Disease Misclassification Research Articles

Age-Period-Cohort Analysis of Chronic Obstructive Pulmonary Disease Mortality in Japan, 1950^|^ndash;2004

BackgroundWe examined trends in chronic obstructive pulmonary disease (COPD) mortality in Japan.MethodsAn age-period-cohort model was used to examine trends in COPD mortality by age, time period, and birth cohort among adults aged 40 years or older from 1950 to 2004.ResultsDuring the study period, the age-standardized rate of COPD mortality substantially decreased from 71.3 per 100 000 to 19.7 in men and from 41.7 to 4.3 in women. The period effects rapidly declined during the early period in both sexes. They have increased in recent years in men but have continued to decrease in women. The cohort effects began increasing in the 1865–1869 birth cohort, peaked in the 1880–1889 cohort, and decreased thereafter among the recent cohorts.ConclusionsAn early decrease in the period effects might have been associated with changes in disease structure and misclassification of COPD in the early period. Changes in cohort effects might have been mostly associated with changes in cigarette consumption and smoking prevalence in the Japanese population. Changes in those effects may also be a proxy for other social changes over time.

Misclassification of disease activity when assessing individual patients with early rheumatoid arthritis using disease activity indices that do not include joints of feet

ObjectiveTo study whether assessment of rheumatoid arthritis (RA) disease activity in individual patients using the disease activity score in 28 joints (DAS28) or other instruments excluding joints of feet may...

Does Exposure to Asbestos Cause Ovarian Cancer? A Systematic Literature Review and Meta-analysis

The asbestos and ovarian cancer relationship is not well understood because of small numbers of women exposed to asbestos, small numbers of cases, and misclassification of peritoneal mesothelioma as ovarian cancer on death certificates. The aim of this study was to conduct a meta-analysis to quantify the evidence that exposure to asbestos causes ovarian cancer. Fourteen cohort and two case-control studies were identified in Medline searches from 1950 to 2008. Statistically significant excess mortality was reported in four of the cohort studies, all of which determined their outcomes from the death certificate. Peritoneal mesotheliomas were reported in these studies, two of which reexamined pathology specimens and reported disease misclassification. Exposure-response relationships were inconsistent. When all studies were included in a meta-analysis, the effect size was 1.75 (95% CI, 1.45-2.10) attenuating to 1.29 (95% CI, 0.97-1.73) in studies with confirmed ovarian cancers. Taken without further analysis, women thought to have ovarian cancer had an increased rate in the meta-analysis if reporting having been exposed to asbestos, compared with reference populations. This result may have occurred because of disease misclassification.

Factors associated with misdiagnosis of COPD in primary care

To assess the misclassification of chronic obstructive pulmonary disease (COPD) in Australian primary care. A cross-sectional study was performed in 31 (19%) practices in one Australian state. 341 patients with COPD (database diagnosis or current use of tiotropium plus GP confirmation) completed spirometry and questionnaires. Predictors of misclassification were investigated with multi-level mixed-effects logistic regression allowing for clustering by practice. Spirometric confirmation of COPD (forced expiratory volume in 1 s/forced vital capacity (FEV1/FVC) ratio <0.7) was not present in 107 (31%) patients; 60 (56%) had normal lung function, seven (7%) had scalloped flow-volume curves and FEV1 <80% predicted, 40 (37%) had restriction (FVC <80% predicted). Among 107 misclassified patients the bronchodilators used were tiotropium in 26% and long-acting β2-agonists in 22%. The likelihood of misclassification increased with overweight/obesity (odds ratio (OR) 2.66; 95% CI 1.50 to 4.70) and self-reported allergic rhinitis/hay fever (OR 1.72; 95% CI 1.13 to 2.64) after adjustment for age, gender, and smoking. Symptom-based diagnosis of COPD in primary care is unreliable, especially if patients are overweight, so diagnostic spirometry is essential to avoid inappropriate management.

Abstract 3189: Misclassification of cervical precancer: Impact of a new colposcopy protocol on disease ascertainment and biomarker performance

Abstract Introduction: Cytology-based screening for and treatment of cervical precancers has led to substantial reduction of cervical cancer incidence and mortality worldwide. Current screening approaches have limited accuracy, however, and lead to misclassification of cervical precancer at the levels of primary screening, colposcopic evaluation, and histological evaluation. We designed the Biopsy Study to systematically evaluate the extent of misclassification of cervical precancer at colposcopy. Methods: Previously-untreated women referred to the University of Oklahoma colposcopy clinic for abnormal screening results were enrolled. Before colposcopy, a specimen was taken for liquid-based cytology, HPV genotyping, and biomarker studies. During colposcopy, a digital image of the cervix was taken and annotated for observed lesions and biopsy sites. Up to four colposcopically directed biopsies were taken from distinct lesions; if less than four targeted biopsies were obtained, a random biopsy was added. All biopsies were ranked by severity based on visual impression and evaluated individually in histology. Three cervical cancer screening tests, Pap cytology, HPV DNA testing, and p16/ki67 cytology, were evaluated against disease endpoints based on the most severely appearing biopsy (representing the current clinical standard) and based on the worst histology result from all four biopsies. Results: To date, more than 450 women have been enrolled in the study; 86% of women had at least three, and 64% had four biopsies taken. For 377 women, a histological diagnosis was available. Forty-one of 377 women (10.8%) had CIN3, 119 (31.6%) had CIN2, 82 (35.3%) had CIN1, and 84 (22.3%) had benign changes or normal results as worst histological result. In 70.7% of women, the precancer was detected in the first biopsy, in 22.0% it was found at the second biopsy and in 7.3% it was detected in the third or fourth biopsies. Cervical cancer screening tests were evaluated in a subset of 241 women. The specificity of HPV DNA, Pap cytology, and p16/ki67 for the detection of CIN2+ increased by 6%, 7.5%, and 11.8%, respectively when the improved gold standard was applied. Meanwhile, the sensitivity of all screening tests was not reduced. Discussion: Current colposcopy-biopsy protocols have limited sensitivity in detecting prevalent cervical precancer. In our study, we quantified the incremental benefit of taking multiple biopsies to detect cervical precancer. In the interim analysis, adding a second targeted biopsy improved disease detection by over 20%. The improved sensitivity of colposcopy resulted in decreased misclassification of cervical disease and a more accurate evaluation of cervical cancer screening tests. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 3189. doi:10.1158/1538-7445.AM2011-3189

The risk of cancer in patients with psoriasis: A population-based cohort study in Taiwan

An association between psoriasis and malignancy has been explored. However, no studies have been reported regarding cancer risk in Asian psoriasis populations. The aim of this study was to investigate the risk of cancer development in association with psoriasis. The effects of age at diagnosis, treatment modalities, and observation time were also evaluated. Data for this retrospective population-based cohort study were obtained from the Taiwan National Health Insurance Research Database (NHIRD). This study included 3,686 patients with first-time diagnosis of psoriasis between 1996 and 2000. Another 200,000 patients without psoriasis served as the comparison group. All enrolled subjects were followed-up until the end of 2007. Cumulative incidences and hazard ratios (HRs) of cancer development were determined. Among the 3,686 psoriasis patients, 116 had incident cancers. The 7-year cumulative incidence of cancer among psoriasis patients was 4.8%. The adjusted HR for developing a cancer in association with psoriasis was 1.66 (95% confidence interval [CI], 1.38-2.00). Cancer risk was higher in male patients than in female patients (adjusted HR 1.86 vs.1.14, respectively). Certain cancers were significantly associated with psoriasis, including those of the urinary bladder and skin, followed by oropharynx/larynx, liver/gallbladder, and colon/rectum. Patients with psoriasis had an increased adjusted HR for cancer that varied by age. Younger patients with psoriasis tended to have the greatest risk of cancer. Finally, systemic phototherapy and oral medication did not significantly increase the risk of cancer. Phototherapy with UVB appeared to reduce the risk of cancer in psoriasis (adjusted HR, 0.52; 95% CI, 0.29-0.95; P = .03). NHIRD did not contain information regarding severity of psoriasis, status of smoking, alcohol use, family history of cancer, or diet. Misclassification of disease cannot be ruled out in a registry-based database. Psoriasis carries an elevated risk of malignancies, especially in younger and in male patients. This effect is independent of systemic treatment for psoriasis. Finally, phototherapy with UVB did not increase, but rather reduced, the risk of cancer in psoriasis.

Acute and Long-term Excess Mortality After Polychlorinated Biphenyls and Polychlorinated Dibenzofurans Mixed Exposure From Contaminated Rice Oil: Yusho

PP-29-073 Background/Aims: In Japan in 1968, rice-oil contaminated by polychlorinated biphenyls and polychlorinated dibenzofurans (PCDFs) caused severe food poisoning, termed “Yusho” (oil disease). Several previous studies of Yusho attempted to evaluate the effects of exposure using standard mortality ratios (SMRs), and showed the relationship with total, liver, and lung cancer mortality in men. The findings, however, may have been compromised by disease misclassification or an inappropriate reference population. Therefore, we conducted an area-based SMR study to overcome several limitations of the previous studies. Methods: We adopted the population of a severely affected area (Tamanoura area) as the exposed group, with a reference population from Nagasaki prefecture in Kyushu, which included the Tamanoura area. A large number of residents in Tamanoura were exposed to the rice oil. We estimated SMRs of cancer and noncancer diseases and their confidence intervals (CIs) for the years 1968–2002. Results: In the first 10 years, SMRs of all causes, diabetes mellitus, heart diseases, hypertensive disease, pneumonia/bronchitis, stomach cancer, and bronchus/lung cancer were elevated. SMRs of heart diseases were 2.0 (95% CI: = 1.1–3.6) in 1968, 2.0 (95% CI: = 1.2–3.6) in 1969, and 1.9 (95% CI: = 1.1–3.6) in 1975. SMRs of bronchus/lung cancer were 3.8 (95% CI: = 1.2–11.8) in 1971 and 3.4 (95% CI: = 1.1–11) in 1974. SMRs of several cancers and intentional self-harm were elevated for decades after exposure. Conclusion: In this study, we present further evidence about health effects due to severe exposure to polychlorinated biphenyls and polychlorinated dibenzofurans in Yusho.

SNPs and coronary angiography: can vascular biology be done from angiograms?

This editorial refers to ‘The chromosome 9p21 risk locus is associated with angiographic severity and progression of coronary artery disease’[†][1], by R.S. Patel et al. , on page 3017 Coronary artery disease (CAD) is well known to be, to some degree, heritable. A family history of premature CAD is an established risk factor, and studies in monozygotic twins indicate that this effect is genetic rather than from a shared environment.1 These well established truths have been the driving force behind genetic studies of patients with CAD. An initial wave of studies ∼15 years ago examining informative mutations in candidate genes produced a large collection of underpowered studies that were characterized by lack of reproducibility when tested in larger cohorts, with the exception of variants in the apo e gene.2 The next wave of genetic studies consisted of family-based, linkage studies usually of prematurely affected individuals. These repeated the experience of lack of reproducibility between studies and introduced a new problem, that of identified areas of the genome thought to be causative having no identifiable gene to which a function could be ascribed. Linkage studies did, however, indicate that the task of identifying genetic markers/causes of CAD was going to be tough, with many genes having small additive or interactive effects. The lack of power in the linkage studies, and their vulnerability to disease misclassification, prompted a return to association-based structures but now with herculean numbers of patients and unbiased screening of the whole genome using increasingly dense maps of identified genetic variants [single nucleotide polymorphisms (SNPs)] scattered across the human genome. These have led to the genome-wide association studies (GWAS), which have led to new and potentially very exciting data. Why so much excitement? First, there has now been replication between studies and, secondly, a handful … [1]: #fn-2

043 Adjusting misclassification of outcome in case-control studies

ObjectiveMisclassification of outcome may cause biased estimation for associations of potential risk factors with important diseases. For example, case-control studies of localised prostate cancer frequently measure blood levels of circulating...

Use of nonsteroidal anti‐inflammatory drugs and prostate cancer risk: A meta‐analysis

The association between use of aspirin and other nonsteroidal anti-inflammatory drugs (NSAIDs) and the risk of prostate cancer remains controversial despite many observational epidemiological studies. We conducted a systematic meta-analysis of these studies to examine both the strength and the consistency of the association, and to explore sources of variability between studies. We searched 12 computerized literature databases for reports published before June 2008 and included any epidemiologic studies where the outcome was prostate cancer incidence or mortality, and the exposure was use of NSAIDs. Studies that met the inclusion criteria comprised 10 case-control and 14 cohort studies with a total of 24,230 prostate cancer cases. Studies that assessed the effect of aspirin use on total prostate cancer had a pooled odds ratio (POR) of 0.83 (95%CI: 0.77-0.89), whereas those that assessed the effect of aspirin on advanced prostate cancer had a POR of 0.81 (0.72-0.92). Studies that examined the effects of non-aspirin NSAIDs or all NSAIDs were less consistent but still suggestive of reduced risks. However, most reviewed studies were limited by exposure and disease misclassification, by inadequate information on dose and duration of use and by the possibility of screening and other biases. In conclusion, the epidemiologic evidence for a protective effect of aspirin and other NSAID use against prostate cancer is suggestive but not conclusive. There is a need for well-designed observational studies with adequate exposure measurements, accurate case definition, attention to latency effects, and careful adjustment for screening and other biases.

Prevalence and risk factors for gastroesophageal reflux disease in an impoverished minority population

An epidemiological link between an increased body mass index and complaints of typical heartburn symptoms has been identified. It appears that increasing waist circumference, rather than overall weight is most important. Studies to date have not included minority, impoverished communities. Our aim was to determine the impact of obesity on the prevalence of reflux disease in an impoverished community while controlling for known confounders. Cross-sectional survey delivered by in-home interviews, convenience sampling, and targeted mailing. Data queried include demographics, medical history, lifestyle habits, and symptoms of reflux disease. Height, weight, hip and waist circumference measured in participating subjects. 503 subjects living in the zip code immediately surrounding Temple University Hospital. Included only adults living in the hospital's zip code for at least 3 years. The highest quartile of waist circumference (≥42 in.) demonstrated a strong association with GERD (AOR = 2.15; 95% CI 1.18-3.90). Smoking increased the odds by 1.72 (95% CI 1.13-2.62). There was no relationship between body mass index, waist-hip ratio, or diet and reflux classification. Increasing waist circumference, but not overall body mass index or waist-hip ratio, and smoking are risk factors for prevalent GERD. No association between reflux disease and lifestyle choices such as coffee drinking and fast food dining were found. Potential for recall bias and disease misclassification. Possible methodological errors in self-measurement of waist and hip circumference.

Performance of rheumatoid arthritis disease activity measures and juvenile arthritis disease activity scores in polyarticular‐course juvenile idiopathic arthritis: Analysis of their ability to classify the American College of Rheumatology pediatric measures of response and the preliminary criteria for flare and inactive disease

AbstractObjectiveTo measure the abilities of the continuous measures of disease activity used in rheumatoid arthritis (RA) and the 3 versions of the Juvenile Arthritis Disease Activity Score (JADAS; based upon 10‐, 27‐, and 71‐joint counts) to accurately classify the American College of Rheumatology (ACR) pediatric measures of response, flare, and inactive disease in polyarticular‐course juvenile idiopathic arthritis (JIA).MethodsWe conducted a secondary analysis of a randomized trial of infliximab in polyarticular‐course JIA. Disease activity was calculated at baseline and weeks 14, 28, and 52 using the Disease Activity Score (DAS), DAS in 28 joints, Simplified Disease Activity Index, Clinical Disease Activity Index, and JADAS. The ability of the RA measures and JADAS to classify each ACR pediatric measure, flare, and inactive disease was measured by areas under the receiver operating characteristic curve (AUCs). Positive predictive values (PPVs) for inactive disease were calculated.ResultsData from 97 participants were available. The AUCs for the RA scores for each ACR pediatric measure were 0.73–0.89. The AUCs of the JADAS for the ACR pediatric measures were 0.75–0.92. The PPVs of the RA scores for inactive disease were 0.33–0.67. The PPVs of the JADAS for inactive disease were each 0.93. Based on the RA and JADAS scores, the percentage of visits misclassified as inactive disease ranged from 7–67%.ConclusionThe RA measures and JADAS versions showed acceptable to excellent ability to classify participants for each pediatric outcome measure, but the clinical significance of differences between AUCs for these scores could not be assessed. Misclassification of active disease versus inactive disease by the RA and JADAS scores was not uncommon in this cohort.

Methods to apply probabilistic bias analysis to summary estimates of association

Bias analysis methods are developed for application to 2 x 2 tables, which may be crude or stratified data. Methods for application to associations adjusted for multiple covariates, such as associations from regression modeling, are rarely seen. We have developed probabilistic methods to evaluate bias from disease misclassification or an unmeasured confounder that can be applied to adjusted estimates of association. Rather than applying bias correction methods that rearrange data within 2 x 2 tables, we have applied them to bias factors directly. We illustrate the methods by application to two pharmacoepidemiology problems. In example one, the adjusted odds ratio associating glucocorticoid use with the rate of basal cell carcinoma was 1.15 (95%CI 1.07, 1.25). With bias analysis to account for differential disease misclassification, the median odds ratio was 1.32 and the 95% simulation limits were 1.16 and 1.56. In example two, the adjusted odds ratio associating concomitant use of clopidogrel and proton pump inhibitors with recurrent myocardial infarction was 1.21 (95%CI 0.90, 1.61). With bias analysis to account for confounding by smoking, which was unmeasured, the median odds ratio was 1.15 with 95% simulation interval 0.85 to 1.55. Methods to apply probabilistic bias analysis to adjusted estimates of association can be implemented if a bias factor can be calculated directly from the bias model. This strategy requires that the bias is independent of confounding by measured variables, or requires that the dependence be incorporated into the bias model, as illustrated in an extension of the second example.

Imprecision of Creatinine-Based GFR Estimates in Uninephric Kidney Donors

To ensure long-term safety of living kidney donors, it is now recommended that they be followed for at least 2 years after donation and that serum creatinine levels be monitored. Such levels are often subjected by clinical laboratories to estimating equations and are reported as estimated GFR (eGFR). The accuracy of such equations in uninephric living donors has yet to be validated. This is especially important in older living donors, who often have senescence-related depression of GFR. We compared urinary creatinine clearance, four-variable Modification of Diet in Renal Disease estimating equation (eGFR), and the recently reported CKD-EPI GFR estimating equation with true GFR measured by the urinary iothalamate clearance (iGFR) in 64 subjects after kidney donation. Creatinine clearance overestimated iGFR. Both creatinine-based estimating equations were poorly correlated with and underestimated iGFR. More than half of kidney donors had eGFR <60 ml/min per 1.73 m(2) after donation, a level that categorized them as having stage 3 chronic kidney disease by our current laboratory reporting, whereas only 25% had iGFR <60 ml/min per 1.73 m(2). This misclassification disproportionately affected older donors age > or =55 years, of whom 80% had eGFR <60 ml/min per 1.73 m(2). Neither significant albuminuria nor hypertension was observed. The current practice of reporting eGFR after donation commonly leads to a misclassification of chronic kidney disease, particularly in older donors. To ensure long-term well-being of living kidney donors, more precise estimates of GFR are required, particularly among older potential donors.

Diagnosis of asthma

A significant weakness of the excellent work by Kaplan and colleagues [1][1] relates to a lack of discussion on the spirometric overlap between asthma and chronic obstructive pulmonary disease (COPD) and how this may lead to disease misclassification. This clarification is important since their

Prostate cancer and vasectomy: a hospital-based case–control study in China, Nepal and the Republic of Korea

Background The study of a possible relationship between vasectomy and prostate cancer has yielded mixed results. Data from developing countries are limited. Study Design We conducted a hospital-based case–control study in China, Nepal and the Republic of Korea to evaluate the risk of prostate cancer after vasectomy. Results Prostate cancer in 294 cases (confirmed by independent pathologists) and 879 matched controls were included. The odds ratio of prostate cancer in men with a history of vasectomy was 1.21 [95% confidence interval (95% CI)=0.79, 1.87]. No significant trend was observed in risk by time since vasectomy or age at vasectomy. The odds ratio for localized disease was 1.02 (95% CI=0.53, 1.95); the odds ratio for later stages was 1.41 (95% CI=0.78, 2.53). No confounding factor was identified. The study illustrated differential misclassification of disease by vasectomy status; reference pathologists determined that 28% of men with a history of vasectomy, compared with 17% of men without a history of vasectomy, were misdiagnosed with prostate cancer by local pathologists. Conclusion Vasectomy is not associated with an increased risk of prostate cancer in developing countries where the rate of the disease is low.

The importance of nomenclature for congenital cardiac disease: implications for research and evaluation.

Administrative databases are often used for congenital cardiac disease research and evaluation, with little validation of the accuracy of the diagnostic codes. Metropolitan Atlanta Congenital Defects Program surveillance records were reviewed and classified using a version of the International Pediatric and Congenital Cardiac Code. Using this clinical nomenclature as the referent, we report the sensitivity and false positive fraction (1 - positive predictive value) of the International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis codes for tetralogy of Fallot, transposition of the great arteries, and hypoplastic left heart syndrome. We identified 4918 infants and foetuses with congenital cardiac disease from the surveillance records. Using only the International Classification of Diseases diagnosis codes, there were 280 records with tetralogy, 317 records with transposition, and 192 records with hypoplastic left heart syndrome. Based on the International Pediatric and Congenital Cardiac Code, 330 records were classified as tetralogy, 163 records as transposition, and 179 records as hypoplastic left heart syndrome. The sensitivity of International Classification of Diseases diagnosis codes was 83% for tetralogy, 100% for transposition, and 95% for hypoplastic left heart syndrome. The false positive fraction was 2% for tetralogy, 49% for transposition, and 11% for hypoplastic left heart syndrome. Analyses based on International Classification of Diseases diagnosis codes may have substantial misclassification of congenital heart disease. Isolating the major defect is difficult, and certain codes do not differentiate between variants that are clinically and developmentally different.

Natural History of Severe Ulcerative Colitis in a Community-Based Health Plan

We sought to evaluate long-term outcomes of patients with severe ulcerative colitis (UC) after their first hospitalization for the disease. A cohort of 656 patients hospitalized for UC during 1996 to 2004 was followed up for 9 years through 2004. Time-to-event was estimated using actuarial methods, and the proportions of those under follow-up evaluation who experienced outcomes at month 3, year 1, and year 5 were determined. Outcome measures studied were time to subsequent colectomy, rehospitalization for inflammatory bowel disease, and restarting steroid medication. We also used survival analysis to evaluate whether patient characteristics predicted the risk of rehospitalization and colectomy. Among 656 patients initially hospitalized for severe UC, 20% (N = 129) underwent colectomy during their initial hospitalization. Of the remaining 527, a total of 95% (N = 498) were discharged on a steroid taper. At 1 year after discharge, 29% of those remaining under observation were rehospitalized for UC, and an additional 10% required colectomy. At 1 and 5 years after discharge, 34% and 26% received at least a 90-day supply of steroid in the preceding 9 months. Risk of rehospitalization and colectomy were unrelated to the patient's age, sex, or race/ethnicity. The risk for colectomy in patients hospitalized for the first time with severe UC is 20%. In the 3 months after hospitalization the risk for colectomy is 6%. After that, risks appear to decrease proportionate to the time since initial hospitalization.

Cancer mortality and oil production in the Amazon Region of Ecuador, 1990–2005

To compare cancer mortality rates in Amazon cantons (counties) with and without long-term oil exploration and extraction activities. Mortality (1990 through 2005) and population census (1990 and 2001) data for cantons in the provinces of the northern Amazon Region (Napo, Orellana, Sucumbios, and Pastaza), as well as the province with the capital city of Quito (Pichincha province) were obtained from the National Statistical Office of Ecuador, Instituto Nacional del Estadistica y Censos (INEC). Age- and sex-adjusted mortality rate ratios (RR) and 95% confidence intervals (CI) were estimated to evaluate total and cause-specific mortality in the study regions. Among Amazon cantons with long-term oil extraction, activities there was no evidence of increased rates of death from all causes (RR = 0.98; 95% CI = 0.95-1.01) or from overall cancer (RR = 0.82; 95% CI = 0.73-0.92), and relative risk estimates were also lower for most individual site-specific cancer deaths. Mortality rates in the Amazon provinces overall were significantly lower than those observed in Pichincha for all causes (RR = 0.82; 95% CI = 0.81-0.83), overall cancer (RR = 0.46; 95% CI = 0.43-0.49), and for all site-specific cancers. In regions with incomplete cancer registration, mortality data are one of the few sources of information for epidemiologic assessments. However, epidemiologic assessments in this region of Ecuador are limited by underreporting, exposure and disease misclassification, and study design limitations. Recognizing these limitations, our analyses of national mortality data of the Amazon Region in Ecuador does not provide evidence for an excess cancer risk in regions of the Amazon with long-term oil production. These findings were not consistent or supportive of earlier studies in this region that suggested increased cancer risks.

Reducing Bias from Test Misclassification in Burden of Disease Studies: Use of Test to Actual Positive Ratio – New Test Parameter

To address the problem of estimating disease frequency identified by a diagnostic test, which may not represent the actual number of persons with disease in a community, but rather the number of persons who tested positive. Those two values may be very different, their relationship depending on the properties of the diagnostic test applied and true prevalence of the disease in a population. We defined a new test parameter, the ratio of Test to Actual Positives (TAP), which summarizes the properties of the diagnostic test applied and true prevalence of the disease in a population, and propose that is the most useful summary measure of the potential for bias in disease frequency estimates. A consideration of the relationship between the sensitivity (Se) and specificity (Sp) of the diagnostic test and the true prevalence of disease in a population can inform study design by highlighting the potential for disease misclassification bias. The effects of a decrease in Sp on the TAP ratio at very low disease prevalence are dramatic, as at 80% Sp (and any Se value including 100%), the measured disease frequency will represent a 25-fold overestimate. At a disease prevalence of 0.10, the Sp needs to be 90% or greater to achieve a TAP ratio of 1.0. However, unlike at lower levels of disease prevalence, the test Se is also an important determinant of the TAP ratio. A TAP ratio of 1.0 can be achieved by a Sp of 95% and intermediate Se (40%-60%); or a Sp of 99% and very high Se (over 90%). This illustrates how a test with poor performance characteristics in a clinical setting can perform well in a disease burden study in a population. In circumstances in which the TAP ratio suggests a potential for a large bias, we suggest correction procedures that limit disease misclassification bias and which are often counter-intuitive. We also illustrate how these methods can improve the power of intervention studies, which define outcomes by use of a diagnostic test. Optimal screening test characteristics for use in a population-based survey are likely to be different to those when the test is used in a clinical setting. Calibrating the test a priori to bring the TAP ratio closer to unity deals with the possible large bias in disease burden estimates based on application of diagnostic (screening) test.