On the bias caused by regression toward the mean in studying the relation between change and initial value

How to Test the Association Between Baseline Performance Level and the Modulatory Effects of Non-Invasive Brain Stimulation Techniques

Behavioral effects of non-invasive brain stimulation techniques (NIBS) can dramatically change as a function of different factors (e.g., stimulation intensity, timing of stimulation). In this framework, lately there has been a growing interest toward the importance of considering the inter-individual differences in baseline performance and how they are related with behavioral NIBS effects. However, assessing how baseline performance level is associated with behavioral effects of brain stimulation techniques raises up crucial methodological issues. How can we test whether the performance at baseline is predictive of the effects of NIBS, when NIBS effects themselves are estimated with reference to baseline performance? In this perspective article, we discuss the limitations connected to widely used strategies for the analysis of the association between baseline value and NIBS effects, and review solutions to properly address this type of question.

The impact of adjusting for baseline in pharmacogenomic genome-wide association studies of quantitative change

In pharmacogenomic studies of quantitative change, any association between genetic variants and the pretreatment (baseline) measurement can bias the estimate of effect between those variants and drug response. A putative solution is to adjust for baseline. We conducted a series of genome-wide association studies (GWASs) for low-density lipoprotein cholesterol (LDL-C) response to statin therapy in 34,874 participants of the Genetic Epidemiology Research on Adult Health and Aging (GERA) cohort as a case study to investigate the impact of baseline adjustment on results generated from pharmacogenomic studies of quantitative change. Across phenotypes of statin-induced LDL-C change, baseline adjustment identified variants from six loci meeting genome-wide significance (SORT/CELSR2/PSRC1, LPA, SLCO1B1, APOE, APOB, and SMARCA4/LDLR). In contrast, baseline-unadjusted analyses yielded variants from three loci meeting the criteria for genome-wide significance (LPA, APOE, and SLCO1B1). A genome-wide heterogeneity test of baseline versus statin on-treatment LDL-C levels was performed as the definitive test for the true effect of genetic variants on statin-induced LDL-C change. These findings were generally consistent with the models not adjusting for baseline signifying that genome-wide significant hits generated only from baseline-adjusted analyses (SORT/CELSR2/PSRC1, APOB, SMARCA4/LDLR) were likely biased. We then comprehensively reviewed published GWASs of drug-induced quantitative change and discovered that more than half (59%) inappropriately adjusted for baseline. Altogether, we demonstrate that (1) baseline adjustment introduces bias in pharmacogenomic studies of quantitative change and (2) this erroneous methodology is highly prevalent. We conclude that it is critical to avoid this common statistical approach in future pharmacogenomic studies of quantitative change.

Prediagnostic Body Mass Index Trajectories in Relation to Prostate Cancer Incidence and Mortality in the PLCO Cancer Screening Trial

Background

Evidence suggests that obesity in adulthood is associated with increased risk of "clinically significant" prostate cancer. However, studies of body mass index (BMI) across the adult life course and prostate cancer risks remain limited.

Methods

In a prospective cohort of 69 873 men in the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial, we examined associations of prediagnostic BMI across the adult life course with risk of incident prostate cancer and fatal prostate cancer (prostate cancer-specific mortality). At 13 years of follow-up, we identified 7822 incident prostate cancer cases, of which 3078 were aggressive and 255 fatal. BMI trajectories were determined using latent-class trajectory modeling. Cox proportional hazards regression was used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs).

Results

BMI at age 20 years, 50 years, and baseline questionnaire (mean age = 63 years) were associated with increased risks of fatal prostate cancer (HRs = 1.27-1.32 per five-unit increase). In five BMI trajectories identified, fatal prostate cancer risk was increased in men who had a normal BMI (HR = 1.95, 95% CI = 1.21 to 3.12) or who were overweight (HR = 2.65, 95% CI = 1.35 to 5.18) at age 20 years and developed obesity by baseline compared with men who maintained a normal BMI. Aggressive and nonaggressive prostate cancer were not associated with BMI, and modest inverse associations were seen for total prostate cancer.

Conclusions

Our results suggest that BMI trajectories during adulthood that result in obesity lead to an elevated risk of fatal prostate cancer.

A multilevel modelling solution to mathematical coupling

Owing to mathematical coupling, statistical analyses relating change to baseline values using correlation or regression are erroneous, where the statistical procedure of testing the null hypothesis becomes invalid. Alternatives, such as Oldham’s method and the variance ratio test, have been advocated, although these are limited in the presence of measurement errors with non-constant variance. Furthermore, such methods prohibit the consideration of additional covariates (e.g., treatment group within trials) or confounders (e.g., age and gender). This study illustrates the more sophisticated approach of multilevel modelling (MLM) which overcomes these limitations and provides a comprehensive solution to the analysis of change with respect to baseline values. Although mathematical coupling is widespread throughout applied research, one particular area where several studies have suggested a strong relationship between baseline disease severity and treatment effect is guided tissue regeneration (GTR) within dental research. For illustration, we use GTR studies where the original data were found to be available in the literature for reanalysis. We contrast the results from an MLM approach and Oldham’s method with the standard (incorrect) approach that suffers from mathematical coupling. MLM provides a robust solution when relating change to baseline and is capable of simultaneously dealing with complex error structures and additional covariates and/or potential confounders.

Testing the relation between percentage change and baseline value

Testing the relation between percentage change and baseline value has been controversial, but it is not clear why this practice may yield spurious results. In this paper, we first explained why the usual testing of the relation between percentage change and baseline value is inappropriate and then demonstrated how the appropriate null hypothesis could be formulated. We also proposed a simple procedure for testing the appropriate null hypothesis based on the assumption that when there is no relation between percentage change and baseline value, the coefficients of variation for repeated measurements of a random variable should remain unchanged. Two examples were used to demonstrate how the usual testing gave rise to misleading results, whilst results from our simple test were in general consistent with those from simulations. We also undertook simulations to investigate the impact of measurement errors on the performance of the proposed test. Results suggested the type-I error rates increased with the magnitude of measurement errors, whilst the statistical power to detect a genuine relation decreased. The usual approach to testing the relation between percentage change and baseline value tended to yield misleading results and should be avoided.

Defining and predicting outcomes of non-surgical periodontal treatment: a 1-yr follow-up study

This study reports on 1-yr outcomes of non-surgical periodontal therapy and compares predictive models resulting from different definitions of treatment success. A total of 149 participants, 30-70 yr of age, provided clinical periodontal data and data on sociodemographic status, health status, symptoms, and oral health-care behaviors at baseline. One week later, clinical attachment level and probing pocket depth were recorded again in 148 patients. Participants underwent non-surgical periodontal therapy, including scaling and root planing, during three to four clinical sessions. Three and 12 months later, clinical attachment level, probing pocket depth, and bleeding on probing (BOP) were recorded in 141 and 137 participants, respectively. Using test-retest data, patients were classified as having 'downhill', 'stable', or 'improved' results on three clinical attachment level and three probing pocket depth outcomes, and their classification was found to vary considerably according to outcome. Although the predictors of treatment outcome varied depending on the variable chosen to represent the treatment outcome, some predictors were more commonly noted as predicting improvement, namely a high baseline percentage of sites with subgingival calculus and the presence of suppuration at baseline. The latter was, however, also predictive for tooth loss during the study. Our findings underline the need for uniformity in defining the outcomes in trials of periodontal therapy.

Assessing and correcting for regression toward the mean in deviance-induced social conformity

Our understanding of the mechanisms underlying social conformity has recently advanced due to the employment of neuroscience methodology and novel experimental approaches. Most prominently, several studies have demonstrated the role of neural reinforcement-learning processes in conformal adjustments using a specifically designed and frequently replicated paradigm. Only very recently, the validity of the critical behavioral effect in this very paradigm was seriously questioned, as it invites the unwanted contribution of regression toward the mean. Using a straightforward control-group design, we corroborate this recent finding and demonstrate the involvement of statistical distortions. Additionally, however, we provide conclusive evidence that the paradigm nevertheless captures behavioral effects that can only be attributed to social influence. Finally, we present a mathematical approach that allows to isolate and quantify the paradigm’s true conformity effect both at the group level and for each individual participant. These data as well as relevant theoretical considerations suggest that the groundbreaking findings regarding the brain mechanisms of social conformity that were obtained with this recently criticized paradigm were indeed valid. Moreover, we support earlier suggestions that distorted behavioral effects can be rectified by means of appropriate correction procedures.

Differential treatment of hypertension by primary care providers and hypertension specialists in a barber-based intervention trial to control hypertension in Black men

Black men have less physician contact than other groups and thus lower rates of hypertension treatment and control. In the Barber-Assisted Reduction in Blood Pressure among Ethnic Residents trial, hypertension control in 8 active-intervention barbershops where barbers offered blood pressure (BP) checks with haircuts and motivated black male patrons with high BP to seek provider follow-up showed a small improvement over that in 7 comparison shops where patrons received hypertension pamphlets but not barber-BP checks. Undertreatment of hypertension, which is common in primary care, may have impacted the outcomes. Thus, in patrons with a baseline systolic BP of ≥140 mm Hg and 10-month follow-up including BP and medication data, we performed post hoc comparison of systolic BP reduction between comparison-arm patrons (n = 68) treated by primary care providers (PCPs) with (1) intervention-arm patrons (n = 37) treated by PCPs or (2) intervention-arm patrons (n = 33) who lacked access to PCPs and were treated by hypertension specialist physicians serving as safety net providers. The latter group had higher baseline systolic BP than the others (162 ± 3 vs 155 ± 2 and 154 ± 2 mm Hg, respectively, p <0.01). After adjustment for baseline systolic BP and other covariates, systolic BP reduction was 21 ± 4 mm Hg greater than in the comparison group (p <0.0001), when barbers referred patrons to hypertension specialists but was no different when they referred to PCPs (4 ± 4 mm Hg, p = 0.31). Specialist-treated patrons received more BP medication and different classes of medication than PCP-treated patrons. In conclusion, the barber-based intervention-if connected directly to specialty-level medical care-could have a large public health impact on hypertensive disease in black men.

Effectiveness of a supportive telephone counseling intervention in type 2 diabetes patients: randomized controlled study

OBJECTIVES

This randomized controlled trial investigated whether a patient-centered supportive counseling intervention comprising monthly telephone-based counseling sessions by practice nurses over 12 months improved diabetes-related medical and psycho-social outcomes above usual care in type 2 diabetes patients with poor glycemic control at baseline (HbA1c >7.5%) in a primary care setting.

RESEARCH DESIGN

Patients were individually randomized into intervention (n = 103) and usual care group (n = 101). The primary outcome was change in HbA1c-concentration after 12 and 18 months. Secondary outcomes were lipid levels, blood pressure, health-related quality of life and symptoms of depression. Follow-up-measurements were carried out after 6, 12 and 18 months to assess potential immediate and maintained effects of the intervention. For the multivariate analysis, hierarchical linear models were computed for each outcome to assess within-group changes in outcomes over time and between-group differences in patterns of change.

RESULTS

HbA1c (in %) decreased significantly from baseline to 12-month follow-up measurement both in the intervention (-0.44) and the usual care group (-0.51), but there was no significant between-group intervention effect. Significant improvements in the intervention group along with significant between-group differences were seen for health-related quality of life and, transiently, for systolic blood pressure and depression.

CONCLUSIONS

Although we found no beneficial effect of the supportive telephone counseling in terms of a reduction of HbA1c above usual care, our findings suggest some beneficial effects on cardiovascular risk factors, quality of life and depression. Continuous efforts might be needed to sustain improvements in patient outcomes.

TRIAL REGISTRATION

ClinicalTrials.gov NCT00742547.

Combined bias suppression in single-arm therapy studies

RATIONALE, AIMS AND OBJECTIVES

For therapy evaluation studies, control groups are sometimes not feasible. In single-arm studies, various bias factors apart from the test therapy can affect clinical outcomes. The objective of this analysis was to improve the methods to minimize bias in single-arm studies.

METHOD

We present a procedure for combined suppression of several bias factors, using two methods: sample restriction to patients unaffected by bias, and score adjustment. The procedure was used for a secondary analysis of disease score (doctors' global rating, 0-10) in a cohort of patients receiving anthroposophic therapies for chronic diseases. Four bias factors were suppressed stepwise: attrition bias (by replacing missing values with the baseline value carried forward), bias from natural recovery (by sample restriction to patients with disease duration of >/=12 months), regression to the mean due to symptom-driven self-selection (by replacing baseline scores with scores three months before enrolment) and bias from adjunctive therapies (by sample restriction to patients not using adjunctive therapies).

RESULTS

In the cohort analysed, these four bias factors could together explain a maximum of 37% of the 0- to 6-month improvement of disease score.

CONCLUSION

Combined bias suppression, using sample restriction and score adjustment, is a transparent procedure to minimize bias in single-arm therapy studies. Further applicability of the procedure should be tested in future studies.

Trends and plot methods in MLR studies

Regression toward the mean effects are presented within the field of quantitative structure-activity relationship modeling and in situations in which multilinear regression techniques are considered for model building. The concept is related to the graphical aspect of some scatter plots (experimental vs fitted and fitted vs experimental values). These graphs demonstrate how the point cloud is not always symmetrically distributed along the so-called "ideal" or "desired" line, that is, the bisector of the first and third quadrants. The deviation from the ideal line is fixed, and it is also related to the coefficient of determination. An extrapolation of these regression effects is also discussed within the context of property predictions obtained via the leave-one-out cross-validation technique.

Revisiting the relation between change and initial value: a review and evaluation

The relation between initial disease status and subsequent change following treatment has attracted great interest in clinical research. However, statisticians have repeatedly warned against correlating/regressing change with baseline due to two methodological concerns known as mathematical coupling and regression to the mean. Oldham's method and Blomqvist's formula are the two most often adopted methods to rectify these problems. The aims of this article are to review briefly the proposed solutions in the statistical and psychological literature, and to clarify the popular misconception that Blomqvist's formula is superior to Oldham's method. We argue that this misconception is due to a failure to recognize that the heterogeneity of individual responses to treatment is a source of regression to the mean in the analysis of the relation between change and initial value. Furthermore, we demonstrate how each method actually answers different research questions, and how confusion arises when this is not always understood.

Clinical Benefits of Subgingival Chlorhexidine Varnish Application as an Adjunct to Same-Day Full-Mouth Root Planing: A Pilot Study

BACKGROUND

Although scaling and root planing are considered the therapeutic standard for periodontitis, weakly responding sites often occur. To improve treatment outcome, several chemomechanical treatment concepts have been developed. Recently, the clinical surplus value of a highly concentrated chlorhexidine varnish has been shown when used as an adjunct to sequential scaling and root planing. The aim of this study was to explore the clinical effects of a treatment strategy for chronic periodontitis based on a combination of same-day full-mouth root planing and subgingival chlorhexidine varnish administration.

METHODS

A randomized, controlled, single-blind, parallel trial was conducted on 12 chronic periodontitis patients. The control group received oral hygiene instructions and same-day full-mouth root planing. The test group received the same instructions and treatment; however, all pockets were additionally disinfected using a chlorhexidine varnish. Clinical response parameters were recorded at baseline and subsequently after 1 and 3 months.

RESULTS

Both treatment strategies showed significant reductions in probing depth at both follow-up visits in comparison with baseline levels (P <or=0.029). Yet, at 3 months, initially diseased sites of the test group showed significant additional pocket reductions between 0.70 and 1.37 mm (P <or=0.021). Furthermore, a significant additive clinical attachment gain of 1.17 mm (P = 0.049) for initially deep pockets (>or=7 mm) was found favoring the test group.

CONCLUSION

These preliminary findings suggest that the outcome of same-day full-mouth root planing may benefit from the subgingival administration of a highly concentrated chlorhexidine varnish.

The problem of analysing the relationship between change and initial value in oral health research

The relationship between initial disease status and subsequent change following treatment has attracted great interest in dental research. However, medical statisticians have repeatedly warned against correlating/regressing change with baseline because of two methodological concerns known as mathematical coupling and regression to the mean. In general, mathematical coupling occurs when one variable directly or indirectly contains the whole or part of another, and the two variables are then analyzed by using correlation or regression. Consequently, the statistical procedure of testing the null hypothesis - that the coefficient of correlation or the slope of regression is zero - may become inappropriate. Regression to the mean occurs with any variable that fluctuates within an individual or a population, either owing to measurement error and/or to physiological variation. The aim of this article was to clarify the conceptual confusion around mathematical coupling and regression to the mean within the statistical literature, and to correct a popular misconception about the correct analysis of the relationship between change and initial value. As examples that use inappropriate methods to analyze the relationship between change and baseline are still found in leading dental journals, this article seeks to help oral health researchers understand these problems and explain how to overcome them.

Correcting for regression to the mean in behavior and ecology

If two successive trait measurements have a less-than-perfect correlation, individuals or populations will, on average, tend to be closer to the mean on the second measurement (the so-called regression effect). Thus, there is a negative correlation between an individual's state at time 1 and the change in state from time 1 to time 2. In addition, whenever groups differ in their initial mean values, the expected change in the mean value from time 1 to time 2 will differ among the groups. For example, birds feeding nestlings lose weight, but initially heavier birds lose more weight than lighter birds, a result expected from the regression effect. In sexual selection, males who remain unmated in the first year are, on average, less attractive than mated males. The regression effect predicts that these males will increase their attractiveness in the second year more than mated males. In well-designed experiments, changes in the experimental and control groups would be compared. In observational studies, however, no such comparison is available, and expected differential effects must be accounted for before they can be attributed to external causes. We describe methods to correct for the regression effect and assess alternative causal explanations.

The relationship between baseline value and its change: problems in categorization and the proposal of a new method

Oral health researchers have shown great interest in the relationship between the initial status of diseases and subsequent changes following treatment. Two main approaches have been adopted to provide evidence of a positive association between baseline values and their changes following treatment. One approach is to use correlation or regression to test the relationship between baseline measurements and subsequent change (correlation/regression approach). The second approach is to categorize the lesions into subgroups, according to threshold values, and subsequently compare the treatment effects across the two (or more) subgroups (categorization approach). However, the correlation/regression approach suffers a methodological weakness known as mathematical coupling. Consequently, the statistical procedure of testing the null hypothesis becomes inappropriate. Categorization seems to avoid the problem of mathematical coupling, although it still suffers regression to the mean. We show, first, how the appropriate null hypothesis may be established to analyze the relationship between baseline values and change in the correlation approach and, second, we use computer simulations to investigate the impact of regression to the mean on the significance testing of the differences in the average treatment effects (or average baseline values) in the categorization approach. Data available from previous literature are reanalyzed by testing the appropriate null hypotheses and the results are compared to those from testing the usual (incorrect) null hypothesis. The results indicate that both the correlation and categorization approaches can give rise to misleading conclusions and that more appropriate methods, such as Oldham's method and our new approach of deriving the correct null hypothesis, should be adopted.

How do statistical properties influence findings of tracking (maintenance) in epidemiologic studies? An example of research in tracking of obesity

There is great interest in studying the tracking (maintenance) of health conditions and risk factors over the life span. Tracking is often defined as the maintenance of a distribution position (e.g., quintile or percentile) in a study population over time. This study investigated how statistical properties might influence research findings of tracking with a special attention on the tracking of extreme ranking. Our results show that when repeated measures over time were positively correlated, the probability of tracking in extreme rankings was greater than other rankings and this was closely influenced by the overall correlation (r) and by the categorization. For example, when r = 0.4, 38% remained in the bottom and upper quintile (Q1, Q5) respectively, while only 22% remained in the middle quintile (Q3); when r = 0.8, the figure became 65% vs. 32%. When r = 0.4 and 0.8, 19 and 50% remained in the upper 95th percentile (or under the 5th percentile), respectively. Our real data show that children in the upper body mass index (= weight(kg)/height(m)(2)) quintile were more likely to maintain their ranking (54%) than others (about 30%), but not significantly higher than the expected (47%, p > 0.05). In conclusion, the overall correlation should be considered when studying tracking. Our proposed methods and predicted probabilities of tracking can help test whether one's observed tracking patterns are different from the statistically predicted ones.

Mathematical coupling can undermine the statistical assessment of clinical research: illustration from the treatment of guided tissue regeneration

OBJECTIVES

Previous periodontal literature has shown that there is a strong relationship between treatment effects, such as guided tissue regeneration (GTR), and baseline disease severity. However, relating change to baseline values using correlation or regression is methodologically flawed due to mathematical coupling, where the statistical procedure of testing the null hypothesis-that the coefficient of correlation or slope of regression is equal to zero-becomes erroneous. The aim of this study is to investigate if baseline disease severity is genuinely associated with the treatment outcome of intrabony defects using GTR after adjustment for mathematical coupling. In particular, we seek to demonstrate the potential effect that mathematical coupling has in distorting the results from the statistical analyses of trials of dental treatment, using data from the periodontal literature on GTR. The erroneous results arising from the use of simple correlation and regression techniques to analyse this association will be demonstrated, also the methodological flaw where the statistical procedure tests the null hypothesis-that the coefficient of correlation or the slope of regression is equal to zero.

METHODS

Three main periodontal journals were electronically and manually searched to extract the data for the clinical outcomes of pocket probing depth (PPD) and lifetime cumulative attachment loss (LCAL) in the studies using GTR. The relationship between clinical outcomes and baseline measurements were reanalysed using Oldham's method and the variance ratio test.

RESULTS

The results of these analyses were compared with those from the papers where the authors used the standard approach of correlation or regression. This shows that mathematical coupling caused spurious correlations between baseline disease severity and treatment effect. Ten out of 12 studies for PPD and nine out of 14 for LCAL initially claimed a significant positive relationship; after using either of the more appropriate statistical methods of adjustment, only three correlations in each group of studies remained significant.

CONCLUSIONS

Previous evidence suggesting an association between baseline disease severity and treatment effect for GTR is challenged and therefore needs to be critically reviewed. All future clinical research should avoid using mathematically coupled data in correlation or regression analysis. In seeking to examine the bivariate association between baseline and subsequent change, Oldham's method is recommended.

The Application of Multilevel Modeling in the Analysis of Longitudinal Periodontal Data – Part I: Absolute Levels of Disease

BACKGROUND

Statistical analyses of periodontal data that average site measurements to subject mean values are unable to explore the site-specific nature of periodontal diseases. Multilevel modeling (MLM) overcomes this, taking hierarchical structure into account. MLM was used to investigate longitudinal relationships between the outcomes of lifetime cumulative attachment loss (LCAL) and probing depth (PD) in relation to potential risk factors for periodontal disease progression.

METHODS

One hundred males (mean age 17 years) received a comprehensive periodontal examination at baseline and at 12 and 30 months. The resulting data were analyzed in two stages. In stage one (reported here), the absolute levels of disease were analyzed in relation to potential risk factors; in stage two (reported in a second paper), changes in disease patterns over time were analyzed in relation to the same risk factors. Each approach yielded substantially different insights.

RESULTS

For absolute levels of disease, subject-level risk factors (covariates) had limited prediction for LCAL/PD throughout the 30-month observation period. Tooth position demonstrated a near linear relationship for both outcomes, with disease increasing from anterior to posterior teeth. Sites with subgingival calculus and bleeding on probing demonstrated more LCAL and PD, and supragingival calculus had an apparently protective effect. Covariates had more "explanatory power" for the variation in PD than for the variation in LCAL, suggesting that LCAL and PD might be generally associated with a different profile of covariates.

CONCLUSION

This study provides, for a relatively young cohort, considerable insights into the factors associated with early-life periodontal disease and its progression at all levels of the natural hierarchy of sites within teeth within subjects.

The Application of Multilevel Modeling in the Analysis of Longitudinal Periodontal Data –Part II: Changes in Disease Levels over Time

BACKGROUND

The aim of this study was to investigate the longitudinal relationships between the outcome measurements of changes in lifetime cumulative attachment loss (cLCAL) and changes in probing depth (cPD) in relation to potential risk factors or other risk markers for periodontal disease progression from a cohort of 100 young males. In order to account for the hierarchical data structure, and to explore explicitly the site, tooth, and subject levels simultaneously, multilevel modeling was undertaken.

METHODS

The analyses were undertaken in two parts. Within a previous article, the absolute levels of disease were analyzed in relation to potential risk factors; within this article, changes in disease are analyzed in relation to these factors. Each analytical approach yielded substantively different insights.

RESULTS

Subject-level risk factors had limited predictive value for cLCAL/cPD throughout the 30-month observation period. Tooth position demonstrated a near linear relationship for both outcomes, with disease increasing from anterior to posterior teeth. Supragingival plaque had no significant effect on cLCAL/cPD, while subgingival calculus and bleeding on probing were negatively associated with cLCAL/cPD. In contrast to the outcomes LCAL/PD, supragingival calculus had no significant protective effect on cLCAL/cPD. There was no significant influence of smoking in this cohort.

CONCLUSIONS

This study provides, for a relatively young cohort, considerable insights into the factors associated with longitudinal patterns of early-life periodontal disease at all levels of the natural hierarchy of sites within teeth within subjects. Furthermore, it is demonstrated how multilevel modeling can provide considerable insight into some of the inconsistencies and controversies found in the previous periodontal literature.

Unification of the "burst" and "linear" theories of periodontal disease progression: a multilevel manifestation of the same phenomenon

Previously, burst and linear theories for periodontal disease progression were proposed based on different but limited statistical methods of analysis. Multilevel modeling provides a new approach, yielding a more comprehensive model. Random coefficient models were used to analyze longitudinal periodontal data consisting of repeated measures (level 1), sites (level 2), teeth (level 3), and subjects (level 4). Large negative and highly significant correlations between random linear and quadratic time coefficients indicated that subjects and teeth with greater-than-average linear change experienced decelerated variation. Conversely, subjects and teeth with less-than-average linear change experienced accelerated variation. Change therefore exhibited a dynamic regression to the mean at the tooth and subject levels. Since no equilibrium was attained throughout the study, changes were cyclical. When considered as a multilevel system, the "linear" and "burst" theories of periodontal disease progression are a manifestation of the same phenomenon: Some sites improve while others progress, in a cyclical manner.

Is reduction of pocket probing depth correlated with the baseline value or is it "mathematical coupling"?

Previous studies using correlation or regression analysis have showed that treatment effects measured by the change in clinical parameters are often associated with baseline values of the same parameters. These studies, however, have a methodological weakness. Correlation/regression between baseline measures and the derived change variable invalidates the statistical procedures of testing the null hypothesis: that the coefficient of correlation/regression is zero. This is due to the phenomenon of mathematical coupling. To investigate the impact that this has on the observed correlation/regression coefficient when in reality this is zero, we used random simulations of hypothetical data to model the treatment of periodontal pockets. Results showed a strong probability of obtaining statistically significant correlation/regression coefficients. To separate this artificial effect of mathematical coupling from the true underlying biological relationship, one must apply appropriate analytical strategies to re-evaluate previous evidence within the periodontal literature.

Plasma beta-carotene and retinol concentrations of children increase after a 30-d supplementation with the fruit Momordica cochinchinensis (gac)

BACKGROUND

In rural Vietnam, vitamin A deficiency is a concern. Among the indigenous fruit and vegetables, Momordica cochinchinensis (gac) fruit has been identified as having the highest beta-carotene concentration. Locally, it is mixed with rice in a preparation called xoi gac.

OBJECTIVE

The purpose of this study was to assess this beta-carotene- rich rice preparation as a source of provitamin A for children in rural Vietnam.

DESIGN

Preschoolers (n = 185) participated in a 30-d controlled supplementation trial. Children with low hemoglobin concentrations were assigned to 1 of 3 groups: a fruit group, who received xoi gac that contained 3.5 mg beta-carotene per serving; a powder group, who received rice mixed with 5.0 mg synthetic beta-carotene powder; and a control group, who received rice without fortification.

RESULTS

The mean increase in plasma beta-carotene concentrations in the fruit and powder groups was significantly greater than that in the control group (P < 0.0001). After supplementation, the mean plasma retinol concentration in the fruit group was significantly higher than that in the control (P = 0.006) and powder (P = 0.0053) groups. Among the children with initial hemoglobin concentrations <or=110 g/L, the mean increase in hemoglobin concentrations in the fruit group was marginally higher than that in the control group (P = 0.017) but was not significantly different from that in the powder group.

CONCLUSIONS

beta-Carotene from xoi gac is a good source of provitamin A carotenoids. Severely anemic children might particularly benefit from routine xoi gac consumption

Gingival recession in smokers and non-smokers with minimal periodontal disease

BACKGROUND/AIMS

Smoking is a major risk factor for destructive periodontal disease. There is limited information with regard to effects of smoking in subjects with minimal periodontal destruction. The aim of the present investigation was to assess the development of gingival recession in young adult smokers and non-smokers.

METHODS

61 systemically healthy young adults, 19 to 30 years of age completed the final examination. 30 volunteers smoked at least 20 cigarettes per day, whereas 31 subjects were non-smokers. Clinical periodontal conditions were assessed 4x within a time period of 6 months. Site-specific analyses considering the correlated structure of data were performed.

RESULTS

At the outset, 50% of subjects presented with gingival recession at 1 or more sites. There was no significant difference in the prevalence of gingival recession between non-smokers and smokers. Severe recession in excess of 2 mm affected about 23% non-smokers but only 7% smokers. Some further gingival recession developed during the 6-month observation period. In a multivariate logistic regression analysis, the risk for recession development appeared not to be influenced by smoking status after adjusting for periodontal probing depth, recession at baseline, tooth brushing frequency, gender, jaw, tooth type and site.

CONCLUSIONS

Present data did not support the hypothesis that smokers are at an increased risk for the development of gingival recession.

Is loss of attachment due to root planning and scaling in sites with minimal probing depths a statistical or real occurrence?

BACKGROUND

Following root planing and scaling many studies have implied an association between a loss of clinical attachment at sites with initially shallow pockets (1 to 3 mm) and gains in attachment level for deeper probing depths. However, these effects are also consistent with a statistical phenomenon referred to as regression towards the mean. This principle suggests that extreme values will moderate the next time they are recorded. The purpose of this report was to estimate the effect that regression towards the mean has on perceived changes in attachment level after root planing and scaling.

METHODS

During the initial examination, 2 different investigators conducted 2 full-mouth probings. Two quadrants were randomly selected to be root planed and scaled until the root surfaces were smooth by tactile touch of an explorer. The 2 remaining quadrants were not treated. At 4 to 6 weeks after treatment, another full mouth probing was done. An examiner who was blind to the quadrants that had been scaled measured attachment level and probing depth after therapy. This study design provided periodontal measurements before and after root planing and scaling, measurements before and after a period of 4 to 6 weeks of no therapy, and duplicate measurements at the beginning of the study.

RESULTS

Using the repeat examination when no true change could occur, shallow sites (< or =3 mm of probing depth) showed average negative differences between repeat attachment level measurements (-0.23 mm), which mimicked loss of periodontal attachment. Deep sites, (>6 mm) showed average positive values (0.40) mimicking gain in attachment level. These results suggest that regression towards the mean is a significant effect in this data set. Both shallow non-scaled and scaled sites had similar differences in repeat measures (-0.28 mm, -0.25 mm) which were also similar to and not statistically different from changes after therapy for both non-scaled (-0.21 mm) and scaled sites (-0.08 mm). Thus not only does this data set exhibit regression towards the mean, but it explains the majority of perceived loss of periodontal attachment after scaling at sites that have minimal probing depth.

CONCLUSIONS

These results suggest that the majority of perceived loss of attachment due to scaling at sites of minimal probing depth that have been reported in many studies may be due a statistical phenomenon called regression towards the mean.

Issues in the evaluation of clinical trials of periodontitis: a clinical perspective

The goals of clinical trials designed to establish claims for equivalency or superiority of treatment for periodontitis must be clearly stated and defined in terms of measurable and meaningful response variables. It is suggested that these clinical trials use designs that compare new treatment methods to basic periodontal therapy which consists of thorough scaling and root planing, oral hygiene instruction, and regular maintenance care. The primary response variable should be clinical attachment level. It is important to document changes in probing depth since this is a meaningful measure to many clinicians. Gingival inflammation and bleeding should be used as secondary response variables because these are not necessarily indicative of progressive periodontal destruction. Radiographic measures of disease may be useful as primary response variables if safe, reproducible and valid methods of measuring change are utilized. Microbiological monitoring should be a secondary response variable because of numerous questions concerning sampling methodology, quantitative expression of data, and meaningful interpretation in terms of relevance to disease activity. The length of periodontitis trials should be set at a minimum of 9 months if claims of superiority or equivalency are made compared to basic periodontal therapy. Calibration trials should be included and measurement error should be expressed in terms that are meaningful to the clinician while retaining statistical validity. Statistical methods for determining change should take into account site and subject heterogeneity, bursts of change and gradual change over time. It is important that statistical techniques be used that detect change as early as possible and that appropriate consideration be given to the clinical implications of the type, magnitude and duration of change in outcome variables.

The angular bony defect as indicator of further alveolar bone loss

The aim of the present retrospective study was to evaluate alterations of the alveolar bone level over a 10-year period at tooth sites with "angular" and "even" patterns of bone loss, in subjects who were not exposed to systematic periodontal therapy. A further objective was to evaluate whether the presence of an angular defect can serve as a predictor of additional bone loss. 201 subjects in ages 25-70 years were examined radiographically on 2 occasions 10 years apart. 194 were dentate on the 2nd examination. The radiographic bone height at the mesial and distal aspect of all teeth was assessed by measuring the distance between the cemento-enamel junction and the bone crest. The morphologic pattern of alveolar bone loss at baseline was assessed for each tooth site. Angular defects were classified as degree 1, 2 and 3, according to increasing defect depth. The change in periodontal bone height over the 10-year period was calculated for each site. The results demonstrated an increased frequency of tooth loss among teeth showing presence of an angular bony defect at baseline; whereas 13% of the teeth with an even pattern of bone loss were lost between the 2 examinations, this proportion was 22%, 46% and 68% for teeth with an angular defect of degree 1, degree 2 and degree 3, respectively. Longitudinal bone loss of greater than or equal to 2 mm occurred more often among sites showing an angular defect when compared to sites with an even alveolar bone morphology.(ABSTRACT TRUNCATED AT 250 WORDS)

Periodontal disease activity: a critical assessment

Review of investigations using longitudinal probing and radiographic assessments indicate that the prevalence, magnitude, rate, and temporal patterns of periodontal destruction as well as the ability of clinical and laboratory tests to detect and predict loss of clinical attachment need further investigation. These unresolved issues are discussed and their impact on the practice of periodontics is examined.

Assessment of relationships between site-specific variables

The assessment of relationships between site-specific variables has been a matter of controversy because of the claim that periodontal sites within individuals can be used as independent observations in statistical models. One problem with this approach is the unreliability of the calculated Type I and Type II error rates. Another problem is that such inappropriate analysis may prohibit a correct assessment of causal relationships between site-specific variables. The host-factor can act as an effect modifier and modulate the magnitude of the site-specific effects and/or the host-factor can act as a confounder by superimposing a patient-effect on the studied site-specific effects leading to bias. As a result, site-specific biological mechanisms of disease progression may be misinterpreted. Sites can be used as the experimental unit of analysis, but the sampling design from which the site-specific data originated should not be ignored.

Reproducibility of microbiological samples from periodontal pockets

Duplicate microbiological samples, were taken 1 week apart using the paper point technique from a total of 112 untreated periodontal pockets greater than 6 mm deep in 16 adult periodontal patients. Duplicate samples were also obtained from these sites 6 months following a therapy of oral hygiene instruction and supra- and subgingival debridement. The reproducibility of the total viable counts and the reproducibility of the proportions of various groups or species of microorganisms were studied from these duplicate samples. The results demonstrated an acceptable degree of reproducibility for the recovery of Actinobacillus actinomycetemcomitans and Bacteroides gingivalis. For the total viable counts and for the other investigated bacterial groups, including Bacteroides intermedius, unacceptable levels of reproducibility were observed.

Clinical effects of root debridement in molar and non-molar teeth. A 2-year follow-up

12 patients were studied longitudinally to monitor the effects of basic periodontal therapy in molar and non-molar teeth. Periodontal sites were grouped into molar furcation sites, molar flat-surface sites and non-molar sites. Clinical measurements were taken at baseline and directly followed by full mouth root debridement. Subsequently, measurements were taken every 3rd month until 24 months. At each of these appointments, the patients were monitored for their oral hygiene performance and given supragingival prophylaxis. The mean results indicated that initially moderately deep and deep molar furcation sites responded less favorably to therapy compared to non-molar sites and molar flat-surface sites of similar probing depth. Initial improvements in probing measurements for moderately deep and deep molar furcation sites were limited and also tended to revert during the observation interval. Identification of individual sites with probing attachment loss disclosed that 25% of molar furcation sites lost probing attachment as compared to 7% for non-molar sites and 10% for molar flat-surface sites. These results corroborate previous findings and call for additional or alternative treatment regimens for periodontal furcation pockets.

The impact of regression towards the mean on probing changes in studies on the effect of periodontal therapy

The purpose of this report was to evaluate the magnitude of the regression towards the mean (RTM) relative to the observed changes in probing measurements following periodontal therapy in 2 groups of patients. Regression lines were calculated for observed changes in probing depths and probing attachment levels related to the initial probing depths of the individual sites. Regression lines, adjusted for the RTM effect, were also calculated. For the extreme subgroups of recorded shallow and deep sites, the RTM effect (mm) was found to amount to a maximum of 0.6 mm. It appears unlikely that erroneous conclusions have been made in previous clinical studies evaluating probing depth changes in deep sites, since the RTM accounts for a limited portion of the observed changes. For probing depth changes in shallow sites and for probing attachment level changes in both shallow and deep sites, the RTM accounts for a larger proportion of the observed mm changes. It therefore seems prudent to be careful about interpretation of these changes in clinical studies, where the RTM effect has not been adjusted for. Whenever possible, duplicate initial recordings should be taken to allow determination of and subsequent adjustment for the RTM effect.

Accuracy of repeated single versus averages of repeated duplicates of probing depth measurements

The present study aims to determine to which extent averages or deepest recordings of repeated duplicates of probing depth measurements may increase the accuracy as compared to repeated single measurements. In total, 1247 sites were recorded in 8 patients with advanced periodontitis. The patients had received initial treatment. Each site was probed 4 times with time intervals of 100 min. The standard deviation of differences between repeated single measurements of 0.97 mm decreased with a factor square root of 2 to 0.69 mm for differences between averages of repeated duplicates of measurements. Differences between the averages and the deepest recordings of repeated duplicates of measurements showed a non-Gaussian distribution. This implies that the type-I error can not be computed on the basis of the standard deviation of the error with the use of parametric statistical analysis. The best estimate for the type-I error is the observed frequency of differences. The type-I error for differences of 3 mm or more between repeated single measurements decreased from 1.5% to 0.9% and 0.2%, when deepest recordings and averages of repeated duplicates of measurements, respectively, were compared.