Compliance With Guidelines for the Medical Care of First Urinary Tract Infections in Infants: A Population-Based Study
http://www.100md.com
《小儿科》
Department of Pediatrics
Department of Epidemiology
Department of Child Health Institute and Children's Hospital and Regional Medical Center, University of Washington, Seattle, Washington
Office of Genomics, Centers for Disease Control and Prevention, Atlanta, Georgia
ABSTRACT
Background. No population-based studies have examined the degree to which practice parameters are followed for urinary tract infections in infants.
Objective. To describe the medical care of children in their first year of life after a first urinary tract infection.
Methods. Using Washington State Medicaid data, we conducted a retrospective cohort study of children with a urinary tract infection during their first year of life to determine how many of these children received recommended care based on the most recent guidelines from the American Academy of Pediatrics. Recommended care included timely anatomic imaging, timely imaging for reflux, and adequate antimicrobial prophylaxis. Multivariate logistic-regression models were used to evaluate if hospitalization for first urinary tract infection, young age at time of diagnosis, gender, race, primary language of parents, having a managed care plan, and rural location of household residence were associated with recommended care.
Results. Less than half of all children diagnosed with a urinary tract infection in their first year of life received the recommended medical care. Children who were hospitalized for their first urinary tract infection were significantly more likely than children who were not hospitalized to receive anatomic imaging (relative risk [RR]: 1.38; 95% confidence interval [CI]: 1.20–1.57) and imaging for reflux (RR: 1.62; 95% CI: 1.34–1.90).
Conclusions. There is poor compliance with guideline-recommended care for first urinary tract infections in infants in a Medicaid population. Given the trend toward increased outpatient management of urinary tract infections, increased attention to outpatient imaging may be warranted.
Key Words: urinary tract infections practice parameters/guidelines Medicaid population-based studies
Abbreviations: ICD-9-CM, International Classification of Diseases, Ninth Revision, Clinical Modification CPT-4, Current Procedural Terminology, Fourth Revision RR, relative risk CI, confidence interval
Urinary tract infections are one of the most common acute diseases of childhood, affecting an estimated 6.5% of girls and 3.3% of boys in their first year of life. 1 Vesicoureteral reflux, which is present in 30% to 40% of children with a urinary tract infection, can put children at risk for recurrent infections and renal scarring. 2,3 Effective evaluation of urinary tract infections can identify patients who may develop long-term renal complications such as hypertension and renal disease. 2,4 Infants and young children are of particular interest, because they are at higher risk for renal damage from urinary tract infections than older children. 5
The most recent recommendations by the American Academy of Pediatrics for the evaluation of children <2 years of age with a first urinary tract infection include imaging studies (both renal ultrasound and either voiding cystourethrography or radionuclide cystourethrography) and antimicrobial prophylaxis for children with unknown or diagnosed vesicoureteral reflux. 1 Renal ultrasound can identify hydronephrosis and other anatomic abnormalities but can miss vesicoureteral reflux. 3 Either voiding cystourethrography or radionuclide cystourethrography can be considered the gold standard for diagnosing reflux. 3 Antimicrobial prophylaxis has been associated with fewer recurrent urinary tract infections 6,7 and renal scarring 8 in children with vesicoureteral reflux. Although recent studies have questioned the currently recommended imaging studies as well as the necessity of prophylactic antimicrobials, 9–11 the current practice parameters are nevertheless generally accepted as recommended care, particularly in high-risk populations in which medical follow-up is uncertain. No population-based studies have examined compliance with guideline-recommended medical care for urinary tract infections in children.
The primary aim of this study was to describe the medical care that children enrolled in Medicaid in their first year of life received after a first urinary tract infection. Our secondary aim was to determine predictors of recommended medical care for a first urinary tract infection.
METHODS
Study Design
This was a retrospective cohort study using Washington State Medicaid data. Washington offers Medicaid insurance to all children living within the state whose families earn <200% of the federal poverty level. Children enrolled in Washington State Medicaid at birth are automatically considered eligible for their first year. Children were enrolled in either managed care or fee-for-service medical plans. During the time of this study, children of undocumented aliens were included in the Medicaid system.
This study was approved by the Washington State Institutional Review Board through the Department of Social and Health Services.
Eligible Patients
Children were included if they were (1) born between January 1, 1999, and September 30, 2000, (2) continuously enrolled in Medicaid from the date they were born through the first 15 months of life, and (3) diagnosed with a urinary tract infection in their first year of life. Continuous enrollment was determined by the Washington State Medicaid program based on Medicaid patient-eligibility files. We included only children who were enrolled in Medicaid for the first 15 months of life to allow a full 3 months after the child's first birthday for possible follow-up imaging studies if a child had his or her first urinary tract infection on or near his or her first birthday. The parents of study subjects may have been born outside of either Washington State or the United States, but the study subjects themselves were born in Washington State.
Diagnosis of urinary tract infection was defined as either (1) an inpatient, provider-billed diagnosis of a urinary tract infection or (2) both an outpatient, provider-billed diagnosis of a urinary tract infection and an antibiotic prescription filled within 1 week of diagnosis. If multiple diagnoses were listed for a given visit or hospitalization, we included only diagnoses of urinary tract infections that were listed either first (primary diagnosis) or second (secondary diagnosis). If urinary tract infection was the secondary diagnosis, we reviewed the primary diagnosis and excluded any children with a condition predisposing them to urinary catheterizations, such as spina bifida and major trauma. We did this to exclude iatrogenic and nosocomial urinary tract infections that may not warrant initial imaging studies, such as those caused by indwelling urinary catheters. We did not have access in our database to the results of specific laboratory results such as urinalyses and urine cultures. We defined children who were hospitalized for their first urinary tract infection as children with a urinary tract infection diagnosis who were hospitalized within 1 week of diagnosis.
Both inpatient and outpatient urinary tract infection diagnoses and associated studies were identified with International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnostic and procedure codes and Current Procedural Terminology, Fourth Revision (CPT-4) codes (Table 1). Antibiotic prescriptions were identified by using Medicaid billing drug classes among outpatient prescription-drug claims.
Outcome Variables
Based on the most recent recommendations by the American Academy of Pediatrics, we used the following 3 outcomes to reflect recommended care for a first urinary tract infection in children in their first year of life 1:
Anatomic imaging: urinary tract ultrasonography or renal scan within 3 months of initial diagnosis;
Imaging for reflux: voiding cystourethrography or renal scan within 3 months of initial diagnosis; and
Antimicrobial prophylaxis: antibiotic therapy after diagnosis until imaging for reflux was performed.
We chose to include all imaging studies that were performed within 3 months of diagnosis of a first urinary tract infection. Although there is no definite recommendation of when to perform imaging studies, 3 months is twice the generally accepted upper limit (6 weeks) of when to obtain both anatomic imaging and imaging for reflux. 10
We defined adequate antimicrobial prophylaxis as sufficient antimicrobial prescriptions to last between the date of diagnosis and the date of imaging for reflux. For children who did not receive imaging for reflux, we did not calculate the adequacy of antimicrobial prophylaxis, because those children did not have a defined end point. These children were excluded from the antimicrobial-prophylaxis outcome analysis. We identified the length of each antibiotic course by multiplying the number of days in the prescription by the number of refills. When we did not know the exact number of days in the prescription, we conservatively estimated the length of the drug course from the quantity of drug prescribed based on at least a once-daily dosing regimen. If multiple antibiotic prescriptions were prescribed at the same time, we summed the total number of days prescribed to conservatively estimate the longest possible length of antimicrobial-prophylaxis coverage.
Predictor Variables and Covariates
We analyzed patients in the study with respect to 7 variables that may be associated with compliance with guidelines for the medical care of first urinary tract infections. One variable was part of the medical management of the patient (hospitalization or outpatient management), and 6 variables were preexisting patient characteristics (age at time of diagnosis [90 or >90 days], gender, parent's identification of the child's race [white, black, Hispanic, Asian American, Native American, or other], primary language of parents [English or non-English], Medicaid plan [managed care or fee-for-service], and location of household residence [rural or urban]). The data for the preexisting patient characteristics were collected at enrollment in Medicaid and were based on parental self-report. We considered patients whose parents reported a primary language of English, American Sign Language, or Braille as having a primary language of English. All other languages were considered non-English. Rural classification was based on the patient's zip code using rural-urban commuting area codes, 12 a census tract–based classification scheme developed by the Health Resource and Service Administration and the University of Washington.
Data Analysis
We identified all children with a first urinary tract infection among the study population and created dichotomous variables for each recommended outcome for a first urinary tract infection based on whether it was completed within 3 months of diagnosis. We conducted separate logistic-regression models to evaluate the association between recommended medical care and the previously listed 7 variables: hospitalization for first urinary tract infection, young age at time of diagnosis, gender, race, primary language of parents, managed care plan, and rural location of household residence. We included the following a priori covariates in each model: age, gender, and hospitalization at diagnosis. We used forward stepwise regression to determine the effect of each of the following covariates on risk estimates: race, primary language of parents, medical care plan, and rural location of household residence. Even if the previously listed covariates did not affect the risk estimates significantly, we included them all in the final models to account for any possible confounding and to keep the multiple separate regression models comparable. Because our outcomes are not rare events (ie, incidence > 10%) for children with a first urinary tract infection, the odds ratios derived from our logistic model were converted to relative risks (RRs) to more clearly convey the strength of the association for our predictors of interest. 13
RESULTS
We identified 780 patients who were diagnosed with a urinary tract infection in their first year of life (2.1% incidence among 38793 eligible subjects in their first year), of whom 302 (38.7%) were hospitalized for their first urinary tract infection. Of children diagnosed with a urinary tract infection in our study, 59.7% were female, 17.7% were 90 days old, 29.7% lived in a rural area, and 64.6% were in a managed care plan (Table 2). Most of our subjects were white (43.5%) or Hispanic (27.8%); 3.3% were black, 3.0% were Asian American, 2.3% were American Indian, and 8.2% were of another race. The race variable was missing for 11.9% of the subjects. Twenty-nine percent of our subjects had parents who spoke a primary language other than English.
The majority (83.0%) of the patients with a urinary tract infection had that diagnosis listed as the primary diagnosis. Of the remaining patients identified by using a secondary diagnosis of urinary tract infection, we did not exclude any patients based on their primary diagnosis. The most common primary diagnoses among children with a secondary diagnosis of urinary tract infection were fever, bronchiolitis, and febrile seizure. There were a total of 448 unique medical providers who diagnosed the patients in the study, and the median number of urinary tract infections diagnosed by specific providers during the 36 months of the study was 1 (range: 1–10).
Less than half of all children diagnosed with a urinary tract infection in their first year of life received either timely anatomic imaging (44.0%) or imaging for reflux (39.5%) (Table 3). Of those diagnosed with urinary tract infections, 220 (28.2%) received both anatomic imaging and imaging for reflux, 123 (15.8%) received only anatomic imaging, and 88 (11.3%) received only imaging for reflux. Of those with imaging for reflux, 51.0% had adequate antibiotics to maintain antimicrobial prophylaxis between diagnosis and imaging for reflux.
After controlling for risk factors such as age, gender, race, primary language of parents, managed care plan, and rural location of residence, multivariate analyses showed that children who were hospitalized for their urinary tract infection were significantly more likely than children who were not hospitalized to receive anatomic imaging (RR: 1.38; 95% confidence interval [CI]: 1.20–1.57) and imaging for reflux (RR: 1.62; 95% CI: 1.34–1.90) (Table 4). Controlling for the same risk factors, children 90 days of age were less likely to receive antimicrobial prophylaxis than children >90 days of age (RR: 0.59; 95% CI: 0.35–0.90). Female children had more imaging for reflux than male children (RR: 1.32; 95% CI: 1.09–1.56). Although race was not associated with our outcomes of interest, ethnicity did seem important: Hispanic children were more likely to receive anatomic imaging than were white children (RR: 1.49; 95% CI: 1.22–1.74). Children of parents whose primary language was not English were 27% less likely than children of parents whose primary language was English to receive anatomic imaging after a first urinary tract infection (RR: 0.73; 95% CI: 0.52–0.97). Children who lived in a rural location were less likely to receive both anatomic imaging (RR: 0.76; 95% CI: 0.60–0.94) and imaging for reflux (RR: 0.71; 95% CI: 0.54–0.90). Enrollment in a managed care plan was not associated with a difference in recommended care.
DISCUSSION
We found that more than half of children enrolled in Medicaid in their first year of life who were diagnosed with a urinary tract infection did not receive recommended medical care as defined by the American Academy of Pediatrics. Previous studies have found less-than-complete compliance with national guidelines for a variety of conditions including pediatric asthma care, 14, 15 vision screening, 16 sexually transmitted infection treatment in adolescents, 17 and treatment of otitis media, hyperbilirubinemia, and fever in infants. 18 In populations of children covered by Medicaid, less than half have been found to receive guideline-recommended health supervision visits and immunizations 19 and follow-up care after treatment with antidepressant medications. 20 Based on this and previously published studies, there is poor compliance with guideline-recommended care in both Medicaid and non-Medicaid populations.
We did find certain predictors of recommended medical care in our population. Children who were hospitalized at diagnosis were more likely to receive imaging studies. There have been no previous studies comparing inpatient and outpatient recommended medical care, but there are several potential reasons why hospitalized children may be more likely to receive recommended care. First, children hospitalized for a urinary tract infection may be more acutely ill than nonhospitalized children, although the practice parameters make no distinction based on illness severity. Second, children who are hospitalized may be more likely to have testing completed during the course of their stay and not require an outpatient visit. Third, hospitals may have clinical pathways that detail and even mandate that recommendations be followed. Finally, care coordination at discharge may be superior to what is achieved in outpatient clinics, and follow-up studies may be scheduled more efficiently.
We also found that children who were 90 days of age were at risk for not receiving antimicrobial prophylaxis. This finding may reflect that we do not have billing claims for inpatient intravenous antibiotic-drug administration. These children may have received adequate prophylactic antibiotics intravenously if they received imaging for reflux during their hospitalization.
Female infants were more likely to receive imaging for reflux but not more likely to receive anatomic imaging. Similarly, we found that children of parents who had a primary language other than English were at increased risk for not receiving anatomic imaging but not at increased risk for not receiving imaging for reflux. For risk factors such as gender and primary language, we expected that most children would either have both timely anatomic imaging and imaging for reflux or have neither. However, we found that receiving anatomic imaging was not highly correlated with receiving imaging for reflux and that many subjects received either one or the other. This may reflect that these tests are often not performed at the same time and that different factors may affect the receipt of each test. This disparity may also be due to providers strongly encouraging imaging for reflux in all children regardless of gender or primary language.
We found no increased association between antimicrobial prophylaxis and any of our tested risk factors. Because we defined the end point for antimicrobial prophylaxis as the date of imaging for reflux, all patients included in the antimicrobial-prophylaxis analyses by definition had received imaging for reflux. The lack of association between antimicrobial prophylaxis and the tested risk factors may reflect that patients who receive imaging for reflux are highly compliant with medical recommendations.
There are some limitations to our study that warrant mention. First, we were limited to patients with continued eligibility who may differ in some ways from those with discontinuous enrollment. All children in our study were enrolled in Medicaid for their first 15 months, and we do not know how many children were not continuously enrolled or moved from the state during the time of the study. However, continuity of enrollment in Medicaid programs is generally quite high in this age group, with only 20% to 25% of children not continuously enrolled. 21,22 Furthermore, there is no a priori reason to believe that more mobile families would receive more guideline-compliant care than what we found. Second, we may not have captured all urinary tract infections accurately, given our reliance on an administrative billing database. However, inpatient claims tend to have high specificity, and all outpatient claims included an antibiotic prescription. We also included only primary and secondary diagnoses of urinary tract infections. Furthermore, the incidence of urinary tract infections in our study was comparable to those found in previous studies. 3 Third, there may have been some misclassification in the primary-language variable, and we grouped all non-English languages together. We recognize that different language groups may be associated with different cultures and ideas about medicine that may affect the use of the Western medical care system. Some language groups may speak English well, which may obscure any relationship between language barriers and lack of guideline adherence.
We were concerned about the possibility that individual provider effects might confound our findings. That is, certain physicians might routinely provide recommended medical care for certain segments of the population, and others may not. It is unfortunate that, given the large number (n = 448) of unique providers and groups of providers in our data, controlling for individual provider, group of physicians, or hospital was not possible. We therefore endeavored to determine the number of urinary tract infections diagnosed by each medical provider. Because this number was small (median: 1; range: 1–10), we have some assurance that any detected effects were not due to specific providers who contributed disproportionately to the data.
Practice parameters are recommendations only and may be tailored for individual patients. However, we need more research on the most appropriate evidence-based practice parameters for evaluation and treatment of urinary tract infections. We found hospitalization to be the most important predictor of recommended care, but we do not recommend hospitalizing infants with a first urinary tract infection to increase the likelihood that they will receive guideline-recommended care. Given the trend toward increased outpatient management of urinary tract infections, 1 providers need to be mindful of recommendations for urinary tract infection care in infants when treating patients on an outpatient basis.
FOOTNOTES
Accepted Oct 19, 2004.
No conflict of interest declared.
REFERENCES
American Academy of Pediatrics, Committee on Quality Improvement, Subcommittee on Urinary Tract Infection. Practice parameter: the diagnosis, treatment, and evaluation of the initial urinary tract infection in febrile infants and young children [published corrections appear in Pediatrics. 2000;105:141; 1999;103 :1052; and 1999;104:118]. Pediatrics. 1999;103:843–852
McKerrow W, Davidson-Lamb N, Jones PF. Urinary tract infection in children. Br Med J (Clin Res Ed). 1984;289 :299 –303
Jodal U. The natural history of bacteriuria in childhood. Infect Dis Clin North Am. 1987;1 :713 –729
Shimada K, Matsui T, Ogino T, Ikoma F. New development and progression of renal scarring in children with primary VUR. Int Urol Nephrol. 1989;21 :153 –158
Marget W, Belohradsky BH, Nickel B. Comparison of long- and short-term treatment of recurrent urinary tract infection. Infection. 1979;7 (suppl 4):S394–S397
Lohr JA, Nunley DH, Howards SS, Ford RF. Prevention of recurrent urinary tract infections in girls. Pediatrics. 1977;59 :562 –565
Winter AL, Hardy BE, Alton DJ, Arbus GS, Churchill BM. Acquired renal scars in children. J Urol. 1983;129 :1190 –1194
Hoberman A, Charron M, Hickey RW, Baskin M, Kearney DH, Wald ER. Imaging studies after a first febrile urinary tract infection in young children. N Engl J Med. 2003;348 :195 –202
Mahant S, To T, Friedman J. Timing of voiding cystourethrogram in the investigation of urinary tract infections in children. J Pediatr. 2001;139 :568 –571
Stapleton FB. Imaging studies for childhood urinary infections. N Engl J Med. 2003;348 :251 –252
Zhang J, Yu KF. What's the relative risk A method of correcting the odds ratio in cohort studies of common outcomes. JAMA. 1998;280 :1690 –1691
Finkelstein JA, Lozano P, Shulruff R, et al. Self-reported physician practices for children with asthma: are national guidelines followed Pediatrics. 2000;106 (4 suppl):886–896
Diette GB, Skinner EA, Markson LE, et al. Consistency of care with national guidelines for children with asthma in managed care. J Pediatr. 2001;138 :59 –64
Wall TC, Marsh-Tootle W, Evans HH, Fargason CA, Ashworth CS, Hardin JM. Compliance with vision-screening guidelines among a national sample of pediatricians. Ambul Pediatr. 2002;2 :449 –455
Beckmann KR, Melzer-Lange MD, Gorelick MH. Emergency department management of sexually transmitted infections in US adolescents: results from the National Hospital Ambulatory Medical Care Survey. Ann Emerg Med. 2004;43 :333 –338
Christakis DA, Rivara FP. Pediatricians' awareness of and attitudes about four clinical practice guidelines. Pediatrics. 1998;101 :825 –830
Hakim RB, Bye BV. Effectiveness of compliance with pediatric preventive care guidelines among Medicaid beneficiaries. Pediatrics. 2001;108 :90 –97
Richardson LP, DiGiuseppe D, Christakis DA, McCauley E, Katon W. Quality of care for Medicaid-covered youth treated with antidepressant therapy. Arch Gen Psychiatry. 2004;61 :475 –480
Cooper WO, Hickson GB, Gray CL, Ray WA. Changes in continuity of enrollment among high-risk children following implementation of TennCare. Arch Pediatr Adolesc Med. 1999;153 :1145 –1149
Christakis D, Feudtner C, Zimmerman F. Risk factors for disenrollment from health insurance . Pediatr Res. 2002;51 (suppl):1132(Adam L. Cohen, MD, Freder)
Department of Epidemiology
Department of Child Health Institute and Children's Hospital and Regional Medical Center, University of Washington, Seattle, Washington
Office of Genomics, Centers for Disease Control and Prevention, Atlanta, Georgia
ABSTRACT
Background. No population-based studies have examined the degree to which practice parameters are followed for urinary tract infections in infants.
Objective. To describe the medical care of children in their first year of life after a first urinary tract infection.
Methods. Using Washington State Medicaid data, we conducted a retrospective cohort study of children with a urinary tract infection during their first year of life to determine how many of these children received recommended care based on the most recent guidelines from the American Academy of Pediatrics. Recommended care included timely anatomic imaging, timely imaging for reflux, and adequate antimicrobial prophylaxis. Multivariate logistic-regression models were used to evaluate if hospitalization for first urinary tract infection, young age at time of diagnosis, gender, race, primary language of parents, having a managed care plan, and rural location of household residence were associated with recommended care.
Results. Less than half of all children diagnosed with a urinary tract infection in their first year of life received the recommended medical care. Children who were hospitalized for their first urinary tract infection were significantly more likely than children who were not hospitalized to receive anatomic imaging (relative risk [RR]: 1.38; 95% confidence interval [CI]: 1.20–1.57) and imaging for reflux (RR: 1.62; 95% CI: 1.34–1.90).
Conclusions. There is poor compliance with guideline-recommended care for first urinary tract infections in infants in a Medicaid population. Given the trend toward increased outpatient management of urinary tract infections, increased attention to outpatient imaging may be warranted.
Key Words: urinary tract infections practice parameters/guidelines Medicaid population-based studies
Abbreviations: ICD-9-CM, International Classification of Diseases, Ninth Revision, Clinical Modification CPT-4, Current Procedural Terminology, Fourth Revision RR, relative risk CI, confidence interval
Urinary tract infections are one of the most common acute diseases of childhood, affecting an estimated 6.5% of girls and 3.3% of boys in their first year of life. 1 Vesicoureteral reflux, which is present in 30% to 40% of children with a urinary tract infection, can put children at risk for recurrent infections and renal scarring. 2,3 Effective evaluation of urinary tract infections can identify patients who may develop long-term renal complications such as hypertension and renal disease. 2,4 Infants and young children are of particular interest, because they are at higher risk for renal damage from urinary tract infections than older children. 5
The most recent recommendations by the American Academy of Pediatrics for the evaluation of children <2 years of age with a first urinary tract infection include imaging studies (both renal ultrasound and either voiding cystourethrography or radionuclide cystourethrography) and antimicrobial prophylaxis for children with unknown or diagnosed vesicoureteral reflux. 1 Renal ultrasound can identify hydronephrosis and other anatomic abnormalities but can miss vesicoureteral reflux. 3 Either voiding cystourethrography or radionuclide cystourethrography can be considered the gold standard for diagnosing reflux. 3 Antimicrobial prophylaxis has been associated with fewer recurrent urinary tract infections 6,7 and renal scarring 8 in children with vesicoureteral reflux. Although recent studies have questioned the currently recommended imaging studies as well as the necessity of prophylactic antimicrobials, 9–11 the current practice parameters are nevertheless generally accepted as recommended care, particularly in high-risk populations in which medical follow-up is uncertain. No population-based studies have examined compliance with guideline-recommended medical care for urinary tract infections in children.
The primary aim of this study was to describe the medical care that children enrolled in Medicaid in their first year of life received after a first urinary tract infection. Our secondary aim was to determine predictors of recommended medical care for a first urinary tract infection.
METHODS
Study Design
This was a retrospective cohort study using Washington State Medicaid data. Washington offers Medicaid insurance to all children living within the state whose families earn <200% of the federal poverty level. Children enrolled in Washington State Medicaid at birth are automatically considered eligible for their first year. Children were enrolled in either managed care or fee-for-service medical plans. During the time of this study, children of undocumented aliens were included in the Medicaid system.
This study was approved by the Washington State Institutional Review Board through the Department of Social and Health Services.
Eligible Patients
Children were included if they were (1) born between January 1, 1999, and September 30, 2000, (2) continuously enrolled in Medicaid from the date they were born through the first 15 months of life, and (3) diagnosed with a urinary tract infection in their first year of life. Continuous enrollment was determined by the Washington State Medicaid program based on Medicaid patient-eligibility files. We included only children who were enrolled in Medicaid for the first 15 months of life to allow a full 3 months after the child's first birthday for possible follow-up imaging studies if a child had his or her first urinary tract infection on or near his or her first birthday. The parents of study subjects may have been born outside of either Washington State or the United States, but the study subjects themselves were born in Washington State.
Diagnosis of urinary tract infection was defined as either (1) an inpatient, provider-billed diagnosis of a urinary tract infection or (2) both an outpatient, provider-billed diagnosis of a urinary tract infection and an antibiotic prescription filled within 1 week of diagnosis. If multiple diagnoses were listed for a given visit or hospitalization, we included only diagnoses of urinary tract infections that were listed either first (primary diagnosis) or second (secondary diagnosis). If urinary tract infection was the secondary diagnosis, we reviewed the primary diagnosis and excluded any children with a condition predisposing them to urinary catheterizations, such as spina bifida and major trauma. We did this to exclude iatrogenic and nosocomial urinary tract infections that may not warrant initial imaging studies, such as those caused by indwelling urinary catheters. We did not have access in our database to the results of specific laboratory results such as urinalyses and urine cultures. We defined children who were hospitalized for their first urinary tract infection as children with a urinary tract infection diagnosis who were hospitalized within 1 week of diagnosis.
Both inpatient and outpatient urinary tract infection diagnoses and associated studies were identified with International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnostic and procedure codes and Current Procedural Terminology, Fourth Revision (CPT-4) codes (Table 1). Antibiotic prescriptions were identified by using Medicaid billing drug classes among outpatient prescription-drug claims.
Outcome Variables
Based on the most recent recommendations by the American Academy of Pediatrics, we used the following 3 outcomes to reflect recommended care for a first urinary tract infection in children in their first year of life 1:
Anatomic imaging: urinary tract ultrasonography or renal scan within 3 months of initial diagnosis;
Imaging for reflux: voiding cystourethrography or renal scan within 3 months of initial diagnosis; and
Antimicrobial prophylaxis: antibiotic therapy after diagnosis until imaging for reflux was performed.
We chose to include all imaging studies that were performed within 3 months of diagnosis of a first urinary tract infection. Although there is no definite recommendation of when to perform imaging studies, 3 months is twice the generally accepted upper limit (6 weeks) of when to obtain both anatomic imaging and imaging for reflux. 10
We defined adequate antimicrobial prophylaxis as sufficient antimicrobial prescriptions to last between the date of diagnosis and the date of imaging for reflux. For children who did not receive imaging for reflux, we did not calculate the adequacy of antimicrobial prophylaxis, because those children did not have a defined end point. These children were excluded from the antimicrobial-prophylaxis outcome analysis. We identified the length of each antibiotic course by multiplying the number of days in the prescription by the number of refills. When we did not know the exact number of days in the prescription, we conservatively estimated the length of the drug course from the quantity of drug prescribed based on at least a once-daily dosing regimen. If multiple antibiotic prescriptions were prescribed at the same time, we summed the total number of days prescribed to conservatively estimate the longest possible length of antimicrobial-prophylaxis coverage.
Predictor Variables and Covariates
We analyzed patients in the study with respect to 7 variables that may be associated with compliance with guidelines for the medical care of first urinary tract infections. One variable was part of the medical management of the patient (hospitalization or outpatient management), and 6 variables were preexisting patient characteristics (age at time of diagnosis [90 or >90 days], gender, parent's identification of the child's race [white, black, Hispanic, Asian American, Native American, or other], primary language of parents [English or non-English], Medicaid plan [managed care or fee-for-service], and location of household residence [rural or urban]). The data for the preexisting patient characteristics were collected at enrollment in Medicaid and were based on parental self-report. We considered patients whose parents reported a primary language of English, American Sign Language, or Braille as having a primary language of English. All other languages were considered non-English. Rural classification was based on the patient's zip code using rural-urban commuting area codes, 12 a census tract–based classification scheme developed by the Health Resource and Service Administration and the University of Washington.
Data Analysis
We identified all children with a first urinary tract infection among the study population and created dichotomous variables for each recommended outcome for a first urinary tract infection based on whether it was completed within 3 months of diagnosis. We conducted separate logistic-regression models to evaluate the association between recommended medical care and the previously listed 7 variables: hospitalization for first urinary tract infection, young age at time of diagnosis, gender, race, primary language of parents, managed care plan, and rural location of household residence. We included the following a priori covariates in each model: age, gender, and hospitalization at diagnosis. We used forward stepwise regression to determine the effect of each of the following covariates on risk estimates: race, primary language of parents, medical care plan, and rural location of household residence. Even if the previously listed covariates did not affect the risk estimates significantly, we included them all in the final models to account for any possible confounding and to keep the multiple separate regression models comparable. Because our outcomes are not rare events (ie, incidence > 10%) for children with a first urinary tract infection, the odds ratios derived from our logistic model were converted to relative risks (RRs) to more clearly convey the strength of the association for our predictors of interest. 13
RESULTS
We identified 780 patients who were diagnosed with a urinary tract infection in their first year of life (2.1% incidence among 38793 eligible subjects in their first year), of whom 302 (38.7%) were hospitalized for their first urinary tract infection. Of children diagnosed with a urinary tract infection in our study, 59.7% were female, 17.7% were 90 days old, 29.7% lived in a rural area, and 64.6% were in a managed care plan (Table 2). Most of our subjects were white (43.5%) or Hispanic (27.8%); 3.3% were black, 3.0% were Asian American, 2.3% were American Indian, and 8.2% were of another race. The race variable was missing for 11.9% of the subjects. Twenty-nine percent of our subjects had parents who spoke a primary language other than English.
The majority (83.0%) of the patients with a urinary tract infection had that diagnosis listed as the primary diagnosis. Of the remaining patients identified by using a secondary diagnosis of urinary tract infection, we did not exclude any patients based on their primary diagnosis. The most common primary diagnoses among children with a secondary diagnosis of urinary tract infection were fever, bronchiolitis, and febrile seizure. There were a total of 448 unique medical providers who diagnosed the patients in the study, and the median number of urinary tract infections diagnosed by specific providers during the 36 months of the study was 1 (range: 1–10).
Less than half of all children diagnosed with a urinary tract infection in their first year of life received either timely anatomic imaging (44.0%) or imaging for reflux (39.5%) (Table 3). Of those diagnosed with urinary tract infections, 220 (28.2%) received both anatomic imaging and imaging for reflux, 123 (15.8%) received only anatomic imaging, and 88 (11.3%) received only imaging for reflux. Of those with imaging for reflux, 51.0% had adequate antibiotics to maintain antimicrobial prophylaxis between diagnosis and imaging for reflux.
After controlling for risk factors such as age, gender, race, primary language of parents, managed care plan, and rural location of residence, multivariate analyses showed that children who were hospitalized for their urinary tract infection were significantly more likely than children who were not hospitalized to receive anatomic imaging (RR: 1.38; 95% confidence interval [CI]: 1.20–1.57) and imaging for reflux (RR: 1.62; 95% CI: 1.34–1.90) (Table 4). Controlling for the same risk factors, children 90 days of age were less likely to receive antimicrobial prophylaxis than children >90 days of age (RR: 0.59; 95% CI: 0.35–0.90). Female children had more imaging for reflux than male children (RR: 1.32; 95% CI: 1.09–1.56). Although race was not associated with our outcomes of interest, ethnicity did seem important: Hispanic children were more likely to receive anatomic imaging than were white children (RR: 1.49; 95% CI: 1.22–1.74). Children of parents whose primary language was not English were 27% less likely than children of parents whose primary language was English to receive anatomic imaging after a first urinary tract infection (RR: 0.73; 95% CI: 0.52–0.97). Children who lived in a rural location were less likely to receive both anatomic imaging (RR: 0.76; 95% CI: 0.60–0.94) and imaging for reflux (RR: 0.71; 95% CI: 0.54–0.90). Enrollment in a managed care plan was not associated with a difference in recommended care.
DISCUSSION
We found that more than half of children enrolled in Medicaid in their first year of life who were diagnosed with a urinary tract infection did not receive recommended medical care as defined by the American Academy of Pediatrics. Previous studies have found less-than-complete compliance with national guidelines for a variety of conditions including pediatric asthma care, 14, 15 vision screening, 16 sexually transmitted infection treatment in adolescents, 17 and treatment of otitis media, hyperbilirubinemia, and fever in infants. 18 In populations of children covered by Medicaid, less than half have been found to receive guideline-recommended health supervision visits and immunizations 19 and follow-up care after treatment with antidepressant medications. 20 Based on this and previously published studies, there is poor compliance with guideline-recommended care in both Medicaid and non-Medicaid populations.
We did find certain predictors of recommended medical care in our population. Children who were hospitalized at diagnosis were more likely to receive imaging studies. There have been no previous studies comparing inpatient and outpatient recommended medical care, but there are several potential reasons why hospitalized children may be more likely to receive recommended care. First, children hospitalized for a urinary tract infection may be more acutely ill than nonhospitalized children, although the practice parameters make no distinction based on illness severity. Second, children who are hospitalized may be more likely to have testing completed during the course of their stay and not require an outpatient visit. Third, hospitals may have clinical pathways that detail and even mandate that recommendations be followed. Finally, care coordination at discharge may be superior to what is achieved in outpatient clinics, and follow-up studies may be scheduled more efficiently.
We also found that children who were 90 days of age were at risk for not receiving antimicrobial prophylaxis. This finding may reflect that we do not have billing claims for inpatient intravenous antibiotic-drug administration. These children may have received adequate prophylactic antibiotics intravenously if they received imaging for reflux during their hospitalization.
Female infants were more likely to receive imaging for reflux but not more likely to receive anatomic imaging. Similarly, we found that children of parents who had a primary language other than English were at increased risk for not receiving anatomic imaging but not at increased risk for not receiving imaging for reflux. For risk factors such as gender and primary language, we expected that most children would either have both timely anatomic imaging and imaging for reflux or have neither. However, we found that receiving anatomic imaging was not highly correlated with receiving imaging for reflux and that many subjects received either one or the other. This may reflect that these tests are often not performed at the same time and that different factors may affect the receipt of each test. This disparity may also be due to providers strongly encouraging imaging for reflux in all children regardless of gender or primary language.
We found no increased association between antimicrobial prophylaxis and any of our tested risk factors. Because we defined the end point for antimicrobial prophylaxis as the date of imaging for reflux, all patients included in the antimicrobial-prophylaxis analyses by definition had received imaging for reflux. The lack of association between antimicrobial prophylaxis and the tested risk factors may reflect that patients who receive imaging for reflux are highly compliant with medical recommendations.
There are some limitations to our study that warrant mention. First, we were limited to patients with continued eligibility who may differ in some ways from those with discontinuous enrollment. All children in our study were enrolled in Medicaid for their first 15 months, and we do not know how many children were not continuously enrolled or moved from the state during the time of the study. However, continuity of enrollment in Medicaid programs is generally quite high in this age group, with only 20% to 25% of children not continuously enrolled. 21,22 Furthermore, there is no a priori reason to believe that more mobile families would receive more guideline-compliant care than what we found. Second, we may not have captured all urinary tract infections accurately, given our reliance on an administrative billing database. However, inpatient claims tend to have high specificity, and all outpatient claims included an antibiotic prescription. We also included only primary and secondary diagnoses of urinary tract infections. Furthermore, the incidence of urinary tract infections in our study was comparable to those found in previous studies. 3 Third, there may have been some misclassification in the primary-language variable, and we grouped all non-English languages together. We recognize that different language groups may be associated with different cultures and ideas about medicine that may affect the use of the Western medical care system. Some language groups may speak English well, which may obscure any relationship between language barriers and lack of guideline adherence.
We were concerned about the possibility that individual provider effects might confound our findings. That is, certain physicians might routinely provide recommended medical care for certain segments of the population, and others may not. It is unfortunate that, given the large number (n = 448) of unique providers and groups of providers in our data, controlling for individual provider, group of physicians, or hospital was not possible. We therefore endeavored to determine the number of urinary tract infections diagnosed by each medical provider. Because this number was small (median: 1; range: 1–10), we have some assurance that any detected effects were not due to specific providers who contributed disproportionately to the data.
Practice parameters are recommendations only and may be tailored for individual patients. However, we need more research on the most appropriate evidence-based practice parameters for evaluation and treatment of urinary tract infections. We found hospitalization to be the most important predictor of recommended care, but we do not recommend hospitalizing infants with a first urinary tract infection to increase the likelihood that they will receive guideline-recommended care. Given the trend toward increased outpatient management of urinary tract infections, 1 providers need to be mindful of recommendations for urinary tract infection care in infants when treating patients on an outpatient basis.
FOOTNOTES
Accepted Oct 19, 2004.
No conflict of interest declared.
REFERENCES
American Academy of Pediatrics, Committee on Quality Improvement, Subcommittee on Urinary Tract Infection. Practice parameter: the diagnosis, treatment, and evaluation of the initial urinary tract infection in febrile infants and young children [published corrections appear in Pediatrics. 2000;105:141; 1999;103 :1052; and 1999;104:118]. Pediatrics. 1999;103:843–852
McKerrow W, Davidson-Lamb N, Jones PF. Urinary tract infection in children. Br Med J (Clin Res Ed). 1984;289 :299 –303
Jodal U. The natural history of bacteriuria in childhood. Infect Dis Clin North Am. 1987;1 :713 –729
Shimada K, Matsui T, Ogino T, Ikoma F. New development and progression of renal scarring in children with primary VUR. Int Urol Nephrol. 1989;21 :153 –158
Marget W, Belohradsky BH, Nickel B. Comparison of long- and short-term treatment of recurrent urinary tract infection. Infection. 1979;7 (suppl 4):S394–S397
Lohr JA, Nunley DH, Howards SS, Ford RF. Prevention of recurrent urinary tract infections in girls. Pediatrics. 1977;59 :562 –565
Winter AL, Hardy BE, Alton DJ, Arbus GS, Churchill BM. Acquired renal scars in children. J Urol. 1983;129 :1190 –1194
Hoberman A, Charron M, Hickey RW, Baskin M, Kearney DH, Wald ER. Imaging studies after a first febrile urinary tract infection in young children. N Engl J Med. 2003;348 :195 –202
Mahant S, To T, Friedman J. Timing of voiding cystourethrogram in the investigation of urinary tract infections in children. J Pediatr. 2001;139 :568 –571
Stapleton FB. Imaging studies for childhood urinary infections. N Engl J Med. 2003;348 :251 –252
Zhang J, Yu KF. What's the relative risk A method of correcting the odds ratio in cohort studies of common outcomes. JAMA. 1998;280 :1690 –1691
Finkelstein JA, Lozano P, Shulruff R, et al. Self-reported physician practices for children with asthma: are national guidelines followed Pediatrics. 2000;106 (4 suppl):886–896
Diette GB, Skinner EA, Markson LE, et al. Consistency of care with national guidelines for children with asthma in managed care. J Pediatr. 2001;138 :59 –64
Wall TC, Marsh-Tootle W, Evans HH, Fargason CA, Ashworth CS, Hardin JM. Compliance with vision-screening guidelines among a national sample of pediatricians. Ambul Pediatr. 2002;2 :449 –455
Beckmann KR, Melzer-Lange MD, Gorelick MH. Emergency department management of sexually transmitted infections in US adolescents: results from the National Hospital Ambulatory Medical Care Survey. Ann Emerg Med. 2004;43 :333 –338
Christakis DA, Rivara FP. Pediatricians' awareness of and attitudes about four clinical practice guidelines. Pediatrics. 1998;101 :825 –830
Hakim RB, Bye BV. Effectiveness of compliance with pediatric preventive care guidelines among Medicaid beneficiaries. Pediatrics. 2001;108 :90 –97
Richardson LP, DiGiuseppe D, Christakis DA, McCauley E, Katon W. Quality of care for Medicaid-covered youth treated with antidepressant therapy. Arch Gen Psychiatry. 2004;61 :475 –480
Cooper WO, Hickson GB, Gray CL, Ray WA. Changes in continuity of enrollment among high-risk children following implementation of TennCare. Arch Pediatr Adolesc Med. 1999;153 :1145 –1149
Christakis D, Feudtner C, Zimmerman F. Risk factors for disenrollment from health insurance . Pediatr Res. 2002;51 (suppl):1132(Adam L. Cohen, MD, Freder)