Longitudinal Mental Health Service and Medication Use for ADHD Among Puerto Rican Youth in Two Contexts

Longitudinal Mental Health Service and MedicationUse for ADHD Among Puerto Rican Youth

in Two ContextsHECTOR R. BIRD, M.D., PATRICK E. SHROUT, PH.D., CRISTIANE S. DUARTE, PH.D., M.P.H.,

SA SHEN, PH.D., JOSE J. BAUERMEISTER, PH.D., AND GLORISA CANINO, PH.D.

ABSTRACT

Objective: The study describes prevalence and rates of services and medication use and associated factors over time

among Puerto Rican youths with attention-deficit/hyperactivity disorder (ADHD). Method: Longitudinal data are obtained

on Puerto Rican children ages 5 through 13 years in the south Bronx in New York (n = 1,138) and two metropolitan areas in

Puerto Rico (n = 1,353). The Diagnostic Interview Schedule for Children-IV is the diagnostic tool. Five composite measures

of risk factors: negative family influences, ineffective structuring, environmental risks, child risks, and maternal acceptance

are constructed to relate services andmedication use to risk variables.Results:ADHD prevalence is similar in Puerto Rico

and the south Bronx. Overall mental health services, medication, and psychostimulant use is lower in Puerto Rico across

three time points. Most participants never received treatment at any time point. More environmental risks, negative child

traits, and low maternal warmth are associated with more services, even after adjusting for comorbidity. When risk

variables are controlled, the effects of ADHD on services use decrease. Previous treatment is a strong predictor of

subsequent treatment.Conclusions: Rates of services and medication use are lower in Puerto Rico. Context seems to be

more important than ethnicity in predicting mental health services and medication use among Puerto Rican children with

ADHD. Other psychiatric diagnoses and general risk variables are important correlates of services and medication use.

J. Am. Acad. Child Adolesc. Psychiatry, 2008;47(8):879Y889. Key Words: epidemiology, attention-deficit/hyperactivity

disorder, psychostimulant use.

Attention-deficit/hyperactivity disorder (ADHD) is oneof the most common psychiatric disorders in the childand adolescent population, with reported prevalencesranging from 4% to 7% across different contexts andcultural settings.1 Symptoms indicating overactivity,

inattention, and impulsivity seem to cluster togethersimilarly in different cultural settings across the world,2,3

and there is no compelling reason to believe that thereare features of the syndrome that are unique to anycultural or ethnic tradition.The existing literature on the diagnosis and manage-

ment of ADHD is inconsistent regarding the rates ofmedication use across different geographic locations orethnic groups. Psychostimulants are widely recognizedas the treatment of choice for ADHD4,5 and are notused in treating other childhood conditions except forrare disorders such as narcolepsy. However, a largenumber of children in the community receive psychos-timulants without meeting full diagnostic criteria forADHD,6 albeit that such children have more symptomsof ADHD (elevated above those of untreated children).Although there are concerns that psychostimulantsare overprescribed, there is evidence to suggest the

Accepted March 10, 2008.Drs. Bird, Duarte, and Shen are with the Department of Psychiatry, Columbia

University; Dr. Shrout is with the Department of Psychology at New YorkUniversity; Drs. Canino and Bauermeister are with the Behavioral SciencesResearch Institute, University of Puerto Rico.

The primary study was supported by the NIMH through grants RO-1MH56401 (Dr. Bird, Principal Investigator). Dr. Canino is supported by grantP60MD002261-01 from the National Center for Minority Health Disparities(Dr. Canino, Principal Investigator). The present study was supported through acontract with McNeil Pediatrics Division of McNeil PPC, Inc.

Correspondence to Dr. Hector R. Bird, 1300 Luchetti Street, San Juan, PR00907; e-mail: [email protected].

0890-8567/08/4708-0879�2008 by the American Academy of Child andAdolescent Psychiatry.

DOI: 10.1097/CHI.0b013e318179963c

WWW.JAACAP.COM 879J . AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 47:8, AUGUST 2008

opposite.7 Still, a number of studies report that themajority of children with ADHD do not receivepsychostimulants.6,8

Some studies suggest that Hispanic populations areparticularly underserved when it comes to medicationuse,8Y10 but evidence of such disparities is notconsistent. This may not apply to Puerto Rican childrenbecause other studies report that no significant differ-ences exist between treatment patterns among PuertoRican children and children in other ethnic groups.6,11

In the Methods for the Epidemiology of Child andAdolescent Mental Disorders (MECA) study,12 nodifferences in the prescribing patterns of psychostimu-lants were noted between Puerto Rico and three othercommunities in the United States.13 More recently,Leslie et al.11 compared prescription patterns of clinicchildren with ADHD from two different mental healthsystems, one in Puerto Rico and the other in San Diego,CA. There were no significant site differences inmedication use by site. All three studies showed thatless than 15% of children meeting diagnostic criteria forADHD (Puerto Rican or otherwise), received medica-tion. This is a low rate compared with what had beenpreviously reported by other studies in the generalpopulation.8,9,14 One possible source of the variationbetween Latino and non-Latino groups is that thecomparisons tend to confound group membership withcontext. Latino attitudes toward medication may affecttreatment rates, but living as a minority member in animpoverished area may also affect treatment patterns. Inthis report, we aim to separate context from PuertoRican ethnic identity. Through secondary analyses ofexisting data, we describe the rates of mental healthservices and medication use and the related risk factorsamong community-based probability samples of PuertoRican youths in two different contexts, one in the southBronx, NY and one in San Juan, PR.The present analyses aim to answer four main

questions: Are there differences in the prevalence ofADHD among Puerto Rican children in two differentcontexts? To what extent are Puerto Rican children inthe two contexts receiving medication for ADHD? Howdo differences in context influence mental healthservices, medication, and psychostimulant use in chil-dren of the same ethnic background with and withoutADHD? What influence do risk factors have on mentalhealth services, medication, and psychostimulant useamong Puerto Rican children in the two contexts? Our

hypotheses are that prevalence will be similar regardlessof context, that a smaller proportion of children inPuerto Rico receives mental health services and medica-tion including psychostimulants than Puerto Ricanchildren in the south Bronx, and that an explanationfor such variation in medication use would be differencesin the levels of associated risks in the two situations.

METHOD

Bird et al.15Y17 provide detailed information about the parentstudy`s methodology and initial findings. In brief, this NIMH-funded longitudinal study was designed to assess the prevalence,associated comorbidities, and correlates of disruptive behaviordisorders among Puerto Rican children in two contexts: the southBronx (n = 1,138) and the standard metropolitan areas in San Juanand Caguas, PR (n = 1,353; total N = 2,491). Children ages 5 to 13years at baseline were followed over three waves of data collection 1year apart (mean 349.6 days, SD 54.6). Each sample is a multistageprobability sample of households of the target population, and eachcan be weighted to represent the populations of Puerto Ricanchildren in the south Bronx and Puerto Rico contexts. A householdwas eligible for the study if there was at least one child residing in thehousehold ages 5 through 13 years identified as being of PuertoRican background and at least one of the child`s parents or primarycaretakers also self-identified as being of Puerto Rican background.All of the eligible children were selected to participate up to amaximum of three children per household.At each assessment, children and caretakers were interviewed

in the language of their choice (English or Spanish), using theDiagnostic Interview Schedule for Children-IV (DISC-IV)18,19 anda wide array of other measures previously described by Bird et al.15

The DISC-IV was not administered to children youngerthan 10 years because the reliability of younger child informantsis questionable.20Y24 For the purposes of our longitudinal analyses,including those in the present report, cases of all ages were definedusing only data obtained from the adult informant on the DISC-IVeven though all of the adult informants and children who were10 years or older were administered the DISC-IV. The followingdiagnoses were ascertained: conduct disorder and oppositionaldefiant disorder, attention deficit hyperactivity disorder (ADHD),major depression, dysthymia, anxiety disorders (separation anxiety,panic, and generalized anxiety disorders, social phobia and PTSD),and substance use disorder.The test-retest reliability of parent reports of ADHD using the

DISC-IV in English or Spanish has been reported.18,19 Test-retestreliability for ADHD was acceptable (W = .49 for the Spanish DISC-IV and W = .60 for the English DISC-IV), representing fair to goodagreement. Agreement between lay interviewer and clinicianadministered DISC-IV as well as with clinical judgment for theADHD schedule has fluctuated from fair to good.24,25

Procedures

Informed consent from the adult and assent from the youths wereobtained. Consent forms and procedures were approved by the NewYork State Psychiatric Institute and the University of Puerto RicoMedical School institutional review boards. Each informant wasinterviewed by separate trained lay interviewers using laptopcomputers. Characteristics of the interviews have been previously

BIRD ET AL.

880 WWW.JAACAP.COM J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 47:8, AUGUST 2008

reported,15,17 Data collection for the three-wave study spanned aperiod of almost 4 years averaging approximately 1 year betweenwaves for each respondent. Retention in the three-wave panel designwas excellent and has been previously reported.17

Diagnostic Status. For the present analyses, we constructed threecategories of diagnostic status: neither current ADHD nor ADHDnot otherwise specified (NOS), current ADHD, and currentADHD-NOS. The ADHD-NOS category included children withinattention and/or hyperactivity-impulsivity having clinically sig-nificant impairment, but whose symptom pattern fell below thesymptomatic threshold for the disorder. For ADHD-NOS, the ageat onset criterion was not considered. To examine the impact ofcomorbid psychiatric disorders, for some analyses, children weresubdivided into those who met criteria for one or more of the otherdisorders considered in the study and those with none of thediagnoses included. All of the positive diagnoses include impairedfunctioning in at least one setting.

Mental Health Services and Medication Use.Mental health servicesuse for emotional, behavioral, alcohol, or drug problems during theprevious year was ascertained. This use included inpatient, out-patient, day treatment, special education, or school counseling. Thenames of psychotropic medications were recorded in the dataset andclassified by the first author. Of approximately 30 medicationsrecorded, five were classified as psychostimulants, two amphetamineproducts (Adderall and Dexedrine) and three methylphenidatepreparations (Ritalin, Concerta, and Focalin). Other medicationsconsisted of selective serotonin reuptake inhibitors or otherantidepressants, major and minor tranquilizers, mood stabilizers oranticonvulsants, and antihypertensives. Caretakers were askedwhether they agreed or would agree to their child receivingmedications prescribed for behavior problems or problems withhyperactivity or paying attention, with three response options(Byes^, Bno,^ or Bmaybe^).

Risk Variables. As described in detail by Bird et al.,15 informationabout parenting and environment- and child-based risks wereobtained at the three time points from the caretaker and fromchildren. We considered all of the risk variables included in the study,except for neighborhood and cultural factors, which we did notconsider to be comparable between sites (e.g., acculturation, culturalstress, neighborhood characteristics). The remaining measures weregrouped in four conceptual dimensions: negative family influences,ineffective structuring variables, child risks, and environmental risks.Data reduction analyses led us to focus on 18 specific risk measuresthat have been previously described15 related to these domains. Belowwe provide the main reference source for each measure.Negative family influences included family functioning,26 parent

social support,27 religiosity,28 and single-parent family. Ineffectivestructuring included parental monitoring,29 parental discipline,30

and parental attitudes toward delinquency.31 Environmental risksincluded exposure to violence,32,33 school environment,15 peerdelinquency,31 stressful life events30,34 and an adaptation of theHome Environment Scale.35 Negative child characteristics includedearly aggressivity, poor social adjustment, poor academic perfor-mance, lack of talents and special skills, and peer relationships.36 Inaddition to these four summary variables, we included maternalacceptance/warmth, based on the Hudson Index of ParentalAttitudes.37 Although this variable did not emerge from the datareduction screen, it was a key variable in cross-sectional analysesreported by Bird et al.16 It includes items related to the childYparentrelationship and to parental burden.All of the risk variables retained for further analysis had Cronbach

" values Q.54. Specific values are reported in Bird et al.15 To

represent negative family influences, ineffective structuring variables,environmental risks, and negative child characteristics, we usedregression-based factor score estimates from factor analyses of bothsamples combined. Across both samples, the mean of these scoreswas centered at zero. To make it comparable, the maternalacceptance/warmth scale was standardized to have mean zero andvariance one in the combined sample. The risk variables are highlycorrelated over the 3 years (median correlation of the risks variablesacross the three waves is 0.64 in Puerto Rico and 0.54 in the southBronx) and tend to represent situations that are relatively stable. Forthis reason, we present the risks as the average across the three waves.

Statistical Analyses

Weadjusted for differences in the probability of selection due to thesample design and differences from the 2000 U.S. Census in theage/sex distribution. For estimates of means, rates, and correlationsamong variables, we used SUDAAN software (release 8)38 to computeweighted estimates and to adjust SEs for intraclass correlationsinduced by multistage sampling, with children nested withinhouseholds and households nested within primary sampling units.Because youths in the two sites may have differed in multiple ways

besides living in different locations, we carried out a propensity scoreadjustment in the analysis.39 This method creates groups of PuertoRican and south Bronx youths that are comparable on a number offactors, including a combination of maternal age, maternaleducation, and family income. We used logistic regression to predictsite. This allowed us to identify south Bronx youths who hadcharacteristics similar to those of Puerto Rican youths and also theopposite, Puerto Rican youths who had characteristics similar tothose of south Bronx youths. We used a total of five groups, whichwe refer to as propensity score strata. Indicators of these strata wereincluded in the analyses to adjust for site differences.At each wave of data, we estimated the rates of mental health

services use, all psychotropic medication use, and specificallypsychostimulant use among the three groups of children. Theseanalyses reveal the extent of unmet need for ADHD in particular.They are carried out separately for Puerto Rico and the south Bronx.We developed descriptive models that relate services and medicationuse to demographic characteristics and risk variables to betterunderstand the patterns of use and possible disparities through time.These models used logistic regression with three binary outcomes:services (versus none), medication (versus none), and psychostimu-lants (versus none). Explanatory variables included child`s age, sex,and diagnostic status and risk variables as defined above. Included inthese variables are constructs relevant to services/medication use,including demographics, parental burden, school problems, externa-lizing problems, impairment, and attitudes toward medication.13,40

The data across time were included in the analysis and linear timetrends were examined. To take into account dependencies amongrepeated measures, we used generalized estimating equations41 asimplemented in the GENMOD procedure of SAS.42 The sameanalysis was also conducted in a subsample that included onlychildren with ADHD or ADHD-NOS.

RESULTS

Descriptive Results

The prevalences of both ADHD and ADHD-NOS(Table 1) are similar in Puerto Rico and the south Bronxfor the three waves. Mental health services use, however,

SERVICE AND MEDICATION FOR ADHD PR YOUTH


is substantially lower among children in Puerto Ricorelative to the south Bronx across all three time points.All of the site differences are significant (p < .0001).Services use among children with ADHD is morethan twice the overall services use rate. Among thosewith ADHD-NOS, the results are similar. There is atendency for ADHD-NOS children to receive fewerservices in Puerto Rico than in the south Bronx, withfewer than half in Puerto Rico receiving services.Overall rates of medication use are also significantly

lower in Puerto Rico relative to the south Bronx at allthree waves (wave 1: Puerto Rico 2.8%, south Bronx5.7%; wave 2: Puerto Rico 3.1%, south Bronx 5.9%;wave 3: Puerto Rico 3.3%, south Bronx 5.8%). Amongchildren with ADHD, there are no statistically sig-nificant site differences in medication use. Despite themoderate rates of general services use among childrenwith ADHD, the proportion of children with ADHDwho receive medications is only about one in four ineach site.

Like general services and medication use, there is lessoverall use of psychostimulants in Puerto Rico relativeto the south Bronx. This difference is significant inonly the first two waves, although the same tendency isapparent in wave 3. Although psychostimulants are thetreatment of choice for children with ADHD, relativelyfew of those with the diagnosis receive psychostimu-lants in either site. Similar findings were obtained forthose with ADHD-NOS; site differences were notsignificant and few of those with ADHD-NOS werereceiving psychostimulants. Indeed, among childrenwith ADHD who received medication of any kind, theproportion whose prescription was for psychostimu-lants was as large or larger in Puerto Rico than that inthe south Bronx (results not shown).

Table 1 also shows site differences of an importantindicator of services/medication use that is discussed ingreater detail below, namely, caretaker attitudes towardmedications for their children. In the south Bronx, themajority of caretakers expressed a negative attitude

TABLE 1Prevalence of ADHD, Service, Medication, Psychostimulant Use, and Caretaker Attitudes in Two Contexts

Puerto Rico South Bronx Site Difference

Wave 1 Wave 2 Wave 3 Wave 1 Wave 2 Wave 3 Wave 1 Wave 2 Wave 3

Est % SE Est % SE Est % SE Est % SE Est % SE Est % SE Diff Diff Diff

Prevalence of ADHD 6.4 0.8 4.7 0.6 4.2 0.7 7.1 0.9 6.0 0.7 4.5 0.7 j0.7 j1.3 j0.3Prevalence of ADHD-NOS 7.3 0.8 5.2 0.7 4.4 0.6 7.3 0.7 5.6 0.6 4.1 0.6 0.0 j0.4 0.3Service use 12.5 1.1 14.3 1.3 13.6 1.3 19.2 1.2 22.5 1.3 21.5 1.4 j6.7y j8.2y j7.9yMedication use 2.8 0.5 3.1 0.5 3.3 0.5 5.7 0.8 5.9 0.7 5.8 0.8 j2.9** j2.8*** j2.5**Psychostimulant use 1.5 0.3 1.9 0.4 1.8 0.4 3.4 0.6 3.0 0.5 2.9 0.6 j1.9** j1.1a j1.1Service use among ADHD 40.7 5.8 49.1 6.8 53.8 8.2 51.0 5.6 60.9 5.8 64.0 6.9 j10.3 j11.8 j10.2Service use amongADHD-NOS

19.5 5.0 41.9 6.2 42.4 7.2 47.0 4.6 56.5 5.9 50.8 6.5 j27.5y j14.6a j8.4

Medication useamong ADHD

15.9 4.6 25.5 6.2 22.2 6.9 22.9 5.3 20.9 5.0 28.1 6.9 j7.0 4.6 j5.9

Medication use amongADHD-NOS

8.2 3.3 10.4 4.1 20.9 5.7 12.9 3.1 21.4 4.9 18.7 5.2 j4.7 j11a 2.2

Psychostimulant useamong ADHD

8.8 3.4 18.2 5.2 9.5 4.1 13.8 3.9 11.9 4.0 17.3 5.3 j5.0 6.3 j7.8

Psychostimulant useamong ADHD-NOS

7.0 3.1 9.2 3.8 11.2 4.1 7.4 2.5 6.0 2.9 7.7 3.8 j0.4 3.2 3.5

Caretaker attitudetoward medication

No 42.9 1.9 40.4 2.1 39.4 2.2 65.7 1.8 57.3 1.7 57.7 1.8 j22.8y j16.9y j18.3yMaybe 27.7 1.8 31.8 2.0 34.2 2.2 14.3 1.1 20.4 1.5 16.3 1.3 13.4y 11.4y 17.9yYes 29.4 2.1 27.8 2.1 26.4 2.5 20.0 1.6 22.3 1.6 26.1 1.5 9.4y 5.5a 0.3

Note: Site differences are tested across each corresponding wave. Est = estimated; Diff = difference; ADHD = attention-deficit/hyperactivitydisorder; NOS = not otherwise specified.**p < .01; ***p < .001; yp < .0001. ap = .06Y.10.

BIRD ET AL.


toward medication use, whereas in Puerto Rico, theproportions were significantly smaller.

Multivariate Models of Longitudinal Data

We next considered services/medication patterns acrossall three waves in multivariate linear models that take intoaccount site, sex, and age (centered at 10), as well as diag-nostic and risk variables. These models help to estimatethe relative importance of these predictor sets on services,medication, and psychostimulant use in each of the sites.In Table 2, we present results from two versions of themodel for each of these outcomes. The first considers onlydemographic and diagnostic variables, whereas the secondadjusts for known child/family risk variables.

After adjusting for demographic variables and con-sidering the data from all time points, the site difference

in overall services use remains statistically significant(" = j.63, SE 0.11, p < .001). Children in the southBronx are 1.88 times more likely to report services usethan those in Puerto Rico. Across both sites, there wasno evidence that services use varied for wave of study,but males were more likely to use services and age atbaseline was related to services use in a nonlinearpattern. Children who were younger at baseline wereless likely to use services, but by age 10, there was littledifference in the likelihood of services use withincreasing age. Those with ADHD and ADHD-NOSwere approximately three times more likely to useservices than those with no ADHD (" = 1.22, SE 0.18,p < .001, odds ratio [OR] 3.4; " = 1.08, SE 0.16, p <.001, OR 2.9), but the ADHD and ADHD-NOSgroups did not differ significantly from each other.

TABLE 2Multivariate Prediction of Service and Medication and Psychostimulant Use in Puerto Rican Children

Predictor

Service Use Medication Use Psychostimulant Use

Model 1 Model 2 Model 1 Model 2 Model 1 Model 2

Estimate SE Estimate SE Estimate SE Estimate SE Estimate SE Estimate SE

SitePuerto Rico j0.63*** 0.11 j0.34** 0.13 j1.12y 0.22 j0.99y 0.25 j0.96*** 0.28 j0.90** 0.30South Bronx 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00

Linear time 0.05 0.04 0.07 0.04 0.02 0.06 0.03 0.07 j0.03 0.10 j0.03 0.13SexFemale j0.71y 0.10 j0.61y 0.11 j1.00y 0.24 j1.00y 0.23 j1.00** 0.32 j0.92** 0.31Male 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00

Age 0.03 0.02 0.03 0.02 0.02 0.05 0.02 0.05 j0.08 0.06 j0.08 0.06Age squared j0.02* 0.01 j0.02* 0.01j0.05** 0.02 j0.05** 0.02 j0.08*** 0.02 j0.09y 0.02ADHDADHD 1.22y 0.18 0.34 0.21 0.75** 0.28 0.19 0.34 0.80* 0.37 0.10 0.46ADHD-NOS 1.08y 0.16 0.68y 0.17 1.13*** 0.30 0.18 0.28 0.92** 0.34 0.50 0.33Never ADHD 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00

Any other diagnosis with impairmentYes 0.71y 0.17 0.19 0.20 0.65* 0.30 0.10 0.32 0.60 0.37 j0.03 0.41No 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00

Caretaker medication attitudeYes 0.48y 0.12 0.34** 0.12 2.33y 0.26 2.23y 0.27 2.53y 0.36 2.47y 0.37Maybe 0.08 0.14 0.01 0.14 0.74* 0.34 0.71* 0.35 0.52 0.50 0.48 0.52No 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00

Risk variablesNegative family influence j0.21** 0.07 j0.32** 0.12 j0.36* 0.16Ineffective parent structuring j0.12 0.06 j0.21 0.13 j0.22 0.14Environmental risks 0.20y 0.05 0.09 0.10 0.02 0.14Negative child characteristics 0.77y 0.07 0.61y 0.13 0.50** 0.16Low maternal warmth 0.27y 0.06 0.50y 0.11 0.50*** 0.14

Note: Propensity scores were adjusted in the models. ADHD = attention-deficit/hyperactivity disorder; NOS = not otherwise specified.*p < .05; **p < .01; ***p < .001; yp < .0001.



Those with another diagnosis besides ADHD alsoreported receiving more services, even after adjusting forcomorbid ADHD (" = .71, SE 0.17, OR 2.0). Table 2also shows that children whose parents would considermedication use for their treatment were also more likelyto receive services (" = 0.48, SE 0.12, p < .001, OR 1.6).The second multivariate model adjusted for the five

risk variables in addition to the variables just reviewed.Four of these five variables were related to reportedservices use. Children in families with increased negativefamily influence scores were less likely to receive services(" = j.21, SE 0.07, p < .01), whereas those exposed toelevated environmental risks (" = .20, SE 0.05, p <.0001), who themselves had increased negative childcharacteristics (" = .77, SE 0.07, p < .0001) and whohad decreased maternal warmth (" = .27, SE 0.06, p <.0001) were more likely to receive mental healthservices. These associations were adjusted for ADHDdiagnosis. However, it is interesting to note that themagnitudes of the associations of ADHD and ADHD-NOS with services use were dramatically reduced whenadjusting for the risk processes. After adjustment,ADHD was no longer significant (" = .34, SE 0.21,not significant [NS], OR 1.4) and ADHD-NOS wasreduced by one third (" = .68, SE 0.17, p < .001, OR2.0). The adjustment also eliminated the association ofother diagnoses (" = .19, SE 0.20, NS) and reduced thedifference between the sites by half (" =j0.34, SE 0.13,p < .01, OR 0.72).Next, we consider multivariate models of medication

use. Model 1 in Table 2 reveals substantial site (" =j1.12, SE 0.22, p < .0001, OR 0.33) and sexdifferences (" = j1.0, SE 0.24, p < .0001, OR 0.37)in medication use, with less use in Puerto Rico and lessuse for females at both sites. There was no change infollow-up time in medication use, but there was apattern with age at time of recruitment that resembledthe pattern for overall services use. Children withADHD (" = .75, SE 0.28, p < .01, OR 2.12) andADHD-NOS (" = 1.13, SE 0.30, p < .001, OR 3.10)were more likely to receive medications, as were thosewith other diagnoses (" = .65, SE 0.30, p < .05, OR1.92). Medication use was strongly related to caretakers`reports of openness to pharmaceutical treatment, bothstated unequivocally (Byes^; " = 2.33, SE 0.26, p < .001,OR 10.28) and ambivalently (Bmaybe^; " = .74, SE0.34, p < .05, OR 2.10). When the five child/family riskvariables were adjusted in model 2, the associations of

medication use to ADHD diagnoses (ADHD: " = .19,SE 0.34, NS; ADHD-NOS: " = .18, SE 0.28, NS) andto other diagnoses (" = .10, SE 0.32, NS) were againreduced to nonsignificance. In contrast, the effects forsite, sex, and caretaker attitudes toward medication werenot affected much by the adjustment. Among the fiverisk variables that were adjusted, three were relatedto medication use: negative family influence (" = j.32,SE 0.12, p < .01), negative child characteristics (" =.61, SE 0.13, p < .0001), and low maternal warmth (" =.50, SE 0.11, p < .0001). Environmental risks, which hadbeen significantly related to overall services use, had noassociation with medication use (" = .09, SE 0.10, NS).The final two columns of Table 2 show results for

psychostimulant use (vs. no psychostimulant use). Likemedication use, psychostimulants are less used in PuertoRico (" = j.96, SE 0.28, p < .001, OR 0.38) and bygirls (" = j1.0, SE 0.32, p < .01, OR 0.37). ADHD(" = .80, SE 0.37, p < .05, OR 2.23) and ADHD-NOS(" = .92, SE 0.34, p < .01, OR 2.51) are significantlyrelated to psychostimulant use, but other diagnosis isnot (" = .60, SE 0.37, NS; OR 1.82). Strong caretakeracceptance of pharmaceutical treatment is stronglyrelated to psychostimulant use (" = 2.53, SE 0.36, p <.0001, OR 12.55), but ambivalent acceptance was not(" = .52, SE 0.50, NS; OR 1.68). In model 2, negativefamily influence (" = j.36, SE 0.16, p < .05), negativechild characteristics (" = .50, SE 0.16, p < .01) and lowmaternal warmth (" = .50, SE 0.14, p < .001) wererelated to psychostimulant use, but ineffective parentstructuring and environmental risks were not. Whenthese risks were adjusted, the associations of ADHD andADHD-NOS with psychostimulant use were reducedin magnitude and were no longer significant.We reexamined these predictors of services and

medication use in the subsample of children whoreceived either the ADHD or ADHD-NOS diagnosis.These results are shown in Table 3. Overall, the patternsof results were similar to what we found in the totalsample. Overall services was more likely to be providedto children with ADHD in the south Bronx than inPuerto Rico, to males, to those who had comorbiddisorders, and to those whose parents expressed open-ness to medication interventions. When individual riskvariables were included, the diagnosis of a comorbiddisorder no longer predicted services use. The same riskvariables predict services use among children withADHD as in the total population: Exposure to

BIRD ET AL.


environmental risks, negative child characteristics, andlow maternal warmth made services use more likely,whereas negative family influence was inversely relatedto the likelihood of services use. Similar patterns werefound for medication use. Important factors were site,sex, and caretaker attitude toward medication. Hereagain, comorbidity was not related to medication useonce the sample was limited to those with ADHD.

The results for the ADHD subgroup were mostdifferent from the general sample for the prescription ofmedication in general and psychostimulants specifically.There is no evidence of a site difference in psychosti-mulant use, and the strongest predictor was caretaker`sattitude toward medication. Medication and psychosti-mulant use seems to be more a function of thecaretaker`s attitude toward medication use than thepsychiatric needs of the child. Children with ADHDwhose caretakers were amenable to medication use(responding Byes^ or Bmaybe^ to their acceptance of

using medications) were 11 times more likely to receivepsychostimulants or other medications than those whowere averse to this type of intervention. Amongcaretakers of children with ADHD receiving medica-tion, an average (for three waves) of 90.5% in PuertoRico and 92.0% in the south Bronx were amenable tomedication use (Bmaybe^ or Byes^ responses), but thesame was true of fewer caretakers (65.8% in Puerto Ricoand 48.0% in the south Bronx) of children with ADHDnot receiving medications. It is evident that thecaretaker`s attitude toward medication use has a majorbearing on whether a child with ADHD receivesmedication.Among the composite risk variables, only two were

predictive. Those with ADHD were less likely to receivepsychostimulants if there was dysfunction in the family(negative family influences), and they were more likelyto receive them if there was an impaired relationshipwith the caretaker (low maternal acceptance/warmth).

TABLE 3Multivariate Prediction of Service, Medication and Psychostimulant Use in Puerto Rican Children With Attention-Deficit/Hyperactivity

Disorder/Attention-Deficit/Hyperactivity Disorder Not Otherwise Specified at Baseline

Predictor

Service Use Medication Use Psychostimulant Use

Model 1 Model 2 Model 1 Model 2 Model 1 Model 2

Estimate SE Estimate SE Estimate SE Estimate SE Estimate SE Estimate SE

SitePuerto Rico j0.89y 0.20 j0.66** 0.24 j0.85* 0.37 j0.88* 0.39 j0.41 0.38 j0.62 0.41South Bronx 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Linear time 0.03 0.07 0.05 0.08 j0.12 0.09 j0.13 0.09 j0.29* 0.12 j0.30* 0.13SexFemale j0.83*** 0.23 j0.69** 0.22 j1.38* 0.56 j1.32** 0.45 j1.05 0.61 j1.06* 0.54Male 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Age 0.06 0.05 0.07 0.05 0.07 0.08 0.03 0.07 0.05 0.09 j0.03 0.10Age squared 0.00 0.02 j0.01 0.02 0.00 0.03 j0.03 0.04 0.00 0.03 j0.03 0.04Any other diagnosis with impairmentYes 0.47* 0.20 j0.19 0.22 0.59 0.32 j0.05 0.37 0.34 0.36 j0.19 0.41No 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Caretaker medication attitudeYes 0.55* 0.22 0.36 0.23 2.53y 0.49 2.42y 0.51 2.49y 0.55 2.40y 0.57Maybe j0.19 0.28 j0.12 0.30 0.34 0.65 0.38 0.69 j0.49 1.02 j0.68 0.89No 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Risk variablesNegative family influence j0.36** 0.11 j0.56** 0.19 j0.41* 0.21Ineffective parent structuring j0.06 0.10 j0.17 0.23 j0.17 0.18Environmental risks 0.25** 0.10 0.17 0.15 0.12 0.15Negative child characteristics 0.67y 0.12 0.43* 0.19 0.00 0.21Low maternal warmth 0.22* 0.10 0.55** 0.17 0.65*** 0.18

Note: Propensity scores were adjusted in the models.*p < .05; **p < .01; ***p < .001; yp < .0001.



There was no indication that negative child character-istics increased the likelihood that psychostimulantswould be prescribed. The sex effect was diminished inthis analysis and was only significant when individualrisk variables were adjusted.

DISCUSSION

There were no significant differences in the pre-valence of ADHD among Puerto Rican children in thesouth Bronx and in Puerto Rico. Prevalence falls withinthe range reported in studies carried out in otherpopulations.2 The finding was consistent at each of thethree waves of the study. Given that ADHD isconsidered to be a chronic disorder that is biologicallydetermined, a similar prevalence across both sites and aprevalence comparable to that found in other settingswas expected.Overall use of mental health services is significantly

lower in Puerto Rico than in the south Bronx, even afteradjusting for demographic variables. This had beenpredicted based on previous research findings.6,12,43

Although arguably language could be a barrier to mentalhealth services use in the south Bronx, it does not seemto be a relevant factor in the south Bronx populationstudied. Despite the fact that all of the interviewers werefully bilingual and most were Hispanic/Puerto Ricanthemselves, all of the Puerto Rican children in the southBronx sample and three fourths of their caretakers choseto respond to the interviews in English, suggesting thatin the south Bronx, we were dealing with a populationthat is fairly assimilated. Previous studies have showndisparities in access to care and in the availability ofmental health services in Puerto Rico as compared toother communities in the mainland.12,43 In our study,we do not have comparable data from other populationsto address this issue, but our findings indicate thatPuerto Rican children who live on the island aresubstantially less likely to receive all levels of treatment,including any mental health services, any medication, orpsychostimulant medication, than Puerto Rican chil-dren who live in the south Bronx. The rates of treatmentwith psychotropic medications in the population as awhole are half as large in Puerto Rico. These differencesremain after adjusting for ADHD spectrum diagnosis,other diagnoses, parental attitudes toward medicationuse, and other risk variables. We consider that thedisparity has to do with services system differences

(greater availability of services in the south Bronx) and iscontrary to the notion that minority status is thedetermining factor, given that a larger proportion of aminority group (Puerto Rican in the south Bronx) aremore likely to receive services than Puerto Ricanchildren in Puerto Rico (where they are the majority).Our results also showed a services use sex disparity; at

both sites males received more mental health servicesthan females, even when controlling for psychopathol-ogy, age, caretaker attitudes, and a number ofpsychosocial risk factors. The literature on treatmentdisparities based on sex has been equivocal withsome studies showing boys receiving more services,44Y47

and others not finding sex differences in servicesutilization.48

There appears to be greater acceptance of medicationsfor treating behavioral problems by caretakers in PuertoRico compared to those in the south Bronx. This isconsistent with the authors` impression that there isvirtually no major negative press (newspaper articles ortelevision reports) regarding the use of stimulants inPuerto Rico. In contrast, there were prominent reportsin the New York press during the study period.49

Despite greater acceptance of medication use in thepopulation as a whole, there is significantly lowermedication use in Puerto Rico (Table 3), and thedifference in psychostimulant use is suggested, althoughnot statistically significant. This finding cannot beassociated with parental attitudes toward medication usebecause parents in Puerto Rico are less averse to the useof medication than those in the south Bronx, and it ismore likely to be related to the disparities in the mentalhealth delivery systems noted above.The Multimodal Treatment Study of Children With

ADHD50 demonstrated that medication, specificallypsychostimulants, is the most effective treatment forreduction of ADHD symptoms at 14- and 24-monthfollow-ups, albeit psychosocial treatments have somepositive impact as well and in fact the efficacy of thepsychosocial treatment arm converges with that ofmedication at 36 months.51 The majority of thechildren with ADHD in the present study are notreceiving medications, but some (still a small propor-tion) are receiving other services. It is unlikely, however,that other services in this population are psychosocialtreatments in any way akin to the manual-basedtreatments offered in the Multimodal TreatmentStudy of Children With ADHD.13 Among children

BIRD ET AL.


with ADHD, there is less services and overall medica-tion use in Puerto Rico than in the south Bronx anda tendency in that direction for psychostimulant use aswell.

Although a smaller proportion of children withADHD in Puerto Rico received services, those whodid were more likely to receive psychotropic medicationsthan their counterparts in the south Bronx. Thisparadoxical site difference, given the higher rate ofservices use in the south Bronx must involve otherfactors. Although contrary to Puerto Rico, the majorityof south Bronx children with ADHD received someservices, less than one third with a diagnosis of ADHD(according to the DISC-IV) receive medication and lessthan one fifth receive psychostimulants to treat it.Besides greater availability of services in the south Bronxand greater acceptance of medication in Puerto Rico,this paradoxical finding may also relate to differences inthe services systems that are qualitative rather thanquantitative. In Puerto Rico, mental health services areregulated so that only psychiatrists are reimbursed forprescribing psychotropic medications. In the southBronx, other physicians are able to prescribe psychos-timulants and receive reimbursement for the visits. Weconsider that psychiatrists in Puerto Rico are more likelyto follow treatment guidelines for ADHD than primaryphysicians or other specialists, thus making it morelikely for Puerto Rican children to receive psychotropicmedications once they have used the services system.

Of greater importance in the overall picture is the factthat only one fourth of the children with ADHD receivemedications at both sites. This is less than half themedication rate reported by the Centers for DiseaseControl and Prevention (September 2, 2005),9 namely,that 56.3% of U.S. children with ADHD were takingmedication. Moreover, among the Puerto Rican chil-dren in our study, few of those with the diagnosis arereceiving psychostimulants. The dearth of appropriatetreatments may relate to a misperception of theproblems among parents and mental health providersor a lack of diagnostic acuity or knowledge oftreatment guidelines among the professionals treatingthese children and to differences in the services deliverysystem.

We want to emphasize the finding that parentalattitudes toward medication strongly determinewhether a child with ADHD received medication.Moreover, parental positive attitudes toward medica-

tion seemed to be a stronger predictor of services useas a whole than the diagnosis itself. Bauermeister52

reported on the results of focus groups with parentsand children in Puerto Rico to explore knowledge andattitudes about ADHD and medication treatment.The salient finding of their qualitative study was thatbeing well informed about ADHD treatment withmedication seemed to be the overriding factor inacceptance of, compliance with, and stability in theuse of medication.Given the large proportion of children with ADHD

and ADHD-NOS not receiving any kind of services ormedication, it is important to ask whether there may bean issue involving our method of ascertainment and thatthe diagnostic instrument used (DISC-IV) may beoverdiagnosing. Important arguments against thispossibility are reports indicating that the parent DISC-IV has good psychometric characteristics for thisdiagnosis, as well as the fact that the prevalence ratesof ADHD obtained at each wave over three waves areconsistent with prevalence rates reported in otherstudies.2 It is therefore unlikely that the untreatedADHD cases in our study are false positives.We found that certain risk variables were also

powerful predictors of services, medication, and psy-chostimulant use. The most consistent finding wasnegative child characteristics, which included poorperformance in school, nonengagement in extracurri-cular activities, and aggressive behaviors. Children whowere increased 1 SD on this risk composite were morethan twice as likely to be receiving services and nearlytwice as likely to be receiving medication. They were64% more likely to be receiving psychostimulants.Similarly, youths whose relationships with their parentsor caretakers were troubled, as measured by lowmaternal acceptance/warmth, were also more likely tobe receiving services, medication, and psychostimulants,although the magnitude of the effect was less strong thanthe risks. When services, medication, and psychostimu-lant use were examined in the total population, thesegeneral quantitative risk variables were more powerfulpredictors of treatment than a diagnosis of ADHD.Indeed, when these risks were adjusted, ADHD,ADHD-NOS, and diagnoses of other disorders wereno longer significant predictors of treatment use,indicating that parental attitudes as well as other riskfactors are more likely to lead to services use than thediagnosis itself.



Puerto Rican children with ADHD both in the southBronx and Puerto Rico are underserved. Cliniciansworking with these populations and agencies involved inservices delivery must consider the importance ofscreening for this disorder in schools and day caresettings and referring children who screen positive forin-depth assessments. The fact that a sizable proportionof caretakers are averse to the use of medication fortreating problems associated with ADHD indicates thatcommunity education is a paramount issue, as is theneed for careful alliance building and psychoeducationin the early phases of work with individual patients andparents. The findings point to the need for bettertraining among providers, school personnel, and ahigher level of community education and outreachefforts. In Puerto Rico, where there are fewer servicesavailable but the treatment of children with ADHD ismore likely to be in the hands of specialized psychiatristsand child and adolescent psychiatrists, a larger propor-tion of children received the treatment of choice. Thisfinding suggests that training of other professionals ordirect involvement by psychiatrists in monitoring thecare of children with ADHD is necessary.As with any epidemiological study that is not of

broader scope, there are limits to the generalizability ofthe study findings. Our results are generalizable toPuerto Rican children in Puerto Rico and the southBronx. Among those, the Puerto Ricans in the southBronx are a fairly assimilated group, judging from thefact that the majority chose to respond to thequestionnaires in English. Study findings would needto be replicated in other Puerto Rican populations in theUnited States and in other Latino populations if they areto be applied more broadly. Comparability with researchfindings on other ethnic groups is limited by the factthat differences in results could always be attributed todifferences in research design, ages under consideration,and other methodological differences. Our use of onlythe parent informant for diagnostic purposes is anotherlimitation because it ignores important information thatthe child could provide. The comorbidity of learningdisorders is high with ADHD and unfortunatelylearning disorders were not addressed or ascertained inthe study. A wide array of risk factors are included in ourstudy, but they are by no means exhaustive. Besidesdemographic variables other known predictors ofservices utilization, such as parental burden, parentYchild relationship and impairment, were considered in

our study as indicators within other measures used, butstandardized measures of some of those knownconstructs (e.g., parental burden) were not. Futurestudies would require a more formal delineation of thesevariables.

CONCLUSIONS

Over the course of 3 years, ADHD was related tomental health services, medication, and stimulant useamong Puerto Rican children living in the south Bronxand Puerto Rico. Although both sites presented similarpatterns of association with ADHD, rates of services,medication, and stimulant use are lower in Puerto Rico.Other psychiatric diagnosis and general risk variablesalso seem to be important correlates of use of servicesand medication, including psychostimulants. Particu-larly, environmental risks were related to greater mentalhealth services use, but not to medication use. Negativechild behavior and low maternal warmth were related toboth services and medication use, whereas negativefamily influence was related to less medication use. Atboth sites, the diagnosis itself was a weaker predictor ofservices and medication use than individual riskvariables. Although children in Puerto Rico receivedfewer services and medication overall, the patterns ofassociations were similar across the two sites. Thedisparities reported in the literature between ethnicminorities and other ethnic groups need to be re-examined. Based on the fact that the same ethnic groupshows differences in services and medication usepatterns in two different contexts, it is likely thatdisparities are more related to context than to ethnicity.

Disclosure: Dr. Bauermeister is on the Eli Lily PharmaceuticalInternational Advisory Board. Drs. Bird, Shrout, and Shen receivedstipends for this work from a grant provided by McNeil Laboratories.The other authors report no conflicts of interest.

REFERENCES

1. Bauermeister J, Canino G, Bird H. Epidemiology of disruptive behaviordisorders. Child Adolesc Psychiatr Clin N Am. 1994;3:177Y194.

2. Bird HR. The diagnostic classification, epidemiology and cross-culturalvalidity of ADHD. In: Jensen P, Cooper J, eds. Attention DeficitHyperactivity Disorder: State of the Science: Best Practices. Kingston, NJ:Civic Research Institute; 2002:2-1Y2-16.

3. Ivanova MY, Achenbach TM, Dumenci L, et al. Testing the 8-syndromestructure of the child behavior checklist in 30 societies. J Clin ChildAdolesc Psychol. 2007;36:405Y417.

4. Dulcan MK, Benson RS. Summary of the practice parameters for theassessment and treatment of children, adolescents, and adults withADHD. J Am Acad Child Adolesc Psychiatry. 1997;36:1311Y1317.

5. American Academy of Child and Adolescent Psychiatry. Children and

BIRD ET AL.


the News, No. 67. Updated January 2002. http://www.aacap.org/publications/factsfam/67.htm . Accessed July 2, 2007.

6. Jensen PS, Kettle L, Roper MT, et al. Are stimulants overprescribed?Treatment of ADHD in four U.S. communities. J Am Acad Child AdolescPsychiatry. 1999;38:797Y804.

7. Goldman LS, Genel M, Bezman RJ, Slanetz PJ. Diagnosis and treatmentof attention-deficit/hyperactivity disorder in children and adolescents.Council on Scientific Affairs, American Medical Association. JAMA.1998;279:1100Y1107.

8. Rowland AS, Umbach DM, Stallone L, Naftel AJ, Bohlig EM, SandlerDP. Prevalence of medication treatment for attention deficit-hyperactiv-ity disorder among elementary school children in Johnston County,North Carolina. Am J Public Health. 2002;92:231Y234.

9. Centers for Disease Control and Prevention (CDC). Mental health in theUnited States. Prevalence of diagnosis and medication treatment forattention-deficit/hyperactivity disorderYUnited States, 2003. MMWRMorb Mortal Wkly Rep. 2005;54:842Y847.

10. Safer DJ, Malever M. Stimulant treatment in Maryland public schools.Pediatrics. 2000;106:533Y539.

11. Leslie LK, Canino G, Landsverk J, et al. ADHD treatment patterns ofyouth served in public sectors in San Diego and Puerto Rico. J EmotBehav Disord. 2005;13:224Y236.

12. Lahey BB, Flagg EW, Bird HR, et al. The NIMH Methods for theEpidemiology of Child and Adolescent Mental Disorders (MECA)Study: background and methodology. J Am Acad Child Adolesc Psychiatry.1996;35:855Y864.

13. Bauermeister JJ, Canino G, Bravo M, et al. Stimulant and psychosocialtreatment of ADHD in Latino/Hispanic children. J Am Acad ChildAdolesc Psychiatry. 2003;42:851Y855.

14. Angold A, Erkanli A, Egger HL, Costello EJ. Stimulant treatment forchildren: a community perspective. J Am Acad Child Adolesc Psychiatry.2000;39:975Y984.

15. Bird HR, Canino GJ, Davies M, et al. A study of disruptive behaviordisorders in Puerto Rican youth: I. Background, design, and surveymethods. J Am Acad Child Adolesc Psychiatry. 2006;45:1032Y1041.

16. Bird HR, Davies M, Duarte CS, Shen S, Loeber R, Canino GJ. A studyof disruptive behavior disorders in Puerto Rican youth: II. Baselineprevalence, comorbidity, and correlates in two sites. J Am Acad ChildAdolesc Psychiatry. 2006;45:1042Y1053.

17. Bird HR, Shrout PE, Davies M, et al. Longitudinal development ofantisocial behaviors in young and early adolescent Puerto Rican childrenat two sites. J Am Acad Child Adolesc Psychiatry. 2007;46:5Y14.

18. Shaffer D, Fisher P, Lucas CP, Dulcan MK, Schwab-Stone ME. NIMHDiagnosticInterviewSchedule forChildrenVersionIV(NIMHDISC-IV):description, differences from previous versions, and reliability of somecommon diagnoses. J Am Acad Child Adolesc Psychiatry. 2000;39:28Y38.

19. Bravo M, Ribera J, Rubio-Stipec M, et al. Test-retest reliability of theSpanish version of the Diagnostic Interview Schedule for Children(DISC-IV). J Abnorm Child Psychol. 2001;29:433Y444.

20. Breton JJ, Bergeron L, Valla JP, Lepine S, Houde L, Gaudet N. Dochildren aged 9 through 11 years understand the DISC Version 2.25questions? J Am Acad Child Adolesc Psychiatry. 1995;34:946Y954.

21. Edelbrock C, Costello AJ, Dulcan MK, Kalas R, Conover NC. Agedifferences in the reliability of the psychiatric interview of the child. ChildDev. 1985;56:265Y275.

22. Fallon T Jr, Schwab-Stone M. Determinants of reliability in psychiatricsurveys of children aged 6Y12. J Child Psychol Psychiatry. 1994;35:1391Y1408.

23. Jensen PS, Rubio-Stipec M, Canino G, et al. Parent and childcontributions to diagnosis of mental disorder: are both informantsalways necessary? J Am Acad Child Adolesc Psychiatry. 1999;38:1569Y1579.

24. Schwab-Stone M, Fallon T, Briggs M, Crowther B. Reliability ofdiagnostic reporting for children aged 6Y11 years: a test-retest study ofthe Diagnostic Interview Schedule for Children-Revised. Am J Psychiatry.1994;151:1048Y1054.

25. Ribera JC, Canino G, Rubio-Stipec M, et al. The Diagnostic InterviewSchedule for Children (DISC-2.1) in Spanish: reliability in a Hispanicpopulation. J Child Psychol Psychiatry. 1996;37:195Y204.

26. Good MD, Smilkstein G, Good BJ, Shaffer T, Aarons T. The FamilyAPGAR indices: a study of construct validity. J Fam Pract. 1979;8:577Y582.

27. Thoits PA. Stress, coping and social support processes: where are we?What next? J Health Soc Behav. 1995;Spec No:53Y79.

28. Miller L, Warner V, Wickramaratne P, Weissman M. Religiosity anddepression: ten-year follow-up of depressed mothers and offspring. J AmAcad Child Adolesc Psychiatry. 1997;36:1416Y1425.

29. Patterson GR, Stouthamer-Loeber M. The correlation of family man-agement practices and delinquency. Child Dev. 1984;55:1299Y1307.

30. Goodman SH, Hoven CW, Narrow WE, et al. Measurement of risk formental disorders and competence in a psychiatric epidemiologiccommunity survey: the National Institute of Mental Health Methodsfor the Epidemiology of Child and Adolescent Mental Disorders(MECA) Study. Soc Psychiatry Psychiatr Epidemiol. 1998;33:162Y173.

31. Loeber R, Farrington DP, Stouthamer-Loeber M, van KammenWB. Anti-social Behavior and Mental Health Problems: Explanatory Factors in Child-hood and Adolescence. Hillsdale, NJ: Lawrence Erlbaum Associates; 1998.

32. Raia JA. Perceived Social Support and Coping as Moderators of Effects ofChildren`s Exposure to Community Violence. Los Angeles: UCLA; 1995.

33. Richters JE, Martinez P. The NIMH community violence project: I.Children as victims of andwitnesses to violence.Psychiatry. 1993;56:7Y21.

34. Johnson JH, McKutcheon SM. Assessing life stress in older children andadolescents: preliminary finding with Life Events Checklist. In: SarasonIG, Spielberger CD, eds. Stress and Anxiety. Washington, DC: Hemi-sphere; 1980:111Y125.

35. Bradley RH, Caldwell BM. Home observation for measurement of theenvironment: a validation study of screening efficiency. Am J Ment Defic.1977;81:417Y420.

36. Hudson WW. The WALMYR Assessment Scales Scoring Manual B57^.Tempe, AZ: WALMYR Publishing; 1992.

37. Hudson WW. Methodological observations on applied behavioralscience. A measurement package for clinical workers. J Appl Behav Sci.1982;18:229Y238.

38. Research Triangle Institute. SUDAAN User`s Manual, Version 8.0.2001. Research Triangle Park, NC.

39. Rosenbaum PR, Rubin DB. The central role of the propensity score inobservational studies for causal effects. Biometrika. 1983;70:41Y55.

40. Alegria M, Canino G, Lai S, et al. Understanding caregivers` help-seekingfor Latino children`s mental health care use. Med Care. 2004;42:447Y455.

41. Diggle PJ, Heagerty PJ, Liang KY, Zeger SL. Analysis of LongitudinalData. 2nd ed. New York: Oxford University Press; 2002.

42. Statistical Analysis System. Cary, NC: SAS Institute; 2007.43. Canino G, Shrout PE, Rubio-Stipec M, et al. The DSM-IV rates of child

and adolescent disorders in Puerto Rico: prevalence, correlates, serviceuse, and the effects of impairment. Arch Gen Psychiatry. 2004;61:85Y93.

44. Cohen P, Hesselbart CS. Demographic factors in the use of children`smental health services. Am J Public Health. 1993;83:49Y52.

45. Burns BJ, Costello EJ, Angold A, et al. Children`s mental health serviceuse across service sectors. Health Aff. 1995;14:147Y159.

46. Verhulst FC, Van der Ende J. Factors associated with child mental healthservice use in the community. J AmAcadChild Adolesc Psychiatry. 1997;36:901Y909.

47. Cabiya JJ, Canino G, Chavez L, et al. Gender disparities in mental healthservice use of Puerto Rican children and adolescents. J Child PsycholPsychiatry. 2006;47:840Y848.

48. Kataoka SH, Zhang L, Wells KB. Unmet need for mental health careamong U.S. children: variation by ethnicity and insurance status. Am JPsychiatry. 2002;159:1548Y1555.

49. Goode E. Study finds jump in children taking psychiatric drugs. NYTimes. 2003;Jan 14:A21.

50. MTA Cooperative Group. National Institute of Mental Health Multi-modal Treatment Study of ADHD follow-up: 24-month outcomes oftreatment strategies for attention-deficit/hyperactivity disorder. Pedia-trics. 2004;113:754Y761.

51. Jensen PS, Arnold LE, Swanson JM, et al. 3-year follow-up of the NIMHMTA study. J Am Acad Child Adolesc Psychiatry. 2007;46:989Y1002.

52. Bauermeister J. Medication treatment of ADHD in Latino/Hispanicchildren. Child Adolesc Psychopharmacol News. 2005;10:7Y11.



Documents

Longitudinal Mental Health Service and Medication Use for ADHD Among Puerto Rican Youth in Two Contexts