OBJECTIVE: To define the demographic, diagnostic, procedural, and episode of care characteristics for children admitted with chronic pain.
METHODS: We used the Pediatric Health Information System database to obtain data on demographic characteristics, length of stay, readmission rates, diagnoses, and procedures for children admitted with chronic pain. Patients with sickle cell disease, cancer, burns, cerebral palsy, transplants, and ventilator-dependent children were excluded.
RESULTS: A total of 3752 patients with chronic pain were identified from 2004 through 2010. Admissions increased by 831% over this time period. The mean age of these patients was 13.5 years, the most common race was white (79%), and female subjects outnumbered male subjects by 2.41 to 1. The most common admission and principal discharge diagnosis was abdominal pain; comorbid diagnoses were common, with a mean of 10 diagnoses per patient. In total, 65% of patients had a comorbid gastrointestinal diagnosis and 44% had a psychiatric diagnosis. The mean length of stay was 7.32 days, with an expected length of stay of 4.24 days; 12.5% were readmitted at least once within 1 year. They underwent a mean of 3.18 procedures per patient.
CONCLUSIONS: The average child admitted with chronic pain is a teenaged female with a wide variety of comorbid conditions, many of which are gastrointestinal and psychiatric in nature. Admissions for chronic pain are rising and account for substantial resource utilization. Future studies should further characterize this population, with the overall objective of improving outcomes and optimizing cost-effective care.
- CRPS —
- complex regional pain syndrome
- ED —
- emergency department
- ICD-9 —
- International Classification of Diseases, Ninth Revision
- NEC —
- not elsewhere classified
- NOS —
- not otherwise specified
- PHIS —
- Pediatric Health Information System
- RSD —
- reflex sympathetic dystrophy
What’s Known on This Subject:
Children with chronic pain complaints seem to represent an increasing portion of general pediatric inpatient services. Few data exist, however, on the characteristics of this population, their length of stay, or the best approach to their evaluation and management.
What This Study Adds:
This study defines the demographic, diagnostic, procedural, and episode of care characteristics for children admitted to the acute care setting with chronic pain syndromes. Admission rates are rising, lengths of stay are substantial, and comorbid diagnoses are common.
Chronic pain is common in the adult population and is often treated in the outpatient setting.1 Increasing numbers of children, however, seem to be presenting both to hospitals and outpatient clinics with complaints of recurrent pain.2–10 Many conditions with an organic etiology such as cancer and sickle cell disease can lead to chronic pain states. However, there are also many patients who present with vague or ill-defined somatic complaints, such as recurrent abdominal pain, which do not correspond with an easily identifiable underlying organic etiology or syndrome. Due to a multitude of poorly understood factors, some of these chronic pain states may persist into adulthood, thus creating the potential for a lifelong impact on patient health and well-being.2,11–13 Many of these patients are admitted either for diagnostic evaluation or for management of their pain when it cannot be controlled in the emergency department (ED) or outpatient setting.10,14 Initial evaluation of these patients often results in an extensive and costly diagnostic evaluation in an attempt to identify an underlying organic etiology. The wide variety of presenting symptoms and complaints, as well as a lack of evidence or standards in practice, has led to significant variation among providers and institutions with regard to their diagnostic approach.5,15–19 These factors, in turn, have resulted in increased resource utilization, health care costs, and missed school days, especially among adolescents.10,20–24 The potential long-term impact on patients with chronic pain is not insignificant. Chronic pain in children has been associated with poor social functioning and peer relationships.25 Few studies exist, however, describing the actual incidence, demographic characteristics of, treatment approach to, and length of stay for children admitted with chronic pain syndromes, especially in the acute care setting. Understanding these characteristics in the inpatient setting should help guide future research, which will lead to better treatment and outcomes for these patients.
The current study was undertaken to define the demographic, diagnostic, procedural, and episode of care characteristics for children admitted with chronic pain who may be cared for on a general pediatric inpatient service. We sought to capture a population of children who presented with complaints of recurrent or chronic pain (eg, recurrent abdominal pain) for which no clear organic etiology was apparent.
Data for this retrospective cohort study were obtained from the Pediatric Health Information System (PHIS), an administrative database that contains inpatient, ED, ambulatory surgery and observation data from 43 not-for-profit, tertiary care pediatric hospitals in the United States. These hospitals are affiliated with the Children’s Hospital Association (Overland Park, KS). Data quality and reliability are assured through a joint effort between the Children’s Hospital Association and participating hospitals. The data warehouse function for the PHIS database is managed by Truven Health Analytics (Ann Arbor, MI). For the purposes of external benchmarking, participating hospitals provide discharge/encounter data, including demographic characteristics, diagnoses, and procedures. Forty-two of these hospitals also submit resource utilization data (eg, pharmaceuticals, imaging, laboratory) into PHIS. Data are de-identified at the time of data submission and are subjected to a number of reliability and validity checks before being included in the database. For the current study, data from all 43 hospitals were included where available. The hospitals represented in the PHIS database are geographically spread across the United States, thus providing a representative sample of inpatient chronic pain patients nationally.
Patients between the ages of 0 and 18 years at the time of admission were included. Chronic pain was defined by using the following diagnoses: psychogenic pain not otherwise specified (NOS) (International Classification of Diseases, Ninth Revision [ICD-9], code 307.80), psychogenic pain not elsewhere classified (NEC) (307.89), reflex sympathetic dystrophy (RSD) NOS (337.20), RSD upper limb/arm (337.21), RSD lower limb/leg (337.22), RSD NEC (337.29), chronic pain NEC (338.29), chronic pain syndrome (338.4), complex regional pain syndrome (CRPS)/causalgia of upper limb (354.4), or CRPS/causalgia lower limb (355.71). Patients who were assigned any of these diagnosis codes during their hospitalization were included. Unfortunately, there is, as yet, no specific ICD-9 code for fibromyalgia. Although the more generic code of myalgia and myositis (729.1) may be used to identify patients with this diagnosis, the lack of specificity in identifying patients with fibromyalgia as a cause of chronic pain prevented us from including it as a principal inclusion criterion for the purposes of this study. Similarly, because there are no specific codes to define chronic abdominal pain, only patients who were assigned one of the aforementioned chronic pain codes were included in the study.
Patients with diagnoses related to chronic pain who may not be treated on a general pediatric inpatient unit or those with underlying organic conditions known to cause recurrent pain were excluded. Thus, we excluded patients with a diagnosis of cancer, sickle cell disease, burns, or cerebral palsy, as well as transplant patients and ventilator-dependent children. Patients with NICU charges were also excluded.
Protocol and Data Collection
This study was approved by the Indiana University School of Medicine institutional review board. Using the PHIS database, we collected information on pediatric chronic pain inpatients discharged between January 1, 2004, and December 31, 2010. All information collected was de-identified. The followed data were evaluated: age, race, ethnicity, gender, source of admission, disposition, length of stay, admission diagnosis, principal diagnosis, secondary diagnoses, principal and secondary procedures, attending and procedural physician specialties, and readmission rates within 365 days. Admission, discharge, and secondary diagnoses as well as procedures were then ranked according to total instance count. Mean number of diagnoses and procedures per patient were determined by using the total diagnosis and procedure counts divided by the total number of patients with chronic pain diagnoses. The readmission rate was determined by taking the total number of readmissions in the chronic pain cohort and dividing it by the total number of chronic pain patients. The total number of chronic pain patients, as defined in this study, was then divided by all other patients in the database aged 0 to 18 years, including those with pain due to underlying medical conditions such as sickle cell disease but excluding NICU patients, to determine the overall incidence of the chronic pain cohort under study.
Table 1 displays a summary of the results. A total of 3752 chronic pain patients were identified, accounting for 0.16% of the overall patient population in the database over the study period. Younger children (0–7 years old) only comprised 235 patients of the cohort. The number of patients with chronic pain diagnoses increased by 831% from 2004 to 2010 (Fig 1). The average child admitted with chronic pain was a 13.54-year-old (median: 14 years) white female with an admission diagnosis of abdominal pain–site NOS. Other nonspecific diagnoses such as headache, backache, or limb pain were also common reasons for admission. Female subjects outnumbered male subjects by 2.41 to 1. White/not Hispanic or Latino race accounted for 79% of chronic pain patients; white/Hispanic or Latino race comprised 9.2% and black race accounted for 6.7%. Figure 2 displays the distribution of race and ethnicity. The demographic information for inpatients with chronic pain seems to vary from the PHIS population as a whole; the mean age of the database population was 9 years (median: 4 years), and male subjects outnumbered female subjects by 1.23 to 1. White/not Hispanic or Latino race comprised 62% of the overall population, white/Hispanic or Latino race 19%, and black race 12%.
Among chronic pain patients, one-half were admitted from the ED compared with 59% of the overall population in the database. In this cohort of chronic pain patients, secondary diagnoses were common, with a mean of 10 total diagnoses per patient. The most common secondary diagnoses were abdominal pain (39% of patients), mood disorders (28%), constipation (20%), nausea/vomiting/diarrhea (18%), anxiety and panic disorders (18%), and headaches (18%). In total, 65% of patients had a gastrointestinal diagnosis, and 44% had a psychiatric diagnosis. Notably, 9.9% of patients in the cohort had a coded medication adverse effect, overdose, or substance abuse. The most common specific diagnoses in this category included adverse effects of opiates occurring in 1.4% of patients, tobacco use in 1%, cannabis abuse in 0.9%, and adverse effects of corticosteroids in 0.8%. Table 2 lists the most common secondary diagnoses in each diagnostic category. At discharge, the most common principal diagnosis was abdominal pain, occurring in 15% of patients, followed by RSD (14%), chronic pain/chronic pain syndrome (13%), headaches (5.5%), and constipation (1.9%). Ninety-three percent were discharged to home, 2.3% to home with skilled nursing or home health care, 0.4% to an inpatient psychiatric unit, 0.3% left against medical advice, and 0.1% were transferred to inpatient rehabilitation.
Patients had a mean length of stay of 7.32 days (median: 4 days); the expected length of stay was 4.24 days (median: 2.81 days) based on All Patient Refined Diagnosis Related Group severity scores. Length of stay has remained relatively constant from 2004 to 2010 but is still higher than national expected averages (Fig 3). In total, 12.5% were readmitted within 1 year of their initial discharge: 9.9% at least once within the 12-month period and 2.6% more than once. Pediatric inpatients with chronic pain underwent a mean of 3.18 procedures per patient, the most common of which was an esophagogastroduodenoscopy with closed biopsy (18%), followed by closed biopsy of the large intestines (11.7%) and insertion of a spinal canal catheter (11.6%). The most common attending physician specialty was pediatrics followed by gastroenterology; the most common procedural physician specialty was gastroenterology followed by anesthesiology.
With an overall prevalence of 16 per 1000 patients admitted and the number of admissions rising annually, chronic pain is becoming a more common reason for admission to the pediatric inpatient setting. Children admitted to the inpatient setting with chronic pain are most commonly white adolescent females who have a wide variety of comorbid conditions, many of which are gastrointestinal, psychiatric, and musculoskeletal in nature. Abdominal pain and headaches occur in a substantial portion of these patients. Anxiety and depression were also frequent comorbid diagnoses. Interestingly, asthma was also a common comorbid diagnosis, but our finding of 17% prevalence is similar to that found in the overall inpatient pediatric population, in which asthma occurred in 13% of admitted children over the study time period.
Outpatient estimates of the prevalence of chronic pain vary widely, from 10% to 46% depending on the definition of chronic pain used and the setting.1,17,24,26–28 Inpatient rates are undoubtedly lower, but we believe our finding of a 0.16% prevalence is a gross underestimate because it is almost certain that many children with these conditions are not assigned a chronic pain diagnosis and are instead coded for nonspecific complaints. Indeed, part of what the current study demonstrates is that patients with chronic pain often have many nonspecific symptoms without an underlying causative diagnosis. The findings in our study are generally consistent with those of previously published outpatient studies, which largely agree that the typical chronic pain patient is an adolescent female between 11 and 15 years old with headaches, recurrent abdominal and musculoskeletal pains, as well as depression and anxiety, suggesting that the inpatient chronic pain population is similar to their outpatient cohorts.6,16,26,27,29,30 This is the only study, however, which demonstrates a relatively high degree of adverse medication effects, likely owing to the types of treatments received. In addition, substance abuse and overdose occurred not infrequently in this cohort and many of the treatments these patients receive, such as opiates, have the potential for abuse and addiction. A history of physical/psychological/sexual abuse and family history of psychiatric conditions were also seen. Clearly, the psychomorbidity in this population is not insignificant and likely contributes to the difficulties in treating these patients. It may also add to the number of diagnostic evaluations performed and increased lengths of stay.
To our knowledge, this is the largest study that has focused on pediatric chronic pain in the inpatient setting and adds to the scant evidence on demographic characteristics and lengths of stay from other studies which have included inpatient data. In 1 study, 6.4% of a cohort of children with chronic pain were admitted, with a mean hospital stay of 4.9 days.10 In another study, 90% of children presenting to a pediatric pain clinic with CRPS were females with a mean age of 11.8 years; 40% of these patients required treatment in the inpatient setting.31 Maynard et al32 included demographic and diagnostic characteristics of 41 children admitted to 1 inpatient pain rehabilitation program. The majority of patients were white (85.4%) female (73.2%) subjects with a mean age of 13.8 years. Forty-six percent had CRPS, 22% had headaches, 20% had abdominal pain, and 10% had functional gait disturbances and traumatic brain injuries. Comorbid psychiatric conditions included 41% with anxiety, 34% with depression, 22% with somatoform disorders, and 7.3% with cluster B traits. The mean length of stay was 27.1 days; however, this study focused specifically on patients in an inpatient rehabilitation unit rather than an acute care pediatric hospital. Our study adds to the existing data that psychiatric comorbidities and racial differences occur in the pediatric chronic pain population. However, our review of the available data used for this study does not provide a clear explanation for why these disparities exist.
Our study has some limitations. Third-party databases such as PHIS are limited by trying to identify certain patient populations with the use of ICD-9 codes, which may not accurately reflect the intended population for study. For example, there are very few ICD-9 codes related to chronic pain states. Fibromyalgia is 1 example in which there is no ICD-9 code. These patients are frequently coded as having myalgia and myositis, but so are patients who present with a simple muscle ache who may not have fibromyalgia; therefore, this condition was not used as a principal inclusion criterion. Because it is rare that a patient will be admitted or discharged with 1 of the chronic pain codes used in this study as their principal diagnosis, we included patients who carried a secondary diagnosis of chronic pain. As such, it is difficult to capture those patients who were admitted solely for chronic pain as they are often only assigned a chronic pain code secondarily while their primary code is often their presenting symptom. For example, a patient may have an admission diagnosis of abdominal pain, discharged with a principal diagnosis of abdominal pain, and be given secondary diagnoses of nausea/vomiting, constipation, and chronic pain; this may be confusing when interpreting which is a comorbid condition and which is a primary condition. However, we believe that the intended population for study was appropriately represented in this cohort because there were virtually no patients with underlying medical disorders such as Crohn disease, which may account for their presenting complaint or discharge diagnosis. Patients with chronic pain often have many symptoms, rather than an underlying medical disorder to account for these symptoms, as can be seen by the multitude of nonspecific secondary diagnoses listed in Table 2. Our intention was simply to outline the most common complaints patients who have a diagnosis of chronic pain present with and the associated symptoms and diagnoses with which they suffer.
Another limitation is not all hospitals submit all of the data elements examined in the current study. We attempted to minimize this effect by excluding those patients with missing data when calculating results for those categories. For example, when a patient did not have data available on procedures, they were excluded in the denominator when determining mean number of procedures per patient. Finally, the apparent rise in admission rates noted in our study may be difficult to interpret in light of the evolving ICD-9 codes used for diagnostic coding of patients with chronic pain. For example, the ICD-9 codes for chronic pain NEC (338.29) and chronic pain syndrome (338.4) did not come into use until October 2006. However, the trend in rising admission rates seems to persist even after this time period, as seen in Fig 1.
Nevertheless, this is the first study to confirm the observation that more children are being admitted with chronic pain syndromes. In addition, this study demonstrates for the first time that readmission rates are substantial and that observed lengths of stay for patients with chronic pain are consistently above expected lengths of stay based on All Patient Refined Diagnosis Related Group severity scores. Perhaps not surprisingly, general pediatricians and hospitalists are most commonly the inpatient providers of care for this patient population. However, pediatric gastroenterologists are also commonly involved in care, likely as a result of the high incidence of nonspecific abdominal pain complaints. Similarly, this study is the only study to confirm that many of these patients undergo invasive procedures in the inpatient setting, either as part of their diagnostic evaluation or for intended therapeutic purposes, with gastroenterology and anesthesia as the most common procedural specialties involved. We suspect that the high incidence of esophagogastroduodenoscopies and intestinal biopsies correlates with diagnostic evaluation for etiologies of the most common admission diagnosis (abdominal pain). As expected, lumbar punctures, injections of regional anesthesia, and rehabilitation therapies were also common procedures, presumably representing treatment approaches to these children. The extent to which inpatient diagnostic evaluations and/or treatments are appropriate for this population remains unclear. One study has suggested that inpatient treatment of recurrent headaches may be beneficial in reducing the frequency and duration of attacks,33 but concern remains that acute care hospitalization with extensive diagnostic testing may perpetuate the “sick role” rather than focusing on functional therapies when little else exists to help these patients and their families. Clearly, more information is needed on the underlying pathologic and social mechanisms that contribute to these patients’ symptoms so that we can develop treatments targeted to those mechanisms rather than use reactive symptomatic therapies such as opiates and benzodiazepines, which often have addictive potential and a tendency for adverse effects.
This is the most comprehensive study to date outlining the characteristics of children admitted with chronic pain. Future research will focus on variations between hospitals in medication use in this population and attempt to correlate that with length of stay and costs to define better treatment approaches.
- Accepted May 13, 2013.
- Address correspondence to Thomas A. Coffelt, MD, Pediatric Hospital Medicine Program, Riley Hospital for Children at Indiana University Health, 705 Riley Hospital Dr, RI-3530, Indianapolis, IN 46202. E-mail:
Dr Coffelt conceptualized and designed the study and drafted the initial manuscript; Dr Bauer conceptualized and designed the study and reviewed and revised the manuscript; Dr Carroll assisted with conceptualization and design of the study and critically reviewed the manuscript; and all authors approved the final manuscript as submitted.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
- Southwood TR
- Clinch J,
- Eccleston C
- Hechler T,
- Dobe M,
- Zernikow B
- Konijnenberg AY,
- De Graeff-Meeder ER,
- Kimpen JL,
- van der Hoeven J,
- Buitelaar JK,
- Uiterwaal CS,
- Pain of Unknown Origin in Children Study Group
- Eccleston C,
- Malleson P
- Murray CS,
- Cohen A,
- Perkins T,
- Davidson JE,
- Sills JA
- Lloyd-Thomas AR,
- Lauder G
- Maynard CS,
- Amari A,
- Wieczorek B,
- Christensen JR,
- Slifer KJ
- Copyright © 2013 by the American Academy of Pediatrics