ARTICLE |
a Division of General Pediatrics, Child and Adolescent Health Research Unit, Department of Pediatrics, and Departments of
b Biostatistics
c Preventive Medicine, Vanderbilt University Medical Center, Nashville, Tennessee
| ABSTRACT |
|---|
|
|
|---|
PATIENTS AND METHODS. We conducted a population-based retrospective cohort study in which we identified children with Down syndrome by using TennCare files and birth certificates. We included 1- to 18-year-olds who were continuously enrolled in TennCare and did not fill a prescription for thyroid medication during a 90-day prestudy period. The rate of medically treated thyroid disease (prescription filled for thyroid medication) was the main outcome. We used Poisson regression to estimate rates of medically treated thyroid disease according to study year, age, gender, race, region of residence, and payer type.
RESULTS. During the 11-year study period, 1257 children with Down syndrome (28% black, 72% white) met inclusion criteria. Overall, 10.8% filled a new prescription for thyroid medication. Rates of medically treated thyroid disease per 1000 child-years were 13.25 (1995–1997), 13.34 (1998–1999), 13.62 (2000–2001), 22.37 (2002–2003), and 22.51 (2004–2005). After adjusting for child age and race, there was an increased rate of medically treated thyroid disease in 2002–2003 and 2004–2005 compared with 1995–1997. In a comparison cohort of children without Down syndrome, there was a smaller increase in the rate of medically treated thyroid disease when comparing 2002–2003 and 2004–2005 with 1995–1997.
CONCLUSIONS. Over the 11-year period, 10.8% of children with Down syndrome filled a new prescription for a thyroid medication. A 73% increase in the incidence of medically treated thyroid disease occurred after rerelease of American Academy of Pediatrics guidelines, which may have influenced screening.
Key Words: Down syndrome thyroid disorders guidelines
Abbreviations: DS—Down syndrome AAP—American Academy of Pediatrics ICD—International Classification of Diseases TANF—Temporary Aid to Needy Families SSI—Social Security Insurance IRR—incidence rate ratio CI—confidence interval
| INTRODUCTION |
|---|
|
|
|---|
10 000 children with DS are born in the United States each year.1 A number of medical conditions are associated with DS, including thyroid dysfunction. Although reports vary, prevalence rates of any thyroid dysfunction in children with DS have been estimated up to
15%.2–4 Hypothyroidism has a subtle presentation and can be particularly challenging to detect in patients with intellectual disabilities and communication and language impairments. Furthermore, symptoms of hypothyroidism overlap with features of DS, including impaired intellectual development in young children, decreased linear growth, dry skin, dentition abnormalities, and decreased physical activity.5–8 Although routine screening for thyroid disease is recommended for children with DS, it is not known how recommendations may influence physician practice.
The American Academy of Pediatrics (AAP) recommends annual screening for thyroid disease for individuals with DS who are
1 year of age.2 Several cross-sectional and longitudinal studies have contributed to our understanding of the association of DS and thyroid disease.3,4,9–12 However, there are no large population-based estimates of the incidence of medically treated thyroid disease in children with DS. It is also not known whether the rerelease of AAP guidelines, recommending routine thyroid screening in children with DS, would be associated with increased identification of new cases of disease. The objectives of the study were to investigate whether an increase in medically treated thyroid disease occurred after rerelease of AAP guidelines in 2001 and to estimate the population-based incidence of medically treated thyroid disease in children with DS enrolled in a state public health insurance plan from 1995 to 2005.
| PATIENTS AND METHODS |
|---|
|
|
|---|
Using previously described methods, we identified eligible children and obtained all of the study data from linked Tennessee Medicaid administrative data files and Tennessee State vital records.13,14 We identified children with DS between ages 1 and 18 years from International Classification of Diseases (ICD) diagnoses (69.77%), birth certificate data (8.27%), or both (21.96%). Inpatient and outpatient claims of visits at
12 years old with ICD-9 codes (758.0) or ICD-8 (759.3 used 1977–1978) were used. Children with a single inpatient ICD-9 or ICD-8 diagnosis of DS or 2 outpatient ICD-9 or ICD-8 diagnoses met our study definition of DS.
Our main outcome variable was the rate of medically treated thyroid disease, defined as a new prescription filled for thyroid medication. To be eligible for the study, individuals could not have filled a prescription for thyroid medication during the 90 days before study entry. Therefore, a new prescription was considered to represent the onset of medically treated thyroid disease. Thyroid replacement medications included liotrix, levothyroxine sodium, thyroglobulin, thyroid, and liothyronine sodium. Prescriptions for antithyroid medications (propylthiouracil and methimazole) were also captured and included in our definition. We followed children until they filled a prescription for thyroid medication, had >30 days of consecutive nonenrollment, had out-of-state enrollment, died, or the study ended.
We assessed the rate of medically treated thyroid disease before and after the February 2001 release of the AAP guidelines for health maintenance for individuals with DS. We combined adjacent years to give 5 categories: 1995–1997, 1998–1999, 2000–2001, 2002–2003, and 2004–2005. Because of an odd number of years, the first category combined 3 years, and the other categories included 2 years. We were also particularly interested in whether there were differences in medically treated thyroid disease by participant race. Because participant race was a variable of interest, children with unknown race (8.7%) were excluded. We conducted sensitivity analyses before the exclusion of children with unknown race. Results of bivariate analyses of race and new thyroid medication prescriptions were similar whether the unknown group was separate or combined with either the white or black group or excluded. In addition, children of other racial or ethnic groups were excluded, because they were too few to study (5.6% combined for Latino, Asian, and other). Other predictor variables examined included participant age, participant gender, region of residence in Tennessee (urban, suburban, or rural), and TennCare enrollment category. TennCare enrollment categories included Temporary Aid to Needy Families (TANF), Social Security Insurance (SSI), or otherwise uninsured. For overlapping enrollment segments with different enrollment categories, if there was any TANF, the category was classified as TANF. If there were overlapping SSI and uninsured categories, enrollment was classified as SSI. For overlapping enrollment segments with different regions, if there were any rural regions, the enrollment was classified as rural. If there were suburban and urban regions, region was classified as suburban. Using TennCare claims (1 inpatient and/or 2 outpatient physician claims), we captured the date of the first ICD-9 diagnosis of specific chronic diagnoses, including congenital cardiac disease (745, 746, and 747.0–747.4), gastrointestinal conditions (751.0–751.5), or leukemia (204–208). ICD-8 codes were used for 1977–1978.
As a parallel comparison group, we assembled a cohort of children without DS enrolled in TennCare in 1995–2005. Children in the non-DS cohort did not have DS indicated on the birth certificate or an ICD-9 diagnosis of DS. As with the primary DS cohort, children in the non-DS comparison group were continuously enrolled during the 90 days before study entry and did not fill a prescription for thyroid medication during this prestudy period.
We determined the rate of medically treated thyroid disease in eligible children with DS enrolled in the Tennessee Medicaid Program and the comparison group of children without DS in 1995–2005. The children's baseline characteristics and the presence of associated chronic conditions present before the diagnosis of thyroid disease were expressed as proportions. In person-time analyses, we compared rates of medically treated thyroid disease before and after the rerelease of the 2001 AAP guidelines. Specifically, we compared the rates of medically treated thyroid disease in the years 1998–1999, 2000–2001, 2002–2003, and 2004–2005 with 1995–1997. Child days were converted to child-years, and results were determined per 1000 child-years. We used Poisson regression to determine predictors of medically treated thyroid disease.15 We included in the final model factors that could potentially confound the relationship between the year of study and the rate of medically treated thyroid disease. On the basis of an incidence of DS of 1 in 1000 in the population, we estimated that we would identify
1300 children with DS.1
| RESULTS |
|---|
|
|
|---|
|
|
|
|
| DISCUSSION |
|---|
|
|
|---|
The primary objective of this investigation was to determine whether an increase in medically treated thyroid disease occurred after rerelease in February 2001 of AAP guidelines recommending annual screening for thyroid disease starting at age 1 year. Compared with 1995–1997, rates of medically treated thyroid disease increased 73% in 2002–2003 and 72% in 2004–2005.
We determined rates of medically treated thyroid disease in a comparison cohort of children without DS, in whom an increase in the identification of new cases of medically treated thyroid disease would not be expected after rerelease of DS-specific guidelines. In a Scottish cohort, which included >100 000 children aged 1 to 22 years, the prevalence of receiving
2 prescriptions for thyroid medication was 0.135% over the 3-year study period.25 In the TennCare cohort of children without DS, the incidence of medically treated thyroid disease over the 11-year study period was 0.21%. In contrast to the 73% increase in the cohort of children with DS, there was a 26% increase in the rate of medically treated thyroid disease in 2002–2003 compared with 1995–1997.
There were also differences in medically treated thyroid disease by race and age. Black children with DS were less likely to have medically treated thyroid disease than white children. It is important to note that the decreased rate in black subjects compared with white subjects was also seen in the parallel cohort of children without DS. Previous studies have also reported lower rates of hypothyroidism in black adolescents and adults compared with whites.26,27 Therefore, we do not know whether racial differences in rates of medically treated thyroid disease in children with DS are because of differences in disease incidence caused by biological factors or in differences in screening rates, which could reflect racial differences in health-seeking behavior, access to medical care, or quality of medical care.26–32 Lastly, there were differences in the risk of medically treated thyroid disease by age group. In children with DS, all of the children in older age groups were less likely to have medically treated thyroid disease than the 1- to <3-year age group. This could be the result of a number of factors, including the highest incidence occurring in this age group or the higher likelihood of undergoing screening because of multiple contacts with the health care system. In addition, physicians may be more likely to prescribe medication for young children with abnormal thyroid function studies, because children less than age 3 years are particularly vulnerable to the adverse effects of untreated hypothyroidism on intellectual development.
Previous investigations have explored physician knowledge of and use of clinical guidelines/practice parameters.33–38 From the current literature, the evidence that clinical practice guidelines change physician behavior is limited; however, practices such as uptake of new immunizations have been more successfully implemented after recommendations.39 Although obtaining a test requiring phlebotomy in children with intellectual disabilities may present challenges, it is possible that the potential perceived benefit and the straightforwardness of ordering testing for thyroid dysfunction may have helped facilitate increased screening.
It is also important to consider other potential factors that could have contributed to the increase in the rate of medically treated thyroid disease, such as systemwide policy changes or awareness campaigns that occurred during the study period. However, across the 11-year study period, there were no major changes in the structure of TennCare, no statewide awareness efforts regarding thyroid disease in children with DS, and no changes in how the primary outcome, prescription filled for thyroid medication, was captured. Furthermore, changes in the sensitivity of thyroid function testing were not temporally related to rerelease of the guidelines. Although there was an increase in the rate of medically treated thyroid disease in the non-DS cohort (26%), the increase was lower than in the DS cohort (73%).
There are several limitations to consider. The outcome of medically treated thyroid disease depends on screening for thyroid disease, which is likely subject to variation in physician practice. We were not able to capture screening for thyroid disease or the results of laboratory tests in this study. In addition, practice patterns may vary among physicians in the treatment of abnormal thyroid function. For example, the treatment of compensated thyroid disease, a condition that may be transient, is a topic of debate.3,10 However, this study was designed to determine the rate of medically treated thyroid disease in this population, not the appropriateness of treatment.
Although we assessed the level of medically treated thyroid disease after the rerelease of the policy guideline statement, we were not able to directly evaluate the impact of a particular guideline on screening for thyroid disease. In addition, because of the retrospective nature of this cohort study, it is possible that study findings were influenced by other unmeasured factors. However, this study does provide a population-based estimation of the incidence of medically treated thyroid disease and investigates the temporal relationship of the change in incidence and the rerelease of guidelines.
The findings may also have limited generalizability. The TennCare population in Tennessee represents individuals with disabilities who are low income or unable to otherwise obtain medical insurance. Therefore, we did not capture participants with only private insurance, the uninsured, or individuals who did not seek medical care. The practice patterns of physicians who serve patients in TennCare or disproportionately provide care for children with DS may not reflect those of providers who primarily work with other patient populations. Therefore, the rate of medically treated thyroid disease in the TennCare cohort may not reflect the rate of children who were privately insured or who did not have any type of insurance. However, because TennCare provides insurance for children with disabilities, the study likely includes the majority of children with DS in the state.
| CONCLUSIONS |
|---|
|
|
|---|
| ACKNOWLEDGMENTS |
|---|
We are indebted to the Tennessee Bureau of TennCare of the Department of Finance and Administration and the Tennessee Department of Health, Office of Policy, Planning and Assessment, for providing the data.
| FOOTNOTES |
|---|
Address correspondence to Kecia N. Carroll, MD, MPH, AA 0220 Medical Center North, Nashville, TN 37232. E-mail: kecia.carroll{at}vanderbilt.edu
The authors have indicated they have no financial relationships relevant to this article to disclose.
| What's Known on This Subject
Although reports vary, prevalence rates of any thyroid dysfunction in children with DS have been estimated at up to
|
| What This Study Adds This study estimates the population-based incidence of medically treated thyroid disease in children with DS and determines whether rates increased after the rerelease of AAP health supervision guidelines in 2001.
|
| REFERENCES |
|---|
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||