Screening in the Era of Economic Crisis : Misperceptions and Misuse from a Longitudinal Study on Greek Women Undergoing Benign Vacuum-assisted Breast Biopsy

Greece is a country where cancer screening is extremely problematic, according to data coming from the recent Delphi consensus study (Skroumpelos et al., 2013) as well as the studies issued by the Panhellenic Association for Continual Medical Research (PACMER) (Kamposioras et al., 2007; Mauri et al., 2009). Specifically concerning mammogram, the PACMER investigators reported a screening rate equal to 22.8% (Mauri et al., 2009) whereas the Hellas Health I Survey reported a rate equal to 53.8% adopting less strict criteria (i.e. mammogram within the past 3 years) in a predefined age group (50-69 years). Accordingly, at the level of Greek physicians, the recent Delphi study demonstrated a considerable lack of consensus among experts, especially regarding colorectal cancer (Skroumpelos et al., 2013).


Introduction
Greece is a country where cancer screening is extremely problematic, according to data coming from the recent Delphi consensus study (Skroumpelos et al., 2013) as well as the studies issued by the Panhellenic Association for Continual Medical Research (PACMER) (Kamposioras et al., 2007;Mauri et al., 2009).Specifically concerning mammogram, the PACMER investigators reported a screening rate equal to 22.8% (Mauri et al., 2009) whereas the Hellas Health I Survey reported a rate equal to 53.8% adopting less strict criteria (i.e.mammogram within the past 3 years) in a predefined age group (50-69 years).Accordingly, at the level of Greek physicians, the recent Delphi study demonstrated a considerable lack of consensus among experts, especially regarding colorectal cancer (Skroumpelos et al., 2013).

Screening in the Era of Economic Crisis: Misperceptions and Misuse from a Longitudinal Study on Greek Women Undergoing Benign Vacuum-assisted Breast Biopsy
Philip John Domeyer 1 *, Theodoros Nikolaos Sergentanis 1 , Vasiliki Katsari 2 , Kyriakos Souliotis 2 , Anargiros Mariolis 3 , Flora Zagouri 1 , George Constantine Zografos 1   It is worth mentioning that underscreening represents solely the bright side of the moon; at the dark side a local particularity emerges: physicians frequently recommend tests which are not useful for screening purposes, such as tumor markers (Vittoraki et al., 2007) or annual chest radiography (Kamposioras et al., 2006).Moreover, Greece is facing an important economic recession; economic crisis may further the underuse of screening tests, particularly those associated with higher costs (Myong et al., 2012).
Vacuum-Assisted Breast Biopsy (VABB) is a recently developed biopsy method, aiming to obtain tissue for histopathological diagnosis of non-palpable mammographic lesions.VABB is characterised by satisfactory underestimation of the lesions and can be performed under stereotactic or ultrasonographic guidance; an 11-Gauge (11G) needle is most commonly used for sampling of the suspicious lesion (Dershaw et al., 2003;Zografos et al., 2008).Our previous research work has shown that VABB is a multifaceted experience in a woman's life, considerably affecting the woman's healthrelated quality of life (Domeyer et al., 2010) and stress (Gounaris et al., 2007); its distinct features compared to other breast biopsy methods result in a distinct pattern of adherence to clinical recommendations for follow-up (Sergentanis et al., 2009).
Irrespectively to the biopsy method, benign breast biopsy may exert multipotent effects upon screeningrelated habits of patients.An increase in the adherence to mammogram has been documented (Zografos et al., 2010) whereas data concerning the effects upon Breast Self-Examination (BSE) remain contradictory (Janz et al., 1990;Zografos et al., 2010), as they may depend on the palpability of the lesion (Zografos et al., 2010) or the way the lesion of the patient was discovered (Janz et al., 1990).
The aim of this study is to: i) evaluate the level of knowledge about the recommended screening tests (mammogram, Pap smear, fecal occult blood testing, sigmoidoscopy) as well as about tests without proven screening value (BSE, breast Magnetic Resonance Imaging (MRI), abdominal ultrasound, barium meal, urine analysis, chest X-Ray, electrocardiogram, cardiac ultrasound) in a Breast Unit population undergoing benign VABB; ii) evaluate risk factors for lack of knowledge, making the distinction between necessary and unnecessary tests; and iii) examine whether benign VABB may modify the practice of all the tests above.

Patients
Exclusion criteria for this study were: previous breast cancer, severe comorbidity (psychiatric conditions, stroke, autoimmune diseases, cancer, severe coronary heart failure, i.e.NYHA stage III or IV).In addition patients diagnosed with precursor (atypical ductal hyperplasia, ADH and lobular neoplasia, LN) lesions, as well as carcinomas (ductal in situ, DCIS or invasive, IDC, lobular carcinomas) were excluded from the study, as the follow-up/treatment of these conditions, respectively, may represent significant deviation from the screening recommendations pertaining to the general population.
Of the 164 consecutive patients who came to our Breast Unit due to non-palpable mammographic lesions requiring VABB, only 81 were eligible for this study (Figure 1).The women were 33-80 years old.
Patients were informed (orally and in written) about the procedure, possibility of pain and complications by the surgeon performing VABB.Written signed informed consent was obtained from all patients.The study was approved by the Local Institutional Review Board.
According to the results of a double-blind study (Zografos et al., 2008), a variable number of cores (24-96 cores) has been excised.
All procedures were performed by the same surgeon, in the same Unit, according to the recommended local anesthesia (Dershaw, 2003); in addition two specialist radiologists assisted at the procedures.The surgeon performing VABB was familiar with this method before the onset of this study, having already performed 350 VABB procedures.For local anesthesia, the two-step approach was adopted: 5 cm 3 1% lidocaine without epinephrine (superficial) and 10 cm 3 1% lidocaine with epinephrine (deep) were administered.The biopsy was performed according to a standard protocol to assure quality control.Compression bandages were applied so as to prevent hematoma.

Structure and administration of questionnaires
All patients were asked to complete the questionnaires, i) in the morning of the VABB procedure day (i.e.1-2 hours prior to biopsy, designated as baseline measurement) and ii) 18 months after VABB.
At baseline the following were asked for each test (mammogram, BSE, breast MRI, Pap smear, fecal occult blood testing, sigmoidoscopy, abdominal ultrasound, barium meal, urine analysis, chest X-Ray, electrocardiogram, cardiac ultrasound): "Is this test performed as a screening test?"(yes/no), "If yes, please indicate: i) the age of screening onset for the test and ii) how often the test should be performed" (open items).The above were designated as knowledge-oriented items.
At both time points the following were asked for each test: "Do you regularly perform the test?" (yes/no) and "How often do you perform the test?" (open item); the above were designated as practice-oriented items.
At the baseline assessment the following information concerning possible risk factors was obtained: i) anthropometric features (height, weight, from which Body Mass Index (BMI) was calculated); ii) sociodemographic parameters i.e. age, place of residence (urban or rural),

Evaluation of responses and statistical analysis
Concerning the knowledge-oriented items the patients' responses were evaluated according to the 2010-2011 Recommendations of the U.S. Preventive Services Task Force (Agency for Healthcare Research andQuality, 2010-2011).The distinction between two levels of knowledge was made: i) crude knowledge (CK), i.e. correct response to the item "Is this test performed as a screening test?"(yes/no) and ii) advanced knowledge (AK), i.e. correct response to all three knowledge-oriented items.In case a knowledge-oriented item was left blank the response was separately reported, as it most probably denotes lack of knowledge.Evidently, solely CK pertained to tests without screening value whereas the evaluation concerning recommended screening tests essentially encompassed AK and CK (Figure 2).Descriptive statistics were calculated for both levels of knowledge.
The associations between possible risk factors and both levels of knowledge were assessed first through univariate analysis; the predictors proven significant in the univariate analysis were included in the multivariate models.Multivariate logistic regression was performed with the lack of knowledge (CK or AK) set as the dependent variables so as to identify risk factors for lack of knowledge; backward selection of variables was performed.
Similarly to the evaluation of knowledge, the assessment of the practice-oriented items was performed according to the 2010-2011 Recommendations of the U.S. Preventive Services Task Force (Agency for Healthcare Research and Quality, 2010-2011).To designate practice as "wrong", the following scenarios were possible: i)   Regular performance of the test without proven screening value or ii) No performance or performance at wrong intervals of a screening test.As a result, a binary (wrong/ correct) variable was generated (Figure 3).
Regarding the effect of VABB, the difference between the practice before vs after VABB was assessed with the exact McNemar test.Breast MRI was not included in the analysis concerning VABB as women may have been prescribed an MRI and subsequently performed that in the context of follow-up.In addition sigmoidoscopy was not included in the analysis concerning VABB as the interval after VABB in our study (18 months) was much shorter than the recommended interval (5 years).
The statistical analysis was performed using STATA 8.0 statistical software (Stata Corporation, College Station, TX, USA).

Results
The majority of the study sample resided in urban settings, while 6 out of 10 patients had completed primary or secondary education and 3 quarters of our study sample did not smoke (Table 1).Baseline values of CK, AK as well as all three constituents of AK (recommended for screening, age of onset, screening frequency) were thoroughly evaluated (Table 2).There was a high rate (6 out of 10) of responders with lack of AK regarding the mammogram and a very high rate (9 out of 10) of responders with lack of CK concerning the BSE.Also, 60.5% showed lack of CK for breast MRI  and approximately 5 out of 10 responders showed lack of CK for barium meal, whereas the rates for abdominal ultrasound, urine analysis, chest X-Ray, electrocardiogram and cardiac ultrasound were considerably higher.With regard to Table 2, it is worth noting that the denominator in the second ("Recommended for screening"), fifth ("Lack of CK") and sixth ("Lack of AK") column refers to the total of the sample; on the contrary the denominators in the third ("Age of onset") and fourth ("Screening frequency") columns refer to patients answering correctly the "Recommended for screening" item.
The risk factors for lack of knowledge about all the examined tests were analyzed (Table 3).Living in Athens was associated with lack of AK for mammogram, increasing age was associated with lack of AK for Pap Smear, while decreasing personal income was associated with lack of AK for sigmoidoscopy.Decreasing number of children was associated with lack of CK for Breast MRI and chest X-Ray, decreasing BMI was associated with lack of CK for abdominal ultrasound, being married was associated with lack of CK for urine analysis and being non-smoker was associated with lack of CK for electrocardiogram.No significant predictors were found concerning mammogram (CK), BSE (CK), Pap smear (CK), fecal occult blood test (CK and AK), sigmoidoscopy (CK) and cardiac ultrasound (CK).
The rates of wrong performance of tests before and after VABB were studied (Table 4).No significant differences were documented with the exception of a statistically significant decrease in the wrong performance of mammogram (p=0.021).

Discussion
This study demonstrates a considerable lack of knowledge regarding breast-related tests in the population of the Breast Unit.Given that these women represent a rather selected population being in contact with the tertiary health services of our Unit, the demonstrated level of poor knowledge may seem rather surprising.Specifically, although women have the rather vague knowledge about the need for mammogram (CK equal to 95%), solely a minority of them (AK equal to 40%) are exactly aware of the recommended screening age and intervals; this seems in accordance with the data provided by the nation-wide Hellas Health I Survey (Dimitrakaki et al., 2009) as well as the studies issued by PACMER (Mauri et al., 2009).Importantly, our study denotes a framework of knowledge considerably poorer than that achieved in Northern European countries, where mammography rates equal to almost 80% have been reported (Hakama et al., 2008).Furthermore, the major economic crisis from which Greece is suffering may aggravate this lack of knowledge.This may be the result of the recession of preventive strategies and corresponding public awareness because of the high costs implicated, which are partially due to the fragmented and cost-ineffective Greek primary care system (Oikonomou and Tountas, 2011).
On the other hand, our study demonstrates a considerable misperception of the need for BSE, as 85% of women believe that BSE is part of the recommended screening strategies (in other words lack of CK equal to 93%).Given that BSE does not offer any benefit in terms of early diagnosis and rather leads to unnecessary benign breast biopsies (Kösters et al., 2003), this result may have significant public health implications.Even more surprisingly, 38% of women reported that breast MRI is a screening test which should be routinely performed, whereas 22% of women did not express any opinion, leaving the item blank.At any case, both results dictate that the education of patients aiming to minimize unnecessary acts (i.e., BSE) and costly tests (i.e., breast MRI) is mandatory; to our knowledge this is the first study reporting such a degree of misperception regarding breast MRI.
With respect to Pap smear, once again, the level of AK was extremely low (41%); once again, the facts in Greece seem far from those reported in Nordic countries, where population coverage reaches 100% (Giorgi and Ronco, 2013).By far, however, the most disappointing conditions demonstrated in our study are those pertaining to colorectal cancer.AK for fecal occult blood testing and sigmoidoscopy were 6% and 5% respectively; this is in accordance with other Greek studies (Kamposioras et al., 2007;Dimitrakaki et al., 2009).At the other end of the spectrum considerable misperception overestimating the need for tests without any proven screening value were documented; the levels of CK were particularly low in the case of urine analysis (10%), cardiac ultrasound (25%), electrocardiogram (26%), abdominal ultrasound (28%) and chest X-Ray (31%).This set of findings largely surpasses and extends the isolated observation by the PACMER investigators according to whom unnecessary chest X-Ray is prevalent in Greece (Kamposioras et al., 2006).
Concerning risk factors for lack of knowledge about recommended screening tests, a variety of associations seem worth commenting.Increasing age was associated with poorer knowledge about Pap smear.This finding may merit comparative evaluation with the nation-wide Hellas I survey, which reported increased likelihood of getting a Pap smear in a specific age group (30-49 vs. 21-29 years), but not in the age 50-69 subgroup (again vs. 21-29 years).Given that the median age in our sample is equal to 50, our finding most probably reflects the relative disadvantage of the 50-69 group described in the Hellas I study.With respect to colorectal cancer, our study confirms the well established link between low income and underscreening (Bronner et al., 2013) but essentially extends it at the level of knowledge.On the other hand, the favourable role of living outside Athens in terms of knowledge about mammogram rather points to a secondary association; women living outside Athens as a rule represent wellinformed, self-selected patients which opt for examination in a tertiary Unit in Athens, such as ours.
Regarding tests without proven screening value, family (as reflected upon number of offspring and marital status) seems to play a multivalent role.Specifically, the protective association implicating chest X-Ray and breast MRI may be inscribed into the well established context depicting married women as better performers of screening (Martin-López et al., 2013).On the other hand the aggravating association pointing to marriage as a risk factor for misconception of urine analysis may reflect a different pattern of sexual activity, the latter being a risk factor for urinary tract infections (Moore et al., 2008); in other words married, sexually active women, may be prescribed urine analyses more frequently and thus seem to have converted this fact to a misperception.Higher BMI and smoking appeared as protective factors for the allocation of screening role to abdominal ultrasound and ECG respectively; it seems more rational to interpret these factors as markers of detachment from screening-related tests in general as demonstrated in the case of obese women (Sangrajrang et al, 2012;Kendall et al, 2013).
Apart from the assessment of knowledge this study also evaluates the practice (performance) of tests.Misperceptions were reflected upon misuse at a certain degree.However, regarding recommended screening tests it is worth commenting that practice rates tended to be more favourable (higher) than AK rates; this may be partly due to the fact that AK encompasses theoretical knowledge about the age of screening onset whereas practice essentially implies following the age-specific physician's recommendations.Regarding tests without proven screening value once again practice rates were more favourable (lower) than CK rates; this may reflect the fact that misperceptions are not actually translated into misuse as inhibiting factors such as fear (Levy et al., 2007), lack of symptoms or interest (Levy et al., 2007), hesitation or reluctance may interfere.Importantly, VABB was not capable of modifying the practice patterns with the exception of mammogram, where improvement in performance was documented; this is in accordance with our previous study regarding the effect of benign breast biopsy in general (Zografos et al., 2010).
This study bears several limitations that should be acknowledged; the study has been performed on a well defined, but selected population i.e. women undergoing VABB.Therefore the results may not be extrapolated to the general population with certainty.In addition the relatively small sample size should be acknowledged but was due to the strict selection criteria of the study.
In conclusion, this study demonstrates considerable lack of knowledge concerning screening tests as well as misperceptions allocating screening role to tests deprived from screening properties.Given that VABB did not modify the main set of performance, the education of women seems mandatory so that the contact with a Breast Unit be not a lost opportunity in terms of public health.