Standards of care within the field of pediatric gastroenterology continue to develop, encompassing new understandings of gastrointestinal disorders and guidelines for practice with children and adolescents. This evolution has been particularly evident in the area of chronic abdominal pain for the past 15 years. Since the Rome Committee's initial work in the early 1990s, the way that chronic abdominal pain in children is understood and classified has shifted in a dramatic fashion. One of the main activities of the Rome Committee (now the Rome Foundation) has been the development of symptom-based diagnostic criteria for chronic abdominal pain, collectively referred to as functional gastrointestinal disorders (FGIDs). These criteria presently are in their third version (ie, Rome III) and are meant both “to promote clinical recognition and legitimization of the FGIDs” and “to develop a scientific understanding of their psychopathological mechanisms to achieve optimal treatment” (1). Certainly, the efforts of the Rome Foundation and the development of the Rome criteria have been associated with increased interest in chronic abdominal pain and an explosion of research in this once-neglected area (1).
However, it remains unclear how this shift in theoretical perspective, initiated by a committee of experts, has affected actual pediatric gastroenterology clinical practice. Concerns exist that the Rome criteria, being designed primarily for research purposes, may not be practical for clinical use despite recent efforts to simplify and streamline the criteria and an acknowledged need for common terminology in clinical practice (1–3). In addition, some have suggested that the criteria have developed unevenly, with better codification for irritable bowel syndrome (IBS) than for other FGID diagnoses (eg, functional dyspepsia) (2,3). At this point in time, it is important to reevaluate how children with chronic abdominal pain are viewed, assessed, and treated across North America to determine where disconnects between theory and practice may exist. This information will help to facilitate the integration of both perspectives in future work. Specifically, this information may highlight areas where further education of practitioners is needed. However, it also may point to possible areas for further research investigation and/or eventual empirically based criteria revision.
To this end, this study used survey methodology to examine how children with chronic abdominal pain are viewed, assessed, and treated by pediatric gastroenterologists across North America, and how perspectives have changed since initial release of the Rome criteria approximately 15 years ago. Finally, present knowledge and use of the Rome criteria were assessed, including physician-perceived benefits and drawbacks of this classification system for clinical and research activities.
PATIENTS AND METHODS
Participants in this study were full members of the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition (NASPGHAN), including 151 members in 1992 and 174 members in 2006.
The Pediatric Gastroenterology Practice Survey was developed by the authors specifically for this study. The survey included questions designed to assess standards of care regarding chronic abdominal pain within the field of pediatric gastroenterology using both multiple-choice and open-ended questions. To ensure comparability of responses from 1992 to 2006, questions from the original survey were unchanged in the 2006 version; however, several new questions were added to the 2006 survey to assess knowledge and use of the Rome criteria. The area of dyspepsia was examined most closely at both time points because some (2,3) have suggested that it has been relatively less well codified and remains more elusive a concept than other FGIDs, such as IBS.
In 1992, all NASPGHAN members listed in the membership directory received a postal version of the original Pediatric Gastroenterology Practice Survey with written directions for completion. One hundred fifty-one members completed and returned this anonymous survey, yielding an overall return rate of 29%. In 2006, all NASPGHAN members with an e-mail address listed in the membership directory received an e-mail invitation with directions for completing a Web-based version of the updated Pediatric Gastroenterology Practice Survey. The NASPGHAN members without a current e-mail address, or whose original e-mail invitation was returned because of technical difficulties, received a postal version of the updated survey with written directions for completion. All surveys again were completed anonymously. In total, 90 members submitted surveys online and an additional 84 members returned completed copies of the postal version of the survey, yielding a final sample of 174 members completing surveys in 2006. The return rates were approximately equal for the 2 recruitment strategies (21% for e-mail/online vs 25% for postal mail), yielding an overall return rate of 22%. Analysis based on postal stamps indicated that participants in both 1992 and 2006 represented all major geographic regions of the United States. However, information regarding geographic region was unavailable for approximately 30% of respondents (some from the postal group and all from the e-mail group), so the number of participants from elsewhere in North America (eg, Canada, Mexico) could not be assessed. All study procedures were approved by the institutional review board of the participating institution.
Descriptive statistics were calculated for multiple-choice items to examine general patterns in the current conceptualization, assessment, and treatment of pediatric abdominal pain, as well as knowledge and use of the Rome criteria in practice. The responses to open-ended questions were qualitatively examined and categorized by content into discrete groups for ease of interpretation. Finally, statistical comparisons were made between present practice and historical data from 1992 using Pearson χ2 tests with significance accepted at P < 0.05.
Beliefs About Pediatric Abdominal Pain
In response to an open-ended question, a high percentage of physicians in 2006 reported the belief that abdominal pain was most commonly “functional” in terms of cause (42%). Constipation was cited most often as the most frequent “organic” cause of abdominal pain (29%), followed by reflux/esophagitis (24%), peptic gastritis/duodenitis (15%), and inflammatory bowel disease (11%). Although many of these causes were also cited at high rates in 1992 (ie, constipation, 21%; peptic gastritis/duodenitis, 19%; and reflux/esophagitis, 12%), inflammatory bowel disease was cited by <5% of physicians, whereas IBS and lactose intolerance were cited at higher rates (12% and 9%, respectively). No significant shift was found from 1992 to 2006 in the estimated proportion of patients presenting with abdominal pain for that the physician believed there was a psychological, psychiatric, or behavioral basis (χ2[5, N = 315] = 9.70, P = ns; Table 1].
Beliefs About Pediatric Dyspepsia
Physicians estimated similar rates of dyspepsia within their patients evaluated for abdominal pain in 1992 and 2006 (χ2[4, N = 314] = 4.05, P = ns). At both time points, approximately half of physicians estimated that dyspepsia was present in 10% to 25% of their patients with abdominal pain and an additional one-quarter estimated that dyspepsia was present in 26% to 50% of these patients. The definition of dyspepsia demonstrated only minor change from 1992 to 2006, with fewer physicians endorsing upper abdominal pain (49% vs 37%; χ2[1, N = 323] = 4.97, P < 0.05) as an appropriate definition of dyspepsia. Similar percentages of physicians endorsed epigastric pain (51% vs 52%; χ2[1, N = 323] = 0.02, P = ns) and abdominal pain with nausea and/or vomiting (60% vs 65%; χ2[1, N = 323] = 2.45, P = ns) as appropriate definitions for dyspepsia at both time points. Approximately 15% to 20% of physicians at each time point reported that they would not use any of these definitions, however, and instead listed alternatives. These open-ended definitions centered on a few specific themes, including connection of symptoms to meals, symptom response to medication, associated features, and relation to other disorders. In addition, a small percentage of physicians (approximately 4%) elected not to provide a definition of dyspepsia because they considered the condition to be too vague, ill-defined, or irrelevant to their patients.
In 2006, between 70% and 75% of physicians reported the belief that biopsy evidence of histologic esophagitis, eosinophilic gastroenteritis, or Helicobacter pylori would be indicative of organic disease likely to explain a patient's pain. Greater disagreement was noted for giardiasis (60%), chronic gastritis (51%), and chronic duodenitis (48%). No comparison was available for 1992. In 2006, physicians most frequently cited “functional” (37%) and reflux/esophagitis (26%) as the most common cause of pediatric dyspepsia, followed by peptic gastritis/duodenitis (8%), psychosocial factors (6%), and IBS (5%). Physician-estimated rates of patients with a perceived psychological, psychiatric, or behavioral basis for dyspepsia increased across time (χ2[5, N = 308] = 12.29, P < 0.05), but remained lower than that for abdominal pain in general (χ2[5, N = 342] = 23.72, P < 0.001).
Assessment and Treatment of Pediatric Abdominal Pain
The number of patients with pediatric abdominal pain evaluated annually per physician significantly increased from 1992 to 2006 (χ2[3, N = 323] = 37.51, P < 0.001), whereas the percent of physicians seeing more than 300 children per year became more than double (20% in 1992 vs 52% in 2006). In 2006, NASPGHAN members reported that they most often used reassurance (87%), referral to a mental health professional (65%), and prescription of tricyclic antidepressants (eg, elavil, imipramine; 62%) for patients presenting with abdominal pain for which no organic cause could be identified. Fewer than 20% of respondents reported recommending the other treatments listed in the Pediatric Gastroenterology Practice Survey (eg, selective serotonin reuptake inhibitors, biofeedback-assisted relaxation training, herbal supplements, osteopathic manipulation). However, as illustrated in Figures 1 and 2, physicians spontaneously identified a wide range of treatments, both medication and nonmedication based, that are used routinely in clinical practice when asked more broadly about their individual treatment approach. The most commonly prescribed medications included antispasmodics (25%), stool softeners/laxatives (23%), and acid suppressors (10%), although the most commonly recommended nonmedication treatments included lifestyle changes (14%) and other complementary medicine approaches (eg, acupuncture, peppermint, meditation; 6%). No comparison was available for 1992.
Assessment and Treatment of Pediatric Dyspepsia
The frequency of physicians using specific evaluation tests as part of the initial workup for dyspepsia changed from 1992 to 2006. As illustrated in Table 2, gastric-emptying studies became more common (χ2[1, N = 321] = 11.10, P = .001) and urine laboratories decreased in use (χ2[1, N = 321] = 7.96, P < 0.01). In addition, whereas the frequency of endoscopy in the evaluation of dyspepsia remained stable (χ2[1, N = 321] = 1.19, P = ns), the rates of esophageal (χ2[1, N = 318] = 14.43, P < 0.001), gastric body (χ2[1, N = 320] = 19.15, P < 0.001), and descending duodenal (χ2[1, N = 320] = 54.20, P < 0.001) biopsies increased (Table 3). Physicians spontaneously reported using several other test types as part of the standard workup, including “diagnostic” trials with specific medications (4% in 1992 vs 10% in 2006), stool tests for parasites (4% in 1992 vs 7% in 2006), tests for H pylori (6% in 1992 vs 7% in 2006), and stool tests for blood (6% in both 1992 and 2006). Approximately 5% of physicians in 2006 reported including psychological assessment and/or consultation from a mental health practitioner as part of the “standard” evaluation for dyspepsia. No comparison was available for 1992.
A high percentage of physicians reported treating children with symptoms of dyspepsia empirically before diagnosis, with this number increasing over time (74% in 1992 vs 83% in 2006; χ2[2, N = 318] = 6.74, P < 0.05). Acid suppressors were the most commonly prescribed empirical treatment in both 1992 (65%) and 2006 (72%), with all other treatments cited by <5% of the physicians responding. Less than half of patients with dyspepsia were considered to have had a “good” clinical response to this empiric treatment, with no improvement in response rates noted from one time point to another (χ2[5, N = 289] = 3.28, P = ns).
Knowledge and Use of the Rome Criteria
Nearly all of the physicians surveyed in 2006 reported having heard of the Rome criteria (96%) and/or being knowledgeable about it (74%), but less than half (39%) indicated using this classification system in practice. Approximately 68% of physicians provided a response to the open-ended question asking about the benefits they see in using the Rome criteria, whereas 61% provided a response about drawbacks. Identified benefits covered a variety of topics, including providing a consistent terminology to use, being helpful in training residents, and providing a “definition” of functional abdominal pain. Identified drawbacks also covered a variety of topics, including the complexity or simplicity of the criteria, lack of validation of the criteria, and logistics of application to pediatric patients. The most commonly cited benefits and drawbacks are listed in Table 4. Approximately 6% of physicians reported seeing no benefits or not using the criteria at all, whereas 3% reported seeing no problems with the existing criteria.
When asked to define a functional gastrointestinal disorder, physicians in 2006 most commonly provided a definition focused on symptoms due to specific underlying pathophysiology (45%), although the specific symptoms and pathophysiological processes that were mentioned did vary. Approximately 16% of physicians either directly or indirectly referenced the Rome criteria in the definition they provided, whereas 9% defined FGIDs based on general symptom presentations that contained no reference to the Rome criteria. The remaining definitions were too vague to be classifiable or were reported by <5% of physicians. Finally, across all of the FGID definitions provided, approximately 70% of physicians included a specific requirement that organic disease be ruled out.
This study primarily examined how children with chronic abdominal pain are viewed, assessed, and treated by pediatric gastroenterologists across North America, and how perspectives have changed since initial release of the Rome criteria approximately 15 years ago. Findings from this study highlight several changes in pediatric gastroenterology practice that have occurred over time, as well as many areas in which actual practice has remained consistent despite theoretical shifts in the conceptualization of chronic abdominal pain and, more specifically, FGIDs.
Although a relatively high percentage of physicians reported the belief that pediatric abdominal pain most often has a “functional” cause, the term “functional” did not have a consistent meaning across the pediatric gastroenterologists surveyed in this study. Substantial variability was evidenced in the definitions provided for a functional gastrointestinal disorder, or FGID. Furthermore, many of the definitions provided appeared inconsistent, or even contradictory, with one another and/or with the Rome criteria.
Physicians also varied significantly in what testing they believe is necessary to rule out biochemical or structural explanations for abdominal pain before making a diagnosis of an FGID, such as dyspepsia. This is consistent with a recent clinical report, written by a committee of experts representing the American Academy of Pediatrics Subcommittee on Chronic Abdominal Pain and the NASPGHAN Committee on Abdominal Pain, which found that no clinical practice guidelines for evaluation could be established based on the available literature at this time (4). As suggested in this clinical report, it remains unclear which specific evaluation tools are most helpful because most have not been studied systematically at this time. Further research in this area clearly is warranted.
Although no “standard” assessment could be identified, some clear group shifts over time were observed with regard to a few tests used to evaluate pediatric dyspepsia. Specifically, urine laboratories are less likely to be used in present evaluations for dyspepsia, whereas the use of gastric emptying studies has more than doubled and several areas (eg, esopghagus, gastric body, descending duodenum) are being biopsied more regularly as part of endoscopy.
Despite this increase in biopsy, findings from this study suggest that conflicting views continue to exist regarding the relation between histologic inflammation and symptoms in FGIDs. This uncertainty may be due to a lack of prospective treatment studies demonstrating clinical correlation or clinical efficacy of treatments aimed at these histologic entities. Several lines of research may help to elucidate the relation between histologic inflammation and symptoms in FGIDs in the near future. These include prospective treatment trials to determine whether specific histologic findings predict treatment response or prognosis, studies evaluating relations between inflammation and mechanosensory function and/or psychological function, and studies evaluating relations between sites and types of inflammation with individual symptoms or symptom clusters.
Despite development and dissemination of a biopsychosocial model for FGIDs for the past 15 years, conflicting views continue to exist regarding the relation of psychological, psychiatric, and behavioral factors to chronic abdominal pain in children. In 2006, physicians were no more likely to endorse 1 of these factors as a basis for their patients' abdominal pain than they were in 1992, with estimates in this area varying widely from one physician to another. The phrasing of this survey question was broad, allowing psychological factors to be endorsed either as a primary cause or merely as a contributory cause and, thus, should be viewed as a higher-end estimate of the perceived importance of psychological factors in the onset and maintenance of pediatric abdominal pain. It is important to note that this survey question, like many others, was intentionally left unchanged from the original 1992 administration to ensure comparability across time points. Although helpful in facilitating comparisons, this approach also has some limitations. For example, different responses could have been elicited by framing this particular question to ask about actual change in the physicians' views regarding etiology during the last 15 years. Similarly, framing this question to ask specifically about the relative contribution of various psychological, psychiatric, behavioral, and/or social factors to symptom expression and symptom-related disability may have yielded more detailed information about current perspectives. Further investigation in this area may be warranted to better elucidate possible subtle shifts in biopsychosocial perspective that could not be detected using the present methodology.
In terms of impact of beliefs about psychological, psychiatric, and behavioral etiology on practice, psychological evaluation was included in “standard” assessment for only 5% of physicians (at least for the subset of children with dyspepsia). However, reassurance, referral to a mental health professional, and prescription of a tricyclic antidepressant were the top 3 treatment strategies endorsed by the physicians completing the 2006 survey. Specifically, 65% of physicians reported routinely referring patients to a mental health provider. This seems appropriate given that cognitive-behavioral therapy is the treatment best supported by presently available evidence and is considered “probably efficacious” according to the widely established empirically supported treatments criteria (5,6). However, the fact that many physicians do not feel that there is a psychological basis for their patients' abdominal pain and do not include psychological evaluation in their assessment, yet routinely use or refer for what are considered to be psychological treatments, may create a potential disconnect for patients/families that could increase treatment resistance and/or seeking of further medical evaluation. Beyond the scope of the present survey, research on rates of adherence and perceived barriers to treatment recommendations from the family and provider perspectives would be helpful in directing future integrative care efforts.
Overall, study findings suggest that the creation of the Rome criteria has not resulted in substantial shifts in conceptualization or practice in the area of pediatric abdominal pain during the last 15 years. Although nearly all physicians have heard of the Rome criteria, and most are knowledgeable about it, a small percentage have incorporated it into their working definition of FGIDs and less than half actually use it in practice. A variety of factors may be limiting the use of the criteria in practice at this time.
Qualitative examination of the physician-identified benefits and drawbacks, as well as reported variability in conceptualization, evaluation, and treatment across the gastroenterologists surveyed, suggest that the criteria may not have been communicated adequately to practicing physicians. For instance, many of the physicians surveyed viewed “functional” disorders as diagnoses of exclusion and reported using a variety of testing procedures, which is at odds with the intent of the Rome criteria. In fact, these criteria were designed to be diagnoses of inclusion, standing alone, and reducing the need for testing in the absence of red flags (7). Difficulty also has been reported in consistent application of the criteria among physicians when presented with the same clinical vignette (8). In addition, reliability of the criteria appears problematic in application to children because of different information provided by families (ie, parent and child) resulting in varying diagnoses or, at a minimum, increased physician subjectivity in determining how to interpret and weight the information provided (9). Uncertainty remains about how best to reconcile and interpret information provided by parents, children, tests, and direct clinical observation in making an FGID diagnosis. The limited use of the Rome criteria in practice may also result from a lack of proven clinical utility at this time. Prospective treatment studies using the Rome criteria will be necessary to determine whether diagnosing by these criteria will benefit patients.
In terms of pediatric dyspepsia, specifically, data from the present study suggest that it is not well defined or agreed upon among gastroenterology physicians. Assessment and treatment practices are highly variable among practitioners despite efforts to better codify this cluster of symptoms via the Rome criteria. Even within the Rome process, the diagnostic entity of pediatric dyspepsia has undergone significant changes, such as losing its original subtyping. Although a different set of subtypes recently were proposed for adults, subtypes were eliminated entirely for pediatric dyspepsia in a recent revision of the Rome criteria because of a current lack of evidence to support their existence (2). Research studies designed to improve understanding of pediatric dyspepsia and its pathophysiology, especially those that simultaneously consider potential biological, psychological, and social contributions, is needed to guide further Rome criteria revisions and clinical practice more generally.
Pediatric gastroenterologists are seeing more children for evaluation of chronic abdominal pain, yet estimates of treatment responsiveness have not improved substantially for the past 15 years. This increase in the number of new patient evaluations per physician may be because of a variety of factors not assessed in this survey. Some possibilities include an increase in the incidence of children with chronic abdominal pain, greater numbers of families seeking treatment for their children, more subspecialty referrals from primary care physicians, or amplified pressure on physicians to see/bill for more patients yearly. Regardless of the cause, this increase in evaluation rates underscores that chronic abdominal pain remains a significant problem for children in North America and one with which pediatric gastroenterologists will commonly be faced in clinical practice. Improving our understanding of the psychophysiology of chronic abdominal pain, and FGIDs more specifically, will be necessary to improve the quality and consistency of assessment and treatment approaches. Although the Rome criteria have helped us to take some initial steps in this direction, the responses provided by practicing pediatric gastroenterologists in this study highlight several important questions that remain at this time and point out the direction for future research efforts. For instance, what evaluations yield “abnormalities” with sufficient frequency to warrant their standard use? Are these “abnormalities” relevant, predicting prognosis or treatment response? Can we reliably predict relevant abnormalities from the patient history? Are red flag symptoms all that really matter when determining whether to conduct formal testing? Equally important, we also must take into account in future research the relevant biological, psychological, and social factors concurrently to better understand how these factors influence one another within each diagnostic category. This comprehensive view will enrich our studies and ultimately help to guide the appropriate integration of psychological and environmental treatments into standard medical care for children with chronic abdominal pain in the future. Finally, just as practicing pediatric gastroenterologists have helped us to point the way for future research via responses to our survey, it will be critical to consider how best to translate the resulting expansion in knowledge base back into practice through improved dissemination efforts geared toward the practitioners most involved in seeing children with chronic abdominal pain on a daily basis.