Abstract
Objective: The study goal was to examine the effects of an initial false positive chest X-ray screening result on subsequent lung cancer screening adherence.
Methods: Adherence rates among 4705 individuals in the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial at the Henry Ford Health System site with an abnormal/suspicious chest X-ray screening result in the first study year that was subsequently determined to be noncancerous (false positive result, n = 1137 exams) were compared with adherence rates among individuals with an initial negative chest X-ray screening result (n = 3568 exams).
Results: Univariate results showed a >50% increase in subsequent nonadherence among individuals with false positive screening results compared with those with negative screening results (17.2% versus 10.3% nonadherence rate, respectively; P < 0.001). Multivariable results showed that statistically significant predictors of nonadherence were false positive cases with current smoking status (P < 0.001) and false positive cases with past smoking status (P < 0.001). Additional predictors of subsequent nonadherence were being African-American (P < 0.01), being female (P < 0.001), and having a high school education or less (P < 0.01).
Conclusion: Our results demonstrate that the impact of previous screening results, smoking status, race, gender, and education on subsequent screening adherence needs to be weighed carefully, particularly for smokers, an at-risk group, when conducting lung cancer screening intervention studies and perhaps should be considered in clinical practice as well.
Introduction
The PLCO3 Trial is a 23-year, multisite, and randomized trial funded by the National Cancer Institute (1, 2, 3, 4). The primary objective of the PLCO Trial is to determine whether screening for the four PLCO Trial cancers decreases mortality caused by these cancers in adults aged 55–74 years at trial entry. Participants in the PLCO Trial, who must be generally healthy and asymptomatic for any of the study cancers at entry into the study, are randomized to an intervention arm (receiving cancer screening) or a control arm (not receiving cancer screening) of the PLCO Trial. The value of chest X-rays in reducing lung cancer mortality continues to be debated, hence the establishment of the PLCO Trial (1).
Implications of Having a False Positive Chest X-ray Screening Result.
One of the implications of having a false positive chest X-ray screening result is the subsequent risk of being exposed to invasive medical procedures and their associated morbidity and mortality (5). In addition, patients may exhibit a great deal of anxiety at being informed that they have had a positive chest X-ray screening result (5, 6, 7, 8), e.g., Padgett et al. (9) found that, in a sample of 45 African-American women with whom open-ended interviews were conducted, increased feelings of anxiety were reported after a false positive screening mammogram.
Despite the reported increased levels of anxiety associated with having a false positive mammography screening result, such negative affect does not appear to hinder subsequent mammography screening adherence (7, 10, 11). In fact, in previous studies, having a prior false positive screening result has been shown to have a positive impact on subsequent screening adherence (7, 10, 11, 12, 13). Although the majority of these studies shows that anxiety related to health concerns increases after a false positive mammogram, the studies also show a positive association between previous false positive results and subsequent screening behavior.
In contrast to many mammography studies conducted previously, our study focuses on lung cancer screening behavior in a sample that is 45% male and includes a screening mechanism whose effectiveness is as yet unknown, many African-Americans (16%, n = 742), and many current and past smokers (63%, n = 501) who have engaged in a well-known and well-accepted behavioral risk factor for lung cancer.
Aim of the Present Study.
Although the results of previous studies have shed light on screening adherence behavior after a false positive mammography screening result, little is known about screening behavior after a false positive lung cancer screening result. It is unknown whether having a false positive result on an earlier chest X-ray screening exam affects subsequent adherence to screening for lung cancer using chest X-rays or, for that matter, to other cancer screening tests. In high-risk populations, such as smokers, the risk of false positive screening results has been estimated to be 5% (5).
Materials and Methods
PLCO Trial Study Participants.
As of July 1999, when the data for the present study were selected for analysis, 8,186 (41%) of the 19,993 participants enrolled at the HFHS site (Detroit, MI) of the PLCO Trial were men, and 11,807 participants (59%) were women. Of the 2,980 African-American study participants, 1,275 (43%) were men, and 1,705 (57%) were women. Among the 10,004 participants in the intervention (screening) arm at the HFHS PLCO Trial site, 644 (6%) were African-American men, and 855 (8.5%) were African-American women. Overall, African-Americans comprised 15% of the PLCO Trial participants at the HFHS site.
Definitions and Screening Intervals.
Participants in the PLCO Trial intervention arm are screened for six consecutive years after their randomization to this study arm. Included in the screenings is lung cancer screening via chest X-ray for the first four study years. In the PLCO Trial, current smokers are defined as individuals who “smoke cigarettes regularly now.” Past smokers are defined as individuals who “ever smoked cigarettes regularly for ≥6 months” but who are not current smokers. Never-smokers are defined as individuals who have never “smoked cigarettes regularly for ≥6 months.” Individuals in the intervention (screening) arm of the PLCO Trial who are current or past smokers receive lung cancer screening via single view postero-anterior chest X-ray at baseline, year 1, year 2, and year 3. In contrast, lung cancer screening via chest X-ray is performed for never-smokers at baseline, year 1, and year 2. Possible outcomes of lung cancer screening via chest X-ray are: (a) negative for lung cancer; (b) inadequate result, abnormal/suspicious for lung cancer; (c) abnormal/not suspicious for lung cancer; and (d) “test not done.” For the purpose of the present study, a false positive result is defined as a result with “abnormal/suspicious” labeling that did not result in a lung cancer diagnosis within 14 months. By the time of the subsequent screening after the previous false positive screening result, study participants knew that they did not have lung cancer.
All chest X-ray screening tests included in the sample were selected based on their completion by April 1, 1999, to allow an adequate follow-up period before the next scheduled screening (≥14 months), so that adherence to screening and diagnosis by follow-up exam could be assessed. Additionally, individuals who died during the follow-up period from unknown causes were excluded from the sample. Only participants who self-identified as being of African-American or Caucasian ethnicity were included (individuals of Hispanic ethnicity were excluded because of their small sample size).
Study Sample.
The study sample consisted of African-American or Caucasian individuals with a ≥14-month follow-up period from their first lung cancer screening via chest X-ray, which occurred in their baseline screening year of the PLCO Trial. An eligible screening result was defined as a screening that resulted in a negative (n = 3568) or false positive (n = 1336) screening outcome over the study period (January 1994 to April 1999). Individuals in whom cancer was later found were excluded from the study.
In the analyses, screening adherence for the next scheduled lung cancer screening via chest X-ray was compared for participants with false positive chest X-ray screening results on their previous screening and those with negative chest X-ray screening results on their previous screening. As stated previously, adherence was defined as completing the next scheduled lung cancer screening via chest X-ray. As part of the PLCO Trial, study participants were contacted by telephone to schedule their yearly screening appointments.
Analysis.
Nonadherence, the outcome variable, was measured as “yes” versus “no,” i.e., expected chest X-ray screening not done versus screening done in the subsequent screening year. Univariate relationships between nonadherence and each of the categorical covariates were examined using χ2 tests, ORs, and 95% CIs. Age was examined as a continuous covariate, and univariate logistic regression was used to evaluate its relationship with nonadherence. Multivariable logistic regression was used to assess the effect of false positive result with nonadherence, adjusting for all other covariates.
Race, gender, age, educational level, smoking status, and false positive results on other screening exams were included as covariates in the multivariable model. Educational level was dichotomized into post-high school education versus high school education or less. Smoking status was classified as current smoker, past smoker, and never-smoker. In all of the analyses, the never-smoker group was used as the reference group. In the multivariable model, interactions between each covariate and the primary independent variable (result on previous chest X-ray screening) were assessed.
Justification of Sample Size.
All sample size calculations were based on two-sided testing and an α level of 0.05. Given a sample size of 4705 screening exams (1137 false positive results and 3568 negative results), ≥90% power was available to detect a 6% difference in estimated probability of screening adherence between the two groups, e.g., if the false positive group had only an 80% adherence rate, a rate of ≥86% in the group with negative results was able to be detected.
Results
The study included 4705 unique individuals who had a false positive or negative result on their baseline chest X-ray screening. Table 1 shows the demographic characteristics of the study sample. As may be seen, the study sample, which included more women than men and more Caucasians than African-Americans, had a mean age of ∼62 years. The sample also included a higher percentage of individuals with post-high school education than those who had not attained this level of education and a higher percentage of current or past smokers compared with never-smokers. In this sample, false positive cases were older (P < 0.001), more likely to be male (P < 0.003), and more likely to be current smokers (P < 0.001) than were negative cases.
There was a significant univariate relationship (OR = 1.8; 95% CI = 1.5–2.2) between having had a previous false positive chest X-ray screening result and subsequent screening adherence. The data suggest that there was a >50% increase in nonadherence among those with a previous false positive screening result, compared with those with a previous negative screening result (Table 2).
Additional covariates that were considered in the study were African-American race (versus Caucasian race), gender, age, educational level, smoking status, and having had a false positive result on other concurrent PLCO Trial screening tests. The univariate relationships of these factors with nonadherence are shown in Table 2. As may be seen, African-American race is significantly related to subsequent nonadherence (OR = 1.5; 95% CI = 1.2–1.9), i.e., African-Americans were more likely than Caucasians to be nonadherent. Women were more likely (OR = 1.4; 95% CI = 1.2–1.7) to be nonadherent than men. Table 2 also shows that the nonadherent participants were significantly older (P < 0.03) than the adherent participants. Current smokers were more likely to be nonadherent as compared with never-smokers (OR = 1.4; 95% CI = 1.1–1.9). As may be seen in this table, participants of lower educational level were more likely than others to be nonadherent (P < 0.001). The results of this analysis also suggest that there is no evidence for a relationship between having had a false positive screening result on a previous prostate, colorectal, or ovarian PLCO Trial screening and subsequent nonadherence to lung cancer screening (P = 0.68).
The next step in the analyses involved addressing the question of whether having a false positive result on baseline chest X-ray screening would have an effect on subsequent screening nonadherence, after adjusting for possible confounding variables. Thus, in the following model, all of the baseline covariates were included, even those which did not show a statistically significant univariate relationship with adherence. This inclusion was done because the sample size was large enough to rule out decreased power as a concern, and possible interactions among these covariates and false positive status could be explored. Therefore, a multivariable logistic regression model that included false positive status, the seven baseline covariates, and the interaction was tested between false positive status and each baseline covariate (six extra tests). The only statistically significant two-way interaction was that of false positive status by smoking status (P < 0.04). None of the other interactions reached statistical significance at the P > 0.2 level.
Because of the interaction between smoking status and false positive status, five indicator variables were included in the final model (Table 3) to report the specific risk ratios for each category. Table 3 shows that current and past smokers who received previously a false positive result on their baseline lung cancer screen were more than twice as likely to be nonadherent compared with never-smokers with a negative baseline lung cancer screen (OR = 2.4; 95% CI = 1.6–3.5 and OR = 2.1; 95% CI = 1.5–2.8, respectively). As may be seen in Table 3, African-American race, being female, and having less than a high school education were also statistically significant predictors of subsequent screening nonadherence.
As noted earlier, PLCO Trial participants were contacted by telephone to schedule their yearly screening appointments. During the scheduling call, participants’ reasons for not agreeing to schedule an appointment were typically recorded. Among the 542 individuals (96% of the 564 nonadherent group) who gave a rationale for their nonadherence to subsequent lung cancer screening via chest X-ray, the main reasons were refusals (n = 316, 58%), not being able to be scheduled during the window of time required by study protocol (n = 144, 27%), having an illness that prevented adherence (n = 26, 5%), being out of the geographic area of HFHS PLCO Trial clinics at the time of the requested screening (n = 26, 5%), and family problems (n = 30, 5%). The remaining 4% were nonadherent for unknown reasons.
Discussion
This study, which used data from the HFHS site of the PLCO Cancer Screening Trial, began with two aims. The first aim involved assessing whether nonadherence rates among individuals with a previous abnormal/suspicious chest X-ray screening result that was later determined to be noncancerous on follow-up exams (false positive result) would be different from the nonadherence rates of individuals with a previous negative chest X-ray screening result. The second aim involved assessing whether race, age, gender, smoking history, or false positive screening results on concurrent other screening tests affected this relationship.
The study results represent an expansion of existing knowledge related to cancer screening. It is clear that a simple extrapolation of findings from mammography screening studies cannot be applied to lung cancer screening studies, because the results of having a previous false positive effect on subsequent screening adherence are inversely related for the two types of cancer screening, i.e., in breast cancer screening using mammography, having a previous false positive screening result appears to enhance subsequent screening adherence (7, 10, 11, 13), e.g., Aro et al. (10) examined psychological distress in a Finnish sample of 50-year-old women. Distress was defined as illness, worry, anxiety, depression, cancer beliefs, and behavior related to early detection. Although exhibiting greater anxiety, intrusive thinking, and worry about breast cancer at 2 months postscreening, women with false positive screening results engaged in higher levels of breast self-examination than the other women in the study. Although actual screening behavior was not examined postscreening, both groups of women reported similarly high rates of intentionality regarding their future participation in mammography screening. These rates were 98.2% for women with negative screening results and 98.7% for women with false positive screening results. This study showed moderate negative short-term psychological consequences as a result of receiving a false positive breast cancer screening result, with no significant impact on stated future screening behavior. However, a limitation of the study by Aro et al. (10) lies in the fact that subsequent screening behavior was not examined. The actual impact of having a previous false positive screening result on subsequent screening behavior could therefore not be assessed.
Such behavior was assessed in a prospective cohort study conducted by Burman et al. (12). The study sample in this case included 5059 women aged ≥40 years, who had no previous history of breast cancer and had obtained either false positive or true negative previous breast cancer screening results. After adjusting for age, breast cancer risk factors, and previous use of mammography, women with previous false positive mammography screening results were found to be more likely than other women to adhere to subsequent mammography screening.
Similarly, Pisano et al. (11), who used a retrospective cohort design with a sample of 43 cases aged ≥50 years with false positive results and 136 randomly selected controls, found that having a false positive mammography screening result actually served to increase intention to undergo further breast cancer screening using mammography. In addition, negative affect such as anxiety or worry did not appear to differ significantly between the cases and controls. Another study conducted by Pisano et al. (11) showed similar results. In contrast, our results show that in lung cancer screening using a chest X-ray, having a false positive screening result appears to significantly inhibit subsequent lung cancer screening adherence.
The relationship between previous screening results and subsequent screening adherence may be explained using the perceived risks/benefits component of the Health Belief Model and beliefs about the value of cancer early detection (14, 15, 16), i.e., the anxiety (risk) a person experiences when he or she is informed of abnormal/suspicious results for lung cancer that are later determined to be negative may lead her or him to feel that the risk associated with subsequent screening outweighs the potential benefit of cancer early detection. This relationship may be particularly strong if the individual also has doubts about the value of cancer early detection in reducing cancer mortality (16).
Perhaps different conceptualizations of breast cancer as opposed to lung cancer are partially responsible for the difference in behaviors after a false positive breast cancer screening result as compared with behavior after a false positive lung cancer screening result. It is possible that early intervention in breast cancer is perceived as being more efficacious than early intervention in lung cancer, in terms of reducing mortality. It would be interesting to develop conceptual models of perceptions of the two types of cancers and of the perceived efficacy of screening, early intervention, and treatment of these cancers. In a future study, interviews with patients with different racial and ethnic, educational, gender, age, and smoking history backgrounds could be used to actually ascertain these perceptions, because our study showed that these factors were associated with nonadherence after a false positive lung cancer screening result.
Our study results also point to the development of targeted interventions designed to facilitate screening adherence among groups of individuals with different risk and socio-demographic characteristics. In particular, special attention needs to be given to those who have experienced a false positive result for a previous screening. It is important to note that “one size does not fit all,” i.e., factors associated with cancer screening adherence appeared to differ for individuals with different risk (here, smoking history) and socio-demographic characteristics, such as age, race, gender, and education level. Our results showed that, compared with never-smokers, current, and past smokers, both groups of whom are at increased risk of developing lung abnormalities, including lung cancer, were less likely to be adherent to lung cancer screening after a false positive lung cancer screening result. It may be that smokers, who are well known to be at risk for developing lung cancer, are more fearful than others of getting this disease. Therefore, having a false positive screening result may have greater emotional impact on subsequent nonadherent behavior among current or past smokers than among those who have never smoked. In combination with an initial fear of being diagnosed with lung cancer, having a false positive screening result may serve to exacerbate nonadherence. The fact that current or past smokers have engaged in an activity known to put them at increased risk for the very disease for which they are getting screened adds an additional layer of complexity to the assessment of screening-related behavior. This issue needs to be examined in greater detail in future studies.
In addition, in our study, African-Americans, women, and those with less than a high school education were found to be more likely than others to exhibit nonadherent behavior after a false positive lung cancer screening result. Perhaps investigators in the growing field of research focused on adherence to mammography after a false positive screening result could take these socio-demographic factors into account when analyzing their data, to assess the effects of these factors on study outcomes. These factors could be incorporated into interventions through the use of methods such as focus groups, intensive individual interviews, and surveys.
Through these methods, the attitudes and perceptions of individuals with various socio-demographic characteristics could be ascertained and used as the building blocks for the development of interventions, e.g., Paskett et al. (17) developed a hierarchic weighted utility model, based on interviews with women who had experienced an abnormal Pap smear, to identify the women’s health beliefs and values related to their decision to receive follow-up treatment. The model showed that the following factors influenced the decision to seek further treatment: (a) the opinions of a physician; (b) the perceived accuracy of the Pap smear test result; (c) the perceived importance of early detection; (d) familiarity with the follow-up treatment procedures; (e) time-related difficulties in obtaining further treatment; (f) fear of cancer; and (g) perceived risk of cancer. As Paskett et al. (17) note, interventions could be developed based on each of these specific types of reasons for, or barriers to, obtaining follow-up treatment, e.g., individuals who value the opinions of physicians could receive group or one-on-one decision counseling from a physician, whereas individuals who have different levels of perceived cancer risk could be given risk information tailored to themselves, to help them better identify their personal risk of cancer. Similar interventions could be incorporated into screening trials to enhance adherence by mitigating the negative effects of a false positive screening result on subsequent screening adherence.
Strengths of the present study include the racial diversity of its sample and its ability to monitor the screening practices of a defined population. Longitudinal screening intervention studies, such as the PLCO Trial, offer a unique opportunity to explore the dynamics of factors affecting adherence to cancer screening.
A limitation of this study is the fact that the information gained provides insight into the possible negative effects of false positive cancer screening results for lung cancer only among individuals who have agreed to participate in a long-term cancer screening trial. These individuals are not likely representative of the general population in that they may be more involved in health promotion activities than nonparticipants. Because it was found that having a false positive result on a cancer screening exam had an effect on a population already committed to health promotion, as indicated by their participation in the PLCO Trial, the effects found in this study will likely be greater in a population of individuals not as committed to their own health promotion.
In addition, the individuals described in the present study represent only participants at the HFHS site of the 10-site PLCO Trial. It is possible that because of variations in populations with regard to cultural factors and health beliefs, these results are not generalizable to other geographic locations. The results shown in this study apply to lung cancer screening using chest X-ray, which is not yet a proven method of lung cancer early detection.
In summary, this study highlights the importance of identifying factors associated with adherence to cancer screening after a previous false positive screening result. It is important to note that false positive results had a stronger effect on nonadherence among ever-smokers than among never-smokers. In the general population, smokers are a group at higher risk of developing lung cancer than are nonsmokers. As Lam et al. (18) point out, long-term heavy smokers have a continued high risk of lung cancer even years after they quit smoking. Developers of targeted interventions designed to facilitate lung cancer screening among members of this at-risk population will need to take into account the factors of previous screening results (if known), race, gender, and education to render these interventions as effective as possible. Randomized trials of the targeted interventions could then be conducted.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Supported by National Cancer Institute Grant N0 1 CN 25512 and National Institute on Aging Grant P 30 AG 5286.
The abbreviations used are: PLCO, Prostate, Lung, Colorectal, and Ovarian Cancer Screening; CI, confidence interval; OR, odds ratio; HFHS, Henry Ford Health System.
Demographic characteristics of sample (n = 4705)
Demographic characteristics . | n . | % . |
---|---|---|
Gender | ||
Female | 2587 | 55.0% |
Male | 2118 | 45.0% |
Race | ||
African-American | 742 | 15.8% |
Caucasian | 3963 | 84.2% |
Mean age (SD), in years | 62.3 | (5.3) |
Post-high school educationa | ||
Yes | 3257 | 69.3% |
Nob | 1443 | 30.7% |
Smoking statusa | ||
Current smoker | 666 | 14.4% |
Past smoker | 2249 | 48.7% |
Never-smoker | 1700 | 36.9% |
Demographic characteristics . | n . | % . |
---|---|---|
Gender | ||
Female | 2587 | 55.0% |
Male | 2118 | 45.0% |
Race | ||
African-American | 742 | 15.8% |
Caucasian | 3963 | 84.2% |
Mean age (SD), in years | 62.3 | (5.3) |
Post-high school educationa | ||
Yes | 3257 | 69.3% |
Nob | 1443 | 30.7% |
Smoking statusa | ||
Current smoker | 666 | 14.4% |
Past smoker | 2249 | 48.7% |
Never-smoker | 1700 | 36.9% |
Some study participants were missing data on this variable.
Of the 1443 participants who provided this response, 71% had completed high school, 25% had completed grades 8–11, and 4% reported having a <8th grade level of education.
Subsequent nonadherence after a previous false positive or negative chest X-ray screening result (n = 4705)
. | Subsequent Nonadherence . | . | OR (95% CI) . | P . | |
---|---|---|---|---|---|
. | Yes n (%) . | No n (%) . | . | . | |
Previous chest X-ray screening result | 1.81 (1.50, 2.19) | <0.001 | |||
False positive | 196 (34.8%) | 941 (22.8%) | |||
Negative | 368 (65.2%) | 3200 (77.2%) | |||
Race | 1.49 (1.20, 1.86) | <0.001 | |||
African-American | 118 (20.9%) | 624 (15.1%) | |||
Caucasian | 446 (79.1%) | 3517 (84.9%) | |||
Gender | 1.44 (1.20, 1.73) | <0.001 | |||
Female | 354 (62.8%) | 2233 (53.9%) | |||
Male | 210 (37.2%) | 1908 (46.1%) | |||
Mean age (SD) | 62.8 (5.6) | 62.3 (5.3) | 1.02 (1.01, 1.04) | 0.029 | |
Smoking statusa | |||||
Current | 105 (19.1%) | 561 (13.8%) | 1.45 (1.12, 1.88)b | 0.004 | |
Past | 250 (45.5%) | 1999 (49.2%) | 0.97 (0.80, 1.18)b | 0.77 | |
Never | 194 (35.4%) | 1506 (37.0%) | |||
Education level completedc | 1.40 (1.17, 1.68) | <0.001 | |||
High school or less | 210 (37.3%) | 1233 (29.8%) | |||
Post-high school | 353 (62.7%) | 2904 (70.2%) | |||
False positive result on nonlung cancer screens in PLCO Trial | 0.96 (0.79, 1.17) | 0.68 | |||
Yes | 156 (27.7%) | 1180 (28.5%) | |||
No | 408 (72.3%) | 2961 (71.5%) |
. | Subsequent Nonadherence . | . | OR (95% CI) . | P . | |
---|---|---|---|---|---|
. | Yes n (%) . | No n (%) . | . | . | |
Previous chest X-ray screening result | 1.81 (1.50, 2.19) | <0.001 | |||
False positive | 196 (34.8%) | 941 (22.8%) | |||
Negative | 368 (65.2%) | 3200 (77.2%) | |||
Race | 1.49 (1.20, 1.86) | <0.001 | |||
African-American | 118 (20.9%) | 624 (15.1%) | |||
Caucasian | 446 (79.1%) | 3517 (84.9%) | |||
Gender | 1.44 (1.20, 1.73) | <0.001 | |||
Female | 354 (62.8%) | 2233 (53.9%) | |||
Male | 210 (37.2%) | 1908 (46.1%) | |||
Mean age (SD) | 62.8 (5.6) | 62.3 (5.3) | 1.02 (1.01, 1.04) | 0.029 | |
Smoking statusa | |||||
Current | 105 (19.1%) | 561 (13.8%) | 1.45 (1.12, 1.88)b | 0.004 | |
Past | 250 (45.5%) | 1999 (49.2%) | 0.97 (0.80, 1.18)b | 0.77 | |
Never | 194 (35.4%) | 1506 (37.0%) | |||
Education level completedc | 1.40 (1.17, 1.68) | <0.001 | |||
High school or less | 210 (37.3%) | 1233 (29.8%) | |||
Post-high school | 353 (62.7%) | 2904 (70.2%) | |||
False positive result on nonlung cancer screens in PLCO Trial | 0.96 (0.79, 1.17) | 0.68 | |||
Yes | 156 (27.7%) | 1180 (28.5%) | |||
No | 408 (72.3%) | 2961 (71.5%) |
n = 90 with missing data on smoking status.
Never-smokers as reference group.
n = 5 with missing data on educational status.
Multivariable logistic regression model for nonadherence results (n = 4610)
. | OR (95% CI) . | P . |
---|---|---|
False positive, current smokera | 2.38 (1.62, 3.49)b | <0.001 |
False positive, past smokera | 2.07 (1.55, 2.76)b | <0.001 |
False positive, never-smokera | 1.32 (0.93, 1.86)b | 0.12 |
Negative result, current smokera | 1.32 (0.95, 1.84)b | 0.10 |
Negative result, past smoker | 0.90 (0.70, 1.14)b | 0.38 |
African-American | 1.39 (1.11, 1.75) | 0.005 |
Female | 1.50 (1.23, 1.82) | <0.001 |
High school education or less | 1.29 (1.07, 1.56) | 0.009 |
False positive result on previous nonlung cancer PLCO Trial screening | 0.94 (0.76, 1.15) | 0.53 |
Age | 1.01 (0.99, 1.03) | 0.23 |
. | OR (95% CI) . | P . |
---|---|---|
False positive, current smokera | 2.38 (1.62, 3.49)b | <0.001 |
False positive, past smokera | 2.07 (1.55, 2.76)b | <0.001 |
False positive, never-smokera | 1.32 (0.93, 1.86)b | 0.12 |
Negative result, current smokera | 1.32 (0.95, 1.84)b | 0.10 |
Negative result, past smoker | 0.90 (0.70, 1.14)b | 0.38 |
African-American | 1.39 (1.11, 1.75) | 0.005 |
Female | 1.50 (1.23, 1.82) | <0.001 |
High school education or less | 1.29 (1.07, 1.56) | 0.009 |
False positive result on previous nonlung cancer PLCO Trial screening | 0.94 (0.76, 1.15) | 0.53 |
Age | 1.01 (0.99, 1.03) | 0.23 |
Results of previous lung cancer screening.
Negative result, never-smokers as reference group.