Maria Fernandez, PHD

Pai-Ho Chen, BA

Abstract

This study assessed correlates of cervical cancer screening behavior among Chinese women in Taiwan (N=125).  Results showed that 30% of the sampled women had never received a Pap test, and that only 58% were adherent to the recommended screening guidelines.  Intention to have a Pap test in the coming year was higher among women reporting a recent Pap test (90%) than women reporting no Pap test in last three years (58%).  Multiple logistic regression analysis showed significant associations between screening adherence and women’s knowledge, perceived pros, cons, and norms of a Pap test. Measurement instrument assessing these factors revealed Cronbach alpha as 0.70 for knowledge scale, 0.88 for pros scale, 0.68 for cons scale, and 0.72 for perceived norms scale.  In addition to identify several psychosocial factors associated with Pap test screening, this study also provided a basis for measuring these factors among Chinese women.

Keywords: Pap test screening, psychosocial factors, instrument development, Chinese women

Background

According to National Institute of Health Consensus Development Panel (NIHCDP, 1996), cervical cancer is the second most common cancer among women worldwide, with approximately 471,000 new cases reported each year.  Data obtained from the Cancer Registry in Taiwan (1995) showed that cervical cancer is the leading malignant neoplasm in Taiwan (incidence rate 32.49/100,000), followed by breast cancer (24.37/100,000) and colon cancer (17.22/100,000).  About 40% of the women in Taiwan die from cancer and about 10% of all the cancer deaths are due to cervical cancer.

Due to the typically long pre-clinical phase of cervical cancer, screening for early detection is very important. Timely detection and treatment of precancerous cervical lesions identified by Papanicolaou (Pap) screening can reduce the possibility of developing cervical cancer, and is the best method for reducing the incidence and mortality of invasive cervical cancer (NIHCDP, 1996).  In the US, regular check-ups and Pap screening have contributed greatly to the 70% decrease in the death rate over the past 50 years (CDC, 1998).

Since there have not been many studies conducted in Taiwan regards women’s Pap screening behavior, other studies investigated Asian populations in the US could provide us some useful insights. Despite the proven efficacy of the Pap test, Pap test screening rates are low among Asian populations in the US. The percentage of Asian women living in the US, who having at least one Pap test, ranges from 46% to 56% (Yi, 1994; Pham & McPhee, 1992; Hiatt & Pasick, 1995).  In contrast, Hiatt and Pasick found that the prevalence of women having at least one Pap test was 98% among whites and blacks, 97% among Latinos. Among women in Taiwan, the very few existing studies reported the similar low rates of Pap screening utilization as Asian women in the US ranged from 58% to 62% (Wang & Lin, 1996; Lee, Kuo, & Chou, 1997).

The reasons that women do not have regular Pap screening according to the CDC’s National Breast and Cervical Cancer Early Detection Program (NBCCEDP), common barriers reported among women in the US including fear, embarrassment, cost, transportation, communication barriers, lack of physician referral, lack of child care, and lack of time.  The reasons that Asian women do not go for a Pap screening might be a little bit different because even the Asian population constituted only about 2~3% in the sample of the study.  Cultural differences could limit generalization of these barriers to Asian women.  In recent years, there have been only a few studies that have identified factors that hinder Asian women in the US from obtaining a Pap test.  These studies showed that some of the factors including the same concerns as other women in the US like fear and embarrassment, but others were different which included marital status, age, knowledge, perceived risk, discomfort, and the gender of physicians (Pham & McPhee, 1992; Kelly et al, 1996; Seow et al, 1995; Yi, 1998).

In Taiwan, National Health Insurance Plan (1996) provides free health care coverage for Pap screening to women aged 30 and over.  Despite universal coverage for these women, most women do not obtain regular Pap smears. Wang and Lin (1996) examined Pap screening service utilization among women aged 20 and older in Taiwan in 1993. Their data indicated that 40% of the sampled women in Taipei had never had a Pap test.  Given the high incidence and mortality rate of cervical cancer in Taiwan, the low screening prevalence, and the availability of an inexpensive screening method, it is important to understand the psychosocial factors that influence women’s Pap screening behavior.  This paper describes the results from a pilot study conducted in spring, 1999 in Taichung, Taiwan.  This study was conducted to: (a) identify factors associated with Pap test screening behavior, and (b) test the scale reliabilities and refine the preliminary survey instrument.

Materials and Methods

Sample

The study population consisted of a convenience sample of female family members of inpatients who were admitted to one of the major teaching hospitals in Taichung, Taiwan in spring, 1999 (N=125). A self-administered questionnaire, consisting of 73 items, was distributed to 160 women.  Female family members of inpatients were asked to return the questionnaire to the nurse’s station after they completed the survey. The response rate was 78% (125/160).  Those who declined participation gave reasons such as they were not interested or had low literacy level.

Measurement

The survey consisted of items in six major categories: (1) prior screening experience; (2) perceived pros and cons of a Pap test; (3) perceived norms; (4) perceived risk (susceptibility) to cervical cancer; (5) knowledge related to cervical cancer and Pap screening; and (6) demographic variables.  The categories described above represent constructs derived from existing models of health behavior.  These included perceived pros/cons from the Transtheoretical Model (Prochaska, Norcross, & DeClemente, 1994), perceived norms from the Theory of Planned Behavior (Ajzen & Fishbein, 1980), and perceived risk (susceptibility) from the Health Belief Model (Rosenstock, 1974).  Some pros/cons items were adapted from previous studies (Pearlman et al., 97; Rakowski et al., ‘92, ’93, ’97), with additional items developed from the focus groups conducted among women in Taiwan.  Some of the knowledge items were adapted and translated from unpublished thesis by Huang (1996) from National Taiwan University. The remainder knowledge items, susceptibility and perceived norms items were developed by the principal investigator, based on theories as well as focus group findings obtained from Chinese women.  Scale items were developed in English, translated into Chinese, and back translated. Items on the two English versions were checked for discrepancies of meaning.

Methods

Prior to administer the pilot survey, three focus groups were conducted to ensure the appropriateness of the content as well as to validate the specific constructs selected in the preliminary questionnaire. Participants of the focus groups were recruited from churches and community centers in Taichung, Taiwan through local community contact persons (N=24).  A semi-structured focus group protocol with open-ended and probing questions was used to explore potential behavior and environmental factors that might influence Pap screening behavior in this particular population. Potential intervention ideas were also collected.  Focus groups were used because of the method’s unique advantage of generating data from group interaction rather than obtaining information produced merely from individual interviews (Krueger, 1994).  Appropriate revisions of the survey items were made, and four new items of perceived pros and two new items of perceived cons were added in the questionnaire based on the feedback from the focus group before the pilot survey was conducted.  The final version of the pilot survey instrument consisted of 11 pros scale items, 11 cons scale items, 4 perceived norms scale items, 2 perceived risk items, and 17 knowledge scale items.

Scoring of the scales

The items making up the perceived pros/cons, perceived norms, and susceptibility scales were 5-point Likert scale items. Item responses ranged from 1 (strongly disagree) to 5 (strongly agree).  The total score calculated for each particular scale was the sum of the item scores in that scale.  Knowledge scale items were scored one (1) if answered correctly, and were scored zero (0) when answered wrong, or when the response was “do not know.”  A total of 16 items were included in the final knowledge scale analysis.  One knowledge item was excluded from the scale analysis because almost every woman answered it correctly.  The possible range for the knowledge scale was from 0 to 16.

Data Analysis

Outcome of interest in the study was Pap test adherence, defined as women who had their Pap screenings within the past three years.  Women who had never had a Pap test, had not had one in the past three years, or could not remember when their last Pap test was, were classified as non-adherent. Categorical variables were analyzed by cross-tabulations using chi-square statistics.  Continuous variables were analyzed using t-tests in two-group comparisons.  Univariate analyses were conducted to examine the effects of demographic variables on screening behavior.  Variables with significant associations (p<0.05) were included in the multivariate model.  Logistic regression modeling was used to analyze the effect of each factor on screening participation, after adjusting for the confounding variables.

Results

Sample Characteristics

A total of 125 women completed the pilot survey. The mean age was 37.93 (SD=10.45). Most women in the sample were married (90%).  About 62% of the women worked full time, and 32% of the women had college education level or higher.  Table 1 describes demographics of women who participated in the study according to their screening adherence.  There were no significant differences between adherent and non-adherent women in age, employment status, or education.  Women in both groups showed similar distribution on most demographic characteristics.  Marital status was the only variable that showed significant association with Pap screening adherence.  A higher proportion of married women (62%) reported a Pap test within the preceding 3 years, as compared to non-married women (31%).

Table 1. Demographics of the sampled women

Scale Reliability

Internal consistency reliability was calculated for the pros, cons, norms, and perceived risk scales.  Items with discrimination (corrected-item-total correlation CITC) less than 0.2 were re-evaluated for their appropriateness. Items with low correlation within the scale and not considered crucial to the scale construct were removed from scale analysis.  Both internal consistency and item difficulty (mean) were used to evaluate knowledge scale items.  Item means that fell in the range of 0.3~0.7 were considered moderately difficult items (Nunnally & Bernstein, 1994).  A summary of the reliabilities of the original and revised scales is presented in Table 2.

Pros scale

All the CITC of the 11 items in the pros scale fell between 0.3~0.7 indicating that all of the pros items showed sufficient correlation with other items in the same scale.  The Cronbach alpha for the pros scale indicated reliable internal consistency (a=0.8751).

Cons scale

After the original cons scale was reviewed (11 items, Cronbach alpha=0.6604), two items, “cancer treatment is worse than the cancer itself” and “I would rather not know if I had cancer,” were dropped from the scale analysis. Data also showed both items had low CITC (<0.2), indicating these two items had low correlation with other cons items.  It was determined that although these two items may represent barriers to screening, they were not specific “costs” or “cons” of a Pap test itself.  Although the item “I would feel more comfortable to obtain a Pap if a female doctor does the procedure” had a low CITC and the internal consistency (a) would increase if this item were deleted, the question was considered an important aspect that could influence a woman’s Pap screening behavior, and was therefore retained. The revised cons scale included 9 items with Chronbach alpha as 0.6817.

Norms scale

All the CITC of the items in the norms scale fell between 0.3~0.7.  Analysis results indicated all the norm items were appropriate and showed high correlation within the scale.  The Cronbach alpha for the norm scale was 0.7158 (Table 2).

Susceptibility items

Items in the survey related to perceive susceptibility included “I might get cervical cancer at some point during my life,” and “I feel that my chance of getting cervical cancer is high.”  Although there were only two items measuring perceived susceptibility of cervical cancer, the Cronbach alpha was acceptable (0.6849).

Knowledge scale

The original knowledge scale contained 17 items (a=0.6971).  After the scale was re-evaluated, the item “Cervical cancer can be detected early with Pap screenings” was removed because it had a negative CITC.  Additionally, since 99% of the women answered this item correctly (mean=0.99), it did not add to the scale’s ability to discriminate between low and high levels of knowledge. Items “Women who are post-menopausal still need a Pap test” and “A Pap test screening only finds problems when they are too far away to be treated” were re-visited because they had CITC less than 0.2.  Although the former item was negatively correlated with five items, the strengths of the correlation were very small.  Both of these items were considered important knowledge that could influence women’s Pap screening behavior.  Keeping the two items in the scale did not compromise the internal consistency significantly.  The remaining items showed sufficient correlation with other items in the knowledge scale (CITC>0.2).  The final revision of the knowledge scale consisted of 16 items with one item removed from the original scale (a=0.7010).

Table 2. Reliability of the pros, cons, norms, perceived risk, and knowledge scales

Scales	Number of items in original scale	Original scale Chronbach α	Number of items in revised scale	Revised scale Chronbach α
Pros	11	0.8751	No revision was made
Cons	11	0.6604	9	0.6817
Norms	4	0.7158	No revision was made
Perceived Risk	2	0.6849	Add more items in the future
Knowledge	17	0.6971	16	0.7010

Knowledge scale item difficulty analysis

Of the 17 knowledge items, 12 (70%) had a mean between 0.8-0.9, indicating that the sampled women answered the majority of the questions on the knowledge scale correctly.  The remaining 5 items showed moderate difficulty, with item means between 0.4-0.6. Table 3 lists the item means (item difficulty) and standard deviation for each knowledge item.  The knowledge scale was considered relatively easy for the sampled women in Taiwan because it mixed 70% easy and 30% moderately difficult items.

Table 3. Item difficulty analysis of the knowledge scale

	Item	Mean (% Correct)	SD
Q42.	Cervical cancer is the most common cancer among women in Taiwan.	0.904	0.2958
Q43.	People having sex without condoms have higher risk of getting cervical cancer.	0.504	0.502
Q44.	Pap screening is done through drawing your blood sample.	0.888	0.3166
Q45.	Women who are postmenopausal do not need Pap screenings.	0.88	0.3263
Q47.	If you have early stage cervical cancer, you will feel pain.	0.536	0.5007
Q48.	Women who do not have regular Pap screenings are more likely to have advanced cervical cancer when they were diagnosed.	0.624	0.4863
Q49.	After women stop having children, they do not need Pap smears.	0.936	0.2457
Q50.	A Pap is most helpful when you have one every year or two.	0.936	0.2457
Q51.	Healthy adult women should have Pap screenings every year.	0.944	0.2308
Q52.	A Pap is not important for a woman at my age.	0.936	0.2457
Q53.	Only women who have had many sex partners need to get a Pap.	0.928	0.2595
Q54.	A Pap screening only finds problems when they are too far along to be treated.	0.408	0.4934
Q55.	A Pap screening is necessary even if there is no family history of cancer.	0.92	0.2724
Q56.	Once I have a negative Pap, I do not need to have any more.	0.952	0.2146
Q57.	I need a Pap smear only when I experience vaginal bleeding other than menstruation.	0.92	0.2724
Q58.	A Pap smear exam is the same as a cervix biopsy.	0.592	0.4934

Prior Pap screening behavior and intention to have a Pap test in the coming year

Seventy percent of the women (88/125) in the sample reported they had previously had a Pap test in the last three years.  Among these women, 60% (53/88) reported they had one Pap test in the last year, and 23% (20/88) reported they had one Pap test in the preceding 1-3 years.  The mean number of prior Pap tests obtained was 2.42 among women had previous Pap tests. Fifty-eight percent of all the sampled women (73/125) were adherent to the Pap screening guidelines (women who had their Pap screenings within the past three years).

Intention to have a Pap test in the coming year was 76% (95/125) among all the women in the study (Table 4).  A higher proportion of adherent women reported an intention to have a Pap test in the coming year compared with women in the non-adherent group. Almost 90% (65/73) of the adherent women reported that they intended to have a Pap, while only 58% (30/52) of the non-adherent women showed the intention (p=0.000).

Table 4. Prior pap screening behavior

	N (%)
Pap Screening
Never	37 (29.6%)
Ever	88 (70.4%)
– Within 1 year	53 (60.2%)
– Between 1-3 years	20 (22.7%)
– More than 3 years ago	8 (9.1%)
– Don’t remember	7 (7.9%)
Total	125 (100%)
	Nonadherent	Adherent	Total
Intention of Pap in coming year
Yes	30 (57.7%)	65 (89%)	95 (76%)
No	22 (42.3%)	8 (11%)	30 (24%)
Total	52 (100%)	73 (100%)	125 (100%)

 

Univariate Analysis

The knowledge scale mean was 13.25 among adherent women and 12.19 among non-adherents. The mean score of each knowledge item was 0.83 (item mean = scale mean/scale item = 13.25/16) among adherent women and 0.76 among non-adherents.  Adherent women had a significantly higher knowledge score than non-adherents (p=0.025) (Table 5).

Table 5. Summary of scale statistics (N=125)

Scale	Number of Items	Scale Mean (SD)	Scale Min-Max	P-Value (t test)
Knowledge
Adherent		13.25 (1.93)
Nonadherent	16	12.19 (2.90)	3–16	0.025*
All Women		12.81 (2.43)
Pros
Adherent		47.37 (4.46)
Nonadherent	11	45.44 (6.37)	30-55	0.069
All women		46.57 (5.39)
Cons
Adherent		25.65 (4.8)
Nonadherent	9	28.76 (5.27)	17-38	0.001*
All women		26.96 (5.22)
Norms
Adherent		16.32 (2.15)
Nonadherent	4	15.48 (2.16)	11-20	0.035*
All women		15.97 (2.18)

The mean score on the Pros scale (positive aspects of a Pap test) was 46.57  (11 items) among all the sampled women, with scale range between 30-55.  Women’s average endorsement level of each pros item was 4.2.  Although the difference was not significant between adherent and non-adherent women, it approached significance (p=0.069).

The mean score on the cons scale (negative aspects of a Pap test) was 25.65  (9 items) among adherents and 28.76 among non-adherents, with scale range between 17-38 (p=0.001).  Women who were non-adherent to Pap test screening tended to perceive more negative aspects of obtaining a Pap test than adherent women.

The mean score on the norm scale (4 items) differed significantly between adherents (scale mean=16.32) and non-adherents (scale mean=15.48) as well (p=0.035). Women’s average endorsement level of each norm item was 4.08 among the adherent group and 3.87 among the non-adherent group.  Women who adhered to screening guidelines, on average, rated “agree” or “strongly agree” on norm items; while women who were non-adherent to Pap test screening often chose the responses of “not sure” or “agree”.

Logistic Regression Analysis

Multiple logistic regression models were used to assess the relationships between independent variables (constructs) and screening behavior (adherent versus non-adherent).  The analysis included marital status as a covariable since it showed a significant association with Pap screening behavior in the univariate analysis.  The analysis revealed significant associations between knowledge, perceived pros, cons, and norms with women’s Pap screening behavior (Table 6).

Although the univariate analysis of pros scale did not show a significant association with Pap screening behavior (p=0.069), after marital status was added to the regression model, the association became significant (p=0.041).  The adjusted odds ratio (OR) was 1.0781 [95% CI (1.0031, 1.1588)].

The cons scale showed significant association with screening status in the univariate analysis.  After adjusted for marital status, the association is even stronger (p=0.0009).  The OR was 0.8713 [95% CI (0.8034, 0.9045)].

Both norms and knowledge scales were strongly associated with women’s adherence in screening.  After adjusted for marital status, the associations were even stronger. Adherent women tended to score higher on the norms scale than non-participants [OR (95%CI): 1.2465 (1.0376, 1.4975)].  They also had higher knowledge levels than non-adherent women (p=0.017); the adjusted OR was 1.2145 [95% CI (1.0354, 1.4246)].

Table 6. Multiple logistic regression: Effect of each factor on screening adherence, adjusted for marital status

P value: p value of Wald F test.

Scales	Odds Ratio	95% Confidence Interval	P value
Pros	1.0781	(1.0031, 1.1588)	0.0410*
Cons	0.8713	(0.8034, 0.9450)	0.0009∗
Norm	1.2465	(1.0376, 1.4975)	0.0186∗
Knowledge	1.1445	(1.0102, 1.2967)	0.0340∗

Discussion

All the scales used in this pilot study reveal acceptable reliability with Cronbach alpha 0.7 or greater.  The cons scale had the lowest internal consistency. Additional items should be developed to strengthen the scale.  Some of the existing items should be reworded.  For example, the item “I would feel more comfortable to obtain a Pap if a female doctor does the procedure” should be revised so that it represents a con statement (i.e. I would not want a male doctor to examine me).

This pilot study showed that 30% of the women had never had a Pap.  Wang and Lin (1996) surveyed women in Taipei in 1993, and found that 40% of the sampled women had never had a Pap.  The prior study was conducted 6 years ago.  The difference may be due to the increases in accessibility of the Pap test through universal coverage of health insurance in Taiwan since 1996.  Still, 30% represents a large proportion of women who had not been screened regularly.

These pilot data indicated that screening adherent rate was 58%.  Other studies have found lower rates (Lee & Chou, 1997).  According to the recorded data of Health Department in Taiwan, the Pap test adherent rate is expected to be 64% by year 2001.  A larger study is required to determine the true rate of screening among a representative sample of women in Taiwan.

Our results indicated that women’s intention to have a Pap test was strongly associated with their screening status. That was, intention to obtain a Pap in the coming year was significantly higher among women who adhered to screening guideline than women did not. This association is consistent with the Theory of Reasoned Action (Ajzen & Fishbein, 1980), which emphasizes the role of intention in health behavior.  Although the direction of the relationship between intention and screening behavior was not explained, preliminary results indicated that future interventions designed to influence women’s intention to have a Pap test could have an impact on screening behavior.

Marital status was the only demographic factor that was significantly associated with women’s compliance with Pap screening.  Similar findings were reported in Yi’s (1998) study.  Women who had never been married were less likely than currently and previously married women to have had a Pap.  It is possible that Chinese women, regardless of their age, believe that Pap test screening is necessary or appropriate only if they are or have been married.  A comment from one of the participants illustrates the feelings of a number of women who participated in the focus group: “… age was not an issue, if I got married, I would have a Pap”.

Another possible explanation regarding the influence of marital status on screening behavior could be that most unmarried women may be less sexually active than married women, and thus less comfortable being exposed or physically touched by others, especially body areas considered private.  Some of the unmarried women in the focus groups said, “I feel very uncomfortable…. I think the procedure intimidates me.  It is invasive, it is private, and it is personal…”.  Many of the unmarried women thought they might feel more comfortable of being exposed after they get married or being sexually active. Although some of the married women expressed similar concerns about being exposed, most married women indicated lack of time or lack of reminder was their major barriers.

Most unmarried women thought Pap screening was only necessary for married or sexually active women.  Some women from the focus groups stated the following: “I am not married, and I am not sexually active, so I don’t think I need a Pap screening”.  “Pap screening is something that I would avoid till I get married.  I won’t obtain a Pap if I am not sexually active”.

The results of the qualitative and quantitative analyses indicated that screening behavior might be influenced by women’s marital status or the sexuality beliefs.  Many of the unmarried women in the focus group made statements such as the following, “I want to keep my body as a whole.  Especially in Chinese culture, we tend to view this as a very important thing ”.  Several women even brought up the issue that many doctors are usually unwilling to perform a Pap screening for women who are not sexually active because the doctors are afraid of being sued by women’s parents.  Similar culture values were found among Vietnamese women in the US (Yi, 1994).  Within the Asian culture, virginity is highly valued and could influence a woman’s social status as well as her own self-esteem.  The results of this study underscore the importance of considering cultural beliefs when designing interventions for cervical cancer screening among this population.

Most women participating in this study had a high level of knowledge related to Pap screening. Nevertheless, knowledge levels were significantly higher among adherent women than non-adherents.  A significant association between knowledge and Pap screening was also found in a previous study conducted among Chinese women (Lee & Chou, 1997).

Women adhered to Pap screening guidelines tended to have higher endorsement on their perceptions of the benefits (pros) of Pap screening than non-adherent women.  The result indicated that there is a relationship between endorsement of benefits of Pap test and screening compliance.  This finding suggests that interventions should include approaches aim at increasing perception of benefits to Pap test screening.

Scores on the items representing “cons” (disadvantages) of a Pap test were lower for adherent women than non-adherents. Intervention efforts should focus on decreasing perceptions of disadvantages of Pap (such as embarrassment, discomfort, nervousness, culture belief of sexuality, gender of physician, lack of doctor’s reminder, and etc.), and increasing perceptions of benefits of a Pap test.

Perceived norms were also significantly associated with women’s adherence in screening.  Perceived norms combines the normative belief concerned with approval or disapproval of a person’s behavior by a referent other, and motivation to comply with the referent’s idea for one’s behavior (Ajzen & Fishbein, 1980). Adherent women perceived higher norms of Pap screening compared with non-adherent women in this pilot study. The result also consisted with Gotay’s study (1998), suggesting that intervention should address women’s perceptions about other women’s behavior and attitudes regards Pap screening in order to improve screening adherence.

Limitations

One of the limitations in the study is the recruitment of the sample population as well as the small sample size.  Since all the women voluntarily agreed to participate in the study, volunteer bias may exist.  Study participants were recruited from female family members of inpatients who admitted to the hospital during the data collection period.  Since women were obtained from a hospital setting, it is possible that these women were more familiar with the health care system than women in the general population.  These women may be different in that they were more open to health related information or may be more likely to have ever been screened.  The small number of the women in the sample could affects power to detect differences between groups, limits the sample representatives, and affects generalizability as well.

Similar to many other studies, this study relies on self-reported Pap test screening and might not reflect actual behavior. Since women could receive care from a wide variety of health care facilities in Taiwan, validation through medical record data of their self-reports of screening behavior was not feasible.  Still, evidence suggests that self-report for Pap test is valid (Suarez, Goldman, & Weiss, 1995).  It is possible that social desirability bias could exist in self-reported screening behavior (Hancock, Sanson-Fisher, & Kentish, 1998).  Self-report measurement might result in classifying some non-adherent women incorrectly thus over estimating the numbers of women in adherent group.  The possibility of classifying adherent women into non-adherent group was low. Previous research has shown that, when women reported they had not had a Pap test within a certain period, the majority was accurate (95%) (Walter et al., 1988).  Since women in the adherent group consisted of some non-adherents, the actual difference between the two-group comparisons should be larger than the study results showed.

This survey only contained two items that measured women’s perceived risk of getting cervical cancer.  Since a reliable measure of a construct, suggested by Nunnally and Bernstein (1994), should consist of at least 3 items.  More items on measuring perceived risk should be developed to make perceived susceptibility into a more reliable scale.

Another limitation of this study is the design of the study.  The cross-sectional survey design restricted the interpretations of the directionality of the relationships between the independent variables and the screening behavior.  Screening may influence attitudes rather than via versa.  For example, women who were recently screened (adherent women) might be more likely to report higher perceived norms or higher knowledge than those non-adherent women.

In summary, this pilot study provides an important foundation to measure psychosocial factors that may be associated with Pap screening behavior among Chinese women.  Larger studies are required to further examining the relationships between Pap screening behavior and these factors.  Findings from this study would help define the approach for further needs assessment, correlates measurement, as well as intervention development for women in this specific population.

Reference

Ajzen I. & Fishbein M. (1980). Understanding attitudes and predicting social behavior. Englewood Cliffs, NJ: Prentice-Hall.

Centers for Disease Control and Prevention. (1998). The National Breast and cervical cancer early detection program; At-A-Glance, 1998.US Department of Health and Human Services, 1-6.

Department of Health, T. E. Y. R. (1995). Public Health in Republic of China.

Gotay C.C. & Wilson M.E. (1998). Social support and cancer screening in African American, Hispanic, and Native American women. Cancer Practice, 6(1), 31-37.

Hancock L., Sanson-Fisher R., & Kentish L. (1998). Cervical cancer screening in rural NSW: health insurance commission data compared to self-report. Australian & New Zealand Journal of Public Health, 22(3), 307-312 (suppl).

Health and Vital Statistics. (1998) Department of Health, the Executive Yuan, Republic of China.

Hiatt R.A. & Pasick R. (1995). The challenge of interventions in multiethnic communities: breast and cervical cancer screening (Meeting abstract). American Society of Preventive Oncology, 19th Annual Meeting, March 8-11, 1995, Houston, TX.

Huang X.L. & Zhang M. (1994).  Effects of two teaching programs on women’s cervical cancer screening knowledge, health belief and behavior. [Chinese]. Unpublished doctoral dissertation, National Taiwan University, School of Nursing, Taiwan.

Kelly A.W., Fores C.M., Wollan P.C., Trapp M.A., Weaver A.L., Barrier P.A., Franz W.B. 3rd, & Kottke T.E. (1996). A program to increase breast and cervical cancer screening for Cambodian women in a mid-western community. Mayo Clinic Proceedings, 71(5), 437-44.

Krueger R.A. (1994). Focus Groups. (2nd ed.). Thousand Oaks, California: Sage Publications, Inc.

Lee T.F., Kuo H.S., & Chou B.S. et al (1997). Factors of Pap screening behavior among women in King-Men, Taiwan. Public Health of Republic of China.16 (3),198-209.

National Institutes of Health Consensus Development Panel. (1996). National Institutes of Health consensus development conference statement: Cervical cancer, April 1-3, 1996. Journal of the National Cancer Institutes Monographs, 21,vii-xix.

Nunnally J.C, & Bernstein I.H. (1994). Psychometric Theory. (3rd ed.). New York: McGraw-Hill, Inc.

Pearlman D.N., Rakowski W., Clark M.A., Ehrich B., Rimer B.K., Goldstein M.G., Woolverton H. III, & Dube C.E. (1997) Why do women’s attitudes toward mammography change over time? Implications for physician-patient communication. Cancer Epidemiology, Biomarkers & Prevention. 6, 451-457.

Pham C.T. & McPhee S.J. (1992). Knowledge, attitudes, and practices of breast and cervical cancer screening among Vietnamese women. Journal of Cancer Education,7(4),305-310.

Prochaska J.O., Norcross J.C., & DiClemente C.C. (1994). Changing for good. (3rd ed.). New York: William Morrow and Company, Inc.

Rakowski W., Dube C.E., Marcus B.H., & Abrams D.B. (1992) Assessing elements of women’s decisions about mammography. Health Psychology. 11(2), 111-118.

Rakowski W., Fulton J.P., & Feldam J.P. (1993) Women’s decision making about mammography: A replication of the relationship between stages of adoption and decisional balance. Health Psychology. 12(3), 209-214.

Rakowski W., Clark M.A., Pearlman D.N., Ehrich B., Rimer B.K., Goldstein M.G., Dube C.E., & Woolverton H. III. (1997) Integrating pros and cons for mammography and pap testing: Extending the construct of decisional balance to two behaviors. Preventive Medicine. 26, 664-673.

Rosenstock I.M. & Krischt J.P.  (1974). The Health Belief Model and personal health behavior.  Health Education Monographs, 2,470-473.

Seow A., Wong M.L., Smith WCS, & Lee H.P. (1995). Beliefs and attitudes as determinants of cervical cancer screening: A community-based study in Singapore. Preventive Medicine, 24, 134-141.

Suarez L., Goldman D.A., & Weiss N.S. (1995). Validity of Pap smear and mammogram self-reports in a low-income Hispanic population. American Journal of Preventive Medicine, 11, 94-98.

Walter S.D., Clarke E.A., Hatcher J., & Stitt L.W. (1988). A comparison of physician and patient reports of Pap smear histories. Journal of Clinic Epidemiology, 41, 401-410.

Wang P.D. & Lin R.S. (1996). Socio-demographic factors of pap smear screening in Taiwan. Public Health,110(2), 123-127.

Yi J.K. (1994). Factors associated with cervical cancer screening behavior among Vietnamese women. Journal of Community Health, 19(3), 189-200.

Yi J.K. (1998). Acculturation and Pap screening practices among college-aged Vietnamese women in the United States. Cancer Nursing,21(5),335-341.