Comprehensive evaluation of cervical cancer screening programs: the case of Colombia

Comprehensive evaluation of cervical cancer screening programs: the case of Colombia Raúl Murillo, MD, MPH, (1) Carolina Wiesner, MD, MPH, (2) Ricardo Cendales, MD, MSc, (2) Marion Piñeros, MD, MSc, (3)
of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Similar Documents
Comprehensive evaluation of cervical cancer screening programs: the case of Colombia Raúl Murillo, MD, MPH, (1) Carolina Wiesner, MD, MPH, (2) Ricardo Cendales, MD, MSc, (2) Marion Piñeros, MD, MSc, (3) Sandra Tovar, Lic en Enf. (2) Murillo R, Wiesner C, Cendales R, Piñeros M, Tovar S. Comprehensive evaluation of cervical cancer screening programs: the case of Colombia. Salud Publica Mex 2011;53: Abstract Objective. To identify critical screening program factors for reducing cervical cancer mortality in Colombia. Material and Methods. Coverage, quality, and screening follow-up were evaluated in four Colombian states with different mortality rates. A case-control study (invasive cancer and healthy controls) evaluating screening history was performed. Results. 3-year cytology coverage was 72.7%, false negative rate 49%, positive cytology follow-up 64.2%. There was no association between screening history and invasive cancer in two states having high cytology coverage but high false negative rates. Two states revealed association between deficient screening history and invasive cancer as well as lower positive-cytology follow-up. Conclusions. Reduced number of visits between screening and treatment is more relevant when low access to health care is present. Improved quality is a priority if access to screening is available. Suitable interventions for specific scenarios and proper appraisal of new technologies are compulsory to improve cervical cancer screening. Comprehensive process-failure audits among invasive cancer cases could improve program evaluation since mortality is a late outcome. Key words: uterine cervical neoplasms; mass screening; program evaluation; cytology; Colombia Murillo R, Wiesner C, Cendales R, Piñeros M, Tovar S. Evaluación integral de los programas de tamizaje de cáncer de cuello uterino: el caso de Colombia. Salud Publica Mex 2011;53: Resumen Objetivo. Identificar factores críticos para reducir la mortalidad por cáncer cervical en Colombia. Material y métodos. Se evaluó cobertura, calidad y seguimiento del tamizaje en cuatro departamentos con tasas de mortalidad diferenciales. Un estudio de casos (cáncer invasor) y controles (sanos) evaluó historia de tamizaje. Resultados. Cobertura 72,7%; falsos negativos 49%; acceso a diagnóstico-tratamiento de HSIL 64,2%. La historia de tamizaje no se asoció con cáncer invasor en dos departamentos con elevada cobertura pero elevada proporción de falsos negativos. Dos departamentos con asociación entre historia de tamizaje deficiente y cáncer invasor tuvieron cobertura aceptable pero bajo acceso a diagnóstico-tratamiento. No hubo relación entre mortalidad y desempeño del programa. Conclusiones. Reducir visitas entre tamizaje y tratamiento es prioritario ante bajo acceso a los servicios. Incrementar calidad es prioritario si hay adecuado acceso al tamizaje. Intervenciones y tecnologías apropiadas a cada contexto son indispensables para obtener mejores resultados. Vigilar integralmente el cáncer invasor contribuye a la evaluación de los programas por ser un desenlace más temprano que la mortalidad. Palabras clave: neoplasias del cuello uterino; tamizaje masivo; evaluación de programas y proyectos en salud; citología; Colombia (1) Subdirección de Investigaciones y Salud Pública, Instituto Nacional de Cancerología de Colombia. Colombia. (2) Grupo de Prevención, Instituto Nacional de Cancerología de Colombia. Colombia. (3) Área de Salud Pública, Instituto Nacional de Cancerología de Colombia. Colombia. Received on: October 22, 2010 Accepted on: September 20, 2011 Corresponding author: Raúl Murillo. Calle 73 No apto Bogotá, Colombia Artículo original Murillo R y col. Cervical cancer incidence and mortality have been reduced up to 80% in developed countries. 1 Despite the absence of clinical trials, the relation between Pap-smear screening and reduction of cervical cancer incidence and mortality has been recognized. 2 The decreased burden of disease has been attributed to screening coverage leading to increased effort for extending conventional cytology to all women worldwide. Although screening programs have been successful in developed nations, they have not fulfilled their objectives in the majority of developing countries. 1,3 In Colombia, cervical cancer remains the first cause of cancer mortality and the second cause of cancer incidence among women. 4 The lack of impact of conventional cytology in low and middle income countries has been linked to social and economic factors as well as deficiencies in program organization. 5 Recent studies revealed that Pap-smear coverage does not correlate with trends of cervical cancer mortality in Latin America; similarly, for countries in the region with decreasing cervical cancer mortality (Mexico, Chile, Costa Rica, Colombia), no clear association between reported organization of screening program and reduction in mortality rates could be determined. 6 The aforementioned situation indicates that screening coverage is not sufficient for cervical cancer control, and highlights the need for better understanding of program components and critical success factors. Accordingly, numerous frameworks for cervical cancer screening program evaluation have been proposed; 7-9 but few have been implemented in Latin America, where most evaluations remain focused on screening coverage. 6 A model on screening components impact on cervical cancer mortality in Colombia revealed greater influence of follow-up of positive screened women (high grade lesion HSIL) than did cytology coverage, since proper cervical intraepitelial neoplasm (CIN) treatment, after screening, is the real factor for preventing invasive cancer. 10 Conversely, a review of evaluative studies demonstrates that evaluation on access to colposcopy and CIN treatment for HSIL cytological diagnosis is scant in Latin America, and inexistent in Colombia. 6,11 In addition, a few countries in the region have reported nationwide information on Pap-smear quality. 6 The analysis of cytology quality is challenging due to the abundant indicators and methodologies; however, two major topics might be considered the most influential on screening effectiveness: smear collection quality and ability of the test for detecting CIN. 9 The available information shows no major differences in Pap-smear sensitivity between Latin America and Europe and North America, thus contributing to the lack of understanding about sustained high mortality rates. Therefore, the National Cancer Institute of Colombia conducted an evaluative study aimed at identifying critical program factors for cervical cancer mortality reduction in the country through analyzing screening coverage, cytology quality, and access to colposcopy and CIN treatment. Every component was evaluated in an independent study; thus, the objective of this report is to combine the results from all components in order to provide a comprehensive analysis of the program. Materials and Methods The protocol was approved by the Ethical Committee at the National Cancer Institute of Colombia (INC). We conducted an evaluative study analyzing four components: screening services supply, screening coverage, conventional cytology quality, and follow-up of positive screened women. Furthermore, a case-control study to evaluate Pap-smear effectiveness was done. The study was carried out in four Colombian states selected by combined levels of mortality and screening organization. Two states with high cervical cancer mortality (Caldas and Tolima), and two states with low cervical cancer mortality (Boyacá and Magdalena) were included. 16 Of the two high mortality states, one revealed satisfactory performance of cervical cancer screening program but the other was deficient, according to Ministry of Health indicators; 17 the same criterion was used for selecting low cervical cancer mortality states (Table I). All four states were included according to feasibility criteria: study affordability, geographical access to municipalities, and security conditions. In total, the four states represent inhabitants and women from different cultural and social backgrounds. Detailed methodology and results for each one of the five abovementioned components has been previously described To analyze screening services supply three kinds of institutions were considered: Pap-smear collection centers, cytology laboratories, and colpsocopy centers. Pap-smear collection centers were included based on an independent simple random sampling for every state. For cytology labs and colposcopy centers total census was reviewed. 18 An audit survey was carried out on human resources availability (nurses, cytotechnologists, pathologists, colposcopists), quality control process, and productivity. Screening coverage was analyzed through the National Survey on Demography and Health (NSDH) A subset of women was interviewed, and questionnaires included questions on Pap-smear screening history. NSDH is a national probabilistic household survey representative of every state including rural and urban areas. Cytology quality was evaluated using a stratified sample of reported Pap-smears in every state (independent universes). 19 Three strata were defined: unsatisfactory smears, negative results, and positive results. Selected Pap-smears and correspondent original reports were sent to INC for a second reading by two blinded expert pathologists, using the Bethesda 2001 system. The second reading recategorized Pap-smears as unsatisfactory, non-positive (normal, low grade squamous intraepithelial lesion -LSIL-, ASCUS), positive (ASC-H, HSIL, cancer), and glandular cell abnormalities. If no agreement was initially obtained, both pathologists defined consensus diagnosis. Agreement between original and INC-expert reports was evaluated by Kappa index and correspondent 95% confidence intervals (95%CI). Additionally, assuming INC-expert report as gold standard, the proportion of negative results disagreed (false negative) and positive results disagreed (false positive) were estimated for every state. A sample of women with HSIL or higher diagnosis was recruited to evaluate access to colposcopy and CIN treatment. A structured survey was conducted and medical records reviewed registering information on compliance with colposcopy-biopsy and CIN treatment during six months after the HSIL report, as well as reasons for lack of attendance. 21 A two-stage sample was designed independently for every state. The first stage selected municipalities in two strata: those with compulsory inclusion (all having colposcopy centers) and those selected by simple random sampling (without replacement). The second stage selected women with HSIL cytological report: among municipalities with compulsory inclusion women were selected by simple random sampling, and for randomly selected municipalities all women were included. For the case-control study, fifty women, ages 25-69, with invasive cervical cancer histological diagnosis were randomly selected in every state from all invasive cancer reports; and fifty neighboring population controls matched by age, and without history of CIN, cervical cancer, or cervical treatment. 22 A structured survey was conducted and medical records reviewed when necessary to determine history of cervical cancer screening within previous 48 months. Cytology exams for followup or diagnosis (symptomatic women) were excluded from analysis. Statistical analysis A sample size of Pap-smear collection centers for analysis of screening services supply was estimated to obtain a variance coefficient lower than 10% for any proportion. For screening coverage component, a proportion near 70-80% (3-year) with relative standard error 1-2% was estimated. 23 For evaluating cytology quality, sample size was calculated expecting a 5% reading error, relative standard error 80%, 95%CI 3-7%, and 20% of Papsmears not to be found. To analyze colposcopy and CIN treatment access, the sample size estimated proportions around 30-40% with variance coefficients 20-40%. For case control study, the sample size was estimated based on McNemar test, assuming 80% power, disparity rate 2:3, and proportion of discordant pairs For screening services supply, screening coverage, and colposcopy-treatment access, all results are presented as expanded data according to specific weighting factors in sample design. Total and ratio estimates and variation coefficients were calculated. For Pap-smear quality, frequencies and percentages are described, as well as non-weighted Kappa with 95%CI. In the casecontrol study, a hierarchic conditional logistic regression model was carried out and adjusted paired odds ratios obtained. All statistical analyses were performed with SPSS software. Results In total 859 Pap-smear collection centers (Boyacá 313, Caldas 128, Magdalena 174, Tolima 244), 52 cytology labs (Boyacá 14, Caldas 7, Magdalena 21, Tolima 10), and 99 colposcopy centers (Boyacá 17, Caldas 40, Magdalena 9, Tolima 33) were identified in the four states. According to number of municipalities, Magdalena and Tolima had highest collection centers rate, Magdalena highest cytology labs rate, and Caldas highest colposcopy centers rate (Table I). No major differences on human resources availability were observed among states. Only Tolima accomplished basic indicators on quality control process. Magdalena and Caldas reported low percentage of positive smears reviewed for quality control (Table I). With the exception of Magdalena, average of smears read annually by cytotechnician surpassed 3000; the highest average of smears by reader was in Caldas (8490.4); however, Caldas revealed lowest average of biopsies per colposcopy (0.4). One-year coverage ranged between 43.3 in Boyacá and 56.8 in Caldas; 3-year coverage from 66.6 in Magdalena and 81.3 in Caldas (Table II). Boyacá and 471 Artículo original Murillo R y col. Table I Study sample characteristics. Colombia Characteristic Boyacá Caldas Magdalena Tolima A. Study population Cervical cancer standardized mortality ratios Screening performance Satisfactory Satisfactory Deficient Deficient Total population Total women Women Municipalities B. Screening resources Infraestructure per municipaliy Pap-smear collection centers Cytology labs Colposcopy centers Human resources Nurses per 1000 women Cytotechnologists per lab Pathologist/cytotechnologist rate Gynecologists per colposcopy center Quality control Cytology labs reporting on Bethesda % 71.4% 61.9% 100.0% Negative smears reviewed 11.3% 9.6% 3.2% 10.9% Positive smears reviewed 76.9% 17.9% 10.3% 100.0% Productivity Pap-smears/nurse (collection) Pap-smears/cytotechnologist (reading) Colposcopies/gynecologist Biopsies/colposcopy Standardized Mortality Ratios use national mortality rates as standard. 15 Screening performance: Optimum performance more than 85% accomplishment of programmed Pap-smears; satisfactory performance 75-85%, deficient performance 40-74%, and very deficient performance less than 40% (Ministry of Health 2000 programmatic goals; Resolución 3384). Population and municipalities: National Department on Statistics DANE; 2005 (available at Cytology labs and colposcopy centers are reported for every 10 municipalities Caldas reported higher coverage among rural population, and Boyacá and Tolima highest percentage of women with no cervical cancer screening history (17.4 and 12.9). Tolima and Magdalena accounted for highest percentage of inadequate smears (13 and 7.5, respectively); and Boyacá and Caldas highest percentage of false negative results (61.1 and 58.8, respectively). Overall percentage of false positive results was low (range 2.4 to 3.8) (Table III). For this component, the sample in Magdalena could not be recruited as planned because only 30.8% of cytology labs stored Pap-smears. Some 19.4% of women with HSIL did not undergo colposcopy in Boyacá and 1.3% did not remember hav- ing colposcopy during the preceding six months. For Caldas these percentages were 19.3 and 9.1 respectively; Magdalena 31.3 and 1.0; Tolima 22.4 and 15.6 (Figure 1A). In total 166 women had indication of treatment in Boyacá, 109 in Caldas, 101 in Magdalena, and 218 in Tolima. From these numbers 17.9%, 5.1%, 4.0%, and 12.4% did not undergo treatment in Boyacá, Caldas, Magdalena, and Tolima, respectively. In summary, in Boyacá 29.4% of women with HSIL cytological report did not access or did not remember confirmatory diagnosis or treatment, Caldas 32.1%, Magdalena 33.5%, and Tolima 41.6% (Figure 1A). Among women who did not attend confirmatory colposcopy-biopsy, in Boyacá 40% reported reasons 472 Table II Urban and rural cervical cancer screening coverage for states in the study. Colombia Coverage Boyacá Caldas Magdalena Tolima 1-year coverage Urban Rural Global year coverage Urban Rural Global Never screened Urban Rural Global Results reported as percentages of women related to health services organization or economic barriers, Caldas 50.9%, Magdalena 59.1%, Tolima 60.8% (Figure 1B). The rate of non-response to questionnaire on reasons for lack of attendance among women with indication of preneoplasic lesions treatment was: Boyacá 25.8%, Caldas 18.2%, Tolima 67.9%, Magdalena 0%. The case-control study showed significant increased risk of cervical cancer according to history of screening in Magdalena and Tolima, but no increased risk in Boyacá and Caldas (Figure 2). There was a significant difference (p 0,01) in average number of Pap-smears between cases and controls during the previous 48 months (Difference 0.45, 0.61, 0.84, and 1.53 for Boyacá, Caldas, Magdalena, and Tolima, respectively). Discussion The reduction of cervical cancer mortality following organization of Pap-smear in European countries has driven implementation of screening programs globally; additionally, some cost-effectiveness models show that program organization might be a more efficient way to decrease burden of disease. 24 Thus, the World Health Organization (WHO) has published guidelines indicating necessary components for an organized program; 6 however, many developed nations have largely reduced cervical cancer mortality without meeting WHO standards. 25 There is abundant literature describing requirements for organized programs, but few studies evaluating impact of organized screening. 26 A meta-analysis revealed that most process failures related to invasive cervical cancer diagnosis are derived from poor screening frequency (low number of Pap-smears or never screened), and only 11.9% are due to poor follow-up of abnormal results; 27 however, these data come from developed nations where follow-up is generally guaranteed. Low screening test sensitivity and lack of correlation between screening coverage and cervical cancer mortality in Latin America 6 suggest that successful screening programs depend upon frequent repetition of Pap-smears and permanent access to confirmatory diagnosis and treatment. 11 Accordingly, our results revealed a significant association between history of screening and risk of invasive cancer only for Magdalena and Tolima; Table III Quality of cytology as percentage of inadequate, positive results disagreement, negative results disagreement, and overall agreement. Colombia State Inadequate smears Negative results disagreement* Positive results disagreement Agreement with initial diagnosis Boyacá ( ) Caldas ( ) Magdalena ( ) Tolima ( ) Inadequate smears based on the second reading at the Instituto Nacional de Cancerología (INC) * Negative results disagreement corresponds to the percentage of negative Pap-smears considered positive in the second reading at INC Positive results disagreement corresponds to the percentage of positive Pap-smears considered negative in the second reading at INC Data correspond to Kappa index, indicating level of agreement between second reading at INC, and original reading by cytotechnologists in the corresponding states Source: Reference Artículo
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!