The factors that drive overtreatment of gastroesophageal reflux disease (GERD) are not well understood, but it has been proposed that the use of the “GERD” disease label could perpetuate use of medication in otherwise healthy infants.
To determine if use of the disease label GERD influences parents’ perceived need to medicate an infant, we surveyed parents in a general pediatric clinic. Parents were given a hypothetical clinical scenario describing an infant who cries and spits up excessively but is otherwise healthy. Using a 2 × 2 factorial design, parents were randomized to receive a scenario in which the doctor either gave a diagnosis of GERD or did not provide a disease label; additionally, half of parents were told that existing medications are probably ineffective, whereas the rest were not given any effectiveness information. We measured parent interest in medication, perception of illness severity, and appreciation of medication offer.
Parents who received a GERD diagnosis were interested in medicating their infant, even when they were told that the medications are likely ineffective. However, parents not given a disease label were interested in medication only when medication effectiveness was not discussed (and hence likely assumed).
Labeling an otherwise healthy infant as having a “disease” increased parents’ interest in medicating their infant when they were told that medications are ineffective. These findings suggest that use of disease labels may promote overtreatment by causing people to believe that ineffective medications are both useful and necessary.
In a recent letter, Dr. Tim J. Ford and colleagues cited a number of problematic issues with our recent article.1 We welcome the opportunity to address each concern in turn.
First, and most importantly, Ford and colleagues erroneously stated that we did not report critical statistics, specifically the 2-way interactions in Figures 2 and 3. We did indeed report these statistics for both Figures 2 and 3 (page 4 of the article, column 3). We also referenced the figures in the text, and explicitly referred to the statistics as representing the results of an interaction.
Second, the authors wrote that the vignettes "inherently bias parents towards medicating their infants irrespective of a disease label." Because ours was a comparative study based on a randomized intervention, our goal was not to examine absolute levels of interest, but rather interest relative to the information that was/was not provided (e.g. GERD label vs. no label). Any systematic bias would not explain the relative differences we found among the intervention arms. The authors also questioned whether a score of 2.5 on a 5-point scale indicates "interest" in medications. Again, examining absolute levels of interest was not our aim.
Third, the authors felt that the vignettes "suggests underlying pathology and (are) not representative of the vast majority of infants that present with GER." We respectfully but emphatically disagree. In all four vignettes the infant was described as "gaining weight," and "otherwise healthy." Weight gain is used by physicians (and parents) as a gauge of health. We also disagree that the vignettes are not representative. The vignettes describe an infant with reflux and irritability and are based directly on symptoms from the North American GERD clinical guidelines. We explicitly refer to this in our article (page 2 column 3).
Fourth, the authors took issue with the medication ineffectiveness information, because the hypothetical physician said that the medication "probably" won't work. When developing the vignette, we thought carefully about whether to include the word "probably" or not. We included it because we reasoned that it would be both inaccurate and unrealistic for the physician to say that the medication "won't work," because this implies 100% certainty.
Finally, the authors would have liked to see more detail about randomization method and distribution of parent age across conditions. This information was omitted due to word count limitations, but is provided below:
1. The different versions of the surveys were printed, shuffled, and then handed out to parents of patients in a pediatric waiting room over the course of several months.
2. There were no significant differences in age of participant across experimental study arms (p values greater than .22 for all comparisons across arms).
3. Number of participants per condition: a. No label, no information about medication effectiveness: 53. b. GERD label, no information about medication effectiveness: 40. c. No label, information about effectiveness: 35. d. GERD label, information about effectiveness: 47
We thank the Editors for the opportunity to respond to these critiques.
1. Ford TJ, Weist MM, Danchin M. The GERD label: Is there more to it? Letter in response to: Scherer et al et al. Influence of ''GERD'' Label on Parents' Decision to Medicate Infants. Pediatrics. 2013;131(5):839-845.
Conflict of Interest:
None declared
The term 'acid-reflux' is commonly used by parents that seek medical care for infants with feeding problems and irritability. This is often followed by a request for an acid-suppressing medication. Although long presumed safe, recent evidence now questions the safety of proton-pump inhibitors and has shown their potential to cause harm.1, 2
With this in mind, we read with interest the recent manuscript by Scherer and colleagues3 that investigated how different levels of information provided by a physician can influence parents' decisions regarding treatment. Specifically, in their study, four separate written vignettes were provided to parents to ascertain whether the provision of a disease label (Gastro Esophageal Reflux Disease (GERD)) and information on medication 'ineffectiveness' influenced parental interest in medicating their infant. We believe the way the vignettes were presented would inherently bias parents towards medicating their infants irrespective of a disease label. In our view the condition of the infant in the vignettes suggests underlying pathology and is not representative of the vast majority of infants that present with GER. Further, information provided regarding medication 'ineffectiveness' actually suggests the medication might be effective given the word 'probably' was used in the vignette. However, despite these reservations, the study is novel and highlights how clever marketing, personal experience and the way the information is provided by a physician can influence decisions regarding treatments.
Despite employing a complex study design, the study unfortunately lacks important information. There is no description of how randomisation was done and the number of parents in each group is not provided. Despite collecting the data, the demographic variation between subgroups is not described and there is no evidence that the distribution of potential confounding factors such as parental age have been considered. A subgroup analysis looking at these variables would have provided insights into the demographic most likely to opt for medication, potentially useful information for the clinician.
Important information regarding data analysis is also omitted. Although the authors make conclusions on the interaction between two variables (disease label and medication ineffectiveness), they only report the overall F-statistic from the two-way ANOVA, not the F-statistic for interaction. The figures presenting the interaction between these variables (Figure 2 and 3) lacks error bars which, in combination with the way the ANOVA was reported, do not allow the reader to ascertain the significance of the difference in group means. The conclusion from Figure 2 that parents who received a GERD diagnosis were interested in medicating the infant is potentially misleading, as we query whether a score of 2.5/5 actually indicates interest.
Overall, despite an innovative approach to an important question, the lack of detail presented on the design and result analysis hinders the reader's ability to draw firm conclusions and limits generalisability of the findings.
References 1. Hassall E. Over-prescription of acid- suppressing medications in infants: how it came about, why it's wrong, and what to do about it. J Pediatr. 2012;160(2):193-198 2. Holbrook J, Wise R, Gold B, Blake K, Brown E, Castro M, Dozor A, Lima J, Mastronarde J, Sockrider M, Teague W. Lansoprazole for children with poorly controlled asthma: a randomized control trial. JAMA. 2012;307(4):373-81. 3. Scherer L, Zikmund-Fisher B, Fagerlin A, Tarini B. Influence of "GERD" label on parents' decision to medicate infants with excessive crying and reflux. Pediatrics. 2013;131(5):839-845.
T Ford, M M Wiest, M Danchin.
Conflict of Interest:
None declared