Behavioral Health Screening among Massachusetts Children Receiving Medicaid

ARTICLE IN PRESS THE JOURNAL OF PEDIATRICS • www.jpeds.com

ORIGINAL ARTICLES

Behavioral Health Screening among Massachusetts Children Receiving Medicaid Judith A. Savageau, MPH1,2, David Keller, MD3, Georgianna Willis, PhD2, Kathleen Muhr, MEd2, Gideon Aweh, MS2, Jack Simons, PhD4, and Emily Sherwood, MPA4 Objective To assess the impact of a Massachusetts Medicaid policy change (the Children’s Behavioral Health Initiative; CBHI, which required and reimbursed behavioral health [BH] screening with standardized tools at well child visits and developed intensive home- and community-based BH services) on primary care practice examining the relationship of BH screening to subsequent BH service utilization. Study design Using a repeated cross-sectional design, our 2010 and 2012 Medicaid study populations each included 2000 children/adolescents under the age of 21 years. For each year, the population was randomly selected and stratified into 4 age groups, with 500 members selected per group. Two data sources were used: medical records and Medicaid claims. Results The CBHI had a large impact on formal BH screening and treatment utilization among children/ adolescents enrolled in Medicaid. Screening increased substantially (73%: 2010; 74%: 2012) since the baseline/ premandate period (2007) when only 4% of well child visits included a formal screen. BH utilization increased among those formally screened but decreased among those with informal assessments. Conclusions CBHI implementation transformed the relationship between primary care and BH services. Changes in regulation and payment resulted in widespread BH screening in Massachusetts primary care practices caring for children/adolescents on Medicaid. (J Pediatr 2016;■■:■■-■■). etween 13% and 24% of all US children have behavioral health (BH) conditions.1-3 Most children with BH conditions do not receive needed treatment,4 and children from racially/ethnically diverse backgrounds are less likely than whites to receive services.5,6 Undertreatment puts children with BH conditions at increased risk of suicide, school dropout, substance abuse, criminal behavior, and risky sexual behaviors.7,8 Early intervention among children with BH problems results in better outcomes.9,10 Pediatric providers can promote early intervention through BH screenings during well child visits (WCVs).11-13 Enhanced systematic screenings maximize health attainment when interventions are begun early. Standardized screening instruments have been shown to be more effective in identifying developmental, behavioral, and psychosocial issues than are clinical assessments alone.14-16 In addition, screening fosters communication between parents and providers with the potential to improve health.17-19 Pediatric providers, however, have been slow to adopt screening14,15,20-23 citing a variety of barriers14,23-27 despite recommendations to screen for behavioral and developmental issues at WCVs.28,29 BH screening rates have risen in response to policy and regulatory changes that mandate use of standardized screening instruments at WCVs.30,31 However, little is known about the impact of BH screening on subsequent service utilization. Rushton et al32 found that fewer than one-half of patients referred by their pediatrician for BH had a BH visit in the subsequent 6 months. Romano-Clarke et al33 examined referrals and BH service utilization after the implementation of mandated BH screening at WCVs. Even though they found an increase in BH services, many children for whom further assessment and/or treatment seemed warranted did not access BH services. In 2007, the Children’s Behavioral Health Initiative (CBHI) implemented by MassHealth (Massachusetts Medicaid) required and reimbursed for BH screenFrom the 1Department of Family Medicine and Community Health, University of Massachusetts Medical ing for children and adolescents during WCVs in response to a 2001 MassachuSchool, Worcester, MA; 2Center for Health Policy and setts class action law suit (Rosie D et al vs Jane Swift et al34) filed on behalf of Research, University of Massachusetts Medical School, Shrewsbury, MA; 3Department of Pediatrics, University of MassHealth-enrolled children under the age of 21 years who had serious emoColorado, Denver, CO; and 4Children’s Behavioral Health Initiative, Massachusetts Executive Office of Health and tional disturbances. The Rosie D judgement (implemented on December 31, 2007) Human Services, Boston, MA required MassHealth providers to offer standardized BH screening at every WCV Funded by the Massachusetts Executive Office of Health

B

and Human Services, which approved this manuscript for submission to a peer-reviewed journal. The authors declare no conflicts of interest.

BH CBHI ED ESP FY MRR WCV

Behavioral health Children’s Behavioral Health Initiative Emergency department Emergency services program Fiscal year Medical record review Well child visit

Portions of the study were presented at the following meetings: Children’s Mental Health Research and Policy Conference, Tampa, FL, March 2-6, 2014; Pediatric Academic Societies, Vancouver, BC, May 3-6, 2014; and AcademyHealth Annual Research Meeting, San Diego, CA, June 8-10, 2014. 0022-3476/$ - see front matter. © 2016 Elsevier Inc. All rights reserved. http://dx.doi.org10.1016/j.jpeds.2016.07.029

1 FLA 5.4.0 DTD ■ YMPD8531_proof ■ August 18, 2016

THE JOURNAL OF PEDIATRICS • www.jpeds.com

Volume ■■

using a formal screening tool from the MassHealth-approved standardized BH tools. The court order established a timetable for rolling out (in 2009) an extensive array of new homeand community-based services designed to support and treat children and adolescents with complex BH needs. A previous chart audit of BH screening practices prior to the implementation of the Rosie D remedy showed that few providers were engaged in formal BH screening.35 We sought to assess the result of this policy change on primary care practice through a statewide chart audit of WCVs and to describe BH screenings and subsequent BH service utilization.

screening/surveillance without a specific tool. Where both formal and informal screens were conducted and abstracted, subsequent analyses prioritized results from the formal screening. In addition, abstractors recorded charted notes and documentation on BH referrals (made at the time of the WCV), patient demographics (ie, age, sex, ethnicity, and primary language spoken at home), interpreter use during the WCV, and use of a non-English BH screening tool. MRR occurred at the practice or remotely (vendor’s office, through secure postal delivery or faxed transmittal of medical record information) depending on the number of abstracted records per practice.

Methods

Claims Data Chart abstraction data were merged with MassHealth enrollment and claims data to assess BH services utilization. Paid claims from July 1, 2009 to December 31, 2012 were extracted to account for a 6-month follow-up period. BH assessment and treatment services were identified and diagnosis/ procedure codes extracted, also identifying the setting where BH services were conducted (ie, inpatient, emergency department [ED], outpatient, and emergency services program [ESP]). Claims for laboratory, radiology, or pharmacy services were excluded. When chart abstraction data indicated a standardized BH screening tool was used but no result documented, we used claims data to identify billing codes indicating the provider conducted BH screening and the screening result. These results supplemented missing chart abstraction data.

A repeated cross-sectional design examined standardized screenings, screening results, referral rates, and service utilization. Our baseline study,35 using fiscal year (FY) 2007 data, guided the current study of 2 follow-up periods spanning the implementation of the CBHI: FYs 2010 (July 1, 2009-June 30, 2010) and 2012 (July 1, 2011-June 30, 2012). Similar to the baseline study, medical record data were obtained to assess screenings, screening results, and service referrals. We used MassHealth enrollment and claims data to identify the study population, supplement chart abstraction screening results, and assess service utilization. The study was approved by the University of Massachusetts Medical School’s Institutional Review Board. The study population for each period consisted of MassHealth-enrolled children under the age of 21 years. The final inclusion criterion, for each period, required children/ adolescents to have a paid claim for a WCV, identified using current procedural terminology and International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis codes. Based on age group stratification identified from American Academy of Pediatrics periodicities for WCVs and recommendations for MassHealth-approved standardized screening tools, stratified random sampling selected 500 members from each of 4 age groups (ie, 6 months-2 years, 3-5 years, 6-11 years, and 12-20 years) in each study period, resulting in a total sample of 2000 members per year. Children under 6 months of age were excluded because the MassHealthapproved BH screening tools had not been validated for the youngest children. Sample size calculations per age group were based on Healthcare Effectiveness Data and Information Set sampling guidelines (n = ~ 411) and increased by 20% to account for potentially absent medical records. An experienced medical record review (MRR) vendor conducted a retrospective MRR of the sample’s 4000 children/ adolescents. Three registered nurses used a chart abstraction tool developed by one of the researchers and a panel of practicing physicians. Before implementation of the MRR, nurse abstractors had to pass Gold Standard Testing and attain interrater reliability scores of 95% or higher. The abstraction tool was piloted in a large community-based practice. From chart notes and documentation, nurse abstractors determined the presence of standardized BH screening, screening results, and referrals. They also detailed the presence of nonMassHealth approved screening tools and notations of informal

Variables Dependent variables included (1) standardized BH screening defined as the percent of WCVs for children/adolescents where a MassHealth-approved BH screening tool was used; (2) positive screening rate defined as the percent of WCVs with a screen for which a positive screen for BH conditions resulted; (3) referral rate defined as the percent of WCVs where a positive screen resulted in a BH referral; and (4) BH service utilization defined as the percent of WCVs with a formal screen (positive or negative) where the child subsequently received BH services as well as services following informal screening or provider surveillance (eg, general observations about BH noted in the medical records). BH utilization was categorized into 4 groups based on service setting: (1) inpatient; (2) ED; (3) outpatient; and (4) ESP. ESP provides critical BH services in the community (primarily in-home) including crisis assessment, intervention, and stabilization. The new array of home- and community-based services created under CBHI were categorized as outpatient services. Analyses combined inpatient/ ED services as well as outpatient/ESP services. Independent variables included age, sex, race, ethnicity, primary language, and MassHealth plan type (ie, managed care vs primary care case management). All sociodemographic variables were obtained from medical records except for plan type, which was derived from MassHealth. With substantial missing data for race, ethnicity, and primary language in the medical records (ie, 18% of our 2010 data and 15% of our 2012 data were absent all 3 sociodemographic variables), we used MassHealth data when available. For race and ethnicity (where

2

Savageau et al FLA 5.4.0 DTD ■ YMPD8531_proof ■ August 18, 2016

■■ 2016

ORIGINAL ARTICLES

40%-45% of each variable was missing in the claims data), we also imputed missing values based on the majority characteristics of our study population.

Table I. Sociodemographic characteristics of MassHealth members, FYs 2007*, 2010, and 2012 MassHealth members

Data Analyses Descriptive statistics were computed for sociodemographic characteristics. Tables I-IV present 2010 and 2012 data as well as the 2007 baseline data for comparison. Only univariate statistics are presented. Bivariate and multivariate analyses are presented in the larger technical report.35 Significance testing is not reported; all evaluations of medical record and utilization data were performed using SAS v 9.3 (SAS Institute, Cary, North Carolina).

Results Baseline data collected in FY2007 (prepolicy mandate) indicated few providers used formal screening tools, although informal screening was common.35 Among those with formal screens, 11.8% received BH services, 87.5% of which were provided in the outpatient setting. Among those formally screening positive, 40.0% received BH services, all in the outpatient setting. Among those informally screened, 20.1% received BH services with 87.9% of services provided in the outpatient setting. Lastly, among those with informal screens and positive results, 65.9% received BH services, nearly all (92.6%) were provided in the outpatient setting.35 In the current study, 3801 (of 4000) medical records were abstracted (95.0% retrieval rate). Excluded from analysis were 160 medical records (representing 280 WCVs) because no WCV visit notes could be identified, leaving 1801 MassHealth members (with 2332 WCVs) in 2010 and 1840 MassHealth members (with 2355 WCVs) in 2012. Table I presents the sociodemographic characteristics of MassHealth members studied. There were no significant differences between the 2010 and 2012 cohorts. Although nearly 20% identified as being non-English speaking, only 4% of encounters documented the use of interpreters (data not shown). Nearly three-quarters of all WCVs abstracted had evidence of formal BH screening (73.0% in 2010; 73.9% in 2012; Table II). Almost all formal BH screenings were conducted using validated instruments from the MassHealth-approved screeners. One-half of the WCVs without evidence of formal screening had charted notes indicating informal screening/ surveillance. The remaining visits had no BH screening documentation. Very few charts (1%) documented “parent refused screening.” Documentation of a positive BH screen was found in 13.5% of the visits in 2010 (where a formal screen occurred at the WCV) and 12.4% in 2012. Many records contained no documentation of screening results. Review of billing codes associated with those screens showed that charts without documented screening results were likely to be negative screens, and encounters were coded as such. Of the 8 MassHealth-approved screening tools, 3 were used most frequently (Table III): the Parents’ Evaluation of Developmental Status (46.3% in 2010; 44.6% in 2012); followed by the Pediatric Symptom Checklist (combined 39.7% in 2010;

Characteristics Age group 6 mo-2 y 3-5 y 6-11 y 12-20 y Sex Male Female Race White Non-White Ethnicity Hispanic Non-Hispanic Primary language English Spanish Other/unknown

2007 (N = 1336)† N (%) 348 347 338 303

(26.0) (26.0) (25.3) (22.7)

2010 (N = 1801)†,‡ N (%) 469 453 449 430

(26.0) (25.2) (24.9) (23.9)

2012 (N = 1840)†,‡ N (%) 467 460 456 457

(25.4) (25.0) (24.8) (24.8)

663 (49.6) 673 (50.4)

937 (52.0) 864 (48.0)

941 (51.1) 899 (48.9)

395 (43.7)‡ 509 (56.3)

825 (57.9) 599 (42.1)

872 (59.5) 594 (40.5)

292 (32.3)§ 612 (67.7)

494 (41.6) 694 (58.4)

522 (41.9) 725 (58.1)

1481 (82.2) 169 (9.4) 151 (8.4)

1498 (81.4) 212 (11.5) 130 (7.1)

1095 (82.0) 125 (9.4) 115 (8.6)

*2007 results are presented herein for contextual/comparison purposes only; evaluation of baseline data was not a component of the current study. †Ns for individual demographic variables may not total to the number of members whose records were abstracted in each of the 2 cohort years because of sporadic missing data. However, missing data were far more prevalent for member race and ethnicity (as described in the Methods section). ‡For sex, race, ethnicity, and language, MassHealth administrative data were used to supplement medical record missing data (predominantly for race, ethnicity, and language). Supplementing medical record data in FY2010 and FY2012 were less prevalent and related to changes in charting. §Only MassHealth administrative data were used (where available) for the FY2007 data as medical records were absent these 3 variables in greater than 50% of the cases.

38.6% in 2012); and the Modified Checklist for Autism in Toddlers (12.2% in 2010; 12.9% in 2012). The use of the other 5 approved screening tools ranged from 0% to 4% in both FYs. Many of the formal MassHealth-approved screening tools are available in non-English languages. However, although nearly 20% of each cohort did not list English as their primary

Table II. BH screening at WCVs and screening results as documented in the medical record, FY 2007*, 2010, and 2010 WCVs

Type of screenings

2007 2010 2012 (N = 1717)† (N = 2332)† (N = 2355)† N (%) N (%) N (%)

Formal screening with validated tool 68 (4.0) 1703 (73.0) 1741 (73.9) Positive 5 (7.4) 230 (13.5) 216 (12.4) Negative/not documented 63 (92.6) 1473 (86.5) 1525 (87.6) Informal surveillance 1337 (77.8) 310 (13.3) 348 (14.8) Positive 187 (14.0) 83 (26.8) 107 (30.7) Negative/not documented 1150 (86.0) 227 (73.2) 241 (69.3) No screening or surveillance 312 (18.2) 319 (13.7) 266 (11.3) N/A 25 (1.1) 13 (0.6) Refused screening‡ *2007 results are presented herein for contextual/comparison purposes only; evaluation of baseline data was not a component of the current study. †The total number of WCVs is categorized by the number of formal screens, informal screens, or visits with no screening/surveillance. ‡The number of visits where screening was refused are not counted in this total. These data were not collected in the baseline study as screenings were not required at that time.

Behavioral Health Screening among Massachusetts Children Receiving Medicaid FLA 5.4.0 DTD ■ YMPD8531_proof ■ August 18, 2016

3

THE JOURNAL OF PEDIATRICS • www.jpeds.com

Volume ■■

Table III. Use of formal BH screening tools, and frequency of positive results, FY 2007*, 2010, and 2012 WCVs 2007 Tool used N (%) MassHealth approved screening tools PEDS PSC Y-PSC M-CHAT ASQ BITSEA CRAFFT PHQ-9 SDQ Non-English version of tool used

2010

Positive results N (%)

†

N = 68‡ 10 4 3 1

(14.3) (5.8) (4.3) (1.4) 0 0 0 2 (2.9) N/A¶ N/A**

Tool used N (%)

2012

Positive results N (%)

N = 1703§ 0 1 0 0 N/A N/A N/A 1 N/A¶ N/A**

789 552 135 211 53

(46.3) (31.9) (7.8) (12.2) (3.1) 0 72 (4.2) 16 (0.9) 0 137 (8.0)

†

Tool used N (%)

Positive results† N (%)

N = 1741§ 125 85 12 7

(15.8) (15.4) (8.9) (3.3) 3 N/A 4 7 N/A N/A

776 526 156 228 77

(44.6) (29.8) (8.8) (12.9) (4.4) 0 78 (4.4) 18 (1.0) 2 (0.1) 138 (7.9)

89 81 26 13

(11.5) (15.4) (16.7) (5.7) 3 N/A 4 3 0 N/A

ASQ, Ages and Stages Questionnaires; BITSEA, Brief Infant-Toddler Social and Emotional Assessment; CRAFFT, Car, Relax, Alone, Forget, Friends, Trouble; M-CHAT, Modified Checklist for Autism in Toddlers; N/A, not applicable; PEDS, Parents' Evaluation of Developmental Status; PHQ-9, Patient Health Questionnaire-9; PSC, Pediatric Symptom Checklist; SDQ, Strengths and Difficulties Questionnaire; Y-PSC, Pediatric Symptom Checklist- Youth Report. References for all screening tools can be found at: http://www.mass.gov/eohhs/gov/commissions-and-initiatives/cbhi/screening-for-behavioral-health-conditions/the-masshealth-approved-screening-tools/. *2007 results are presented herein for contextual/comparison purposes only; evaluation of baseline data was not a component of the current study. †Positive screening results (as a percentage of formal screens) not reported in FY2007 because of too few instances of tool being used; positive screening results not reported for infrequently used tools in FY2010 and FY2012 because of too few instances of tool being used. ‡Totals from the individual tools do not total to 68 because there were 50 formal tools used that were not on the MassHealth approved tools list (not required prior to December 31, 2007 (eg, Home, Education, Activities, Drugs, Suicidality, and Sex); a total of 70 formal screens were conducted among the 68 WCVs as some visits had more than 1 screening tool used at that time. §Totals from the individual tools used may be higher than the number of WCVs where a formal screen was completed because some visits may have had more than 1 screening tool used at that time. ¶Not assessed in baseline study; Child Behavior Checklist was identified as a MassHealth approved tool in FY2007 but dropped in subsequent years. **Data were not abstracted from the medical records in the baseline study because use of formal screening tools was not yet required by MassHealth.

language, the use of non-English tools was uncommon (8.0% in 2010; 7.9% in 2012; Table III). BH service utilization in 2010 and 2012 increased when compared with our baseline study (12% of WCVs compared with less than 2% at baseline).36 In 2010, 3669 claims for BH services were recorded, representing 506 children/adolescents with an average of 7.2 BH claims/member over the 6-month postWCV period. The majority of BH claims (72.2%) followed formal BH screening. In 2012, the number of claims increased by over 60% to 6067, representing an average of 10.7 claims/member for the 567 children and adolescents receiving BH services. More than one-half of these services (59.0%) were provided following a formal BH screen. One-third of the services (32.1%) were subsequent to an informal screening/ surveillance, and a small fraction of the services (8.9%) were provided absent any screening documentation. The percentage of children receiving BH services increased from 26.5% in FY2010 to 29.4% in FY2012 (P = .052; Table IV) largely due to an increase in the percentage of children receiving BH services after informal screening (P = .005). Among those who received formal screening, the rate of receipt of BH services following a positive screen increased from 30.7% in FY2010 to 43.6% in FY2012 (P = .008). However, the rate of receiving BH services was higher for children who had positive informal screens in both FY2010 and FY2012 (67.1% and 71.4%, respectively). Substantial proportions of children with negative formal or informal screens also received subsequent BH services. Nearly all claims in both years were for services provided in the outpatient setting (which includes home- and

community-based CBHI services): 99.8% in 2010 and 99.7% in 201236; these percentages did not vary dependent on whether the services followed formal, informal, or no BH screening. Regardless of whether services followed screening or not, compared with our 2007 baseline data, claims for inpatient services all but disappeared (ie, from 11% to 12% in FY 2007 to <1% in FYs 2010 and 2012).35,36

Discussion BH screening practices and subsequent BH service utilization changed dramatically after the 2008 implementation of the CBHI. Child health providers in Massachusetts now routinely screen a majority of children/adolescents for BH conditions using standardized tools. BH service utilization during the 6-month period after a WCV dramatically increased compared with pre-CBHI baseline data, and the use of ED/ inpatient BH services decreased substantially.35,36 Although the increase in service utilization may have resulted from pentup demand for services from pre-CBHI, it should be noted that there was also an increase in the number of BH services available, the number of providers, and extensive education and outreach to MassHealth families identifying greater accessibility to home- and community-based services. Similar changes in screening patterns were seen by others who have assessed pediatric BH screening interventions. When formal tools and screening protocols were introduced, BH screening rates increased from 10%-20% to 54%-80%.13,17 Hacker et al19 analyzed FY2009 claims data and found a 46% screening rate 1 year after screening initiation. Romano-Clarke

4

Savageau et al FLA 5.4.0 DTD ■ YMPD8531_proof ■ August 18, 2016

■■ 2016

ORIGINAL ARTICLES

Table IV. BH service utilization 6 months after WCV by screening type and screening outcome—FYs 2007*, 2010, and 2012 MassHealth members with BH encounters† Number of MassHealth members (% with BH encounters)‡ Screenings

FY2007 N (%)

Type of screening Formal/standardized

53

Informal/surveillance

1059

No screening Total

267 1336

Positive screens§ Formal/standardized

5

Informal/surveillance

170

Negative screens/not documented§ Formal/standardized Informal/surveillance

N/A¶ N/A¶

FY2010 N (%)

FY2012 N (%)

383/1361 (28.1%) 83/275 (30.2%) 40/270 (14.8%) 506/1906 (26.5%)

Test of proportions z-score (2-tailed)

P value

404/1413 (28.6%) 125/301 (41.5%) 38/217 (17.5%) 567/1931 (29.4%)

−0.26

.795

−2.83

.005

−0.81

.418

−1.94

.052

61/199 (30.7%) 55/82 (67.1%)

85/195 (43.6%) 75/105 (71.4%)

−2.66

.008

−0.64

.522

172/692 (24.9%) 26/169 (15.4%)

179/737 (24.3%) 39/174 (22.4%)

0.25

.803

−1.66

.097

*2007 results are presented herein for contextual/comparison purposes only; evaluation of baseline data was not a component of the current study. †Member count is a unique count per row; members may be included in more than 1 row. ‡For 2010 and 2012 only; percentages of members with encounters based on number of MassHealth members in the respective row. §Not all screens had results (positive or negative) documented in the medical record. Number of members represented by formal screens without documentation of results have been included in the category of negative screening results; percentages are based on those screens with recorded results in the medical record as well as on encounter form billed from the provider's practice for FY2010 and FY2012 formal screens only. ¶Data are not available for BH services received nor the number of members represented by those services, in our baseline study, subsequent to a negative screen.

et al33 found that 83% of WCVs conducted in 2 Massachusetts primary care practices had evidence of BH assessments. Our study shows similar changes occurring statewide through changes in Medicaid policy and payment. This study adds to the literature through the use of both medical record and claims data, which provides for a more complete picture than what each source details alone. Our study additionally was able to show a temporal relationship between screening practices and subsequent service utilization as well as the location of service delivery. Although barriers to screening were not formally addressed, we found little evidence of parental refusal. Others have reported time, reimbursement, training, confidence, availability of validated tools, and availability of referral resources as screening barriers.24 During implementation, MassHealth focused primarily on time and reimbursement. They provided numerous trainings in using the approved tools, identifying resources for screening, interpreting results, and implementing screening protocols. MassHealth paid providers approximately an additional $10 for each screen at WCVs. Future efforts to increase screening rates include enabling parents and adolescents to complete screening tools prior to the WCV, using capabilities inherent in electronic health records, and patient portals. MassHealth allowed the use of any of 8 validated screening instruments. Primary care providers seemed to prefer simpler, 1-page instruments to more comprehensive tools. Ad hoc inquiries to providers throughout the state noted that, in

a busy practice, it is easier to implement self-administered protocols needing minimal scoring. MassHealth has expanded the list of approved screeners, adding more “1-page,” nonproprietary screeners in response to provider feedback. Few providers used standardized screening tools in nonEnglish languages despite their availability in many languages. Nearly 20% of our sample noted a primary language other than English. We looked for potential disparities in screening by comparing member’s age, sex, race, ethnicity, and primary language among screened vs not screened. No consistent differences were observed in both bivariate and multivariate analyses (data not shown). Still, it is likely that language remains a substantial barrier to BH screening and treatment. Future efforts at quality improvement should focus on the elimination of this potential disparity in access. Screening is the first step in assuring access to comprehensive services. It may result in detecting conditions that require intervention in the primary care setting or more intensive work within the BH system. Hacker et al, 37 using FY2008-FY2010 MassHealth claims data, found that BH services postscreening were received among 57% of children screening positive (as identified through administrative data), but only among 22% of children who screened negative. However, our study, combining data from both medical records and MassHealth claims, showed that BH services were frequently accessed within 6 months following a WCV in children who screened negative as well as positive screens (using formal assessment tools). The CBHI was a comprehensive initiative that included enhanced

Behavioral Health Screening among Massachusetts Children Receiving Medicaid FLA 5.4.0 DTD ■ YMPD8531_proof ■ August 18, 2016

5

THE JOURNAL OF PEDIATRICS • www.jpeds.com screening as well as public health messaging and enhanced availability of BH services. It is possible that screening served to enhance access to BH services by destigmatizing parental concerns around child mental health or highlighting new programs to increase access to mental health services, rather than by identifying those most in need of mental health services. Further study is needed to tease out the ways in which such interventions increase access to and utilization of mental health services. The rate of positive screens (as well as follow-up service utilization) was higher among those WCVs with informal screening vs the use of a standardized tool. Although this might call into question the value of formal screening, it may speak more to the use of informal screening/surveillance among those children: known to the provider to have a prior diagnosis; with a suspicion of a BH problem; and/or who are already receiving BH services. Even with a mandate to offer formal BH screening at every WCV, providers may not strongly advocate for screening but rather provide informal surveillance through ‘watchful waiting’. Along with the introduction of formal BH screening into the WCV, the CBHI introduced new home- and communitybased services statewide as reflected in changes in the services accessed in 2010 and 2012. Overall, service utilization increased and ED/inpatient BH services became much less common as the use of conventional and the new homeand community-based services became more prevalent. These data suggest that BH services can be delivered in the outpatient, home, and community settings for the majority of children/adolescents. This was a retrospective study based on claims data and chart review, each of which is subject to limitations. Basing our sample on paid claims may have resulted in undercounting because unpaid claims can be rebilled. Some data elements were missing in both datasets, which could lead to information bias. Although we conducted bivariate and multivariate sensitivity analyses, we are unsure if those with missing demographic information more often represented minority or majority populations. Strengths of the study include detailed data abstraction; random sample selection; sample size; high interrater reliability; and the comparative nature of our results to other published studies. Our study also was able to assess issues of language, the specifics of screening tools used, as well as screening rates and subsequent use of BH services among those formally screening as well as children/adolescents informally assessed at the time of the WCV. Each of these assessments could only be accomplished through the use of both medical records and administrative claims data, which enhances the previously published literature where many studies used claims data as a primary data source. The initiation of universal screening for BH conditions initiated by the Commonwealth’s CBHI affords a unique opportunity to examine screening practices and subsequent service utilization. Our data show that primary care pediatric practices are able to incorporate formal screenings and that MassHealth can support the infrastructure to provide an enhanced level of service. Further research could aid understanding whether

Volume ■■ screening rates can be increased beyond 75%, why particular tools are used and whether or not new tools are needed, and how decision-making occurs regarding referrals to BH services absent a formal screening and/or a positive finding. Ultimately, we need to decide whether screenings and referrals to specialty care result in earlier intervention and improved health outcomes among MassHealth children and adolescents at risk for BH conditions. Those analyses may indicate whether or not improving access to care also improves health outcomes, as well as patient and family experiences, at lower cost. In summary, the implementation of the CBHI has fundamentally transformed the relationship between primary care services and BH services within Massachusetts. The current study shows (1) changes in regulation and payment have resulted in the implementation of widespread BH screening in primary care practices that care for children/adolescents on Medicaid; (2) primary care physicians chose simpler screening tools that identify children/adolescents at risk over complex tools with greater diagnostic specificity; (3) the initiation of the CBHI within MassHealth resulted in both increased formal BH screening in primary care and increased BH service utilization, although the relationship between screening at a WCV and service utilization is complex; and (4) a specific area for quality improvement activity should focus on improving access to BH screening and services among families where English is not the primary language. ■ We thank YouFu Li, MD, MPH, for programming and analytic support and for selecting the samples for chart abstraction; Amy Leary for assistance with coding; and New York County Health Services Research Organization/MedReview who conducted the chart abstraction (as a contracted vendor for this project). Submitted for publication Mar 22, 2016; last revision received Jun 8, 2016; accepted Jul 14, 2016

References 1. Kuo AA, Etzel RA, Chilton LA, Watson C, Gorski PA. Primary care pediatrics and public health: meeting the needs of today’s children. Am J Public Health 2012;102:e17-23. 2. Lucenko BA, He L, Mancuso D, Huber A, Felver BEM. Needs and services for children and youth with Medicaid coverage in Washington State. RDA Report 3.40. Olympia (WA): Washington State Department of Social and Health Services; 2013. 3. Perou R, Bitsko RH, Blumberg SJ, Pastor P, Ghandour RM, Gfroerer JC, et al. Mental health surveillance among children—United States, 20052011. MMWR Surveill Summ 2013;62 (Suppl 2):1-35. 4. Kataoka SH, Zhang L, Wells KB. Unmet need for mental health care among U.S. children: variation by ethnicity and insurance status. Am J Psychiatry 2002;159:1548-55. 5. Merikangas KR, He J, Burstein M, Swendsen J, Avenevoli S, Case B, et al. Service utilization for lifetime mental disorders in U.S. adolescents: results of the National Comorbidity Survey–Adolescent Supplement (NCS-A). J Am Acad Child Adolesc Psychiatry 2011;50:32-45. 6. Pires S, Grimes K, Gilmer T, Allen K, Mahadevan R, Hendricks T. Identifying opportunities to improve children’s behavioral health care: an analysis of Medicaid utilization and expenditures. Center for Health Care Strategies, Inc; 2013. 7. Copeland W, Miller-Johnson S, Keeler G, Angold A, Costello E. Childhood psychiatric disorders and young adult crime: a prospective, population-based study. Am J Psychiatry 2007;164:1668-75.

6

Savageau et al FLA 5.4.0 DTD ■ YMPD8531_proof ■ August 18, 2016

■■ 2016

ORIGINAL ARTICLES

8. Lehrer JA, Shrier LA, Gortmaker S, Buka S. Depressive symptoms as a longitudinal predictor of sexual risk behaviors among U.S. middle and high school students. Pediatrics 2006;118:189-200. 9. Belelieu J. Rosie D. and mental health screening: a case study in providing mental health screening at the Medicaid EPSDT visit. New York (NY): Teen Screen National Center for Mental Health Checkups at Columbia University; 2010. 10. Shonkoff JP, Phillips DA, eds. From neurons to neighborhoods: the science of early childhood development. Washington DC: National Academy Press. Board on Children,Youth, and Families, Institute of Medicine, and National Research Council; 2000. http://www.nap.edu/read/9824/chapter/1. 11. Croghan TW, Brown JD. Integrating mental health treatment into the patient centered medical home. Rockville (MD): 2010. p. AHRQ Publication No. 10-0084-EF. 12. Keller D, Sarvet B. Is there a psychiatrist in the house? Integrating child psychiatry into the pediatric medical home. J Am Acad Child Adolesc Psychiatry 2013;52:3-5. 13. Kuhlthau K, Jellinek M, White G, Vancleave J, Simons J, Murphy M. Increases in behavioral health screening in pediatric care for Massachusetts Medicaid patients. Arch Pediatr Adolesc Med 2011;165:660-4. 14. Sand N, Silverstein M, Glascoe FP, Vidya GB, Tonniges TP, O’Connor KG. Pediatricians’ reported practices regarding developmental screening: do guidelines work? Do they help? Pediatrics 2005;116:174-9. 15. Reijneveld SA, Brugman E, Verhulst FC, Verloove-Vanhorick SP. Identification and management of psychosocial problems among toddlers in Dutch preventive child health care. Arch Pediatr Adolesc Med 2004;158:811-7. 16. Guevara JP, Gerdes M, Localio R, Huang YV, Pinto-Martin J, Minkovitz CS, et al. Effectiveness of developmental screening in an urban setting. Pediatrics 2013;131:30-7. 17. Schonwald A, Huntington N, Chan E, Risko W, Bridgemohan C. Routine developmental screening implemented in urban primary care settings: more evidence of feasibility and effectiveness. Pediatrics 2009;123:660-8. 18. Garg A, Butz AM, Dworkin PH, Lewis RA, Thompson RE, Serwint JR. Improving the management of family psychosocial problems at lowincome children’s well-child care visits: the WE CARE project. Pediatrics 2007;120:547-58. 19. Hacker KA, Penfold R, Arsenault L, Zhang F, Murphy M, Wissow L. Screening for behavioral health issues in children enrolled in Massachusetts Medicaid. Pediatrics 2014;133:46-54. 20. Brown JD, Wissow LS, Riley AW. Physician and patient characteristics associated with discussion of psychosocial health during pediatric primary care visits. Clin Pediatr (Phila) 2007;46:812-20. 21. Cooper S, Valleley RJ, Polaha J, Begeny J, Evans JH. Running out of time: physician management of behavioral health concerns in rural pediatric primary care. Pediatrics 2006;118:e132-8. 22. Sices L, Feudtner C, McLaughlin J, Drotar D, Williams M. How do primary care physicians manage children with possible developmental delays? A national survey with an experimental design. Pediatrics 2004;113:27482.

23. Williams J, Klinepeter K, Palmes G, Pulley A, Meschan Foy J. Diagnosis and treatment of behavioral health disorders in pediatric practice. Pediatrics 2004;114:601-6. 24. Hacker K, Goldstein J, Link D, Sengupta N, Bowers R, Tendulkar S, et al. Pediatric provider processes for behavioral health screening, decision making, and referral in sites with colocated mental health services. J Dev Behav Pediatr 2013;34:680-7. 25. Hart CN, Kelleher KJ, Drotar D, Scholle SH. Parent-provider communication and parental satisfaction with care of children with psychosocial problems. Patient Educ Couns 2007;68:179-85. 26. Pinto-Martin JA, Dunkle M, Earls M, Fliedner D, Landes C. Developmental stages of developmental screening: steps to implementation of a successful program. Am J Public Health 2005;95:1928-32. 27. Weitzman CC, Leventhal JM. Screening for behavioral health problems in primary care. Curr Opin Pediatr 2006;18:641-8. 28. American Academy of Pediatrics Task Force on Mental Health. Appendix S4: the case for routine mental health screening. Pediatrics 2010;125(Suppl):S133-9. 29. Hagan JF, Shaw JS, Duncan PM. Bright futures: guidelines for health supervision of infants, children, and adolescents. 3rd ed. Pocket guide. Elk Grove Village (IL): American Academy of Pediatrics; 2008. 30. Klein S, McCarthy D. North Carolina’s ABCD program: using community care networks to improve the delivery of childhood developmental screening and referral to early intervention services. Issue Brief (Commonw Fund) 2009;66:1-28. 31. Earls MF, Hay SS. Setting the stage for success: implementation of developmental and behavioral screening and surveillance in primary care practice—the North Carolina Assuring Better Child Health and Development (ABCD) project. Pediatrics 2006;118:183-8. 32. Rushton J, Bruckman D, Kelleher K. Primary care referral of children with psychosocial problems. Arch Pediatr Adolesc Med 2002;156:592-8. 33. Romano-Clarke G, Tang MH, Xerras DC, Egan HS, Pasinski RC, Kamin HS, et al. Have rates of behavioral health assessment and treatment increased for Massachusetts children since the Rosie D. decision? A report from two primary care practices. Clin Pediatr (Phila) 2014;53:2439. 34. Rosie D., Swift J., et al., v. J. Swift, et al. Civil Action No. 01-30199-MAP. http://www.rosied.org. Accessed August 4, 2016. 35. Savageau J, Cabral L, Gettens J, O’Connell E. Clinical topic review: behavioral health screening for children with well visits. http://www.mass.gov/ eohhs/docs/masshealth/cbhi/reports/clinical-topic-review-ctr-10-29 -09.pdf. Accessed November 9, 2015. 36. Savageau J, Willis G, Muhr K, Keller D, Aweh G, O’Connell E. Clinical topic review 2013- behavioral health screening among MassHealth children and adolescents. http://www.mass.gov/eohhs/docs/masshealth/cbhi/ reports/bh-screening-among-mh-children-feb-2015.pdf. Accessed November 9, 2015. 37. Hacker KA, Penfold RB, Arsenault LN, Zhang F, Murphy M, Wissow LS. Behavioral health services following implementation of screening in Massachusetts Medicaid children. Pediatrics 2014;134:737-46.

Behavioral Health Screening among Massachusetts Children Receiving Medicaid FLA 5.4.0 DTD ■ YMPD8531_proof ■ August 18, 2016

7