Psychosomatics 2015:56:495–503
Published by Elsevier Inc. on behalf of The Academy of Psychosomatic Medicine.
Original Research Reports Clinical and Sociodemographic Factors Associated With Attention-Deficit/Hyperactivity Disorder in Patients With Cystic Fibrosis Efe Eworuke, Ph.D., Qian Ya Lensa Zeng, Pharm D., Almut G. Winterstein, Ph.D.
Background: There is scarce evidence on the epidemiology of attention-deficit/hyperactivity disorder (ADHD) in patients with cystic fibrosis (CF). Objective: We employed stepwise logistic regression to examine the association between ADHD diagnosis and selected patient characteristics. Methods: This was a cross-sectional analysis of inpatient and outpatient billing data for Medicaid beneficiaries with CF ages 3–18 years to obtain ADHD diagnosis prevalence and incidence estimates from 1999–2006. Results: Annual ADHD prevalence increased 1.55-fold from 5.26% (95% CI: 5.25–5.27) to 8.16% (8.15–8.17), and annual ADHD incidence rose slightly from 1.70% (1.70–1.71) to 2.01% (2.00–2.01). As in the general population, males were significantly more likely to have a diagnosis of ADHD compared with females (odds ratio: 1.97 [CI: 1.49–2.60]), as were children with recent diagnoses of anxiety, emotional disorder,
depression, adjustment disorder, and learning, motor, and communication disorders. Patients with ADHD diagnoses were also more likely to be in foster care (odds ratio ¼ 4.36 [CI: 2.26–8.40]). Except for recent DNase use (odds ratio ¼ 0.64 [CI: 0.43–0.93]), CF severity indicators and medications including pancreatic enzymes, inhaled tobramycin, inhaled or oral corticosteroids, inhaled bronchodilators, and oral antibiotics had no association with ADHD diagnosis. Conclusion: ADHD prevalence in CF increased during the study period. Clinical and sociodemographic determinants of ADHD diagnosis were similar to the general population, whereas treatment and severity of CF appeared to have little influence. Our findings warrant future research evaluating diagnostic protocols and assessment of safety and efficacy of ADHD treatment in children with CF. (Psychosomatics 2015; 56:495–503)
INTRODUCTION Cystic fibrosis (CF) is a rare but life-threatening genetic disease that affects approximately 30,000 people in the United States.1 Varying autosomal recessive genetic mutations translates into this multiorgan disease, which is characterized by respiratory infections, pulmonary exacerbations, pancreatic insufficiency, and liver impairment.2 Advanced clinical research and novel treatment strategies have extended life expectancy from 2 years when the disease was first discovered in 1958 to approximately 41.1 years as of 2012.1 Psychosomatics 56:5, September/October 2015
Received August 6, 2014; revised August 26, 2014; accepted September 2, 2014. From Department of Pharmaceutical Outcomes & Policy, College of Pharmacy, University of Florida, Gainesville, FL (EE, QYLZ, AGW); Department of Epidemiology, Colleges of Medicine and Public Health and Health Professions, University of Florida, Gainesville, FL (AGW). Send correspondence and reprint requests to Efe Eworuke, Ph.D., Division of Epidemiology II, Office of Pharmacovigilance and Epidemiology, Office of Surveillance and Epidemiology, Center for Drug Evaluation and Research, Food and Drug Administration, Silver Spring, MD; e-mail: [email protected] Published by Elsevier Inc. on behalf of The Academy of Psychosomatic Medicine.
www.psychosomaticsjournal.org
495
Risk Factors for ADHD in Cystic Fibrosis With a longer life expectancy, factors affecting long-term CF treatment, CF prognosis, and quality of life, as well as the management of concomitant diseases are becoming increasingly important. Patients with chronic diseases such as CF are at greater risk for psychiatric disorders, given the complexity surrounding therapy and burden of disease.3,4 With 8.8% of children in the United States currently diagnosed with attention-deficit/hyperactivity disorder (ADHD), it is plausible that this psychiatric condition may be as prevalent in the CF population.5,6 However, there are currently no epidemiologic estimates for ADHD in the CF population. ADHD impedes normal functioning and ability to carry out daily activities; therefore, for patients with CF, with an already burdened treatment regimen, ADHD could be detrimental to attaining overall treatment goals for the following reasons.7 First, psychiatric disorders have often been cited as a factor for poor treatment adherence in patients with CF.3 This is important because prognosis is contingent on adherence, which is already less than 50% in youths with CF.8 Secondly, with an increased life expectancy, it is important that both quality of life and self-worth be maintained throughout adulthood. Conditions such as ADHD not only affect day-to-day activities but also the ability to lead a functioning and successful life. Lastly, untreated ADHD can be associated with significant consequences, including an increased risk of academic failure, substance abuse, and delinquent and antisocial behavior.7,9 Thus, epidemiologic studies examining ADHD diagnosis and treatment pattern among the CF population will be important in understanding both the disease burden and respective treatment approaches. Therefore, we aimed to determine the prevalence and incidence of ADHD in a national cohort of Medicaid youths with CF and to evaluate the association between ADHD diagnoses and selected patient, clinical, and sociodemographic characteristics. METHODS Data Source The study data set comprises Medicaid administrative claims data, referred to as Medicaid Extract Files, from 28 US states (Appendix), which were selected based on Medicaid enrollment numbers, 496
www.psychosomaticsjournal.org
extent of managed care penetration, and CF prevalence as reported by the CF foundation.1 Medicaid Extract Files, provided by the Centers for Medicare and Medicaid Services, consists of person-level data files containing information on monthly eligibility, including reason for eligibility and health care coverage spectrum, detail on diagnoses and procedures associated with inpatient and outpatient care, as well as detail on dispensed medications. Because medical encounter detail for Medicaid beneficiaries in managed care plans is incomplete, we restricted our source population to patients who receive Medicaid benefits in the fee-for-service or primary care case management program. Patients who were enrolled in a behavioral or dental managed care plan were included if they were registered in the fee-for-service or primary care case management program. The institutional review and privacy boards of Centers for Medicare and Medicaid Services and the University of Florida approved the study. Study Population We included Medicaid fee-for-service recipients 3–18 years of age between January 1, 1999 and December 31, 2006, with at least 2 inpatient or outpatient diagnoses with International Classification of Disease version 9, Clinical Modification codes for CF (277.0x). Patients were also required to have a minimum of 12 months continuous Medicaid insurance for prevalence estimates. For incidence estimates, patients were required to have at least 24 months of continuous Medicaid eligibility covering the assessed and the previous year. Data Analysis Annual ADHD Incidence and Prevalence For each calendar year from 2000–2006, ADHD incidence was defined as the proportion of patients with CF with at least 2 outpatient claims for ADHD (International Classification of Disease version 9, Clinical Modification code 314.xx) in the sample of patients with CF who were not diagnosed with ADHD in the year before the year being assessed. Therefore, the source population from 1999 was used to determine the at-risk sample for the incidence estimates in 2000. ADHD prevalence from 1999–2006 was calculated as the proportion of patients with at least Psychosomatics 56:5, September/October 2015
Eworuke et al. 2 outpatient ADHD claims during the calendar year being assessed, using the total number of patients with CF with continuous Medicaid Extract Files eligibility in the measurement year as the denominator. Incidence rate ratios and prevalence ratios were obtained by comparing estimates of the last study year to the respective first study year to examine longitudinal trends. Given that patients could present in both the first year and the last year, it was important to take into account the potential correlation between yearly estimates. Therefore, we used the asymptotic standard error calculating the 95% CI using the Delta method.
included the following covariates: number of unique pharmacy prescriptions filled, number of CF-related outpatient visits, and the presence of at least 1 CFrelated hospitalization in the 6 months before the index date. We employed a multivariate conditional logistic regression to examine the association of each covariate and initial ADHD diagnosis. The model used a forward stepwise inclusion process with an α criterion of 0.50 to select covariates into the final model. RESULTS Source Population
Risk Factors Associated With ADHD For this analysis, ADHD cases were defined as patients with at least 2 outpatient claims for ADHD anytime during the study period. For each case, a control defined as patients without any ADHD claim during the study period was randomly selected from the CF cohort and matched on date of birth (⫾5 d) and study year. Both incident cases and controls were required to have at least 180 days of continuous eligibility before the index date; the index date for the cases was the date of second ADHD claim, whereas the controls adopted the index date for the matching case. The following sociodemographic characteristics were obtained from 6-month period before the index date: sex, race categorized as whites and others, state of residence, cash assistance, and Medicaid eligibility status categorized as poverty, disability, and foster care. Cash assistance is defined as receiving financial aid through state programs. The states were further classified into categories based on national estimates of ADHD prevalence (r7%, 7.1%– 9%, 9.1%–11%, and 11.1%–13%; Appendix).10 Clinical determinants of ADHD considered for analysis included the presence of at least 1 outpatient claim for any of the following psychiatric conditions: conduct disorder/opposition defiant disorder; emotional/ behavioral disorder; anxiety; adjustment disorder; depression; bipolar disorder; intellectual disability; learning, motor, and communication disorders; and schizophrenia (Appendix for operational definitions). Other considered factors comprise the use of commonly prescribed CF medications including use of inhaled tobramycin, pancreatic enzymes, inhaled bronchodilators, inhaled or oral corticosteroids, and oral antibiotics. To stage CF disease severity, we Psychosomatics 56:5, September/October 2015
We identified 19,760 patients with at least 2 inpatient or outpatient CF claims anytime during the study period. For the prevalence calculations, 6940 patients with a minimum of 12 months continuous Medicaid eligibility were examined. In each study year, males and whites accounted for slightly more than 50% and 65% of the population, respectively (Table 1). Mean age was stable across the study period, ranging from 9.79–9.99 years, as were the number of outpatient visits and hospital stays during follow-up. ADHD Incidence and Prevalence From 2000–2006, ADHD incidence rates increased slightly from 1.70% (95% CI: 1.70–1.71) in 2000 to 2.01% (CI: 2.00–2.01) in 2006. This increase in new ADHD cases was reflected in increasing ADHD prevalence from 5.26% (CI: 5.25–5.27) in 1999 to 8.16% (CI: 8.15–8.17) in 2006 (Figure). Although the ADHD incidence in 2006 was not statistically different from the incidence obtained in 2000 (incidence rate ratio: 1.17 [CI: 0.78–1.77]), the ADHD prevalence ratio was found to be statistically significant at 1.55 (CI: 1.29–1.87). Selected Factors Associated With ADHD Diagnoses During the study period, we identified 658 patients with 2 outpatient diagnoses of ADHD from the prevalence cohort with 6 months of continuous Medicaid eligibility before the second ADHD claim date. Of the 658 cases, we matched 99.1% (n ¼ 652) to a control patient. The following demographic differences emerged: cases comprised 16% more males compared with controls, reflecting the national trends for ADHD estimates.5,10 Although there were negligible differences in the racial composition of cases and www.psychosomaticsjournal.org
497
498
2.11 ⫾ 1.68 2.09 ⫾ 1.67 2.04 ⫾ 1.61 2.03 ⫾ 1.68 2.02 ⫾ 1.70
www.psychosomaticsjournal.org
IP ¼ inpatient; OP ¼ outpatient; SD ¼ standard deviation.
2.09 ⫾ 1.72 2.15 ⫾ 1.94 2.18 ⫾ 1.93
2001 3082 9.96 ⫾ 4.16 1631 (52.9) 1994 (64.7) 17.19 ⫾ 45.29 2000 2689 9.86 ⫾ 4.13 1417 (52.7) 1700 (63.2) 17.49 ⫾ 47.35 1999 2433 9.86 ⫾ 4.13 1281 (52.5) 1565 (64.3) 18.12 ⫾ 48.72
Variable Sample size Mean age Males, n (%) Whites, n (%) Mean number of OP visits (restricted to medical visits) (n ⫾ SD) Mean number of IP stays (restricted to unique hospital stays) (n ⫾ SD)
TABLE 1.
Patient and Health Care Utilization Characteristics, Stratified by Study Year
2002 3265 9.86 ⫾ 4.21 1761 (53.9) 2138 (65.5) 16.42 ⫾ 43.47
2003 3884 9.79 ⫾ 4.30 2102 (54.1) 2573 (66.3) 18.12 ⫾ 46.75
2004 3937 9.80 ⫾ 4.34 2145 (54.5) 2632 (66.9) 18.03 ⫾ 44.12
2005 4082 9.84 ⫾ 4.34 2268 (55.6) 2679 (65.6) 17.46 ⫾ 38.56
2006 4057 9.99 ⫾ 4.32 2262 (55.7) 2659 (65.5) 19.14 ⫾ 43.02
Risk Factors for ADHD in Cystic Fibrosis controls, the most prominent age group matched included 5–9-year-old children. There were also noted geographic differences between the states. States with the lowest ratio of cases to controls were Tennessee (3.76% [for cases] vs 7.99% [for controls]), Massachusetts (1.88% vs 3.92%), Pennsylvania (3.44% vs 4.55%), and Wisconsin (1.88% vs 2.04%). Conversely, states with the highest ratio comparison were Ohio (13.15% vs 6.11%), South Carolina (5.01% vs 4.08%) Georgia (9.08% vs 5.80%), and West Virginia (2.35% vs 1.41%). Lastly, there was a higher proportion of cases living in poverty (8.14% vs 5.49%) and in foster care (9.08% vs 2.19%) compared with controls (Table 2). In addition to demographic differences, we also observed clinical differences, specifically regarding psychiatric diagnoses, the use of certain CF medications, and CF-related hospitalizations. In general, cases had a greater proportion of psychiatric diagnoses compared with controls (Table 2). Regarding commonly-used CF medications, a lower proportion of cases received pancreatic enzymes, inhaled tobramycin, and DNase compared with controls. Cases were also less likely to be recently hospitalized, with a lower average of outpatient visits compared with controls (Table 2). Of the 32 considered patient characteristics, 21 entered into the final logistic regression model (Table 3). Males were more likely to be diagnosed with ADHD, as were children in foster care and with any of the considered psychiatric illnesses. Patients with a diagnosis of ADHD were also more likely to have a recent diagnosis of anxiety (odds ratio [OR] ¼ 3.94, CI: 1.46–10.63); adjustment disorder (OR ¼ 4.70, CI: 2.23–9.90); learning, motor, and communication disorders (OR ¼ 2.48, CI: 1.53–4.03); depression (OR ¼ 3.99, CI: 1.87–8.52); and emotional disorder (OR ¼ 4.36, CI ¼ 1.58–12.0). Among social factors, being in foster care was strongly associated with recent ADHD diagnosis (OR ¼ 6.13, CI: 2.91– 12.91). Although univariate comparisons suggest lesser disease severity among the cases, severity indicators including CF-related hospital stay were not significant. Among the commonly used CF medications, the sole negative determinant of ADHD was recent use of DNase (OR ¼ 0.66, CI: 0.44–0.98). DISCUSSION To our knowledge, this is the first large observational study adequately powered to examine ADHD Psychosomatics 56:5, September/October 2015
Eworuke et al.
FIGURE.
Incidence and Prevalence Estimates for ADHD in Patients With CF.
incidence and prevalence in CF with the ability to build a comprehensive predictive model determining the association between ADHD diagnosis and patient, clinical, and sociodemographic factors. The analysis reveals several key findings: first, ADHD prevalence in patients with CF increased during the study period and was comparable with the general pediatric population.5,10 Although prevalence estimates reported in our study are slightly lower than previously reported in a CF population, it is noteworthy to mention that the referenced study was conducted in a single center comprising 188 patients with CF.6 Our case definition of 2 ADHD-related claims may have also resulted in the exclusion of less severe forms of ADHD. Second, ADHD incidence estimates were slightly higher than rates reported in a large cohort study of Medicaid beneficiaries representing the general publicly insured pediatric population.10 It is plausible that differences in actual estimates are related to differences in the study period and population age. Although the large cohort study included patients ranging from 6 months to 20 years from 1995–2004, our study included patients 3–18 years of age from 1999–2006. Third, we observed significant demographic differences among patients with CF diagnosed with ADHD. Akin to previous studies, males were 2 times more likely to be diagnosed with ADHD compared with females.5,10 Although previous studies observed ADHD prominence among whites, our study did not find race a significant predictor of ADHD diagnosis. Consistent with previous studies, patients living in foster care were more likely to be diagnosed with Psychosomatics 56:5, September/October 2015
ADHD.5,11 A significant variation in the proportion of cases by state was observed, signaling that the patient’s state of residence may influence the likelihood of being diagnosed. Notably, states with the highest proportion of cases to controls (West Virginia, South Carolina, and Ohio) also had state-based prevalence rates that were higher than the national average in the general population presented in a study conducted by the Centers for Disease Control and Prevention and Health Resources and Services Administration.5 Patients diagnosed with ADHD were more likely to have other psychiatric diagnoses in the 6 months period before ADHD diagnosis. Although there is evidence suggesting higher prevalence of depression and anxiety in patients with CF owing to the psychologic and emotional toll of the disease, a causal relationship between ADHD diagnosis and other psychiatric conditions cannot be inferred from our study.10,12–14 It is important to note the lack of association between CF disease severity and ADHD diagnosis. Although patients newly diagnosed with ADHD were less likely to receive DNase, no other diagnosis determinants were found, even though univariate comparisons suggested lower disease severity among ADHD cases. In particular, recent use of other CF medications, CF-related hospitalizations, outpatient visits, and number of medications were not found to be significant. Further studies with the ability to operationalize disease severity using traditional indicators such as the forced expiratory volume in 1 second (FEV) and body mass index will be needed to confirm this finding. www.psychosomaticsjournal.org
499
Risk Factors for ADHD in Cystic Fibrosis TABLE 2.
Demographic, Patient, and Social Characteristics Comparing ADHD Cases and Controls
Covariates
ADHD cases (n ¼ 652)
Controls (n ¼ 652)
Males, n (%)
465 (71.32)
362 (55.52)
465 (71.32)
423 (64.88)
35 (5.37) 336 (51.53) 190 (29.14) 91 (13.95)
34 (5.21) 336 (51.53) 190 (29.14) 92 (14.11)
Proportion of patients with at least 1 inpatient or outpatient diagnosis during the 6-mo look-back period, n (%) Conduct disorder/opposition defiant disorder Anxiety Emotional/behavioral disorder Adjustment disorder Bipolar disorder Depression Learning, motor, and communication disorders Intellectual disability (mental retardation) Schizophrenia and other psychotic disorders Substance abuse Tics
18 (2.76) 31 (4.75) 39 (5.98) 66 (10.12) 26 (3.99) 44 (6.75) 82 (12.58) 26 (3.99) 4 (0.61) 5 (0.77) 5 (0.77)
1 (0.15) 7 (1.07) 6 (0.92) 11 (1.69) 1 (0.15) 11 (1.69) 36 (5.52) 22 (3.37) 0 (0.00) 0 (0.00) 0 (0.00)
Proportion of patients with chronic medication use during the 6-mo look-back period, n (%) Pancreatic enzymes Inhaled tobramycin DNase Inhaled/oral corticosteroids Inhaled bronchodilators Oral antibiotics (macrolides, ampicillins, cephalosporins, and quinolones)
258 106 141 255 379 435
326 144 194 241 396 417
Number of unique pharmacy claims filled during the 6-mo look-back period (mean ⫾ SD) Number of CF-related outpatient visits during the 6-mo look-back period (mean ⫾ SD)
21.46 ⫾ 18.97 10.67 ⫾ 20.94
23.00 ⫾ 21.33 12.34 ⫾ 24.01
State of residence, n (%)† Alabama Arkansas Florida Georgia Iowa Idaho Illinois Indiana Kansas Louisiana Massachusetts Minnesota Missouri Mississippi North Carolina Nebraska New Hampshire New Jersey New York Ohio Pennsylvania South Carolina Tennessee Texas
3 (0.46) 4 (0.61) 72 (11.04) 62 (9.51) 9 (1.38) 6 (0.92) 28 (4.29) 31 (4.75) 7 (1.02) 28 (4.29) 12 (1.84) 12 (1.84) 21 (3.22) 5 (0.77) 30 (4.60) 2 (0.31) 2 (0.31) 11 (1.69) 42 (6.44) 83 (12.73) 21 (3.22) 31 (4.75) 24 (3.68) 51 (7.82)
3 (0.46) 13 (2.04) 61 (9.36) 38 (5.83) 5 (0.77) 4 (0.61) 37 (5.62) 25 (3.83) 12 (1.84) 32 (4.91) 25 (3.83) 13 (2.61) 17 (2.61) 11 (1.69) 40 (6.13) 9 (1.38) 3 (0.46) 9 (1.38) 55 (8.44) 72 (6.44) 28 (4.29) 26 (3.99) 48 (7.36) 52 (7.98)
Race, n (%) Whites Age category, n (%) 45 y 5–9 y 10–14 y 14–18 y
500
*
www.psychosomaticsjournal.org
(39.57) (16.26) (21.63) (39.11) (58.13) (66.72)
(50.00) (22.09) (29.75) (36.96) (51.10) (63.96)
Psychosomatics 56:5, September/October 2015
Eworuke et al. TABLE 2. Continued Covariates
ADHD cases (n ¼ 652)
Controls (n ¼ 652)
6 (0.92) 24 (3.68) 10 (1.53) 15 (2.30)
5 (0.77) 19 (2.91) 12 (1.84) 8 (1.23)
Disease history, n (%) Recent CF-related hospital stay Cash assistance
74 (11.35) 339 (51.99)
84 (12.88) 340 (52.15)
Eligibility status Poverty Disability Foster care
497 (76.23) 267 (40.95) 54 (8.28)
485 (74.39) 311 (47.70) 14 (2.15)
Virginia Vermont Wisconsin West Virginia
ADHD ¼ attention-deficit/hyperactivity disorder; CF ¼ cystic fibrosis; SD ¼ standard deviation. n
†
Age was not included in the model, as it was a matching variable. Categories for states were included in the model.
Our study brings into light several issues for further discussion within the CF community. First, further examination will have to be undertaken to examine how clinicians decide on which patient with CF to treat for ADHD. This is especially important given the rising prevalence of ADHD among TABLE 3.
Logistic Regression Model Comparing ADHD Incident Cases With Non-ADHD Controls
Effect
Point 95% Wald estimate confidence limits
Race (Whites vs others) Gender (males vs females) Anxiety Emotional disorder Adjustment disorder Depression Intellectual disability Learning disorder Use of pancreatic enzymes Use of inhaled tobramycin Use of DNase Use of inhaled corticosteroids Use of antibiotics CF-related hospitalization stay Cash assistance Foster care Disability Poverty
1.27 2.00 3.94 4.36 4.70 3.99 0.76 2.48 0.85 0.87 0.66 1.23 1.20 0.77 1.49 6.13 0.69 1.36
0.94 1.52 1.46 1.58 2.23 1.87 0.37 1.53 0.61 0.58 0.44 0.93 0.90 0.51 0.96 2.91 0.47 0.90
1.71 2.64 10.63 12.01 9.90 8.52 1.56 4.03 1.20 1.30 0.99 1.63 1.61 1.16 2.30 12.91 1.02 2.07
State categories Category 2 vs category 1 Category 3 vs category 1 Category 4 vs category 1
0.75 1.38 1.06
0.54 0.93 0.66
1.05 2.06 1.70
CF ¼ cystic fibrosis.
Psychosomatics 56:5, September/October 2015
children with CF which is comparable to prevalence estimates reported in the general population. Because CF severity seems to play an insignificant role in deciding on a diagnosis for ADHD, it is crucial to investigate how clinicians decide on when to commence treatment and if and how risk-benefit assessments of ADHD treatment are made. Second, to inform risk-benefit decisions, concrete evidence on the safety and effectiveness of ADHD treatment in children with CF is needed. Children with CF face profound psychologic challenges, resulting in complex mental comorbidities, which may in turn affect the need to address ADHD. Although the effectiveness of stimulants in patients with CF has not been investigated, potential beneficial effects might include improved social functioning and parent or teacher relations as well as improved coping and medication adherence to their life-preserving CF treatment regimen. On the contrary, the complexity of CF medication regimen and the CF pathology itself offer the potential for a variety of drug interactions and adverse effects of ADHD treatment. Second, little is known about the influence of anorexic effects of stimulants, given the established correlation between weight and both CF progression and respiratory function. Other important considerations are potentiated cardiac side effects, bearing in mind an already increased cardiac risk, including early onset of diabetes mellitus in patients with CF, the potential cardiac risk of long-term macrolide therapy, and potential sympathomimetic drug-drug www.psychosomaticsjournal.org
501
Risk Factors for ADHD in Cystic Fibrosis interactions. A comprehensive understanding of both the effect of ADHD treatment of CF progression and life expectancy and on mental health and quality of life is critical to optimize care for this vulnerable population. In light of the prominence of ADHD in patients with CF, further research is urgently needed to evaluate the safety and effectiveness of ADHD treatment in children with CF. There are noteworthy limitations related to the data. Owing to the nature of administrative data, undercoding or miscoding of health care encounter information are inherent. To reduce diagnostic uncertainty, we required at least 2 ADHD claims to identify cases. In addition, gold standard markers for nutritional status and respiratory function in patients with CF, body mass index and forced expiratory volume, were not available and not included in the analysis. Instead, recent hospitalization stay, number of outpatient visits, and medications were used as proxy for disease severity. Finally, patient characteristics and differences in the care for patients with CF in states or private insurance not represented should be considered when interpreting our findings.
CONCLUSION Similar to the general pediatric population, we noted an increasing emergence of ADHD in this population of children with CF. Demographic and psychiatric determinants of ADHD were also similar to reports in the general population. Both CF treatment and severity appeared to have limited effect on ADHD diagnosis. Our study findings necessitate the need for future research to evaluate the prevalence of treatment of ADHD in CF and describe subpopulations that initiate treatment with the eventual goal of examining the safety and effectiveness of ADHD treatment in children with CF. This work was completed while all authors were affiliated with the University of Florida. The views expressed are those of the authors and not necessarily those of the US Department of Health and Human Services or the FDA. Disclosure: The authors disclosed no proprietary or commercial interest in any product mentioned or concept discussed in this article.
APPENDIX. Cohort Restrictions
Excluded Medicaid beneficiaries from Virginia from 1999–2002 owing to truncation of ICD9-CM codes Tennessee before January 2002 owing to the extent of managed care penetration Pennsylvania before June 2000 owing to the extent of managed care penetration.
Disease Severity At least one corresponding ICD9-CM diagnosis claim in the 6-mo period before the index date CF-related hospitalization stay 277.xx in primary or secondary diagnoses field in inpatient claims file Psychiatric disorder At least one ICD9-CM diagnosis claim for the correlating psychiatric disorder Adjustment disorder ICD9-CM codes: 309–309.20, 309.22–309.80, 309.82–309.99, and 313.89 Attention-deficit/hyperactivity disorder (ADHD) ICD9-CM codes: 314, 314.0, 314.00, 314.01, and 314.9 Anxiety ICD9-CM codes: 300.0–300.09, 300.2–300.29, 300.3, 309.21, 309.81, 313.0, 313.2x, 308.0, and 308.2x–308.9x Bipolar disorder ICD9-CM codes: 296.0–296.19, 296.4–296.81, 296.83–296.99, and 301.13 Conduct disorder ICD9-CM codes: 312.0, 312.1, 312.2, 312.3, 312.00, 312.01, 312.03, 312.10, 312.11, 312.12, 312.13, 312.20, 312.21, 312.22, 312.23, 312.30, 312.31, 312.32, 312.33, 312.34, 312.35, 312.39, 312.4, 312.8, 312.81, 312.82, 312.89, and 312.9 Depression ICD9-CM codes: 296.2x, 296.3x, 296.82, 300.4, and 311 Intellectual disabilities ICD9-CM codes: 317, 318, 318.0, 318.1, 318.2, and 319 Learning, motor, and communication disorders ICD9-CM codes: 307.1, 307.4x, 307.5x, 307.6–307.79, 347, 780.52–780.59, and 787.6 Obsessive-compulsive disorder (OCD) and ICD9-CM codes: 312–312.29, 312.4x–312.9x, 312.30, 312.34, 312.35, 313, 313.8, oppositional defiant disorder (ODD) 313.81, and 313.9x
502
www.psychosomaticsjournal.org
Psychosomatics 56:5, September/October 2015
Eworuke et al. APPENDIX.
Continued
Schizophrenia and other psychotic disorders National ADHD prevalence Category 1: Very low prevalence (r7.0%) Category 2: Low prevalence (7.1%–9%) Category 3: Average prevalence (9.1%–11%) Category 4: High prevalence (11.1%–13%)
ICD9-CM codes: 295.xx, 298.xx, and 297.xx Georgia, Idaho, Illinois, Minnesota, Nebraska, Texas, and Alabama Florida, Iowa, Kansas, Massachusetts, Mississippi, Missouri, New Hampshire, New York, New Jersey, Pennsylvania, Tennessee, Virginia, Vermont, and Wisconsin Arkansas, Indiana, Ohio, and South Carolina Louisiana, North Carolina, and West Virginia
References 1. Foundation CF. Cystic Fibrosis Foundation Annual Data Report. Available from: http://www.cff.org/uploadedfiles/ aboutcffoundation/annualreport/2011-annual-report.pdf. 2011. Accessed August 8, 2013. 2. Yankaskas JR, Marshall BC, Sufian B, Simon RH, Rodman D: Cystic fibrosis adult care: consensus conference report. Chest 2004; 125(1 Suppl):1S–39S 3. Smith BA, Wood BL: Psychological factors affecting disease activity in children and adolescents with cystic fibrosis: medical adherence as a mediator. Curr Opin Pediatr 2007; 19(5):553–558 4. Wells KB, Golding JM, Burnam MA: Psychiatric disorder in a sample of the general population with and without chronic medical conditions. Am J Psychiatry 1988; 145(8):976–981 5. (CDC) CfDCaP: Increasing prevalence of parent-reported attention-deficit/hyperactivity disorder among children— United States, 2003 and 2007. MMWR Morb Mortal Wkly Rep 2010; 59(44):1439–1443 6. Georgiopoulos AM, Hua LL: The diagnosis and treatment of attention deficit-hyperactivity disorder in children and adolescents with cystic fibrosis: a retrospective study. Psychosomatics 2011; 52(2):160–166 7. Harpin VA: The effect of ADHD on the life of an individual, their family, and community from preschool to adult life. Arch Dis Child 2005; 90(1 Suppl):i2–i7
Psychosomatics 56:5, September/October 2015
8. White T, Miller J, Smith GL, McMahon WM: Adherence and psychopathology in children and adolescents with cystic fibrosis. Eur Child Adolesc Psychiatry 2009; 18 (2):96–104 9. Teplin LA, Abram KM, McClelland GM, Dulcan MK, Mericle AA: Psychiatric disorders in youth in juvenile detention. Arch Gen Psychiatry 2002; 59(12):1133–1143 10. Winterstein AG, Gerhard T, Shuster J, et al: Utilization of pharmacologic treatment in youths with attention deficit/ hyperactivity disorder in Medicaid database. Ann Pharmacother 2008; 42(1):24–31 11. Froehlich TE, Lanphear BP, Epstein JN, Barbaresi WJ, Katusic SK, Kahn RS: Prevalence, recognition, and treatment of attention-deficit/hyperactivity disorder in a national sample of US children. Arch Pediatr Adolesc Med 2007; 161(9):857–864 12. Quittner AL, Barker DH, Snell C, Grimley ME, Marciel K, Cruz I: Prevalence and impact of depression in cystic fibrosis. Curr Opin Pulm Med 2008; 14(6):582–588 13. Cruz I, Marciel KK, Quittner AL, Schechter MS: Anxiety and depression in cystic fibrosis. Semin Respir Crit Care Med 2009; 30(5):569–578 14. Quon BS, Aitken ML: Cystic fibrosis: what to expect now in the early adult years. Paediatr Respir Rev 2012; 13(4): 206–214
www.psychosomaticsjournal.org
503
Published by Elsevier Inc. on behalf of The Academy of Psychosomatic Medicine.
Original Research Reports Clinical and Sociodemographic Factors Associated With Attention-Deficit/Hyperactivity Disorder in Patients With Cystic Fibrosis Efe Eworuke, Ph.D., Qian Ya Lensa Zeng, Pharm D., Almut G. Winterstein, Ph.D.
Background: There is scarce evidence on the epidemiology of attention-deficit/hyperactivity disorder (ADHD) in patients with cystic fibrosis (CF). Objective: We employed stepwise logistic regression to examine the association between ADHD diagnosis and selected patient characteristics. Methods: This was a cross-sectional analysis of inpatient and outpatient billing data for Medicaid beneficiaries with CF ages 3–18 years to obtain ADHD diagnosis prevalence and incidence estimates from 1999–2006. Results: Annual ADHD prevalence increased 1.55-fold from 5.26% (95% CI: 5.25–5.27) to 8.16% (8.15–8.17), and annual ADHD incidence rose slightly from 1.70% (1.70–1.71) to 2.01% (2.00–2.01). As in the general population, males were significantly more likely to have a diagnosis of ADHD compared with females (odds ratio: 1.97 [CI: 1.49–2.60]), as were children with recent diagnoses of anxiety, emotional disorder,
depression, adjustment disorder, and learning, motor, and communication disorders. Patients with ADHD diagnoses were also more likely to be in foster care (odds ratio ¼ 4.36 [CI: 2.26–8.40]). Except for recent DNase use (odds ratio ¼ 0.64 [CI: 0.43–0.93]), CF severity indicators and medications including pancreatic enzymes, inhaled tobramycin, inhaled or oral corticosteroids, inhaled bronchodilators, and oral antibiotics had no association with ADHD diagnosis. Conclusion: ADHD prevalence in CF increased during the study period. Clinical and sociodemographic determinants of ADHD diagnosis were similar to the general population, whereas treatment and severity of CF appeared to have little influence. Our findings warrant future research evaluating diagnostic protocols and assessment of safety and efficacy of ADHD treatment in children with CF. (Psychosomatics 2015; 56:495–503)
INTRODUCTION Cystic fibrosis (CF) is a rare but life-threatening genetic disease that affects approximately 30,000 people in the United States.1 Varying autosomal recessive genetic mutations translates into this multiorgan disease, which is characterized by respiratory infections, pulmonary exacerbations, pancreatic insufficiency, and liver impairment.2 Advanced clinical research and novel treatment strategies have extended life expectancy from 2 years when the disease was first discovered in 1958 to approximately 41.1 years as of 2012.1 Psychosomatics 56:5, September/October 2015
Received August 6, 2014; revised August 26, 2014; accepted September 2, 2014. From Department of Pharmaceutical Outcomes & Policy, College of Pharmacy, University of Florida, Gainesville, FL (EE, QYLZ, AGW); Department of Epidemiology, Colleges of Medicine and Public Health and Health Professions, University of Florida, Gainesville, FL (AGW). Send correspondence and reprint requests to Efe Eworuke, Ph.D., Division of Epidemiology II, Office of Pharmacovigilance and Epidemiology, Office of Surveillance and Epidemiology, Center for Drug Evaluation and Research, Food and Drug Administration, Silver Spring, MD; e-mail: [email protected] Published by Elsevier Inc. on behalf of The Academy of Psychosomatic Medicine.
www.psychosomaticsjournal.org
495
Risk Factors for ADHD in Cystic Fibrosis With a longer life expectancy, factors affecting long-term CF treatment, CF prognosis, and quality of life, as well as the management of concomitant diseases are becoming increasingly important. Patients with chronic diseases such as CF are at greater risk for psychiatric disorders, given the complexity surrounding therapy and burden of disease.3,4 With 8.8% of children in the United States currently diagnosed with attention-deficit/hyperactivity disorder (ADHD), it is plausible that this psychiatric condition may be as prevalent in the CF population.5,6 However, there are currently no epidemiologic estimates for ADHD in the CF population. ADHD impedes normal functioning and ability to carry out daily activities; therefore, for patients with CF, with an already burdened treatment regimen, ADHD could be detrimental to attaining overall treatment goals for the following reasons.7 First, psychiatric disorders have often been cited as a factor for poor treatment adherence in patients with CF.3 This is important because prognosis is contingent on adherence, which is already less than 50% in youths with CF.8 Secondly, with an increased life expectancy, it is important that both quality of life and self-worth be maintained throughout adulthood. Conditions such as ADHD not only affect day-to-day activities but also the ability to lead a functioning and successful life. Lastly, untreated ADHD can be associated with significant consequences, including an increased risk of academic failure, substance abuse, and delinquent and antisocial behavior.7,9 Thus, epidemiologic studies examining ADHD diagnosis and treatment pattern among the CF population will be important in understanding both the disease burden and respective treatment approaches. Therefore, we aimed to determine the prevalence and incidence of ADHD in a national cohort of Medicaid youths with CF and to evaluate the association between ADHD diagnoses and selected patient, clinical, and sociodemographic characteristics. METHODS Data Source The study data set comprises Medicaid administrative claims data, referred to as Medicaid Extract Files, from 28 US states (Appendix), which were selected based on Medicaid enrollment numbers, 496
www.psychosomaticsjournal.org
extent of managed care penetration, and CF prevalence as reported by the CF foundation.1 Medicaid Extract Files, provided by the Centers for Medicare and Medicaid Services, consists of person-level data files containing information on monthly eligibility, including reason for eligibility and health care coverage spectrum, detail on diagnoses and procedures associated with inpatient and outpatient care, as well as detail on dispensed medications. Because medical encounter detail for Medicaid beneficiaries in managed care plans is incomplete, we restricted our source population to patients who receive Medicaid benefits in the fee-for-service or primary care case management program. Patients who were enrolled in a behavioral or dental managed care plan were included if they were registered in the fee-for-service or primary care case management program. The institutional review and privacy boards of Centers for Medicare and Medicaid Services and the University of Florida approved the study. Study Population We included Medicaid fee-for-service recipients 3–18 years of age between January 1, 1999 and December 31, 2006, with at least 2 inpatient or outpatient diagnoses with International Classification of Disease version 9, Clinical Modification codes for CF (277.0x). Patients were also required to have a minimum of 12 months continuous Medicaid insurance for prevalence estimates. For incidence estimates, patients were required to have at least 24 months of continuous Medicaid eligibility covering the assessed and the previous year. Data Analysis Annual ADHD Incidence and Prevalence For each calendar year from 2000–2006, ADHD incidence was defined as the proportion of patients with CF with at least 2 outpatient claims for ADHD (International Classification of Disease version 9, Clinical Modification code 314.xx) in the sample of patients with CF who were not diagnosed with ADHD in the year before the year being assessed. Therefore, the source population from 1999 was used to determine the at-risk sample for the incidence estimates in 2000. ADHD prevalence from 1999–2006 was calculated as the proportion of patients with at least Psychosomatics 56:5, September/October 2015
Eworuke et al. 2 outpatient ADHD claims during the calendar year being assessed, using the total number of patients with CF with continuous Medicaid Extract Files eligibility in the measurement year as the denominator. Incidence rate ratios and prevalence ratios were obtained by comparing estimates of the last study year to the respective first study year to examine longitudinal trends. Given that patients could present in both the first year and the last year, it was important to take into account the potential correlation between yearly estimates. Therefore, we used the asymptotic standard error calculating the 95% CI using the Delta method.
included the following covariates: number of unique pharmacy prescriptions filled, number of CF-related outpatient visits, and the presence of at least 1 CFrelated hospitalization in the 6 months before the index date. We employed a multivariate conditional logistic regression to examine the association of each covariate and initial ADHD diagnosis. The model used a forward stepwise inclusion process with an α criterion of 0.50 to select covariates into the final model. RESULTS Source Population
Risk Factors Associated With ADHD For this analysis, ADHD cases were defined as patients with at least 2 outpatient claims for ADHD anytime during the study period. For each case, a control defined as patients without any ADHD claim during the study period was randomly selected from the CF cohort and matched on date of birth (⫾5 d) and study year. Both incident cases and controls were required to have at least 180 days of continuous eligibility before the index date; the index date for the cases was the date of second ADHD claim, whereas the controls adopted the index date for the matching case. The following sociodemographic characteristics were obtained from 6-month period before the index date: sex, race categorized as whites and others, state of residence, cash assistance, and Medicaid eligibility status categorized as poverty, disability, and foster care. Cash assistance is defined as receiving financial aid through state programs. The states were further classified into categories based on national estimates of ADHD prevalence (r7%, 7.1%– 9%, 9.1%–11%, and 11.1%–13%; Appendix).10 Clinical determinants of ADHD considered for analysis included the presence of at least 1 outpatient claim for any of the following psychiatric conditions: conduct disorder/opposition defiant disorder; emotional/ behavioral disorder; anxiety; adjustment disorder; depression; bipolar disorder; intellectual disability; learning, motor, and communication disorders; and schizophrenia (Appendix for operational definitions). Other considered factors comprise the use of commonly prescribed CF medications including use of inhaled tobramycin, pancreatic enzymes, inhaled bronchodilators, inhaled or oral corticosteroids, and oral antibiotics. To stage CF disease severity, we Psychosomatics 56:5, September/October 2015
We identified 19,760 patients with at least 2 inpatient or outpatient CF claims anytime during the study period. For the prevalence calculations, 6940 patients with a minimum of 12 months continuous Medicaid eligibility were examined. In each study year, males and whites accounted for slightly more than 50% and 65% of the population, respectively (Table 1). Mean age was stable across the study period, ranging from 9.79–9.99 years, as were the number of outpatient visits and hospital stays during follow-up. ADHD Incidence and Prevalence From 2000–2006, ADHD incidence rates increased slightly from 1.70% (95% CI: 1.70–1.71) in 2000 to 2.01% (CI: 2.00–2.01) in 2006. This increase in new ADHD cases was reflected in increasing ADHD prevalence from 5.26% (CI: 5.25–5.27) in 1999 to 8.16% (CI: 8.15–8.17) in 2006 (Figure). Although the ADHD incidence in 2006 was not statistically different from the incidence obtained in 2000 (incidence rate ratio: 1.17 [CI: 0.78–1.77]), the ADHD prevalence ratio was found to be statistically significant at 1.55 (CI: 1.29–1.87). Selected Factors Associated With ADHD Diagnoses During the study period, we identified 658 patients with 2 outpatient diagnoses of ADHD from the prevalence cohort with 6 months of continuous Medicaid eligibility before the second ADHD claim date. Of the 658 cases, we matched 99.1% (n ¼ 652) to a control patient. The following demographic differences emerged: cases comprised 16% more males compared with controls, reflecting the national trends for ADHD estimates.5,10 Although there were negligible differences in the racial composition of cases and www.psychosomaticsjournal.org
497
498
2.11 ⫾ 1.68 2.09 ⫾ 1.67 2.04 ⫾ 1.61 2.03 ⫾ 1.68 2.02 ⫾ 1.70
www.psychosomaticsjournal.org
IP ¼ inpatient; OP ¼ outpatient; SD ¼ standard deviation.
2.09 ⫾ 1.72 2.15 ⫾ 1.94 2.18 ⫾ 1.93
2001 3082 9.96 ⫾ 4.16 1631 (52.9) 1994 (64.7) 17.19 ⫾ 45.29 2000 2689 9.86 ⫾ 4.13 1417 (52.7) 1700 (63.2) 17.49 ⫾ 47.35 1999 2433 9.86 ⫾ 4.13 1281 (52.5) 1565 (64.3) 18.12 ⫾ 48.72
Variable Sample size Mean age Males, n (%) Whites, n (%) Mean number of OP visits (restricted to medical visits) (n ⫾ SD) Mean number of IP stays (restricted to unique hospital stays) (n ⫾ SD)
TABLE 1.
Patient and Health Care Utilization Characteristics, Stratified by Study Year
2002 3265 9.86 ⫾ 4.21 1761 (53.9) 2138 (65.5) 16.42 ⫾ 43.47
2003 3884 9.79 ⫾ 4.30 2102 (54.1) 2573 (66.3) 18.12 ⫾ 46.75
2004 3937 9.80 ⫾ 4.34 2145 (54.5) 2632 (66.9) 18.03 ⫾ 44.12
2005 4082 9.84 ⫾ 4.34 2268 (55.6) 2679 (65.6) 17.46 ⫾ 38.56
2006 4057 9.99 ⫾ 4.32 2262 (55.7) 2659 (65.5) 19.14 ⫾ 43.02
Risk Factors for ADHD in Cystic Fibrosis controls, the most prominent age group matched included 5–9-year-old children. There were also noted geographic differences between the states. States with the lowest ratio of cases to controls were Tennessee (3.76% [for cases] vs 7.99% [for controls]), Massachusetts (1.88% vs 3.92%), Pennsylvania (3.44% vs 4.55%), and Wisconsin (1.88% vs 2.04%). Conversely, states with the highest ratio comparison were Ohio (13.15% vs 6.11%), South Carolina (5.01% vs 4.08%) Georgia (9.08% vs 5.80%), and West Virginia (2.35% vs 1.41%). Lastly, there was a higher proportion of cases living in poverty (8.14% vs 5.49%) and in foster care (9.08% vs 2.19%) compared with controls (Table 2). In addition to demographic differences, we also observed clinical differences, specifically regarding psychiatric diagnoses, the use of certain CF medications, and CF-related hospitalizations. In general, cases had a greater proportion of psychiatric diagnoses compared with controls (Table 2). Regarding commonly-used CF medications, a lower proportion of cases received pancreatic enzymes, inhaled tobramycin, and DNase compared with controls. Cases were also less likely to be recently hospitalized, with a lower average of outpatient visits compared with controls (Table 2). Of the 32 considered patient characteristics, 21 entered into the final logistic regression model (Table 3). Males were more likely to be diagnosed with ADHD, as were children in foster care and with any of the considered psychiatric illnesses. Patients with a diagnosis of ADHD were also more likely to have a recent diagnosis of anxiety (odds ratio [OR] ¼ 3.94, CI: 1.46–10.63); adjustment disorder (OR ¼ 4.70, CI: 2.23–9.90); learning, motor, and communication disorders (OR ¼ 2.48, CI: 1.53–4.03); depression (OR ¼ 3.99, CI: 1.87–8.52); and emotional disorder (OR ¼ 4.36, CI ¼ 1.58–12.0). Among social factors, being in foster care was strongly associated with recent ADHD diagnosis (OR ¼ 6.13, CI: 2.91– 12.91). Although univariate comparisons suggest lesser disease severity among the cases, severity indicators including CF-related hospital stay were not significant. Among the commonly used CF medications, the sole negative determinant of ADHD was recent use of DNase (OR ¼ 0.66, CI: 0.44–0.98). DISCUSSION To our knowledge, this is the first large observational study adequately powered to examine ADHD Psychosomatics 56:5, September/October 2015
Eworuke et al.
FIGURE.
Incidence and Prevalence Estimates for ADHD in Patients With CF.
incidence and prevalence in CF with the ability to build a comprehensive predictive model determining the association between ADHD diagnosis and patient, clinical, and sociodemographic factors. The analysis reveals several key findings: first, ADHD prevalence in patients with CF increased during the study period and was comparable with the general pediatric population.5,10 Although prevalence estimates reported in our study are slightly lower than previously reported in a CF population, it is noteworthy to mention that the referenced study was conducted in a single center comprising 188 patients with CF.6 Our case definition of 2 ADHD-related claims may have also resulted in the exclusion of less severe forms of ADHD. Second, ADHD incidence estimates were slightly higher than rates reported in a large cohort study of Medicaid beneficiaries representing the general publicly insured pediatric population.10 It is plausible that differences in actual estimates are related to differences in the study period and population age. Although the large cohort study included patients ranging from 6 months to 20 years from 1995–2004, our study included patients 3–18 years of age from 1999–2006. Third, we observed significant demographic differences among patients with CF diagnosed with ADHD. Akin to previous studies, males were 2 times more likely to be diagnosed with ADHD compared with females.5,10 Although previous studies observed ADHD prominence among whites, our study did not find race a significant predictor of ADHD diagnosis. Consistent with previous studies, patients living in foster care were more likely to be diagnosed with Psychosomatics 56:5, September/October 2015
ADHD.5,11 A significant variation in the proportion of cases by state was observed, signaling that the patient’s state of residence may influence the likelihood of being diagnosed. Notably, states with the highest proportion of cases to controls (West Virginia, South Carolina, and Ohio) also had state-based prevalence rates that were higher than the national average in the general population presented in a study conducted by the Centers for Disease Control and Prevention and Health Resources and Services Administration.5 Patients diagnosed with ADHD were more likely to have other psychiatric diagnoses in the 6 months period before ADHD diagnosis. Although there is evidence suggesting higher prevalence of depression and anxiety in patients with CF owing to the psychologic and emotional toll of the disease, a causal relationship between ADHD diagnosis and other psychiatric conditions cannot be inferred from our study.10,12–14 It is important to note the lack of association between CF disease severity and ADHD diagnosis. Although patients newly diagnosed with ADHD were less likely to receive DNase, no other diagnosis determinants were found, even though univariate comparisons suggested lower disease severity among ADHD cases. In particular, recent use of other CF medications, CF-related hospitalizations, outpatient visits, and number of medications were not found to be significant. Further studies with the ability to operationalize disease severity using traditional indicators such as the forced expiratory volume in 1 second (FEV) and body mass index will be needed to confirm this finding. www.psychosomaticsjournal.org
499
Risk Factors for ADHD in Cystic Fibrosis TABLE 2.
Demographic, Patient, and Social Characteristics Comparing ADHD Cases and Controls
Covariates
ADHD cases (n ¼ 652)
Controls (n ¼ 652)
Males, n (%)
465 (71.32)
362 (55.52)
465 (71.32)
423 (64.88)
35 (5.37) 336 (51.53) 190 (29.14) 91 (13.95)
34 (5.21) 336 (51.53) 190 (29.14) 92 (14.11)
Proportion of patients with at least 1 inpatient or outpatient diagnosis during the 6-mo look-back period, n (%) Conduct disorder/opposition defiant disorder Anxiety Emotional/behavioral disorder Adjustment disorder Bipolar disorder Depression Learning, motor, and communication disorders Intellectual disability (mental retardation) Schizophrenia and other psychotic disorders Substance abuse Tics
18 (2.76) 31 (4.75) 39 (5.98) 66 (10.12) 26 (3.99) 44 (6.75) 82 (12.58) 26 (3.99) 4 (0.61) 5 (0.77) 5 (0.77)
1 (0.15) 7 (1.07) 6 (0.92) 11 (1.69) 1 (0.15) 11 (1.69) 36 (5.52) 22 (3.37) 0 (0.00) 0 (0.00) 0 (0.00)
Proportion of patients with chronic medication use during the 6-mo look-back period, n (%) Pancreatic enzymes Inhaled tobramycin DNase Inhaled/oral corticosteroids Inhaled bronchodilators Oral antibiotics (macrolides, ampicillins, cephalosporins, and quinolones)
258 106 141 255 379 435
326 144 194 241 396 417
Number of unique pharmacy claims filled during the 6-mo look-back period (mean ⫾ SD) Number of CF-related outpatient visits during the 6-mo look-back period (mean ⫾ SD)
21.46 ⫾ 18.97 10.67 ⫾ 20.94
23.00 ⫾ 21.33 12.34 ⫾ 24.01
State of residence, n (%)† Alabama Arkansas Florida Georgia Iowa Idaho Illinois Indiana Kansas Louisiana Massachusetts Minnesota Missouri Mississippi North Carolina Nebraska New Hampshire New Jersey New York Ohio Pennsylvania South Carolina Tennessee Texas
3 (0.46) 4 (0.61) 72 (11.04) 62 (9.51) 9 (1.38) 6 (0.92) 28 (4.29) 31 (4.75) 7 (1.02) 28 (4.29) 12 (1.84) 12 (1.84) 21 (3.22) 5 (0.77) 30 (4.60) 2 (0.31) 2 (0.31) 11 (1.69) 42 (6.44) 83 (12.73) 21 (3.22) 31 (4.75) 24 (3.68) 51 (7.82)
3 (0.46) 13 (2.04) 61 (9.36) 38 (5.83) 5 (0.77) 4 (0.61) 37 (5.62) 25 (3.83) 12 (1.84) 32 (4.91) 25 (3.83) 13 (2.61) 17 (2.61) 11 (1.69) 40 (6.13) 9 (1.38) 3 (0.46) 9 (1.38) 55 (8.44) 72 (6.44) 28 (4.29) 26 (3.99) 48 (7.36) 52 (7.98)
Race, n (%) Whites Age category, n (%) 45 y 5–9 y 10–14 y 14–18 y
500
*
www.psychosomaticsjournal.org
(39.57) (16.26) (21.63) (39.11) (58.13) (66.72)
(50.00) (22.09) (29.75) (36.96) (51.10) (63.96)
Psychosomatics 56:5, September/October 2015
Eworuke et al. TABLE 2. Continued Covariates
ADHD cases (n ¼ 652)
Controls (n ¼ 652)
6 (0.92) 24 (3.68) 10 (1.53) 15 (2.30)
5 (0.77) 19 (2.91) 12 (1.84) 8 (1.23)
Disease history, n (%) Recent CF-related hospital stay Cash assistance
74 (11.35) 339 (51.99)
84 (12.88) 340 (52.15)
Eligibility status Poverty Disability Foster care
497 (76.23) 267 (40.95) 54 (8.28)
485 (74.39) 311 (47.70) 14 (2.15)
Virginia Vermont Wisconsin West Virginia
ADHD ¼ attention-deficit/hyperactivity disorder; CF ¼ cystic fibrosis; SD ¼ standard deviation. n
†
Age was not included in the model, as it was a matching variable. Categories for states were included in the model.
Our study brings into light several issues for further discussion within the CF community. First, further examination will have to be undertaken to examine how clinicians decide on which patient with CF to treat for ADHD. This is especially important given the rising prevalence of ADHD among TABLE 3.
Logistic Regression Model Comparing ADHD Incident Cases With Non-ADHD Controls
Effect
Point 95% Wald estimate confidence limits
Race (Whites vs others) Gender (males vs females) Anxiety Emotional disorder Adjustment disorder Depression Intellectual disability Learning disorder Use of pancreatic enzymes Use of inhaled tobramycin Use of DNase Use of inhaled corticosteroids Use of antibiotics CF-related hospitalization stay Cash assistance Foster care Disability Poverty
1.27 2.00 3.94 4.36 4.70 3.99 0.76 2.48 0.85 0.87 0.66 1.23 1.20 0.77 1.49 6.13 0.69 1.36
0.94 1.52 1.46 1.58 2.23 1.87 0.37 1.53 0.61 0.58 0.44 0.93 0.90 0.51 0.96 2.91 0.47 0.90
1.71 2.64 10.63 12.01 9.90 8.52 1.56 4.03 1.20 1.30 0.99 1.63 1.61 1.16 2.30 12.91 1.02 2.07
State categories Category 2 vs category 1 Category 3 vs category 1 Category 4 vs category 1
0.75 1.38 1.06
0.54 0.93 0.66
1.05 2.06 1.70
CF ¼ cystic fibrosis.
Psychosomatics 56:5, September/October 2015
children with CF which is comparable to prevalence estimates reported in the general population. Because CF severity seems to play an insignificant role in deciding on a diagnosis for ADHD, it is crucial to investigate how clinicians decide on when to commence treatment and if and how risk-benefit assessments of ADHD treatment are made. Second, to inform risk-benefit decisions, concrete evidence on the safety and effectiveness of ADHD treatment in children with CF is needed. Children with CF face profound psychologic challenges, resulting in complex mental comorbidities, which may in turn affect the need to address ADHD. Although the effectiveness of stimulants in patients with CF has not been investigated, potential beneficial effects might include improved social functioning and parent or teacher relations as well as improved coping and medication adherence to their life-preserving CF treatment regimen. On the contrary, the complexity of CF medication regimen and the CF pathology itself offer the potential for a variety of drug interactions and adverse effects of ADHD treatment. Second, little is known about the influence of anorexic effects of stimulants, given the established correlation between weight and both CF progression and respiratory function. Other important considerations are potentiated cardiac side effects, bearing in mind an already increased cardiac risk, including early onset of diabetes mellitus in patients with CF, the potential cardiac risk of long-term macrolide therapy, and potential sympathomimetic drug-drug www.psychosomaticsjournal.org
501
Risk Factors for ADHD in Cystic Fibrosis interactions. A comprehensive understanding of both the effect of ADHD treatment of CF progression and life expectancy and on mental health and quality of life is critical to optimize care for this vulnerable population. In light of the prominence of ADHD in patients with CF, further research is urgently needed to evaluate the safety and effectiveness of ADHD treatment in children with CF. There are noteworthy limitations related to the data. Owing to the nature of administrative data, undercoding or miscoding of health care encounter information are inherent. To reduce diagnostic uncertainty, we required at least 2 ADHD claims to identify cases. In addition, gold standard markers for nutritional status and respiratory function in patients with CF, body mass index and forced expiratory volume, were not available and not included in the analysis. Instead, recent hospitalization stay, number of outpatient visits, and medications were used as proxy for disease severity. Finally, patient characteristics and differences in the care for patients with CF in states or private insurance not represented should be considered when interpreting our findings.
CONCLUSION Similar to the general pediatric population, we noted an increasing emergence of ADHD in this population of children with CF. Demographic and psychiatric determinants of ADHD were also similar to reports in the general population. Both CF treatment and severity appeared to have limited effect on ADHD diagnosis. Our study findings necessitate the need for future research to evaluate the prevalence of treatment of ADHD in CF and describe subpopulations that initiate treatment with the eventual goal of examining the safety and effectiveness of ADHD treatment in children with CF. This work was completed while all authors were affiliated with the University of Florida. The views expressed are those of the authors and not necessarily those of the US Department of Health and Human Services or the FDA. Disclosure: The authors disclosed no proprietary or commercial interest in any product mentioned or concept discussed in this article.
APPENDIX. Cohort Restrictions
Excluded Medicaid beneficiaries from Virginia from 1999–2002 owing to truncation of ICD9-CM codes Tennessee before January 2002 owing to the extent of managed care penetration Pennsylvania before June 2000 owing to the extent of managed care penetration.
Disease Severity At least one corresponding ICD9-CM diagnosis claim in the 6-mo period before the index date CF-related hospitalization stay 277.xx in primary or secondary diagnoses field in inpatient claims file Psychiatric disorder At least one ICD9-CM diagnosis claim for the correlating psychiatric disorder Adjustment disorder ICD9-CM codes: 309–309.20, 309.22–309.80, 309.82–309.99, and 313.89 Attention-deficit/hyperactivity disorder (ADHD) ICD9-CM codes: 314, 314.0, 314.00, 314.01, and 314.9 Anxiety ICD9-CM codes: 300.0–300.09, 300.2–300.29, 300.3, 309.21, 309.81, 313.0, 313.2x, 308.0, and 308.2x–308.9x Bipolar disorder ICD9-CM codes: 296.0–296.19, 296.4–296.81, 296.83–296.99, and 301.13 Conduct disorder ICD9-CM codes: 312.0, 312.1, 312.2, 312.3, 312.00, 312.01, 312.03, 312.10, 312.11, 312.12, 312.13, 312.20, 312.21, 312.22, 312.23, 312.30, 312.31, 312.32, 312.33, 312.34, 312.35, 312.39, 312.4, 312.8, 312.81, 312.82, 312.89, and 312.9 Depression ICD9-CM codes: 296.2x, 296.3x, 296.82, 300.4, and 311 Intellectual disabilities ICD9-CM codes: 317, 318, 318.0, 318.1, 318.2, and 319 Learning, motor, and communication disorders ICD9-CM codes: 307.1, 307.4x, 307.5x, 307.6–307.79, 347, 780.52–780.59, and 787.6 Obsessive-compulsive disorder (OCD) and ICD9-CM codes: 312–312.29, 312.4x–312.9x, 312.30, 312.34, 312.35, 313, 313.8, oppositional defiant disorder (ODD) 313.81, and 313.9x
502
www.psychosomaticsjournal.org
Psychosomatics 56:5, September/October 2015
Eworuke et al. APPENDIX.
Continued
Schizophrenia and other psychotic disorders National ADHD prevalence Category 1: Very low prevalence (r7.0%) Category 2: Low prevalence (7.1%–9%) Category 3: Average prevalence (9.1%–11%) Category 4: High prevalence (11.1%–13%)
ICD9-CM codes: 295.xx, 298.xx, and 297.xx Georgia, Idaho, Illinois, Minnesota, Nebraska, Texas, and Alabama Florida, Iowa, Kansas, Massachusetts, Mississippi, Missouri, New Hampshire, New York, New Jersey, Pennsylvania, Tennessee, Virginia, Vermont, and Wisconsin Arkansas, Indiana, Ohio, and South Carolina Louisiana, North Carolina, and West Virginia
References 1. Foundation CF. Cystic Fibrosis Foundation Annual Data Report. Available from: http://www.cff.org/uploadedfiles/ aboutcffoundation/annualreport/2011-annual-report.pdf. 2011. Accessed August 8, 2013. 2. Yankaskas JR, Marshall BC, Sufian B, Simon RH, Rodman D: Cystic fibrosis adult care: consensus conference report. Chest 2004; 125(1 Suppl):1S–39S 3. Smith BA, Wood BL: Psychological factors affecting disease activity in children and adolescents with cystic fibrosis: medical adherence as a mediator. Curr Opin Pediatr 2007; 19(5):553–558 4. Wells KB, Golding JM, Burnam MA: Psychiatric disorder in a sample of the general population with and without chronic medical conditions. Am J Psychiatry 1988; 145(8):976–981 5. (CDC) CfDCaP: Increasing prevalence of parent-reported attention-deficit/hyperactivity disorder among children— United States, 2003 and 2007. MMWR Morb Mortal Wkly Rep 2010; 59(44):1439–1443 6. Georgiopoulos AM, Hua LL: The diagnosis and treatment of attention deficit-hyperactivity disorder in children and adolescents with cystic fibrosis: a retrospective study. Psychosomatics 2011; 52(2):160–166 7. Harpin VA: The effect of ADHD on the life of an individual, their family, and community from preschool to adult life. Arch Dis Child 2005; 90(1 Suppl):i2–i7
Psychosomatics 56:5, September/October 2015
8. White T, Miller J, Smith GL, McMahon WM: Adherence and psychopathology in children and adolescents with cystic fibrosis. Eur Child Adolesc Psychiatry 2009; 18 (2):96–104 9. Teplin LA, Abram KM, McClelland GM, Dulcan MK, Mericle AA: Psychiatric disorders in youth in juvenile detention. Arch Gen Psychiatry 2002; 59(12):1133–1143 10. Winterstein AG, Gerhard T, Shuster J, et al: Utilization of pharmacologic treatment in youths with attention deficit/ hyperactivity disorder in Medicaid database. Ann Pharmacother 2008; 42(1):24–31 11. Froehlich TE, Lanphear BP, Epstein JN, Barbaresi WJ, Katusic SK, Kahn RS: Prevalence, recognition, and treatment of attention-deficit/hyperactivity disorder in a national sample of US children. Arch Pediatr Adolesc Med 2007; 161(9):857–864 12. Quittner AL, Barker DH, Snell C, Grimley ME, Marciel K, Cruz I: Prevalence and impact of depression in cystic fibrosis. Curr Opin Pulm Med 2008; 14(6):582–588 13. Cruz I, Marciel KK, Quittner AL, Schechter MS: Anxiety and depression in cystic fibrosis. Semin Respir Crit Care Med 2009; 30(5):569–578 14. Quon BS, Aitken ML: Cystic fibrosis: what to expect now in the early adult years. Paediatr Respir Rev 2012; 13(4): 206–214
www.psychosomaticsjournal.org
503
