| Sign In to gain access to subscriptions and/or personal tools. |
Periodontal Treatment and Glycemic Control in Diabetic Patients: the Problem of a Possible Hawthorne EffectDepartment of Periodontology and Preventive Dentistry, King’s College London Dental Institute at Guy’s, King’s and St Thomas’ Hospitals, Floor 21, Guy’s Tower, London Bridge, SE1 9RT, UK; trevor.watts{at}kcl.ac.uk To the Editor: Janket et al.(2005) were unable to reject their null hypothesis of no effect of periodontal treatment upon diabetic control, and included studies by myself and colleagues in their meta-analysis. In the Table listing the 10 studies analyzed, only our two small intervention studies randomized diabetic patients to periodontal treatment or no treatment (Aldridge et al., 1995). All our controls were given significant attention, emphasizing the reasons for our trials and making it clear that they would also be offered periodontal treatment in due course, but received no direct periodontal intervention during the trials. We hoped to pick up any Hawthorne effect on diabetic control, and distinguish it from any possible effect of periodontal treatment. In the event, we found a Hawthorne effect on oral hygiene, but no significant change in metabolic control. Periodontal and diabetic treatment both depend on patient adherence to professional advice. The Hawthorne effect may occur when a behavioral variable is affected by a factor not directly related to it. The emphasis on personal health in periodontal cohort studies, particularly when diabetes is highlighted, may have an effect on the patients’ compliance in their metabolic control regime. Future studies of the effect of periodontal treatment on diabetic health should be conducted similarly, as randomized single-blind trials, with significant attention given to control patients during the study to concentrate their thoughts on matters of health. Without this feature, there will be no way of inferring that periodontal health influences diabetic metabolic control, if such a link exists. REFERENCES
The authors replyBoston University, Goldman School of Dental Medicine; skjanket{at}bu.edu Dr. Trevor Watts has raised a very valid and important issue in dental research. We agree with Dr. Watts in that it is possible that the observed effects might have been attributable to the patients’ compliance in their metabolic control regimen. We also concur with him in that there should be a mechanism in place to control for dietary input in glycemic control, considering the relationship of glucose metabolism and oral infection (Janket et al., 2006). Although we acknowledge Dr. Watts’ concern for a possible Hawthorne effect, we observed that the evidence for this effect was not strong in the studies in which he participated (Aldridge et al., 1995). First, type 1 diabetes mellitus (DM) patients may not be an appropriate target population, as presented in our paper (p. 1157) (Janket et al., 2005). Type 1 DM is of autoimmune origin and is usually tightly controlled by insulin administration. Second, unless insulin requirements are presented, glycemic control itself may not show any changes under these circumstances. As shown in our Fig. 2, studies that examined type 1 DM patients show markedly less reduction in Hemoglobin A1c (HbA1c) subsequent to periodontal treatment (p. 1157) (Janket et al., 2005). There were too many other factors in Dr. Watts’ study that might have contributed to the null results to attribute the non-significant results to the Hawthorne effect, as delineated in our recommendations for future study (p. 1157) (Janket et al., 2005). If we could not observe any changes in glycemic control when all these conditions are met, then we may indeed be able to attribute the lack of evidence to the Hawthorne effect. Another caveat is that not all randomized trials will have a balanced distribution of confounding factors. To achieve this balance, the sample size should be fairly large (in hundreds or thousands). One way to avoid an unbalanced distribution of confounding factors may be by utilizing a crossover design with appropriate measures in place to control for dietary input to HbA1c level, and with an adequate washout period. In conclusion, well-designed future research is needed to elucidate the true effects of periodontal treatment on glycemic control. REFERENCES
Journal of Dental Research, Vol. 85, No. 4,
294-295 (2006)
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
