- Email: [email protected]
Nonspecific Abdominal Pain in Pediatric Primary Care: Evaluation and Outcomes
Nonspecific Abdominal Pain in Pediatric Primary Care: Evaluation and Outcomes Elizabeth M. Wallis, MD, MSHP; Alexander G. Fiks, MD, MSCE From the Division of General Pediatrics, Medical University of South Carolina, Charleston, SC (Dr Wallis); and Pediatric Research Consortium (PeRC), Center for Pediatric Clinical Effectiveness and PolicyLab at The Children’s Hospital of Philadelphia, Philadelphia, Pa (Dr Fiks) The authors declare that they have no conflict of interest. Address correspondence to Elizabeth M. Wallis, MD, MSHP, Medical University of South Carolina, 135 Rutledge Ave, MSC 561, Charleston, SC 29425 (e-mail: [email protected]). Received for publication April 14, 2014; accepted September 13, 2014.
ABSTRACT OBJECTIVE: To describe the characteristics of children with
diagnosis for their pain. Seventeen percent and 14% of children had laboratory and radiology testing at the index visit, respectively. Only 3% of laboratory evaluations helped to yield a diagnosis. Among variables considered, only preceding pain of more than 7 days at the index visit was associated with persistent pain (odds ratio 2.15, 95% confidence interval 1.19– 3.89). None of the variables considered was associated with receiving a specific diagnosis. CONCLUSIONS: Most children with AP do not receive a diagnosis, many have persistent pain, and very few receive a functional AP diagnosis. Results support limited use of diagnostic testing and conservative management consistent with national policy statements.
nonspecific abdominal pain (AP) in primary care, their evaluation, and their outcomes. METHODS: Between 2007 and 2009, a retrospective cohort of children from 5 primary care practices was followed from an index visit with AP until a well-child visit 6 to 24 months later (outcome visit). Using International Classification of Disease, 9th Revision (ICD-9), codes and chart review, we identified afebrile children between 4 and 12 years old with AP. Use of diagnostic testing was assessed. Multivariable logistic regression was used to model the association of index visit clinical and demographic variables with persistent pain at the outcome visit, and receipt of a specific diagnosis. RESULTS: Three hundred seventy-five children presented with AP, representing 1% of the total population of 4- to 12-year-olds during the study period. Eighteen percent of children had persistent pain, and 70% of the study cohort never received a specific
KEYWORDS: abdominal pain; functional gastrointestinal disorders; primary care ACADEMIC PEDIATRICS 2014;-:1–7
WHAT’S NEW
or recurrent pain. Chronic AP can cause significant impairment in children and increase the risk of functional disorders in adulthood if not adequately addressed.10,11 A growing body of literature has examined the care of children with functional AP, but data are based primarily on studies in subspecialty settings.12,13 Furthermore, clinicians have been criticized for the overly aggressive diagnostic evaluation and subspecialty referral of children with AP.14 Testing and subspecialty referrals for AP have also been reported as inconsistent15 and without an evidence base.16,17 The Rome III criteria were developed to help classify functional AP disorders and better guide management. Specifically, there are 5 pediatric AP-related FGIDs, including childhood functional AP, which is pain for at least 2 months without evidence of an inflammatory, anatomic, metabolic, or neoplastic process.18,19 Additionally, guidelines from the American Academy of Pediatrics (AAP) and the North American Society of Gastroenterology, Hepatology and Nutrition (NASPGHAN) support a conservative approach to children with chronic AP.20,21 Despite this, FGIDs are underdiagnosed, and it is unclear how frequently primary care physicians are using the Rome III criteria to characterize functional AP disorders. Multiple small studies of psychological treatments for FGIDs have shown efficacy in reducing pain and limiting functional
In this large study of children with nonspecific abdominal pain in primary care, we found persistent pain is common, and clinicians may underrecognize functional gastrointestinal disorders in this population of children.
ABDOMINAL PAIN (AP) is one of the most common complaints to pediatricians and accounts for 2% to 4% of primary care visits annually.1 Almost all children experience AP at some point, and the estimated prevalence of chronic AP in school children is 10% to 19%.2–4 The most common causes of AP in children include functional gastrointestinal disorders (FGIDs), constipation, and gastroenteritis, but many children present to primary care physicians with AP with no immediately identifiable cause. A few small studies have examined the diagnoses that children with AP have received when presenting to primary care and emergency departments and report that 15% to 35% receive a nonspecific AP diagnosis.5–8 Despite the prevalence of nonspecific AP, we lack longitudinal studies that examine the clinical course of children with AP beyond the initial visit, the evaluations these children receive, or their outcomes. In most children, AP is benign and self-limited.9 However, some children develop chronic ACADEMIC PEDIATRICS Copyright ª 2014 by Academic Pediatric Association
1
Volume -, Number -–- 2014
2
WALLIS AND FIKS
ACADEMIC PEDIATRICS
impairment,22–25 but these treatments are used infrequently, in part because primary care physicians may not be adequately classifying functional disorders and matching children to appropriate treatments. An improved understanding of the clinical course of children with AP is necessary in order to accurately diagnose children with FGIDs, avoid unnecessary testing, and match children with appropriate treatment. To help achieve this goal, we described the population of children presenting with AP in primary care, their management, and their clinical course.
METHODS STUDY POPULATION AND SETTING Eligible patients were drawn from 5 primary care practices from The Children’s Hospital of Philadelphia (CHOP) Pediatric Research Consortium (PeRC), a primary care practice–based research network.26,27 Children aged 4 to12 years at the time of study entry were identified using AP-related International Classification of Disease, 9th Revision (ICD-9), codes at an index visit between January 1, 2007, and December 31, 2009. We searched using ICD-9 codes for AP not otherwise specified and chronic AP (789.00 to 789.09). Because of our broad interest in children with AP, we also searched for children with the 5 AP-related FGID diagnoses at the index visit (functional dyspepsia, 536.8; FGID, 536.9; functional constipation, 564.09, irritable bowel syndrome, 564.1, abdominal migraine, 346.2) but found none; nor did we find any other functional digestiverelated codes (564). For simplicity, and consistent with prior work,5 we refer to this cohort of children in our study as having nonspecific AP. We focused on children ages 4 to 12 years because they are the group most commonly affected by AP.2 By manual chart review, we performed a 6-month look back to ensure that none of the cohort had a visit for AP in the preceding 6 months and that our data set accurately captured the index visit. To ensure that any follow-up was also captured, children included in the study cohort needed to have at least one well-child visit 6 to 24 months after the index visit (outcome visit). Well visits were chosen as the outcome visit because they include a detailed review of systems and pain assessment. In contrast, sick visits are variable in their content. Children with fever or a secondary diagnosis at the index visit that indicated a likely cause of the AP were excluded. The Figure outlines the eligible patient pool and the ICD-9 codes used to exclude patients to obtain the final study cohort. DATA COLLECTION Data were collected using chart review and supplemented by the automated extraction of data from electronic health records. Using a standardized abstraction form, clinical information was collected regarding children’s AP and associated symptoms; physical examination, laboratory, and radiology studies and results; and primary care follow-up. Chart review included visits 6 months before
Figure. Narrowing of practice population to study cohort.
the index visit, the index visit, and all subsequent care through June 30, 2012. Charts were reviewed by the research team and any ambiguous cases were resolved by consensus. Study data were collected and managed using REDCap electronic data capture tools hosted at The Children’s Hospital of Philadelphia.28 OUTCOME VARIABLES The 2 primary outcomes were persistent AP (defined as pain at the outcome visit) and receipt of a specific diagnosis for the AP at or before the outcome visit. We also evaluated the number of children in the cohort who went on to receive a diagnosis consistent with the Rome III criteria. The secondary outcomes were the frequency of laboratory and radiology evaluations at the index visit and their results (classified as normal, incidental finding, abnormal but no impact on clinical management, abnormal and contributed to diagnosis). We considered patients diagnosed with constipation via radiograph if the clinician’s documentation referenced the radiograph as diagnostic. INDEPENDENT VARIABLES Independent variables considered included patient age, race (black, white, other), insurance type (Medicaid, non-Medicaid), gender, and duration of pain at the index visit (<7 days, $7 days). Seven days was chosen as a clinically meaningful duration. To ensure that results were not biased on the basis of the specification of the exposure (pain duration), we conducted a sensitivity analysis including pain durations of 14 and 21 days as cut points. Additional covariates included associated symptoms, pain score at the visit (0, 1–10), findings of physical examination (diffuse pain, localized pain, normal abdominal examination findings), and the presence of complex chronic conditions as defined by Feudtner et al29 in any visit in the study window or in the child’s chronic problem
ACADEMIC PEDIATRICS
NONSPECIFIC ABDOMINAL PAIN
list. Missing data were rare (<5%) for all variables used in the analyses. STATISTICAL ANALYSES We characterized the population of children with AP using descriptive statistics and compared them to the overall population of children of the same age at the study practices. To avoid colinearity in multivariable models, the correlation between independent variables and among covariates was assessed. If 2 variables were highly correlated (r > 0.8), the one deemed most clinically meaningful was retained. Bivariate analysis was used to assess the association between each covariate and each outcome. In adjusted models, all covariates were retained if associated with the outcome at P < .3 (prespecified). Logistic regression models were then created to measure the association of independent variables with the outcomes of persistent pain (primary) and receipt of a specific diagnosis explaining the AP. In a secondary analysis, we described the frequency of laboratory and radiology evaluations and whether the results changed clinical management. All analyses were completed by Stata 12 software (StataCorp, College Station, TX). The CHOP institutional review board approved this study. Waiver of consent was granted.
RESULTS STUDY POPULATION Table 1 summarizes the demographic characteristics of the 375 children in the final study cohort compared to all children aged 4 to 12 years seen over the same time period. The study cohort represented 1% of the total population of children 4 to 12 years old. The overall population at 5 urban Table 1. Demographic Characteristics of Total and Study Populations* Characteristic N Race White Black Other Gender Female Male Payor Medicaid Non-Medicaid Age, y Mean 4–9 y 10–12 y Complex chronic condition Yes No
Study Population, n (%)
Total Population, n (%)
375
34202
P <.001
65 (17) 253 (68) 57 (15)
4017 (12) 26943 (79) 3242 (9) .018
207 (55) 168 (45)
16779 (49) 17423 (51)
199 (53) 176 (47)
21070 (62) 13132 (38)
7.9 256 (68) 119 (32)
8.2 20942 (61) 13260 (39)
.018
.005
.11 69 (19) 292 (81)
5466 (16) 28736 (84)
*Total population is all patients aged 4 to 12 years, seen in the same time period in the 5 practices from which the study cohort was derived.
3
study practices was varied, with 79% black and 62% Medicaid insured. Children evaluated with AP were more likely to be girls (55%, P ¼ .018), non-Medicaid insured (53%, P ¼ .018), and white (17%, P < .001) than the larger population. Children with AP and those without had similar rates of complex chronic conditions (19% vs 16% P ¼ .11). Table 2 summarizes the clinical characteristics of children with AP in the study cohort. Of note, all the patients in our cohort had an AP not otherwise specified code at the index visit, and there were none with the other diagnoses searched. Among children with AP, 42% reported a history of diffuse pain and 28% had localized pain (data not shown), though only 34% of children had abdominal tenderness at physical examination. Almost three quarters (72%) reported intermittent pain. Children presented with a median pain duration of 5 days (range 1–1095 days), with 42% presenting with pain for >7 days and 23% presenting with pain for >30 days. The most common associated symptoms at presentation were vomiting (26% of the cohort), decreased appetite (18%), nausea (17%), diarrhea (13%), difficulty stooling (13%), and headache (12%). More than one quarter of children (26%) reported no associated symptoms. PERSISTENT AP AND RECEIPT OF A SPECIFIC DIAGNOSIS Persistent AP was common in our sample. Eighteen percent of children in the cohort reported persistent pain at the outcome visit. In unadjusted analyses, patients with Table 2. Abdominal Pain Characteristics at Index Visit* Clinical Characteristic Pain duration at visit 1 #7 d >7 d Tenderness at physical examination Localized tenderness Diffuse tenderness No tenderness Pain pattern Constant Intermittent Missing/not reported Associated symptoms† Vomiting Decreased appetite Nausea Difficulty stooling Diarrhea Headache Dysuria/urinary symptoms Cough Congestion/rhinorrhea Chest pain Fatigue Weight loss Joint pain Skin changes Bloody stools (melena) None *Pain characteristics are patient or family reported. †Patients may report >1 associated symptom.
n (%) 218 (58) 157 (42) 93 (25) 32 (9) 239 (66) 31 (8) 270 (72) 74 (20) 97 (26) 67 (18) 65 (17) 47 (13) 47 (13) 45 (12) 30 (8) 30 (8) 28 (8) 15 (4) 12 (3) 9 (2) 9 (2) 7 (2) 4 (1) 96 (26)
4
WALLIS AND FIKS
ACADEMIC PEDIATRICS
persistent pain at the outcome visit were more likely to present at the index visit with pain for >7 days preceding the visit (odds ratio [OR] 2.33, 95% confidence interval [CI] 1.36–3.97, P ¼ .002). In adjusted models, the association with pain duration was maintained (OR 2.15, 95% CI 1.19–3.89, P ¼ .01). In our sensitivity analyses, with pain duration of either 14 or 21 days as the cut points, results were similar. No other clinical or demographic characteristics were associated with persistent pain in either adjusted or unadjusted analyses (Table 3).
Table 3. Clinical and Demographic Characteristics and Their Associations With Persistent Pain*
Characteristic Payer Medicaid Non-Medicaid Gender Female Male Age 4–9 y 10–12 y Race Black White Other Duration of pain at presentation (index visit) <7 d $7 d Abdominal pain location at physical examination Localized pain Diffuse pain No pain Pain score at index visit† 0 1–10 Complex chronic condition‡,§ No Yes Laboratory studies completed at index visit No Yes Radiology studies completed at index visit No Yes
n
Unadjusted Proportion
Adjusted Analyses OR (95% CI)
In the majority of children (70%) presenting with AP, no specific cause of the AP was determined. In addition, no clinical pain characteristics or demographic variables considered were significantly associated with receipt of a specific diagnosis for the pain (Table 4). Adjusted analysis of radiograph testing was associated with receipt of a diagnosis but not statistically significant (P ¼ .06). This represents cases in which constipation was diagnosed using abdominal radiographs. Of the 30% of children who received a diagnosis (n ¼ 113), the most common diagnoses were constipation (57%), functional AP (16%), and gastroesophageal reflux disease (5%). Among the 262 Table 4. Clinical and Demographic Characteristics and Their Associations With Receipt of a Specific Diagnosis for Abdominal Pain*
P Characteristic
199 176
0.49 0.51
207 168
0.64 0.36
256 119
0.72 0.28
253 65 57
0.66 0.19 0.15
218 157
0.40 0.60
Ref 2.15 (1.19–3.89) .01
93 32 239
0.19 0.17 0.64
Ref 0.51 (0.18–1.42) .20 0.46 (0.18–1.19) .11
171 194
292 69
313 62
324 51
0.52 0.48
0.75 0.25
Ref 0.87 (0.49–1.56) .64
Ref 0.95 (0.87–1.05) .35
Ref 1.44 (0.73–2.87) .30
0.85 0.15
0.87 0.13
CI indicates confidence interval; OR, odds ratio. *Persistent pain defined as abdominal pain still present at a well-child visit 6 months—2 years after the index visit for abdominal pain. †Ten patient records had no pain scores recorded. ‡As defined by Feudtner et al.29 §Data missing for 14 patients (4%).
Payer Medicaid Non-Medicaid Gender Female Male Age 4–9 y 10–12 y Race Black White Other Duration of pain at presentation† <7 d $7 d Abdominal pain location‡,§ Localized pain Diffuse pain No pain Pain scorejj 0 1–10 Complex chronic condition No Yes Laboratory studies completed at index visit No Yes Radiology studies completed at index visit No Yes
Adjusted Analyses
n
Unadjusted Proportion
OR (95% CI)
P
199 176
0.48 0.52
Ref 1.28 (0.80–2.06)
.31
207 168
0.54 0.46
256 119
0.70 0.30
253 65 57
0.62 0.18 0.20
Ref 0.92 (0.55–1.95) 2.16 (0.98–4.75)
.92 .06
218 157
0.53 0.47
93 32 239
0.33 0.07 0.60
171 194
0.49 0.51
292 69
0.75 0.25
Ref 1.72 (0.98–3.00)
.06
313 62
0.85 0.15
324 51
0.82 0.18
Ref 1.85 (0.99–3.45)
.06
CI indicates confidence interval; OR, odds ratio. *Receipt of a specific diagnosis defined using chart review as a diagnosis that adequately explains the etiology of previously defined abdominal pain. †Reported by child/family. ‡At provider physical examination. §A small portion of records (n ¼ 11, 3% of total) were missing any mention of the abdominal physical examination findings. jjTen patient records had no pain scores recorded.
ACADEMIC PEDIATRICS
children with no specific diagnosis, only 12% had any mention of an AP-related functional disorder as part of a differential diagnosis, though 21% of this group of children met one or more criteria for one of the AP-related functional disorders (Rome III criteria) at the index visit. Very few children presenting with afebrile AP were subsequently diagnosed with an emergent or potentially life-threatening diagnosis. In our cohort of 375 children, 3 children were diagnosed with inflammatory bowel disease, 2 with appendicitis, 1 with gall bladder disease requiring cholecystectomy, 1 with sickle cell pain causing AP, and 2 with pancreatitis. In all of these cases, these children presented with red flag–associated symptoms or physical findings based on the AAP policy statement on the evaluation and management of children with chronic AP.11 These symptoms and physical findings include bloody stools, significant weight loss, family history of inflammatory bowel disease, or peritoneal signs at physical examination. All of these children were recognized as having red flag symptoms by their pediatricians and were diagnosed shortly after the index visit. LABORATORY AND RADIOLOGY EVALUATION Although >10% of children had laboratory or radiology testing at the index visit (17% and 14%, respectively), these evaluations rarely changed clinical management or contributed to a specific diagnosis. The majority of laboratory testing was normal (61%). In only 3% of cases did testing help to determine the cause of a child’s AP. Much more commonly, incidental findings (5%) or abnormal findings requiring follow-up (31%) were reported. Examples of incidental findings included a mild electrolyte abnormality on a basic metabolic panel, while abnormal findings requiring follow-up included abnormal autoimmune panels that did not yield an autoimmune diagnosis. A similar pattern was noted in radiology evaluations, with the majority (66%) having normal results. Although 20% (n ¼ 9) of radiology studies aided physicians in diagnosing a patient’s AP, the only diagnosis made via radiology evaluation was constipation. The demographic and clinical characteristics of children who underwent laboratory or radiology testing were similar to children who did not (P > .15 for all variables). Pain for more than 7 days at presentation was associated with an increased likelihood of laboratory (OR 3.05, 95% CI 1.73–5.38, P ¼ .0001) and radiology testing (OR 2.43, 95% CI 1.33–4.46, P ¼.004), and children without abdominal tenderness at physical examination were less likely to receive radiology imaging (OR 0.38, 95% CI 0.37–2.59, P ¼ .04) than those with either diffuse or localized pain. Other variables considered were not associated with testing. Laboratory testing was not associated with persistent pain or receipt of a specific diagnosis. Radiology testing was not associated with persistent pain, but clinicians used x-rays to diagnose constipation-related pain (P ¼ .06 for the comparison of whether or not radiology studies were ordered at the index and the association with receipt of a specific AP diagnosis).
NONSPECIFIC ABDOMINAL PAIN
5
DISCUSSION The literature lacks longitudinal studies of children with AP in pediatric primary care. In our study, we found that nearly one fifth of children in our cohort of 375 children diagnosed with AP had persistent pain at a well-child visit 6 to 24 months after the index visit. Children presenting with pain for >7 days were more likely to have persistent AP but were not more likely to receive a specific diagnosis for their AP. Less than one third of all children received a specific diagnosis, and while many patients with AP may have FGIDs, this was uncommonly documented as a potential diagnosis. Diagnostic codes for FGIDs were never used in this sample. When used, laboratory and radiology testing rarely changed clinical management. Studies of hospitalized children and those referred to gastroenterologists with functional abdominal complaints have documented a similar demographic distribution to our study, with those affected more likely to be girls, to have non-Medicaid insurance, and to be white.7,30 In total, AP complaints account for an estimated 2% to 4% of primary care visits annually.1 Consistent with our results, prior studies have estimated a significant proportion of these to be nonspecific or functional complaints.23 We found that persistent AP was common among those presenting with AP. Eighteen percent of children in our cohort with AP had persistent pain at the outcome visit, and children with pain for >7 days at the index visit were more than 2 times more likely to have persistent AP. To our knowledge, no previous study has recognized this association. Our findings may help primary care clinicians identify children who are likely to develop a FGID. In the majority of children who presented with AP, a cause for their AP was not determined despite persistent pain. These results support a high index of suspicion for a functional diagnosis among children with AP. Given the high prevalence of persistent AP, many of the children in our cohort may have an AP-related functional disorder, although only 5% were identified as such. From previous studies, it is known that pediatricians often fail to recognize functional disorders,11,31 likely accounting for the low number of children diagnosed in our cohort. Although clinical care does not always mirror diagnostic coding, both the clinical decision making in reviewed charts and the ICD-9 coding did not reflect consideration of a functional diagnosis. The Rome III criteria provide specific diagnostic classifications and an algorithm to assist in making a functional AP diagnosis.32 However, even gastroenterologists have reported inconsistent knowledge and use of the Rome III criteria,31 and although the criteria have been validated, their clinical usefulness has not been well examined in primary care or subspecialty practices. Our results indicate that primary care clinicians may not be using these criteria for children likely to have a functional disorder. These results suggest that work is needed to support the implementation of the Rome III criteria and the AAP policy statement on chronic AP into primary care settings. Clinician training, feedback, and decision support, ideally delivered as part
6
WALLIS AND FIKS
of a quality improvement initiative, may facilitate this process. The AAP policy statement on chronic AP supports judicious testing in this group of children, and our results strengthen the evidence base for these recommendations.20 Specifically, the AAP policy statement recommends laboratory or radiology studies only in children with one or more red flags raised at the history or physical examination. Consistent with these recommendations, we found that testing rarely changed clinical management and that children who underwent testing were not more likely to receive a specific diagnosis. Further restriction of radiographic testing may be possible. The use of abdominal radiograph to diagnose constipation, the major clinical contribution of radiograph use in our study, is not generally recommended; providers should instead rely on a stooling history and physical examination.33 Additionally, all of the children in our cohort with an emergent or potentially lifethreatening diagnosis had one or more red flag findings at the index visit. Several studies on children suspected of having functional AP have identified reassurance of the family as a motivator for physicians ordering tests.15 Given evidence that testing does not provide the reassurance intended,15 and with the incidence of hospitalization for chronic pain rising,30 it is important that primary care providers limit unnecessary testing and match children with evidence-based treatments. Our findings including the prevalence of persistent pain underscore the need for further work to identify bestpractice behavioral health interventions and integrate these interventions into clinical practice. Effective psychological treatments have been identified for children with persistent functional symptoms and impairment in daily activities. Psychological treatments including cognitive behavioral therapy, family therapy, biofeedback, gut-directed hypnotherapy and psychoeducational therapies have been piloted in referred patients, with improvements in pain symptoms and functional impairment in children, and improved coping by parents and children.10,22,34,35 However, most behavioral health interventions for functional AP have yet to be introduced into primary care practice and are not widely available. This study had several limitations. Although children treated in multiple primary care sites were studied, all the sites were part of one larger health system. In addition, our study sample was predominantly urban and minority patients, and further study is needed to assess how these results generalize to other health settings and populations. It is also possible that the practices studied, as part of an academic medical center, may have different familiarity with the Rome criteria than community practitioners. Our study was limited to information accessible in the medical record. We were unable to assess the interaction between the clinician and family, which may have influenced testing choice or treatment plan, as it was beyond the scope of this study. We were also not able to assess factors related to when a family seeks care (eg, pain tolerance and parental anxiety). Variables including social stressors, family history of mental health disorders or chronic pain syndromes,
ACADEMIC PEDIATRICS
missed school, and other psychological factors that were not consistently documented may have helped to explain our findings. It is also possible that patients received care outside of the health system and receipt of this care was not captured in the study. In addition, outcomes such as functional impairment could not be assessed in this retrospective review. A prospective study that captures these variables as well as outcomes including functional impairment for children with AP treated in primary care is warranted.
CONCLUSIONS In this longitudinal study of children with AP in primary care, we found that persistent pain is common, but few children are given a functional diagnosis (using ICD-9 codes for FGIDs or based on clinician documentation of their assessment and differential diagnosis). Our results support the AAP policy statement on chronic AP and indicate that laboratory and radiology studies are most often normal and rarely change clinical management. Additional training is needed to improve recognition of FGIDs by providers in order to better manage children with persistent pain and avoid unnecessary testing.
ACKNOWLEDGMENTS We would like to thank James P. Guevara, MD, Robert W. Grundmeier, MD, and William T. Basco, MD, for their thoughtful review. We would like to thank the network of primary care clinicians and their patients and families for their contributions to clinical research through the Pediatric Research Consortium at CHOP. Data were managed using REDCap (Research Electronic Data Capture), a secure, Web-based application designed to support data capture for research studies.28 Supported by training grant HRSA T32 HP10026 from the National Institutes of Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
REFERENCES 1. Starfield B, Hoekelman RA, McCormick M, et al. Who provides health care to children and adolescents in the United States? Pediatrics. 1984;74:991–997. 2. Apley J, Naish N. Recurrent abdominal pains: a field survey of 1,000 school children. Arch Dis Child. 1958;33:165–170. 3. Schwille IJ, Giel KE, Ellert U, Zipfel S, Enck P. A community-based survey of abdominal pain prevalence, characteristics, and health care use among children. Clin Gastroenterol Hepatol. 2009;7:1062–1068. 4. Saps M, Seshadri R, Sztainberg M, Schaffer G, Marshall BM, Di Lorenzo C. A prospective school-based study of abdominal pain and other common somatic complaints in children. J Pediatr. 2009; 154:322–326. 5. Gieteling MJ, Lisman-van Leeuwen Y, van der Wouden JC, et al. Childhood nonspecific abdominal pain in family practice: incidence, associated factors, and management. Ann Fam Med. 2011;9:337–343. 6. Reynolds SL, Jaffe DM. Diagnosing abdominal pain in a pediatric emergency department. Pediatr Emerg Care. 1992;8:126–128. 7. Loening-Baucke V, Swidsinski A. Constipation as cause of acute abdominal pain in children. J Pediatr. 2007;151:666–669. 8. Scholer SJ, Pituch K, Orr DP, Dittus RS. Clinical outcomes of children with acute abdominal pain. Pediatrics. 1996;98(4 pt 1):680–685. 9. Chitkara DK, Rawat DJ, Talley NJ. The epidemiology of childhood recurrent abdominal pain in Western countries: a systematic review. Am J Gastroenterol. 2005;100:1868–1875.
ACADEMIC PEDIATRICS 10. Wendland M, Jackson Y, Stokes LD. Functional disability in paediatric patients with recurrent abdominal pain. Child Care Health Dev. 2010;36:516–523. 11. Walker LS, Dengler-Crish CM, Rippel S, Bruehl S. Functional abdominal pain in childhood and adolescence increases risk for chronic pain in adulthood. Pain. 2010;150:568–572. 12. Lisman-van Leeuwen Y, Spee LA, Benninga MA, et al. Prognosis of abdominal pain in children in primary care—a prospective cohort study. Ann Fam Med. 2013;11:238–244. 13. Gieteling MJ, Bierma-Zeinstra SM, Passchier J, Berger MY. Prognosis of chronic or recurrent abdominal pain in children. J Pediatr Gastroenterol Nutr. 2008;47:316–326. 14. Miele E, Simeone D, Marino A, et al. Functional gastrointestinal disorders in children: an Italian prospective survey. Pediatrics. 2004; 114:73–78. 15. Bonilla S, Deli W, Saps M. The prognostic value of obtaining a negative endoscopy in children with functional gastrointestinal disorders. Clin Pediatr (Phila). 2011;50:396–401. 16. Lane MM, Weidler EM, Czyzewski DI, Shulman RJ. Pain symptoms and stooling patterns do not drive diagnostic costs for children with functional abdominal pain and irritable bowel syndrome in primary or tertiary care. Pediatrics. 2009;123:758–764. 17. van Tilburg MA, Venepalli N, et al. Parents’ worries about recurrent abdominal pain in children. Gastroenterol Nurs. 2006;29:50–55. 18. Baber KF, Anderson J, Puzanovova M, Walker LS. Rome II versus Rome III classification of functional gastrointestinal disorders in pediatric chronic abdominal pain. J Pediatr Gastroenterol Nutr. 2008; 47:299–302. 19. Drossman DA, Dumitrascu DL. Rome III: new standard for functional gastrointestinal disorders. J Gastrointestin Liver Dis. 2006;15: 237–241. 20. Di Lorenzo C, Colletti RB, Lehmann HP, et al. Chronic abdominal pain in children: a clinical report of the American Academy of Pediatrics and the North American Society for Pediatric Gastroenterology, Hepatology and Nutrition. J Pediatr Gastroenterol Nutr. 2005;40: 245–248. 21. Chronic abdominal pain in children. Pediatrics. 2005;115: e370–e381. 22. Eccleston C, Palermo TM, Williams AC, et al. Psychological therapies for the management of chronic and recurrent pain in children and adolescents. Cochrane Database Syst Rev. 2009;CD003968.
NONSPECIFIC ABDOMINAL PAIN
7
23. von Baeyer CL. Understanding and managing children’s recurrent pain in primary care: a biopsychosocial perspective. Paediatr Child Health. 2007;12:121–125. 24. Levy RL, Langer SL, Walker LS, et al. Twelve-month follow-up of cognitive behavioral therapy for children with functional abdominal pain. JAMA Pediatr. 2013;167:178–184. 25. Huertas-Ceballos A, Logan S, Bennett C, Macarthur C. Psychosocial interventions for recurrent abdominal pain (RAP) and irritable bowel syndrome (IBS) in childhood. Cochrane Database Syst Rev. 2008;CD003014. 26. Fiks AG, Grundmeier RW, Margolis B, et al. Comparative effectiveness research using the electronic medical record: an emerging area of investigation in pediatric primary care. J Pediatr. 2012;160:719–724. 27. Grundmeier RW, Swietlik M, Bell LM. Research subject enrollment by primary care pediatricians using an electronic health record. AMIA Annu Symp Proc. 2007;289–293. 28. Harris PA, Taylor R, Thielke R, et al. Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42:377–381. 29. Feudtner C, Silveira MJ, Christakis DA. Where do children with complex chronic conditions die? Patterns in Washington State, 1980–1998. Pediatrics. 2002;109:656–660. 30. Coffelt TA, Bauer BD, Carroll AE. Inpatient characteristics of the child admitted with chronic pain. Pediatrics. 2013;132:e422–429. 31. Schurman JV, Hunter HL, Friesen CA. Conceptualization and treatment of chronic abdominal pain in pediatric gastroenterology practice. J Pediatr Gastroenterol Nutr. 2010;50:32–37. 32. Chogle A, Dhroove G, Sztainberg M, et al. How reliable are the Rome III criteria for the assessment of functional gastrointestinal disorders in children? Am J Gastroenterol. 2010;105:2697–2701. 33. Pensabene L, Buonomo C, Fishman L, et al. Lack of utility of abdominal x-rays in the evaluation of children with constipation: comparison of different scoring methods. J Pediatr Gastroenterol Nutr. 2010;51: 155–159. 34. Levy RL, Langer SL, Walker LS, et al. Cognitive-behavioral therapy for children with functional abdominal pain and their parents decreases pain and other symptoms. Am J Gastroenterol. 2010;105: 946–956. 35. Duarte MA, Penna FJ, Andrade EM, et al. Treatment of nonorganic recurrent abdominal pain: cognitive-behavioral family intervention. J Pediatr Gastroenterol Nutr. 2006;43:59–64.
ABSTRACT OBJECTIVE: To describe the characteristics of children with
diagnosis for their pain. Seventeen percent and 14% of children had laboratory and radiology testing at the index visit, respectively. Only 3% of laboratory evaluations helped to yield a diagnosis. Among variables considered, only preceding pain of more than 7 days at the index visit was associated with persistent pain (odds ratio 2.15, 95% confidence interval 1.19– 3.89). None of the variables considered was associated with receiving a specific diagnosis. CONCLUSIONS: Most children with AP do not receive a diagnosis, many have persistent pain, and very few receive a functional AP diagnosis. Results support limited use of diagnostic testing and conservative management consistent with national policy statements.
nonspecific abdominal pain (AP) in primary care, their evaluation, and their outcomes. METHODS: Between 2007 and 2009, a retrospective cohort of children from 5 primary care practices was followed from an index visit with AP until a well-child visit 6 to 24 months later (outcome visit). Using International Classification of Disease, 9th Revision (ICD-9), codes and chart review, we identified afebrile children between 4 and 12 years old with AP. Use of diagnostic testing was assessed. Multivariable logistic regression was used to model the association of index visit clinical and demographic variables with persistent pain at the outcome visit, and receipt of a specific diagnosis. RESULTS: Three hundred seventy-five children presented with AP, representing 1% of the total population of 4- to 12-year-olds during the study period. Eighteen percent of children had persistent pain, and 70% of the study cohort never received a specific
KEYWORDS: abdominal pain; functional gastrointestinal disorders; primary care ACADEMIC PEDIATRICS 2014;-:1–7
WHAT’S NEW
or recurrent pain. Chronic AP can cause significant impairment in children and increase the risk of functional disorders in adulthood if not adequately addressed.10,11 A growing body of literature has examined the care of children with functional AP, but data are based primarily on studies in subspecialty settings.12,13 Furthermore, clinicians have been criticized for the overly aggressive diagnostic evaluation and subspecialty referral of children with AP.14 Testing and subspecialty referrals for AP have also been reported as inconsistent15 and without an evidence base.16,17 The Rome III criteria were developed to help classify functional AP disorders and better guide management. Specifically, there are 5 pediatric AP-related FGIDs, including childhood functional AP, which is pain for at least 2 months without evidence of an inflammatory, anatomic, metabolic, or neoplastic process.18,19 Additionally, guidelines from the American Academy of Pediatrics (AAP) and the North American Society of Gastroenterology, Hepatology and Nutrition (NASPGHAN) support a conservative approach to children with chronic AP.20,21 Despite this, FGIDs are underdiagnosed, and it is unclear how frequently primary care physicians are using the Rome III criteria to characterize functional AP disorders. Multiple small studies of psychological treatments for FGIDs have shown efficacy in reducing pain and limiting functional
In this large study of children with nonspecific abdominal pain in primary care, we found persistent pain is common, and clinicians may underrecognize functional gastrointestinal disorders in this population of children.
ABDOMINAL PAIN (AP) is one of the most common complaints to pediatricians and accounts for 2% to 4% of primary care visits annually.1 Almost all children experience AP at some point, and the estimated prevalence of chronic AP in school children is 10% to 19%.2–4 The most common causes of AP in children include functional gastrointestinal disorders (FGIDs), constipation, and gastroenteritis, but many children present to primary care physicians with AP with no immediately identifiable cause. A few small studies have examined the diagnoses that children with AP have received when presenting to primary care and emergency departments and report that 15% to 35% receive a nonspecific AP diagnosis.5–8 Despite the prevalence of nonspecific AP, we lack longitudinal studies that examine the clinical course of children with AP beyond the initial visit, the evaluations these children receive, or their outcomes. In most children, AP is benign and self-limited.9 However, some children develop chronic ACADEMIC PEDIATRICS Copyright ª 2014 by Academic Pediatric Association
1
Volume -, Number -–- 2014
2
WALLIS AND FIKS
ACADEMIC PEDIATRICS
impairment,22–25 but these treatments are used infrequently, in part because primary care physicians may not be adequately classifying functional disorders and matching children to appropriate treatments. An improved understanding of the clinical course of children with AP is necessary in order to accurately diagnose children with FGIDs, avoid unnecessary testing, and match children with appropriate treatment. To help achieve this goal, we described the population of children presenting with AP in primary care, their management, and their clinical course.
METHODS STUDY POPULATION AND SETTING Eligible patients were drawn from 5 primary care practices from The Children’s Hospital of Philadelphia (CHOP) Pediatric Research Consortium (PeRC), a primary care practice–based research network.26,27 Children aged 4 to12 years at the time of study entry were identified using AP-related International Classification of Disease, 9th Revision (ICD-9), codes at an index visit between January 1, 2007, and December 31, 2009. We searched using ICD-9 codes for AP not otherwise specified and chronic AP (789.00 to 789.09). Because of our broad interest in children with AP, we also searched for children with the 5 AP-related FGID diagnoses at the index visit (functional dyspepsia, 536.8; FGID, 536.9; functional constipation, 564.09, irritable bowel syndrome, 564.1, abdominal migraine, 346.2) but found none; nor did we find any other functional digestiverelated codes (564). For simplicity, and consistent with prior work,5 we refer to this cohort of children in our study as having nonspecific AP. We focused on children ages 4 to 12 years because they are the group most commonly affected by AP.2 By manual chart review, we performed a 6-month look back to ensure that none of the cohort had a visit for AP in the preceding 6 months and that our data set accurately captured the index visit. To ensure that any follow-up was also captured, children included in the study cohort needed to have at least one well-child visit 6 to 24 months after the index visit (outcome visit). Well visits were chosen as the outcome visit because they include a detailed review of systems and pain assessment. In contrast, sick visits are variable in their content. Children with fever or a secondary diagnosis at the index visit that indicated a likely cause of the AP were excluded. The Figure outlines the eligible patient pool and the ICD-9 codes used to exclude patients to obtain the final study cohort. DATA COLLECTION Data were collected using chart review and supplemented by the automated extraction of data from electronic health records. Using a standardized abstraction form, clinical information was collected regarding children’s AP and associated symptoms; physical examination, laboratory, and radiology studies and results; and primary care follow-up. Chart review included visits 6 months before
Figure. Narrowing of practice population to study cohort.
the index visit, the index visit, and all subsequent care through June 30, 2012. Charts were reviewed by the research team and any ambiguous cases were resolved by consensus. Study data were collected and managed using REDCap electronic data capture tools hosted at The Children’s Hospital of Philadelphia.28 OUTCOME VARIABLES The 2 primary outcomes were persistent AP (defined as pain at the outcome visit) and receipt of a specific diagnosis for the AP at or before the outcome visit. We also evaluated the number of children in the cohort who went on to receive a diagnosis consistent with the Rome III criteria. The secondary outcomes were the frequency of laboratory and radiology evaluations at the index visit and their results (classified as normal, incidental finding, abnormal but no impact on clinical management, abnormal and contributed to diagnosis). We considered patients diagnosed with constipation via radiograph if the clinician’s documentation referenced the radiograph as diagnostic. INDEPENDENT VARIABLES Independent variables considered included patient age, race (black, white, other), insurance type (Medicaid, non-Medicaid), gender, and duration of pain at the index visit (<7 days, $7 days). Seven days was chosen as a clinically meaningful duration. To ensure that results were not biased on the basis of the specification of the exposure (pain duration), we conducted a sensitivity analysis including pain durations of 14 and 21 days as cut points. Additional covariates included associated symptoms, pain score at the visit (0, 1–10), findings of physical examination (diffuse pain, localized pain, normal abdominal examination findings), and the presence of complex chronic conditions as defined by Feudtner et al29 in any visit in the study window or in the child’s chronic problem
ACADEMIC PEDIATRICS
NONSPECIFIC ABDOMINAL PAIN
list. Missing data were rare (<5%) for all variables used in the analyses. STATISTICAL ANALYSES We characterized the population of children with AP using descriptive statistics and compared them to the overall population of children of the same age at the study practices. To avoid colinearity in multivariable models, the correlation between independent variables and among covariates was assessed. If 2 variables were highly correlated (r > 0.8), the one deemed most clinically meaningful was retained. Bivariate analysis was used to assess the association between each covariate and each outcome. In adjusted models, all covariates were retained if associated with the outcome at P < .3 (prespecified). Logistic regression models were then created to measure the association of independent variables with the outcomes of persistent pain (primary) and receipt of a specific diagnosis explaining the AP. In a secondary analysis, we described the frequency of laboratory and radiology evaluations and whether the results changed clinical management. All analyses were completed by Stata 12 software (StataCorp, College Station, TX). The CHOP institutional review board approved this study. Waiver of consent was granted.
RESULTS STUDY POPULATION Table 1 summarizes the demographic characteristics of the 375 children in the final study cohort compared to all children aged 4 to 12 years seen over the same time period. The study cohort represented 1% of the total population of children 4 to 12 years old. The overall population at 5 urban Table 1. Demographic Characteristics of Total and Study Populations* Characteristic N Race White Black Other Gender Female Male Payor Medicaid Non-Medicaid Age, y Mean 4–9 y 10–12 y Complex chronic condition Yes No
Study Population, n (%)
Total Population, n (%)
375
34202
P <.001
65 (17) 253 (68) 57 (15)
4017 (12) 26943 (79) 3242 (9) .018
207 (55) 168 (45)
16779 (49) 17423 (51)
199 (53) 176 (47)
21070 (62) 13132 (38)
7.9 256 (68) 119 (32)
8.2 20942 (61) 13260 (39)
.018
.005
.11 69 (19) 292 (81)
5466 (16) 28736 (84)
*Total population is all patients aged 4 to 12 years, seen in the same time period in the 5 practices from which the study cohort was derived.
3
study practices was varied, with 79% black and 62% Medicaid insured. Children evaluated with AP were more likely to be girls (55%, P ¼ .018), non-Medicaid insured (53%, P ¼ .018), and white (17%, P < .001) than the larger population. Children with AP and those without had similar rates of complex chronic conditions (19% vs 16% P ¼ .11). Table 2 summarizes the clinical characteristics of children with AP in the study cohort. Of note, all the patients in our cohort had an AP not otherwise specified code at the index visit, and there were none with the other diagnoses searched. Among children with AP, 42% reported a history of diffuse pain and 28% had localized pain (data not shown), though only 34% of children had abdominal tenderness at physical examination. Almost three quarters (72%) reported intermittent pain. Children presented with a median pain duration of 5 days (range 1–1095 days), with 42% presenting with pain for >7 days and 23% presenting with pain for >30 days. The most common associated symptoms at presentation were vomiting (26% of the cohort), decreased appetite (18%), nausea (17%), diarrhea (13%), difficulty stooling (13%), and headache (12%). More than one quarter of children (26%) reported no associated symptoms. PERSISTENT AP AND RECEIPT OF A SPECIFIC DIAGNOSIS Persistent AP was common in our sample. Eighteen percent of children in the cohort reported persistent pain at the outcome visit. In unadjusted analyses, patients with Table 2. Abdominal Pain Characteristics at Index Visit* Clinical Characteristic Pain duration at visit 1 #7 d >7 d Tenderness at physical examination Localized tenderness Diffuse tenderness No tenderness Pain pattern Constant Intermittent Missing/not reported Associated symptoms† Vomiting Decreased appetite Nausea Difficulty stooling Diarrhea Headache Dysuria/urinary symptoms Cough Congestion/rhinorrhea Chest pain Fatigue Weight loss Joint pain Skin changes Bloody stools (melena) None *Pain characteristics are patient or family reported. †Patients may report >1 associated symptom.
n (%) 218 (58) 157 (42) 93 (25) 32 (9) 239 (66) 31 (8) 270 (72) 74 (20) 97 (26) 67 (18) 65 (17) 47 (13) 47 (13) 45 (12) 30 (8) 30 (8) 28 (8) 15 (4) 12 (3) 9 (2) 9 (2) 7 (2) 4 (1) 96 (26)
4
WALLIS AND FIKS
ACADEMIC PEDIATRICS
persistent pain at the outcome visit were more likely to present at the index visit with pain for >7 days preceding the visit (odds ratio [OR] 2.33, 95% confidence interval [CI] 1.36–3.97, P ¼ .002). In adjusted models, the association with pain duration was maintained (OR 2.15, 95% CI 1.19–3.89, P ¼ .01). In our sensitivity analyses, with pain duration of either 14 or 21 days as the cut points, results were similar. No other clinical or demographic characteristics were associated with persistent pain in either adjusted or unadjusted analyses (Table 3).
Table 3. Clinical and Demographic Characteristics and Their Associations With Persistent Pain*
Characteristic Payer Medicaid Non-Medicaid Gender Female Male Age 4–9 y 10–12 y Race Black White Other Duration of pain at presentation (index visit) <7 d $7 d Abdominal pain location at physical examination Localized pain Diffuse pain No pain Pain score at index visit† 0 1–10 Complex chronic condition‡,§ No Yes Laboratory studies completed at index visit No Yes Radiology studies completed at index visit No Yes
n
Unadjusted Proportion
Adjusted Analyses OR (95% CI)
In the majority of children (70%) presenting with AP, no specific cause of the AP was determined. In addition, no clinical pain characteristics or demographic variables considered were significantly associated with receipt of a specific diagnosis for the pain (Table 4). Adjusted analysis of radiograph testing was associated with receipt of a diagnosis but not statistically significant (P ¼ .06). This represents cases in which constipation was diagnosed using abdominal radiographs. Of the 30% of children who received a diagnosis (n ¼ 113), the most common diagnoses were constipation (57%), functional AP (16%), and gastroesophageal reflux disease (5%). Among the 262 Table 4. Clinical and Demographic Characteristics and Their Associations With Receipt of a Specific Diagnosis for Abdominal Pain*
P Characteristic
199 176
0.49 0.51
207 168
0.64 0.36
256 119
0.72 0.28
253 65 57
0.66 0.19 0.15
218 157
0.40 0.60
Ref 2.15 (1.19–3.89) .01
93 32 239
0.19 0.17 0.64
Ref 0.51 (0.18–1.42) .20 0.46 (0.18–1.19) .11
171 194
292 69
313 62
324 51
0.52 0.48
0.75 0.25
Ref 0.87 (0.49–1.56) .64
Ref 0.95 (0.87–1.05) .35
Ref 1.44 (0.73–2.87) .30
0.85 0.15
0.87 0.13
CI indicates confidence interval; OR, odds ratio. *Persistent pain defined as abdominal pain still present at a well-child visit 6 months—2 years after the index visit for abdominal pain. †Ten patient records had no pain scores recorded. ‡As defined by Feudtner et al.29 §Data missing for 14 patients (4%).
Payer Medicaid Non-Medicaid Gender Female Male Age 4–9 y 10–12 y Race Black White Other Duration of pain at presentation† <7 d $7 d Abdominal pain location‡,§ Localized pain Diffuse pain No pain Pain scorejj 0 1–10 Complex chronic condition No Yes Laboratory studies completed at index visit No Yes Radiology studies completed at index visit No Yes
Adjusted Analyses
n
Unadjusted Proportion
OR (95% CI)
P
199 176
0.48 0.52
Ref 1.28 (0.80–2.06)
.31
207 168
0.54 0.46
256 119
0.70 0.30
253 65 57
0.62 0.18 0.20
Ref 0.92 (0.55–1.95) 2.16 (0.98–4.75)
.92 .06
218 157
0.53 0.47
93 32 239
0.33 0.07 0.60
171 194
0.49 0.51
292 69
0.75 0.25
Ref 1.72 (0.98–3.00)
.06
313 62
0.85 0.15
324 51
0.82 0.18
Ref 1.85 (0.99–3.45)
.06
CI indicates confidence interval; OR, odds ratio. *Receipt of a specific diagnosis defined using chart review as a diagnosis that adequately explains the etiology of previously defined abdominal pain. †Reported by child/family. ‡At provider physical examination. §A small portion of records (n ¼ 11, 3% of total) were missing any mention of the abdominal physical examination findings. jjTen patient records had no pain scores recorded.
ACADEMIC PEDIATRICS
children with no specific diagnosis, only 12% had any mention of an AP-related functional disorder as part of a differential diagnosis, though 21% of this group of children met one or more criteria for one of the AP-related functional disorders (Rome III criteria) at the index visit. Very few children presenting with afebrile AP were subsequently diagnosed with an emergent or potentially life-threatening diagnosis. In our cohort of 375 children, 3 children were diagnosed with inflammatory bowel disease, 2 with appendicitis, 1 with gall bladder disease requiring cholecystectomy, 1 with sickle cell pain causing AP, and 2 with pancreatitis. In all of these cases, these children presented with red flag–associated symptoms or physical findings based on the AAP policy statement on the evaluation and management of children with chronic AP.11 These symptoms and physical findings include bloody stools, significant weight loss, family history of inflammatory bowel disease, or peritoneal signs at physical examination. All of these children were recognized as having red flag symptoms by their pediatricians and were diagnosed shortly after the index visit. LABORATORY AND RADIOLOGY EVALUATION Although >10% of children had laboratory or radiology testing at the index visit (17% and 14%, respectively), these evaluations rarely changed clinical management or contributed to a specific diagnosis. The majority of laboratory testing was normal (61%). In only 3% of cases did testing help to determine the cause of a child’s AP. Much more commonly, incidental findings (5%) or abnormal findings requiring follow-up (31%) were reported. Examples of incidental findings included a mild electrolyte abnormality on a basic metabolic panel, while abnormal findings requiring follow-up included abnormal autoimmune panels that did not yield an autoimmune diagnosis. A similar pattern was noted in radiology evaluations, with the majority (66%) having normal results. Although 20% (n ¼ 9) of radiology studies aided physicians in diagnosing a patient’s AP, the only diagnosis made via radiology evaluation was constipation. The demographic and clinical characteristics of children who underwent laboratory or radiology testing were similar to children who did not (P > .15 for all variables). Pain for more than 7 days at presentation was associated with an increased likelihood of laboratory (OR 3.05, 95% CI 1.73–5.38, P ¼ .0001) and radiology testing (OR 2.43, 95% CI 1.33–4.46, P ¼.004), and children without abdominal tenderness at physical examination were less likely to receive radiology imaging (OR 0.38, 95% CI 0.37–2.59, P ¼ .04) than those with either diffuse or localized pain. Other variables considered were not associated with testing. Laboratory testing was not associated with persistent pain or receipt of a specific diagnosis. Radiology testing was not associated with persistent pain, but clinicians used x-rays to diagnose constipation-related pain (P ¼ .06 for the comparison of whether or not radiology studies were ordered at the index and the association with receipt of a specific AP diagnosis).
NONSPECIFIC ABDOMINAL PAIN
5
DISCUSSION The literature lacks longitudinal studies of children with AP in pediatric primary care. In our study, we found that nearly one fifth of children in our cohort of 375 children diagnosed with AP had persistent pain at a well-child visit 6 to 24 months after the index visit. Children presenting with pain for >7 days were more likely to have persistent AP but were not more likely to receive a specific diagnosis for their AP. Less than one third of all children received a specific diagnosis, and while many patients with AP may have FGIDs, this was uncommonly documented as a potential diagnosis. Diagnostic codes for FGIDs were never used in this sample. When used, laboratory and radiology testing rarely changed clinical management. Studies of hospitalized children and those referred to gastroenterologists with functional abdominal complaints have documented a similar demographic distribution to our study, with those affected more likely to be girls, to have non-Medicaid insurance, and to be white.7,30 In total, AP complaints account for an estimated 2% to 4% of primary care visits annually.1 Consistent with our results, prior studies have estimated a significant proportion of these to be nonspecific or functional complaints.23 We found that persistent AP was common among those presenting with AP. Eighteen percent of children in our cohort with AP had persistent pain at the outcome visit, and children with pain for >7 days at the index visit were more than 2 times more likely to have persistent AP. To our knowledge, no previous study has recognized this association. Our findings may help primary care clinicians identify children who are likely to develop a FGID. In the majority of children who presented with AP, a cause for their AP was not determined despite persistent pain. These results support a high index of suspicion for a functional diagnosis among children with AP. Given the high prevalence of persistent AP, many of the children in our cohort may have an AP-related functional disorder, although only 5% were identified as such. From previous studies, it is known that pediatricians often fail to recognize functional disorders,11,31 likely accounting for the low number of children diagnosed in our cohort. Although clinical care does not always mirror diagnostic coding, both the clinical decision making in reviewed charts and the ICD-9 coding did not reflect consideration of a functional diagnosis. The Rome III criteria provide specific diagnostic classifications and an algorithm to assist in making a functional AP diagnosis.32 However, even gastroenterologists have reported inconsistent knowledge and use of the Rome III criteria,31 and although the criteria have been validated, their clinical usefulness has not been well examined in primary care or subspecialty practices. Our results indicate that primary care clinicians may not be using these criteria for children likely to have a functional disorder. These results suggest that work is needed to support the implementation of the Rome III criteria and the AAP policy statement on chronic AP into primary care settings. Clinician training, feedback, and decision support, ideally delivered as part
6
WALLIS AND FIKS
of a quality improvement initiative, may facilitate this process. The AAP policy statement on chronic AP supports judicious testing in this group of children, and our results strengthen the evidence base for these recommendations.20 Specifically, the AAP policy statement recommends laboratory or radiology studies only in children with one or more red flags raised at the history or physical examination. Consistent with these recommendations, we found that testing rarely changed clinical management and that children who underwent testing were not more likely to receive a specific diagnosis. Further restriction of radiographic testing may be possible. The use of abdominal radiograph to diagnose constipation, the major clinical contribution of radiograph use in our study, is not generally recommended; providers should instead rely on a stooling history and physical examination.33 Additionally, all of the children in our cohort with an emergent or potentially lifethreatening diagnosis had one or more red flag findings at the index visit. Several studies on children suspected of having functional AP have identified reassurance of the family as a motivator for physicians ordering tests.15 Given evidence that testing does not provide the reassurance intended,15 and with the incidence of hospitalization for chronic pain rising,30 it is important that primary care providers limit unnecessary testing and match children with evidence-based treatments. Our findings including the prevalence of persistent pain underscore the need for further work to identify bestpractice behavioral health interventions and integrate these interventions into clinical practice. Effective psychological treatments have been identified for children with persistent functional symptoms and impairment in daily activities. Psychological treatments including cognitive behavioral therapy, family therapy, biofeedback, gut-directed hypnotherapy and psychoeducational therapies have been piloted in referred patients, with improvements in pain symptoms and functional impairment in children, and improved coping by parents and children.10,22,34,35 However, most behavioral health interventions for functional AP have yet to be introduced into primary care practice and are not widely available. This study had several limitations. Although children treated in multiple primary care sites were studied, all the sites were part of one larger health system. In addition, our study sample was predominantly urban and minority patients, and further study is needed to assess how these results generalize to other health settings and populations. It is also possible that the practices studied, as part of an academic medical center, may have different familiarity with the Rome criteria than community practitioners. Our study was limited to information accessible in the medical record. We were unable to assess the interaction between the clinician and family, which may have influenced testing choice or treatment plan, as it was beyond the scope of this study. We were also not able to assess factors related to when a family seeks care (eg, pain tolerance and parental anxiety). Variables including social stressors, family history of mental health disorders or chronic pain syndromes,
ACADEMIC PEDIATRICS
missed school, and other psychological factors that were not consistently documented may have helped to explain our findings. It is also possible that patients received care outside of the health system and receipt of this care was not captured in the study. In addition, outcomes such as functional impairment could not be assessed in this retrospective review. A prospective study that captures these variables as well as outcomes including functional impairment for children with AP treated in primary care is warranted.
CONCLUSIONS In this longitudinal study of children with AP in primary care, we found that persistent pain is common, but few children are given a functional diagnosis (using ICD-9 codes for FGIDs or based on clinician documentation of their assessment and differential diagnosis). Our results support the AAP policy statement on chronic AP and indicate that laboratory and radiology studies are most often normal and rarely change clinical management. Additional training is needed to improve recognition of FGIDs by providers in order to better manage children with persistent pain and avoid unnecessary testing.
ACKNOWLEDGMENTS We would like to thank James P. Guevara, MD, Robert W. Grundmeier, MD, and William T. Basco, MD, for their thoughtful review. We would like to thank the network of primary care clinicians and their patients and families for their contributions to clinical research through the Pediatric Research Consortium at CHOP. Data were managed using REDCap (Research Electronic Data Capture), a secure, Web-based application designed to support data capture for research studies.28 Supported by training grant HRSA T32 HP10026 from the National Institutes of Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
REFERENCES 1. Starfield B, Hoekelman RA, McCormick M, et al. Who provides health care to children and adolescents in the United States? Pediatrics. 1984;74:991–997. 2. Apley J, Naish N. Recurrent abdominal pains: a field survey of 1,000 school children. Arch Dis Child. 1958;33:165–170. 3. Schwille IJ, Giel KE, Ellert U, Zipfel S, Enck P. A community-based survey of abdominal pain prevalence, characteristics, and health care use among children. Clin Gastroenterol Hepatol. 2009;7:1062–1068. 4. Saps M, Seshadri R, Sztainberg M, Schaffer G, Marshall BM, Di Lorenzo C. A prospective school-based study of abdominal pain and other common somatic complaints in children. J Pediatr. 2009; 154:322–326. 5. Gieteling MJ, Lisman-van Leeuwen Y, van der Wouden JC, et al. Childhood nonspecific abdominal pain in family practice: incidence, associated factors, and management. Ann Fam Med. 2011;9:337–343. 6. Reynolds SL, Jaffe DM. Diagnosing abdominal pain in a pediatric emergency department. Pediatr Emerg Care. 1992;8:126–128. 7. Loening-Baucke V, Swidsinski A. Constipation as cause of acute abdominal pain in children. J Pediatr. 2007;151:666–669. 8. Scholer SJ, Pituch K, Orr DP, Dittus RS. Clinical outcomes of children with acute abdominal pain. Pediatrics. 1996;98(4 pt 1):680–685. 9. Chitkara DK, Rawat DJ, Talley NJ. The epidemiology of childhood recurrent abdominal pain in Western countries: a systematic review. Am J Gastroenterol. 2005;100:1868–1875.
ACADEMIC PEDIATRICS 10. Wendland M, Jackson Y, Stokes LD. Functional disability in paediatric patients with recurrent abdominal pain. Child Care Health Dev. 2010;36:516–523. 11. Walker LS, Dengler-Crish CM, Rippel S, Bruehl S. Functional abdominal pain in childhood and adolescence increases risk for chronic pain in adulthood. Pain. 2010;150:568–572. 12. Lisman-van Leeuwen Y, Spee LA, Benninga MA, et al. Prognosis of abdominal pain in children in primary care—a prospective cohort study. Ann Fam Med. 2013;11:238–244. 13. Gieteling MJ, Bierma-Zeinstra SM, Passchier J, Berger MY. Prognosis of chronic or recurrent abdominal pain in children. J Pediatr Gastroenterol Nutr. 2008;47:316–326. 14. Miele E, Simeone D, Marino A, et al. Functional gastrointestinal disorders in children: an Italian prospective survey. Pediatrics. 2004; 114:73–78. 15. Bonilla S, Deli W, Saps M. The prognostic value of obtaining a negative endoscopy in children with functional gastrointestinal disorders. Clin Pediatr (Phila). 2011;50:396–401. 16. Lane MM, Weidler EM, Czyzewski DI, Shulman RJ. Pain symptoms and stooling patterns do not drive diagnostic costs for children with functional abdominal pain and irritable bowel syndrome in primary or tertiary care. Pediatrics. 2009;123:758–764. 17. van Tilburg MA, Venepalli N, et al. Parents’ worries about recurrent abdominal pain in children. Gastroenterol Nurs. 2006;29:50–55. 18. Baber KF, Anderson J, Puzanovova M, Walker LS. Rome II versus Rome III classification of functional gastrointestinal disorders in pediatric chronic abdominal pain. J Pediatr Gastroenterol Nutr. 2008; 47:299–302. 19. Drossman DA, Dumitrascu DL. Rome III: new standard for functional gastrointestinal disorders. J Gastrointestin Liver Dis. 2006;15: 237–241. 20. Di Lorenzo C, Colletti RB, Lehmann HP, et al. Chronic abdominal pain in children: a clinical report of the American Academy of Pediatrics and the North American Society for Pediatric Gastroenterology, Hepatology and Nutrition. J Pediatr Gastroenterol Nutr. 2005;40: 245–248. 21. Chronic abdominal pain in children. Pediatrics. 2005;115: e370–e381. 22. Eccleston C, Palermo TM, Williams AC, et al. Psychological therapies for the management of chronic and recurrent pain in children and adolescents. Cochrane Database Syst Rev. 2009;CD003968.
NONSPECIFIC ABDOMINAL PAIN
7
23. von Baeyer CL. Understanding and managing children’s recurrent pain in primary care: a biopsychosocial perspective. Paediatr Child Health. 2007;12:121–125. 24. Levy RL, Langer SL, Walker LS, et al. Twelve-month follow-up of cognitive behavioral therapy for children with functional abdominal pain. JAMA Pediatr. 2013;167:178–184. 25. Huertas-Ceballos A, Logan S, Bennett C, Macarthur C. Psychosocial interventions for recurrent abdominal pain (RAP) and irritable bowel syndrome (IBS) in childhood. Cochrane Database Syst Rev. 2008;CD003014. 26. Fiks AG, Grundmeier RW, Margolis B, et al. Comparative effectiveness research using the electronic medical record: an emerging area of investigation in pediatric primary care. J Pediatr. 2012;160:719–724. 27. Grundmeier RW, Swietlik M, Bell LM. Research subject enrollment by primary care pediatricians using an electronic health record. AMIA Annu Symp Proc. 2007;289–293. 28. Harris PA, Taylor R, Thielke R, et al. Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42:377–381. 29. Feudtner C, Silveira MJ, Christakis DA. Where do children with complex chronic conditions die? Patterns in Washington State, 1980–1998. Pediatrics. 2002;109:656–660. 30. Coffelt TA, Bauer BD, Carroll AE. Inpatient characteristics of the child admitted with chronic pain. Pediatrics. 2013;132:e422–429. 31. Schurman JV, Hunter HL, Friesen CA. Conceptualization and treatment of chronic abdominal pain in pediatric gastroenterology practice. J Pediatr Gastroenterol Nutr. 2010;50:32–37. 32. Chogle A, Dhroove G, Sztainberg M, et al. How reliable are the Rome III criteria for the assessment of functional gastrointestinal disorders in children? Am J Gastroenterol. 2010;105:2697–2701. 33. Pensabene L, Buonomo C, Fishman L, et al. Lack of utility of abdominal x-rays in the evaluation of children with constipation: comparison of different scoring methods. J Pediatr Gastroenterol Nutr. 2010;51: 155–159. 34. Levy RL, Langer SL, Walker LS, et al. Cognitive-behavioral therapy for children with functional abdominal pain and their parents decreases pain and other symptoms. Am J Gastroenterol. 2010;105: 946–956. 35. Duarte MA, Penna FJ, Andrade EM, et al. Treatment of nonorganic recurrent abdominal pain: cognitive-behavioral family intervention. J Pediatr Gastroenterol Nutr. 2006;43:59–64.