No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Diagnosing Intramammary Infection: Controlling Misclassification Bias in Longitudinal Udder Health Studies
Abstract
Using imperfect tests may lead to biased estimates of disease frequency and of associations between risk factors and disease. For instance in longitudinal udder health studies, both quarters at risk and incident intramammary infections (IMI) can be wrongly identified, resulting in selection and misclassification bias, respectively. Diagnostic accuracy can possibly be improved by using duplicate or triplicate samples for identifying quarters at risk and, subsequently, incident IMI.
The objectives of this study were to evaluate the relative impact of selection and misclassification biases resulting from IMI misclassification on measures of disease frequency (incidence) and of association with hypothetical exposures. The effect of improving the sampling strategy by collecting duplicate or triplicate samples at first or second sampling was also assessed.
Data sets from a hypothetical cohort study were simulated and analyzed based on a separate scenario for two common mastitis pathogens representing two distinct prevailing patterns. Staphylococcus aureus, a relatively uncommon pathogen with a low incidence, is identified with excellent sensitivity and almost perfect specificity. Coagulase negative staphylococci (CNS) are more prevalent, with a high incidence, and with milk bacteriological culture having fair Se but excellent Sp. The generated data sets for each scenario were emulating a longitudinal cohort study with two milk samples collected one month apart from each quarter of a random sample of 30 cows/herd, from 100 herds, with a herd-level exposure having a known strength of association. Incidence of IMI and measure of association with exposure (odds ratio; OR) were estimated using Markov Chain Monte Carlo (MCMC) for each data set and using different sampling strategies (single, duplicate, triplicate samples with series or parallel interpretation) for identifying quarters at risk and incident IMI.
For S. aureus biases were small with an observed incidence of 0.29 versus a true incidence of 0.25 IMI/100 quarter-month. In the CNS scenario, diagnostic errors in the two samples led to important selection (40 IMI/100 quarter-month) and misclassification (23 IMI/100 quarter-month) biases for estimation of IMI incidence, respectively. These biases were in opposite direction and therefore the incidence measure obtained using single sampling on both the first and second test (29 IMI/100 quarter-month) was exactly the true value.
In the S. aureus scenario the OR for association with exposure showed little bias (observed OR of 3.1 versus true OR of 3.2). The CNS scenario revealed the presence of a large misclassification bias moving the association towards the null value (OR of 1.7 versus true OR of 2.6). Little improvement could be brought using different sampling strategies aiming at improving Se and/or Sp on first and/or second sampling or using a two out of three interpretation for IMI definition.
Increasing number of samples or tests can prevent bias in some situations but efforts can be spared by holding to a single sampling approach in others. When designing longitudinal studies, evaluating potential biases and best sampling strategy is as critical as the choice of test.
... Then during the follow-up period, case identification should use a highly specific test having a high positive predictive value (33). However (34) have shown that a more prevalent and incident disease diagnosed with an imperfect Se and/or Sp will give biased measure of association despite attempts to improve its diagnosis. ...
... Acknowledgement of these biases and possible corrective measures are important when designing longitudinal studies when gold standard measurement of the outcome might not be readily available, like for bacterial diseases (for example subclinical intramammary infection (39), viral diseases (40) or more complex outcome evaluations (e.g., bovine respiratory disease complex (41). Efforts should be made to improve outcome evaluation but absence or limitation of bias is not always granted in some situation (34). demonstrated that for some specific disease incidences and prevalences bias could not be avoided by improving outcome measurements. ...
Using imperfect tests may lead to biased estimates of disease frequency and measures of association. Many studies have looked into the effect of misclassification on statistical inferences. These evaluations were either within a cross-sectional study framework, assessing biased prevalence, or for cohort study designs, evaluating biased incidence rate or risk ratio estimates based on misclassification at one of the two time-points (initial assessment or follow-up). However, both observations at risk and incident cases can be wrongly identified in longitudinal studies, leading to selection and misclassification biases, respectively. The objective of this paper was to evaluate the relative impact of selection and misclassification biases resulting from misclassification, together, on measures of incidence and risk ratio. To investigate impact on measure of disease frequency, data sets from a hypothetical cohort study with two samples collected one month apart were simulated and analyzed based on specific test and disease characteristics, with no elimination of disease during the sampling interval or clustering of observations. Direction and magnitude of bias due to selection, misclassification, and total bias was assessed for diagnostic test sensitivity and specificity ranging from 0.7 to 1.0 and 0.8 to 1.0, respectively, and for specific disease contexts, i.e., disease prevalences of 5 and 20%, and disease incidences of 0.01, 0.05, and 0.1 cases/animal-month. A hypothetical exposure with known strength of association was also generated. A total of 1,000 cohort studies of 1,000 observations each were simulated for these six disease contexts where the same diagnostic test was used to identify observations at risk at beginning of the cohort and incident cases at its end. Our results indicated that the departure of the estimates of disease incidence and risk ratio from their true value were mainly a function of test specificity, and disease prevalence and incidence. The combination of the two biases, at baseline and follow-up, revealed the importance of a good to excellent specificity relative to sensitivity for the diagnostic test. Small divergence from perfect specificity extended quickly to disease incidence over-estimation as true prevalence increased and true incidence decreased. A highly sensitive test to exclude diseased subjects at baseline was of less importance to minimize bias than using a highly specific one at baseline. Near perfect diagnostic test attributes were even more important to obtain a measure of association close to the true risk ratio, according to specific disease characteristics, especially its prevalence. Low prevalent and high incident disease lead to minimal bias if disease is diagnosed with high sensitivity and close to perfect specificity at baseline and follow-up. For more prevalent diseases we observed large risk ratio biases towards the null value, even with near perfect diagnosis.
... mL definition for some pathogens, such as the non-aureus staphylococci, but would yield substantially higher sensitivity (Dohoo et al., 2011). Nevertheless, for these latter pathogens, Haine et al. (2018) highlighted that, in a cohort study using the ≥1 cfu/0.01 mL case definition, the resulting bias would be toward the null value, and that using IMI definitions that are less sensitive and more specific (such as the ≥2 cfu/0.01 ...
The objective of this randomized controlled trial was to assess the efficacy of an on-farm culture system using Petrifilm (3M, London, ON, Canada) for targeted treatment decisions at the quarter level at dry-off and its effects on dry period intramammary infections (IMI) and udder health and milk production in the subsequent lactation. A total of 568 cows (2,247 quarters) from 9 dairy herds with bulk tank somatic cell count <250,000 cells/mL in Québec, Canada, were systematically enrolled and randomly allocated to 4 groups: 2 quarter-based selective (QSDCT) groups, using results of quarter-milk culture on Petrifilm, and 2 blanket dry cow therapy (BDCT) groups. The 2 QSDCT groups consisted of (1) antimicrobial to infected quarters and internal teat sealant (ITS) to healthy quarters (QSDCT/ITS); and (2) antimicrobial and ITS to infected quarters and ITS to healthy quarters (QSDCT+ITS/ITS). The 2 BDCT groups were (1) antimicrobial alone to all quarters (BDCT); and (2) antimicrobial and ITS to all quarters (BDCT+ITS). Quarter milk samples were collected at dry-off and after calving for routine bacteriological culture at the laboratory to monitor IMI; data on milk production, somatic cell count, and clinical mastitis recorded up to 120 d in milk were retrieved from health and DHI records. The probability of avoiding antimicrobial treatment in QSDCT groups was estimated at 48.3% (95% confidence interval: 35.7, 60.9). There was no significant difference between the 4 treatment groups regarding acquisition of new IMI (15.9, 13.2, 15.8, and 15.1% probability for BDCT, BDCT+ITS, QSDCT/ITS, and QSDCT+ITS/ITS, respectively) or persistence of existing IMI (3.2, 2.1, 3.4, and 2.7% probability, respectively) over the dry period. In the subsequent lactation, there was no difference between groups regarding incidence of clinical mastitis (2.4, 3.7, 2.9, and 1.7% respectively for BDCT, BDCT+ITS, QSDCT/ITS, and QSDCT+ITS/ITS), mean milk somatic cell score (1.7, 2.0, 2.0, and 2.0 respectively), or mean daily milk production (43.8, 44.2, 43.2, and 42.6 kg/d, respectively) during the first 120 d in milk. In conclusion, QSDCT using the Petrifilm on-farm culture system to detect infected quarters at dry-off is an interesting option to decrease antibiotic use without any negative effects on udder health or milk production in the first 120 d of the subsequent lactation compared with BDCT.
Mastitis (intramammary inflammation) caused by infectious pathogens is still considered a devastating condition of dairy animals affecting animal welfare as well as economically incurring huge losses to the dairy industry by means of decreased production performance and increased culling rates. Bovine mastitis is the inflammation of the mammary glands/udder of bovines, caused by bacterial pathogens, in most cases. Routine diagnosis is based on clinical and subclinical forms of the disease. This underlines the significance of early and rapid identification/detection of etiological agents at the farm level, for which several diagnostic techniques have been developed. Therapeutic regimens such as antibiotics, immunotherapy, bacteriocins, bacteriophages, antimicrobial peptides, probiotics, stem cell therapy, native secretory factors, nutritional, dry cow and lactation therapy, genetic selection, herbs, and nanoparticle technology-based therapy have been evaluated for their efficacy in the treatment of mastitis. Even though several strategies have been developed over the years for the purpose of managing both clinical and subclinical forms of mastitis, all of them lacked the efficacy to eliminate the associated etiological agent when used as a monotherapy. Further, research has to be directed towards the development of new therapeutic agents/techniques that can both replace conventional techniques and also solve the problem of emerging antibiotic resistance. The objective of the present review is to describe the etiological agents, pathogenesis, and diagnosis in brief along with an extensive discussion on the advances in the treatment and management of mastitis, which would help safeguard the health of dairy animals.
Aim
primarily to estimate the sensitivity (Se) and specificity (Sp) of the commercially available Mastit4 qPCR assay and bacterial culture (BC) for diagnosis of intramammary infections (IMI) and teat apex colonization (TAC) with coagulase‐negative staphylococci (CNS) at different cut‐offs for qPCR cycle threshold (Ct) values using Bayesian latent class analysis (LCA). A secondary objective was to evaluate two cut‐offs of BC for diagnosis of IMI and TAC with CNS.
Methods and Results
we randomly selected 13 to 20 cows with subclinical mastitis from 8 dairy herds. Teat skin samples and aseptically collected foremilk samples were collected from the right hindquarters (n = 149) for BC and qPCR analysis. The Se of qPCR was always higher than BCSe in diagnosis of IMI, however; the Sp of BC was higher than qPCRSp. BCSe and BCSp showed no substantial difference between the tested BC cut‐offs. In contrast to IMI, estimates of BC and qPCR in diagnosing TAC were different. BCSe was higher than qPCRSe at all tested cut‐offs, however; qPCRSp was higher than BCSp.
Conclusion
the overall performance of qPCR is higher than BC in the diagnosis of IMI however; the performance of BC is better than qPCR in diagnosis of TAC. The qPCR and BC are valid diagnostics for bovine IMI with CNS. Whereas for TAC, both techniques require further investigation to reduce the uncertainty of the true status of the quarter and teat skin.
Significance and Impact of the Study
we reported, for the first time, the diagnostic performance of new mastitis technology (Mastit4 PCR) and culture for detection of CNS in milk and non‐milk samples in dairy herds with automatic milking systems. Our findings will improve the interpretation of the test results of culture and qPCR assay and subsequently, will strengthen the control of IMI with CNS in dairy cows.
Criteria for diagnosing intramammary infections (IMI) have been debated for many years. Factors that may be considered in making a diagnosis include the organism of interest being found on culture, the number of colonies isolated, whether or not the organism was recovered in pure or mixed culture, and whether or not concurrent evidence of inflammation existed (often measured by somatic cell count). However, research using these criteria has been hampered by the lack of a "gold standard" test (i.e., a perfect test against which the criteria can be evaluated) and the need for very large data sets of culture results to have sufficient numbers of quarters with infections with a variety of organisms. This manuscript used 2 large data sets of culture results to evaluate several definitions (sets of criteria) for classifying a quarter as having, or not having an IMI by comparing the results from a single culture to a gold standard diagnosis based on a set of 3 milk samples. The first consisted of 38,376 milk samples from which 25,886 triplicate sets of milk samples taken 1 wk apart were extracted. The second consisted of 784 quarters that were classified as infected or not based on a set of 3 milk samples collected at 2-d intervals. From these quarters, a total of 3,136 additional samples were evaluated. A total of 12 definitions (named A to L) based on combinations of the number of colonies isolated, whether or not the organism was recovered in pure or mixed culture, and the somatic cell count were evaluated for each organism (or group of organisms) with sufficient data. The sensitivity (ability of a definition to detect IMI) and the specificity (Sp; ability of a definition to correctly classify noninfected quarters) were both computed. For all species, except Staphylococcus aureus, the sensitivity of all definitions was <90% (and in many cases<50%). Consequently, if identifying as many existing infections as possible is important, then the criteria for considering a quarter positive should be a single colony (from a 0.01-mL milk sample) isolated (definition A). With the exception of "any organism" and coagulase-negative staphylococci, all Sp estimates were over 94% in the daily data and over 97% in the weekly data, suggesting that for most species, definition A may be acceptable. For coagulase-negative staphylococci, definitions B (2 colonies from a 0.01-mL milk sample) raised the Sp to 92 and 95% in the daily and weekly data, respectively. For "any organism," using definition B raised the Sp to 88 and 93% in the 2 data sets, respectively. The final choice of definition will depend on the objectives of study or control program for which the sample was collected.
This study was designed to identify risk factors for intramammary infections with Streptococcus uberis and Staphylococcus aureus under field conditions. An 18-mo survey with sampling of all quarters of all lactating cows at 3-wk intervals was carried out in three Dutch dairy herds with medium bulk milk somatic cell count (200,000 to 300,000 cells/ml). Quarter milk samples were used for bacteriology and somatic cell counting. Data on parity, lactation stage, and bovine herpesvirus 4-serology were recorded for each animal. During the last year of the study, body condition score, and teat-end callosity scores were recorded at 3-wk intervals. A total of 93 new infections with Strep. uberis were detected in 22,665 observations on quarters at risk for Strep. uberis infection, and 100 new infections with Staph. aureus were detected in 22,593 observations on quarters at risk for Staph. aureus infection. Multivariable Poisson regression analysis with clustering at herd and cow level was used to identify risk factors for infection. Rate of infection with Strep. uberis was lower in first- and second-parity cows than in older cows, and depended on stage of lactation in one herd. Quarters that were infected with Arcanobacterium pyogenes or enterococci, quarters that had recovered from Strep. uberis- or Staph. aureus-infection in the past, and quarters that were exposed to another Strep. uberis infected quarter in the same cow had a higher rate of Strep. uberis infection. Teat-end callosity and infection with coagulase-negative staphylococci or corynebacteria were not significant as risk factors. Rate of Staph. aureus infection was higher in bovine herpesvirus 4-seropositive cows, in right quarters, in quarters that had recovered from Staph. aureus or Strep. uberis infection, in quarters exposed to other Staph. aureus infected quarters in the same cow, and in quarters with extremely callused teat ends. Infection with coagulase-negative staphylococci was not significant as a risk factor. The effect of infection with corynebacteria on rate of infection with Staph. aureus depended on herd, stage of lactation, and teat-end roughness. Herd level prevalence of Strep. uberis or Staph. aureus, and low quarter milk somatic cell count were not associated with an increased rate of infection for Strep. uberis or Staph. aureus.
Misclassification errors caused by imperfect sensitivity (U) and specificity (V) can affect statistical inferences in epidemiology. Such errors can lead to biases and increased standard errors in estimates of rates. Furthermore, low U and V can have a catastrophic effect on the power of a test to detect a change in rate, and, if U and V change even slightly as the rate changes, the effect on power may be dramatic.
Objectives of this study were to identify the manageable risk factors associated with the lactational incidence, elimination, and prevalence of coagulase-negative staphylococci (CNS) intramammary infections (IMI) while taking into account the difficulties inherent to their diagnosis. A second objective was to evaluate the effect of CNS IMI misclassification in mastitis research. A cohort of 90 Canadian dairy herds was followed throughout 2007 to 2008. In each herd, series of quarter milk samples were collected from a subsample of cows and bacteriological culture was performed to identify prevalent, incident, and eliminated CNS IMI. Practices used on farms were captured using direct observations and a validated questionnaire. The relationships between herd CNS IMI prevalence and herd incidence and elimination rates were explored using linear regression. Manageable risk factors associated with the prevalence, incidence, or elimination of CNS IMI were identified via Bayesian analyses using a latent class model approach, allowing adjustment of the estimates for the imperfect sensitivity and specificity of bacteriological culture. After adjustment for the diagnostic test limitations, a mean CNS IMI quarter prevalence of 42.7% [95% confidence interval (CI): 34.7, 50.1] and incidence and elimination rates of 0.29 new IMI/quarter-month (95% CI: 0.21, 0.37) and 0.79 eliminated IMI/quarter-month (95% CI: 0.66, 0.91), respectively, were observed. Considerable biases of the estimates were observed when CNS IMI misclassification was ignored. These biases were important for measures of association with risk factors, were almost always toward the null value, and led to both type I and type II errors. Coagulase-negative staphylococci IMI incidence appeared to be a stronger determinant of herd IMI prevalence than IMI elimination rate. The majority of herds followed were already using blanket dry cow treatment and postmilking teat disinfection. A holistic approach considering associations with all 3 outcomes was used to interpret associations between manageable risk factors and CNS IMI. Sand and wood-based product bedding showed desirable associations with CNS IMI compared with straw bedding. Quarters of cows that had access to pasture during the sampling period had lower odds of acquiring a new CNS IMI and of having a prevalent CNS IMI. Many practices showed an association with only one of the CNS outcomes and should, therefore, be considered with caution.
Staphylococcus aureus intramammary infections (IMI) are a major cause of mastitis on farms worldwide. Incidence and elimination rates are the key determinants of prevalence of Staph. aureus, and risk factors associated with these rates must be identified, prioritized, and controlled to obtain long-term reduction in prevalence. The objectives of this study were to identify manageable risk factors associated with the lactational incidence, elimination, and prevalence of Staph. aureus IMI. A cohort of 90 Canadian dairy farms was recruited and followed in 2007 and 2008. Quarter milk samples were collected repeatedly from a selection of cows, and bacteriological culture was realized to assess incidence, elimination, and prevalence of Staph. aureus IMI. Practices used on farms were measured using direct observations and a validated questionnaire. A linear regression model was used to explore the relationship between herd IMI prevalence and incidence and elimination rates. Multilevel logistic regression models were used to compute measures of associations between practices used on farms and IMI incidence, elimination, and prevalence. The herd incidence rate was the most important predictor of herd IMI prevalence: a reduction of the incidence rate equivalent to its interquartile range (0.011 new IMI/quarter-month) was associated with a prevalence reduction of 2.2 percentage points; in comparison, an equivalent increase of the elimination rate by its interquartile range (0.36 eliminated IMI/quarter-month) resulted in a prevalence reduction of 0.4 percentage points. Postmilking teat disinfection and blanket dry-cow therapy were already implemented by most herds. Most of the practices associated with Staph. aureus IMI incidence were related to milking procedures. Among these, wearing gloves during milking showed desirable associations with IMI incidence, elimination, and prevalence. Similarly, adequate teat-end condition and use of premilking teat disinfection were associated with lower IMI incidence and prevalence. The initial herd prevalence of Staph. aureus IMI was positively associated with subsequent IMI incidence. This indicates that, in some situations, an initial reduction of the pool of infected quarters could be justified. Some housing practices were associated with IMI incidence, elimination, or prevalence. The effects of these latter practices, however, were often influenced by specific cow characteristics such as parity or days in milk. These results highlight the importance of good milking practices to prevent Staph. aureus IMI acquisition and, therefore, reduce their prevalence.
The objective was to examine the potential benefits of using different combinations of multiple quarter milk samples compared with a single sample for diagnosing intramammary infections (IMI) in dairy cattle. Data used in the analyses were derived from 7,076 samples from 667 quarters in 176 cows in 8 herds in 4 locations (Minnesota/Wisconsin, n=4; Prince Edward Island, n=2; Ontario, n=1; New York, n=1). Duplicate quarter milk samples were collected at morning milking for 5 consecutive days. Cows were evenly distributed between early postparturient and mid- to late-lactation cows. All samples were frozen for shipping and storage, thawed once, and cultured in university laboratories using standardized procedures consistent with National Mastitis Council guidelines. The presence of specific pathogens was confirmed and identified using the API identification system (bioMerieux, Marcy l'Etoile, France) in each laboratory. A previously developed gold standard was applied to the first sample from d 1, 3, and 5 to classify infected quarters. The data were analyzed separately for coagulase-negative staphylococci (CNS) and Streptococcus spp. Various combinations of test results from d 2 and 4 were used in the test evaluation. These consisted of single samples (n=4), 2 sets of duplicate samples (2 samples collected on the same day), 2 sets of consecutive samples (2 samples collected 2 d apart), and 2 sets of triplicate samples (2 samples on the same day and a third sample 2 d apart). Series interpretation of duplicate or consecutive samples (i.e., positive=same pathogen isolated from both samples) resulted in the highest specificity (Sp; CNS Sp=92.1-98.1%; Streptococcus spp. Sp=98.7-99.6%), but lowest sensitivity (Se; CNS Se=41.9-53.3%; Streptococcus spp. Se=7.7-22.2%). Parallel interpretation of duplicate or consecutive samples (i.e., positive=pathogen isolated from either) resulted in the highest Se (CNS Se=70.8-80.6%; Streptococcus spp. Se=31.6-48.1%), but lowest Sp (CNS Sp=72.0-77.3%; Streptococcus spp. Sp=89.5-93.3%). The difference in estimates between single and duplicate samples was larger than between single and consecutive samples. Overall, triplicate samples provided the best combination of Se and Sp, but compared with a single sample, provided only a modest gain in Sp and little or no gain in Se.
The most common practice to analyze epidemiological follow-up studies is to analyze risk factors of new, i.e. incident, cases of disease. However, analysis of incidence assumes that diseases exist in true dichotomies, which is unlikely to be true. It has also recently been shown that in many typical situations it is very difficult to separate the association between risk factors of disease at baseline and during follow-up using analyses of incidence. Situation is especially problematic for diseases that have large misclassification and low incidence, like asthma. We suggest that reliance on analysis of incidence may be a major obstacle into discovering causes of such disease. Only with greater attention into how to define and how to analyze prospective studies are we likely to learn sufficiently of risk factors of such disease to finally arrive at means for their prevention.
This paper deals with some basis properties of screening tests. Such tests purport to separate people with disease from people without. Minimal criteria for such a process to be a test are discussed. Various ways of judging the goodness of a test are examined. A common use of tests is to estimate prevalence of disease; frequency of positive tests is shown to be a bad estimate, and the necessary adjustmants are given.
Lack of bias in the estimation of relative effect in epidemiologic studies depends on the internal validity of the study. This paper conveys in graphic and tabular form the direction and magnitude of bias due to misclassification of study subjects. A series of computer-generated graphs shows that the departure of the estimate of effect (relative risk or odds ratio) from its true value is a function of sensitivity and specificity (measures of classification validity), disease frequency, and exposure frequency. The discussion of bias emphasizes misclassification of the "outcome" variable; i.e., disease occurrence in a cohort study and exposure rate in a case-control study. Examples are used to illustrate that the magnitude of the bias can be large under circumstances which occur readily in epidemiologic research. When misclassification is equal for the two compared groups, the estimate is biased toward the null value, and in some instances beyond; when differential misclassification occurs (as in selective recall in case-control studies) the bias can be in either direction, and may be great. Formulas are derived to estimate the underlying true value of the relative risk or odds ratio using the investigator's observations together with the estimated sensitivity and specificity of the classification procedure.
Measures of association for cross classifications are frequently used in expressing the relationship between traits in the medical literature. These indices are often employed to measure the strength of association between different characteristics, and to hierarchically order the relationship between different traits and the incidence of a disease. The effect of systematic errors of classification on the index known as relative risk is presented together with an approach of use in the presence of misclassification.
It is often required to evaluate the accuracy of a new diagnostic test against a standard test with unknown error rates. If the two tests are applied simultaneously to the same individuals from two populations with different disease prevalences, then assuming conditional independence of the errors of the two tests, the error rates of both tests and the true prevalences in both populations can be estimated by a maximum likelihood procedure. Generalizations to several tests applied in several populations are also possible.
The effects of misclassification on analyses involving a discrete covariate are examined. The following points are illustrated: 1) Analogous to the 2 X 2 table case, unbiased misclassification of the study exposure leads to reduction in the observed strength of the association of exposure with disease. 2) Both biased and unbiased misclassification will tend to distort the degree of heterogeneity in the measure of association being considered. 3) Misclassification of a confounder leads to a partial loss of ability to control confounding.
Misclassification problems of the disease status often arise in large epidemiologic cohort studies in which the outcome is classified on the basis of record linkage with routinely collected error-prone data sources, such as cancer registries or mortality statistics. If the misclassification is nondifferential, i.e., independent of the exposure status, this leads to bias toward the null in estimates of relative risk. A variety of methods have been proposed to correct for this bias. Most approaches are based on estimates of the sensitivity and specificity of disease classification from validation studies, which typically require invasive and time-consuming diagnostic procedures. For ethical and practical reasons, such procedures may often not be applied on individuals classified as not having the disease, in which case estimates of sensitivity and specificity cannot be obtained. In this paper, an alternative correction method is proposed based on estimates of the positive predictive value, which requires validation of the diagnosis among samples of individuals classified as having the disease only. The method is applicable in situations with either differential or nondifferential specificity of disease classification as long as the sensitivity is nondifferential. Point estimates and large-sample interval estimates of the corrected relative risk are algebraically derived. The performance of the method is assessed by extensive simulations and found to be satisfactory even for small sample sizes.
In epidemiologic research, logistic regression is often used to estimate the odds of some outcome of interest as a function of predictors. However, in some datasets, the outcome of interest is measured with imperfect sensitivity and specificity. It is well known that the misclassification induced by such an imperfect diagnostic test will lead to biased estimates of the odds ratios and their variances. In this paper, the authors show that when the sensitivity and specificity of a diagnostic test are known, it is straightforward to incorporate this information into the fitting of logistic regression models. An EM algorithm that produces unbiased estimates of the odds ratios and their variances is described. The resulting odds ratio estimates tend to be farther from the null but have greater variance than estimates found by ignoring the imperfections of the test. The method can be extended to the situation where the sensitivity and specificity differ for different study subjects, i.e., nondifferential misclassification. The method is useful even when the sensitivity and specificity are not known, as a way to see the degree to which various assumptions about sensitivity and specificity affect one's estimates. The method can also be used to estimate sensitivity and specificity under certain assumptions or when a validation subsample is available. Several examples are provided to compare the results of this method with those obtained by standard logistic regression. A SAS macro that implements the method is available on the World Wide Web at http:@som1.ab.umd.edu/Epidemiology/software.h tml.
We present methods for binomial regression when the outcome is determined using the results of a single diagnostic test with imperfect sensitivity and specificity. We present our model, illustrate it with the analysis of real data, and provide an example of WinBUGS program code for performing such an analysis. Conditional means priors are used in order to allow for inclusion of prior data and expert opinion in the estimation of odds ratios, probabilities, risk ratios, risk differences, and diagnostic test sensitivity and specificity. A simple method of obtaining Bayes factors for link selection is presented. Methods are illustrated and compared with Bayesian ordinary binary regression using data from a study of the effectiveness of a smoking cessation program among pregnant women. Regression coefficient estimates are shown to change noticeably when expert prior knowledge and imperfect sensitivity and specificity are incorporated into the model.
The concept of bias is the lack of internal validity or incorrect assessment of the association between an exposure and an effect in the target population in which the statistic estimated has an expectation that does not equal the true value. Biases can be classified by the research stage in which they occur or by the direction of change in a estimate. The most important biases are those produced in the definition and selection of the study population, data collection, and the association between different determinants of an effect in the population. A definition of the most common biases occurring in these stages is given.
When estimating incidence risk ratios in follow-up studies, subjects testing positive for the disease at baseline are excluded. Although the effect of disease misclassification on estimated incidence risk ratios has otherwise been extensively explored, the effect of disease misclassification at baseline has not previously been analyzed.
The design was theoretical calculations assuming dichotomous disease and a follow-up study with a baseline and a follow-up examination, analyzed using cumulative incidence. Calculations consider nondifferential misclassification of disease mainly at baseline, but no misclassification of exposure.
Nondifferential misclassification of disease at baseline can lead to bias either away or toward null in estimated cumulative incidence risk ratios. This bias is mainly a function of sensitivity at baseline, because imperfect sensitivity leads to failure to exclude all diseased subjects from the follow-up. Imperfect specificity at baseline has less effect. Bias is increased with high true prevalence of disease and low true incidence. Bias is also increased with large differences in true risk ratios at baseline and at follow-up, because observed incidence risk ratios in the presence of misclassification reflect both the true association at baseline and at follow-up.
Nondifferential disease misclassification at baseline examination of a follow-up study can lead to over- or underestimation of the cumulative incidence risk ratios. The bias can be substantial for disease with low incidence and high prevalence, such as asthma or myocardial infarction. The results underscore the need to select a highly sensitive test for disease at baseline to exclude all diseased subjects from the follow-up.