A Planned Care Approach and Patient Registry

to Improve Adherence to Clinical Guidelines for the

Diagnosis and Management of Attention-Deficit/Hyperactivity
Disorder
Paul L. Geltman, MD, MPH; Lise E. Fried, DSc, MS; Lisa N. Arsenault, PhD;
Alice M. Knowles, MS; David A. Link, MD; Joel N. Goldstein, MD; James M. Perrin, MD;
Karen A. Hacker, MD, MPH
From the Department of Pediatrics (Dr Geltman and Dr Link), Institute for Community Health (Dr Fried, Dr Arsenault, and Ms Knowles),
Division of Child Psychiatry (Dr Goldstein), Cambridge Health Alliance (Cambridge, MA) and Harvard Medical School, Boston, Mass;
Department of Pediatrics, Massachusetts General Hospital and Harvard Medical School, Boston, Mass (Dr Perrin); and Allegheny County
Health Department and University of Pittsburgh Graduate School of Public Health, Pittsburgh, Pa (Dr Hacker)
The authors declare that they have no conflict of interest.
Address correspondence to Paul L. Geltman, MD, MPH, 1493 Cambridge St, Macht 317C, Cambridge, MA 02139 (e-mail: pgeltman@challiance.org).
Received for publication August 12, 2014; accepted December 8, 2014.

ABSTRACT
BACKGROUND: Attention-deficit/hyperactivity disorder
(ADHD) affects almost 2.4 million US children. Because American Academy of Pediatrics guidelines for ADHD recommend
use of standardized diagnostic instruments, regular follow-up
and the chronic care model, this pilot project sought to implement and assess an electronic registry of patients with ADHD
combined with care coordination by a planned care team.
METHODS: This quality improvement project was structured
with 2 intervention and 2 control clinics to facilitate evaluation
of the use of a planned care system for management of ADHD.
Care teams included a pediatrician, nurse, medical assistant,
and care coordinator and tracked patients using an electronic
registry with data drawn from the EMR. Clinical work flows
were pilot tested to facilitate use of the Vanderbilt scales and
their incorporation into the EMR at intervention sites. Outcome
measures included 2 recommended clinical follow-ups based on
HEDIS measures as well as use of the Vanderbilt rating scales.
Initiation phase measure was for follow-up after initiating medication, while the continuation phase measure was for subsequent follow-up during the first year of treatment. Measures
were monitored during the project year and then also in the
ensuing period of spread of the intervention to other sites.

RESULTS: Although the modified HEDIS initiation phase measure for patients newly on medication remained static at approximately 50% throughout the project period, the continuation
phase measure showed improvement from 35% at baseline to
45% at the end of the project assessment year, a 29% increase.
Follow-up for patients stable on medications also remained unchanged during the project period, but during subsequent
spreading of the intervention to nonproject sites, follow-up of
these patients improved to over 90%. In adjusted analyses, patients with ADHD at intervention sites were over 2 times
more likely than patients at control sites to have had a Vanderbilt score documented in their records.
CONCLUSIONS: The project achieved modest improvements in
the diagnostic and treatment process for patients with ADHD.
The use of a planned care system and electronic patient registry
shows promise for improving the diagnosis and treatment process for patients with ADHD.
KEYWORDS: ADHD; planned care; primary care; quality
improvement; registry
ACADEMIC PEDIATRICS 2015;15:289296

WHATS NEW

medication management using stimulants compared to

behavioral modification, a combination approach, and
community comparison, while the combination approach
conferred a modest advantage in nonsymptom and
functional outcomes.3,4 On the basis of this and other
evidence,57 the American Academy of Pediatrics (AAP)
released clinical practice guidelines for ADHD,8,9 which
were updated in 2011.10 These guidelines emphasize a
standardized diagnostic approach using validated instruments to assess diagnostic criteria, comorbid conditions,
and functional impairment in multiple settings (eg, the
Vanderbilt ADHD rating scales).11 Treatment should
follow a chronic care model approach with medication

This project demonstrates the effective implementation

of elements of the chronic care model in improving
care for children with attention deficit/hyperactivity
disorder.

ATTENTION-DEFICIT/HYPERACTIVITY

DISORDER

(ADHD) affects approximately 8% to 8.7%1,2 of US

children, or almost 2.4 million, and is frequently managed
by primary care pediatricians. The large Multimodal
Treatment Study of ADHD showed significant
improvement in ADHD symptoms with intensive
ACADEMIC PEDIATRICS
Copyright 2015 by Academic Pediatric Association

289

Volume 15, Number 3

MayJune 2015

290

GELTMAN ET AL

and behavioral management focused on specific target

symptoms with close monitoring. This approach has been
promoted by the AAP for quality improvement (QI) efforts
with an available tool kit.12
Despite the existence of these tools, less than half of
children in one national sample who met diagnostic criteria
for ADHD received either a diagnosis or regular medication treatment.2,13 Disparities also exist in diagnosis and
treatment based on gender and socioeconomic status.
Similar disparities can be seen in use of services as
well.14 Compounding these failings, the high rate of comorbid conditions15 and clinical complexity makes use
of a collaborative chronic care model challenging.16
Strategies to improve adherence to guidelines have
included components of the chronic care model,17 which
is increasingly considered the cornerstone of effective
management of conditions like asthma and obesity.
Within this model are clinical decision supports (such as
evidence-based guidelines, computer-assisted management, and standard assessment tools), registries for
tracking patients, and novel approaches to care delivery.18,19 To this end, Cambridge Health Alliance (CHA)
embarked on development of a registry-based system of
planned care with clinical decision supports to improve
adherence to AAP guidelines for ADHD. The project
aimed to implement an electronic registry linked to the
electronic medical record (EMR); evaluate recommended
ADHD clinical follow-up measures (modified from the
Healthcare Effectiveness Data and Information Set
[HEDIS]);20 and examine the impact of the project on utilization of the Vanderbilt scales. Here we report on lessons
learned during development and implementation of the
registry and planned care system and initial results of those
efforts.

METHODS
SETTING
CHA is a large, integrated health system serving Cambridge and Somerville, Massachusetts, as well as Bostons
Metro-North area with a patient population of 25,000 children and approximately 80,000 ambulatory visits per year
at 10 sites. More than half of patients speak a language
other than English at home. Over 50% of the population
is insured by public payers. CHA has previously undertaken QI projects for asthma, immunizations, and mental
health screening. Past projects focused on systems changes
using elements of the chronic care model (registries and
planned care teams) rather than focusing on individual
physician behaviors. Before this ADHD project, CHA
had increased general mental health screening at wellchild visits to over 70% of all patients at every CHA primary care site.21
These past QI projects were implemented by working
with site-based teams and focused on a systems and
process-oriented approach with electronic tools and data
reporting corresponding to the Six Sigma approach to QI.
Six Sigma favors a measurement-based strategy that
focuses on process improvement and reduction in varia-

ACADEMIC PEDIATRICS

tion.22 The ADHD project leaders served as experts to

help facilitate data reporting and standardization of care
processes (ie, the DMAIC process: define, measure,
analyze, improve, control)22 for incremental improvement
of existing processes, in this case the existing planned care
system.
DEFINITION OF NEED AND DESIGN OF THE PROJECT
Before the start of the project, a random review of records of 60 patients with ADHD revealed that only 5%
had documentation of a standardized diagnostic instrument, and 57% had documentation of any assessment of
comorbid conditions. In addition, only about one-third of
patients receiving ADHD medications received recommended follow-up visits. In contrast, providing evidence
of the potential impact of electronic tools and standardization of care, 95% had documentation of body mass index
(BMI) (to monitor for growth impairment as an important
side effect of stimulant medication) because of the EMRs
automatic BMI calculation and previous QI efforts to
improve timeliness and standardization of well care,
including obesity assessment.
In order to evaluate the impact of the clinic-wide systems
changes, this ADHD QI project involved 2 pilot intervention clinics and 2 control clinics. Each pair of sites included
a large-volume site where a majority of patients spoke English (but also included a smaller, more linguistically and
ethnically heterogeneous population) and a small-volume
site that included one dominant, non-English-speaking
population (Spanish, 42% of patients at the intervention
site; Portuguese, 73% at the control site). Although not a
randomized trial, use of this quasiexperimental strategy allowed project leaders to assess the impact of the whole
package of clinical process changes as distinct from tools
added to the EMR. This was necessary before spreading
the intervention to other clinics. The project was approved
and monitored by the CHA institutional review board.
INTERVENTION: THE IMPROVEMENT PROCESS
The physician leaders of the project team served as experts in ADHD (based on experience and interest) who
worked to standardize and improve care processes. At
intervention sites during the first 2 months of the projects
clinical implementation, they provided limited coaching to
pediatricians. Project staff held a single hour-long session
at each intervention site that was supplemented with
detailed email instructions on using new electronic tools
and clinical work flows as they became available. Coaching focused on instruction in the following areas to reduce
variation in practice: 1) adapting clinical work flows and
the planned care system for use with patients with
ADHD; 2) utilization of new electronic tools for diagnosis
and management of ADHD (such as multilingual copies of
Vanderbilt scales that autopopulate demographic fields, instructions for completing schoolhome daily report cards,
and patient educational materials data as well as data entry
flow sheets with graphing functionality for monitoring
Vanderbilt scores over time); and 3) review of clinical process measures derived from AAP guidelines and HEDIS

ACADEMIC PEDIATRICS

measures. Although coaches emphasized adherence to the

AAP guidelines, they did not focus on other aspects of recommended care, such as brief behavioral interventions and
therapy, use of motivational interviewing, and strategies
for self-management.
Before introducing the clinical work flows (for using the
Vanderbilt scales and monitoring patient care and followup) to intervention teams, they were piloted and refined
by the projects lead clinicians. The incorporation of the
Vanderbilt scores into the EMR allowed clinicians to track
patients progress as measured by symptom and performance scores. The Vanderbilt scores were added to the
EMR as editable documents that autopopulated demographic fields and could be printed and provided to patients caregivers and teachers. Before the project, use of
an ADHD rating scale was via paper and depended on
the clinician maintaining copies of any scale that he or
she normally chose to use, if any. Through the incorporation of electronic copies of the Vanderbilt questionnaires
into the EMR, analyses could quantify their provision
and completion. Templates for progress notes documentation as well as educational materials also were made available in the EMR. At control sites, clinicians were informed
of the availability of these electronic practice tools but
otherwise were not coached to use the planned care system,
registry, or work flows.
In addition to being able to monitor Vanderbilt scores,
pediatricians were matched in teams with one nurse and
one medical assistant to meet monthly for 1 hour with a
coordinator for planned care. The planned care coordinator
prepared registry reports and other patient lists from the
EMR. (Although residents and medical students are present in many CHA clinics, they were not specifically
included in this project and are not included in planned
care team meetings generally.) The team reviewed each patients report and determined the need for follow-up or clinical monitoring such as BMI or blood pressure. Between
meetings, the planned care coordinator or another team
member (usually the medical assistant) would contact patients in need of follow-up.
DATA COLLECTION AND CLEANING FOR MEASUREMENTS
ANALYSES OF THE PROJECT
The electronic patient registry was created as a database
external to the EMR. The registry draws patients from the
EMR using an algorithm based on prescriptions of stimulants or atomoxetine or presence in the problem list of an
ICD-9 code starting with 314. The registry also draws clinical information such as dates of visits with an ADHD diagnosis, date of last well care, and details of medication
prescriptions.
During the first 2 months of this 1-year project, clinicians and planned care coordinators updated the EMR
and the registry. To better understand which patients
were being managed by the primary care physician
(PCP), clinicians were asked to indicate in the problem
list whether ADHD was managed by the PCP or a subspecialist, such as a pediatric neurologist or child psychiatrist.

PLANNED CARE APPROACH AND PATIENT REGISTRY

291

The purpose of this was not to exclude patients but to allow

teams to focus efforts on children who were actively
receiving care from their PCP.
OUTCOMES
On the basis of AAP guidelines, the project set clinical
goals that reflected the core aspects of the diagnostic
process and early treatment period as follows:

Administration of the Vanderbilt scales for parent and
teacher during diagnosis, with documentation as
completed and scored in the EMR.

Follow AAP guidelines for medications, facilitated by
standardized medication orders in the EMR.

After initiating medications, reevaluate the child within
1 month and then at least semiannually.

Assess, document, and address comorbid behavioral
conditions.

Measure and document the patients height, weight,
BMI, and blood pressure at initial and all follow-up
visits.
Once the registry was live, reports and data analyses
were designed to assess performance on measures that
reflected the emphasis on the diagnostic and treatment process: 2 HEDIS measures modified for the project, utilization of the Vanderbilt rating scales, and receipt of
semiannual follow-up visits by established patients
receiving stimulant medications. Normally, HEDIS measures for ADHD are applied to children 6 to 12 years of
age who are prescribed medication. For this project, to
reflect the reality that pediatricians are also managing
adolescents, HEDIS measures were modified to include
children up to 18 years of age. In addition, telephone encounters were not encouraged or included as meeting the
follow-up requirements. This was done to emphasize the
need for clinical monitoring of growth and blood pressure.
The HEDIS measures are as follows:
INITIATION PHASE MEASURE
The percentage of patients with an ambulatory prescription dispensed for ADHD medication and who had 1
follow-up visit with practitioner during the 30-day period
after initiation of medication.

AND

CONTINUATION AND MAINTENANCE PHASE MEASURE

The percentage of patients with an ambulatory prescription dispensed for ADHD medication, who continued to
receive the medication for at least 210 days, and who, in
addition to the visit in the initiation phase, had at least 2
follow-up visits with a practitioner within 270 days (9
months) after the initiation phase ended.
It is important to note that HEDIS measures are specifically designed to assess whether patients receiving medications are receiving appropriate follow-up care. Although
patients with ADHD may receive other clinical interventions besides medication, it was beyond the scope of this
pilot project to measure whether such interventions were
delivered and, if delivered, their level of quality or impact
on patient outcomes.

292

GELTMAN ET AL

ACADEMIC PEDIATRICS

Figure. Project timeline.

In setting the goals and outcomes described above, the

continuation and maintenance phase measure was also
approximated in an additional outcome measure: receipt
of semiannual visits. This was done to ensure appropriate
on-going monitoring of patients diagnosed before the start
of the project, a common quality measure used in contracts
by third-party payers in Massachusetts.
Background variables used in descriptive and bivariate
analyses included age, gender, race/ethnicity, language
spoken, insurance type, presence of diagnosis and/or medication in the registry, numbers and timing of visits billed
with an ADHD diagnosis, evidence of completion of a Vanderbilt rating scale, presence of BMI measurement, and
classification code for provider of care. Race was defined
as white, black, Hispanic, other, or missing. Language categories included English, Spanish, Portuguese, and other.
EVALUATION, ANALYSES, AND PROJECT TIMELINE
The project began in October 2009. The first 2 months
(October and November) were devoted to project planning,
development, and piloting of work flows, as well as finalization of tools in the EMR. The next phase entailed 2
months of data cleaning and provider coaching (December
2009 to January 2010). The actual intervention period
began in February 2010 and lasted for 6 months until
August 2010. Evaluation data for demographic characteristics and completion of Vanderbilt scales were collected in
the last 2 months of the 1-year project period, starting in
August 2010. Because the clinical implementation of the
project began roughly at the beginning of calendar year
2010, data used to define the baseline population included
a look-back period of 1 year, starting in January 2009. Data
for patients newly diagnosed during the project came from
the FebruaryAugust 2010 time period.
HEDIS measures for the preceding 12 months were
reported quarterly starting with the transition into the intervention period in January 2010. Thus, the first calculation
of HEDIS measures was for JanuaryMarch 2010, the
period when the project intervention was phased in. To
monitor trends in the clinical process goal of semiannual
visits, another report included a quarterly look back at

receipt of 2 semiannual visits billed with an ADHD diagnosis in the preceding 12 months. After the project intervention period ended in July 2010, the project team
monitored all reporting metrics through July 2012 to assess
the impact as the intervention was spread to nonproject
clinical sites. Reports continued to be used to the present
by planned care teams in their routine clinical work
(Figure).
Analyses of data were conducted both to help define the
patient population and to detect any demographic variations
in the sample, as well as to assess the impact of the project
on follow-up and use of the Vanderbilt scores. As such, analyses included comparisons between the overall (baseline)
patient populations at the 4 clinics with those in the ADHD
registry and between those at intervention and control sites.
Analyses included bivariate cross-tabs with chi-square tests
for categorical variables and Students t tests for continuous
variables, as well as multivariable logistic regression models
to control for potential confounding. Regression analyses
focused on identifying background factors associated with
the clinical care outcome variable of completion of a Vanderbilt rating scale. On the basis of bivariate analyses, independent variables in the regression models included study
site, gender, age, race/ethnicity, language spoken, and insurance type. Analyses did not control for clustering of effects
by clinician.

RESULTS
PATIENT CHARACTERISTICS
The baseline sample of all children with ADHD diagnosed at the 4 project sites before the annual project period
was 706. Overall, the patients with ADHD differed from
the total pediatric population (n 10,713) with respect
to sex, race/ethnicity, preferred language, and age (all
P < .0001). When compared with the total pediatric population at the 4 clinics, patients with ADHD were more
likely to be male (76% vs 49%) and non-Hispanic white
(49% vs 36%). In addition, patients with ADHD were
more likely to speak English (72% vs 54%) and to be
slightly older (mean age 12.8 years vs 11.9 years).

ACADEMIC PEDIATRICS

PLANNED CARE APPROACH AND PATIENT REGISTRY

293

Table 1. Comparative Characteristics of Patients Newly Diagnosed With ADHD During Project Period by Clinical Site*
All
Characteristic

Total
Sex
Male
Female
Age
Years, mean (SD)
Race/ethnicity
White
Black
Hispanic
Other
Missing
Language of medical care
English
Portuguese
Spanish
Other
Insurance
Public or self-pay
Private
Current ADHD diagnosis
By DX only
By medication only
By DX and medication
Number of Visits
One
Two
Three or more
Mean (SD)
Any ADHD assessment tool in chart
Vanderbilt score (teacher, parent, or both)
Other (Connor or unspecified)
None
BMI measurement in chart during project period
Yes
No

73
51
22

Intervention Sites
%

33
69.9
30.1

11.3 (3.9)

19
14

Control Sites
n

80.0
20.0

.038

40
57.6
42.4

10.9 (3.8)

32
8
11.6 (4.0)

.41

31
17
12
11
2

43.7
23.9
16.9
15.5
.

11
7
10
3
.

35.5
22.6
32.3
9.7
.

20
10
2
8
.

50.0
25.0
5.0
20.0
.

.02

57
9
4
3

78.1
12.3
5.5
4.1

25
2
3
3

75.8
6.1
9.1
9.1

32
7
1
0

80.0
17.5
2.5
0.0

.06

44
29

60.3
39.7

17
16

51.5
48.5

27
13

67.5
32.5

.16

45
11
17

61.6
15.1
23.3

18
5
10

54.6
15.2
3.3

27
6
7

67.5
15.0
17.5

.41

19
19
35
2.2 (0.84)

26.0
26.0
48.0

6.1
27.3
66.7

17
10
13
1.9 (0.87)

42.5
25.0
32.5

.001

25
1
47

34.3
1.4
64.4

16
0
17

48.5
.0
51.5

9
1
30

23.1
2.5
76.9

.024

50
23

68.5
31.5

23
10

69.7
30.3

27
13

67.5
32.5

.84

2
9
22
2.6 (0.61)

.0001

ADHD indicates attention-deficit/hyperactivity disorder; BMI, body mass index.

*Inclusion criteria: all children aged 5 to 18 years with visits at 4 study sites and new diagnosis of ADHD between February 1, 2010, and
August 31, 2010.
Chi-square was calculated after excluding Other due to low cell counts.

After data cleaning at the onset of the project, 256 children were in the target age range, were actively receiving
medications for ADHD, and were managed by a PCP at
the 4 sites (as opposed to a child psychiatrist). In addition,
73 more children were diagnosed and managed by the PCP
during the project period, for a total of 329 subjects. Newly
diagnosed patients at intervention sites had more office
visits during the project time period (mean of 2.6 vs 1.9,
P .0001), and in unadjusted analyses, a larger proportion
had completed Vanderbilt assessments compared to the
patients diagnosed at control sites (48.5% vs 23.1%,
P .024) (Table 1).
MODIFIED HEDIS MEASURES AND SEMIANNUAL VISITS
Initial assessment of the modified HEDIS measures from
data at the end of the intervention period in June 2010 at the
2 intervention sites revealed 53% of pediatricians to have
met the initiation phase requirement and 35% to have met
the continuation phase requirement. In the next review of

HEDIS measures, for December 2010, the initiation phase

measure had declined to 49%, while the continuation phase
increased to 40%. In the next report for June 2011, the measures were 51% and 45% respectively. For the continuation
phase measure, that corresponded to a 29% increase.
In assessing the prevalence of 2 or more visits (semiannual) in the preceding 12 months for ongoing care of patients with established diagnoses of ADHD, reports were
available for the 2 calendar years that overlapped with
the project period. In the baseline year (2009), control
sites ranged from 47% to 64% in 4 quarterly measurements, while intervention sites ranged from 35% to 53%.
In the second year, corresponding to the final threequarters of the annual project period and the remaining
quarter of the year 2010, control sites actually declined
to a range of 20% to 38%, while intervention sites remained fairly consistent from 33% to 52%. During the
formal project evaluation period, neither group showed
improvement.

294

GELTMAN ET AL

ACADEMIC PEDIATRICS

Table 2. Adjusted Odds of Having Vanderbilt Score Completion in Patients Diagnosed With ADHD Stratified by Timing of Diagnosis in Relation to QI Project Period, FebruaryAugust 2010
All Patients With
ADHD (n 321)

Previously Diagnosed
(n 250)

Newly Diagnosed
(n 71)

Independent Variable

OR

95% CI

OR

95% CI

OR

95% CI

Site: Intervention vs control

ADHD: New vs preexisting
Age, y
Sex: Female vs male
Insurance: Public/self-pay vs private
Race: Black vs white
Hispanic vs white
Other vs white
Language of medical care: Non-English vs English

2.43**
2.55**
0.78****
0.92
0.94
0.72
2.84*
2.16
1.48

1.254.76
1.295.07
0.700.86
0.441.94
0.491.81
0.301.73
1.206.68
0.855.49
0.732.98

2.25*
NA
0.80***
0.81
1.29
0.96
2.54
2.65
1.33

1.024.99

1.91
NA
0.70**
2.41
0.33
0.31
5.04
1.28
2.11

0.487.68

0.710.91
0.322.01
0.582.89
0.342.67
0.917.08
0.897.86
0.583.04

0.550.88
0.4413.2
0.091.30
0.051.93
0.7932.1
0.1610.5
0.459.81

ADHD indicates attention-deficit/hyperactivity disorder; QI, quality improvement; OR, odds ratio; and CI, confidence interval.
*P < .05.
**P < .01.
***P < .001.
****P < .0001.

VANDERBILT SCALE COMPLETION

In multivariate analysis of Vanderbilt score completion,
logistic regression models were run for those diagnosed
before and during the project period and the aggregate patient group. Eight children were excluded from multivariate analyses due to missing data, for a sample size of
321 (250 with preexisting diagnoses and 71 newly diagnosed). In unadjusted results, newly diagnosed patients
with ADHD at intervention sites were over 3 times as likely
to have a completed Vanderbilt score compared to patients
at control sites; however, after adding covariates to the
model (age, gender, insurance, and race), the odds ratio
was attenuated and was no longer statistically significant.
Age was a strong predictor of the outcome; older children
were less likely to have the Vanderbilt score completed.
The association with site appeared to be primarily
confounded by race. In adjusted analyses of all patients
with ADHD in the overall sample, those at intervention
sites were over 2 times as likely to have a completed Vanderbilt score compared to patients at control sites. New
ADHD diagnosis, younger age at time of diagnosis, and
Hispanic race were significant predictors of having a
completed Vanderbilt score (Table 2).

DISCUSSION
This ADHD planned care project using a patient registry
resulted in increased follow-up care for patients newly
prescribed medication over the course of the project period
as well as improved use of the Vanderbilt rating scale.
Given the relatively low rates of follow-up at the end of
the intervention period, however, more room for improvement existed. During the 3 years after the formal QI
project, in which the planned care process was spread to
other CHA sites, we have documented notable ongoing
improvement. The most recent rate of having 2 visits in
the preceding 12-month period for children who receive
medication for ADHD prescribed by their PCP at all
CHA clinical sites was 93% for the AprilJune 2014 quarterly measurement period.

The modified HEDIS continuation phase measure

showed an increase of 29% in patients who had the recommended visits during the project period and also continued
to improve after formal conclusion of the project. The increase in follow-up visits during the project evaluation
period reflected the increased visits of newly diagnosed children at intervention sites compared to previously diagnosed
children and children at control sites. This finding is consistent with previous work that showed improved follow-up
of children with ADHD when using computer-based
clinical supports.18,19 The increased odds of Vanderbilt
score use with all patients at the intervention sites likely
reflects increased use for monitoring of progress and
administration to patients with preexisting diagnoses that
had never completed them before. The lack of statistical
significance of increased use among newly diagnosed
patients at the intervention sites (despite a clear trend in
that direction, with an adjusted odds ratio of 1.9) is likely
due to the small sample size and an increase in use of the
Vanderbilt score for newly diagnosed patients at control
sites. The latter is supported by the increased odds of use
with newly diagnosed patients at all sites.
This QI project, although not a clinical trial, suggests
that use of elements of the chronic care model, namely
an electronic registry for tracking patients and use of standardized diagnostic and treatment work flows, may result
in improvement in adherence with AAP guidelines for
diagnosis and management of ADHD. This finding is supported by some recent research that looked at similar use of
electronic tools to aid in care of patients with ADHD. Lavigne et al23 reported on the successful use of a computerassisted medication management system for patients with
ADHD. In a study relevant to this project, Carroll et al24
used a computerized decision aid for ADHD diagnosis
that resulted in higher documented medication adjustments, mental health referrals, and visits to mental health
specialists. Finally, in a study of clinical decision aids,
Brinkman et al25 found that use of written materials
describing treatment modalities and medication choices
significantly facilitated physicians involving caregivers in

ACADEMIC PEDIATRICS

shared decision making, an important part of the chronic

care model but not evaluated in this project.
LIMITATIONS AND LESSONS LEARNED
This QI project had several limitations. It was not a randomized trial, and it therefore cannot show any causal relationship between specific components of the intervention
package and improved outcomes. Given the continued
improvements after the conclusion of the formal project
evaluation period, it would seem likely that standardization
of work flows and clinical processes as was done required a
more extended time period before becoming institutionalized and a routine part of clinical care. In addition, analyses
did not control for the impact of individual clinicians on the
outcome measures. Although one unstated goal of the project was to encourage PCPs to manage their own patients
with ADHD, some clinicians clearly remained more
comfortable than others in the diagnosis and treatment of
ADHD regardless of the projects efforts. Such clinician
variation could not be controlled in this limited QI project
which did not include an assessment of clinician fidelity.
Despite the improved use of the Vanderbilt scales, the
project revealed variations in the diagnosis based on racial
and linguistic differences in the population that could also
reflect clinician practice variation. This finding supports
the need for better standardization of diagnostic and treatment processes. Some racial and linguistic variation likely
reflected the anticipated distributions of these variables
among sites; however, the project revealed underrepresentation of all linguistic and some racial minorities in the baseline ADHD population at the onset of the project. Although
blacks were proportionately represented in the ADHD
population, Hispanics, Asians, and those categorized as
other were underrepresented. This racial profile roughly
corresponds to the non-English-speaking population being
underrepresented as well. Other studies have reported
such underdiagnosing of minorities2629 as well as the
underutilization of care (including medications).14 By making the Vanderbilt scores available in multiple languages in
the EMR, the project had hoped to improve the linguistic
proportionality of diagnoses. Although not reaching statistical significance, a clear trend in use of the Vanderbilt score
favored patients from families of non-English speakers. This
may have reflected the greater need to use such an instrument with non-English speakers as opposed to relying on
reports from other evaluators such as school neuropsychological assessments and educational plans.
In conclusion, a planned care system and electronic patient registry show promise for improving diagnostic and
treatment processes for patients with ADHD through
improved follow-up and use of a standardized diagnostic
instrument. Although important improvements can be
achieved in these areas, ensuring initial engagement in
care and subsequent follow-up of patients can be challenging, even with the assistance of a planned care coordinator. Furthermore, this project did not emphasize shared
decision making and personal goal settingtwo important
elements of the chronic care model. Given the importance

PLANNED CARE APPROACH AND PATIENT REGISTRY

295

of the educational system in the treatment of children with

ADHD, additional coordination with local school systems
and community behavioral health clinicians is likely to be
necessary to improve initiation of care. Subsequent work in
this area could be enhanced by facilitating communication
with schools and parents around early engagement in care
and use of shared decision making, as suggested in AAP
guidelines and other research.10,25,30,31

ACKNOWLEDGMENTS
This project was funded by a pilot grant from the Harvard Catalyst, The
Harvard Clinical and Translational Science Center (DA Link, PI). The
funder had no role in the design and conduct of the project or preparation
of this document.

REFERENCES
1. Centers for Disease Control and Prevention. Mental health in the
United States: prevalence of diagnosis and medication treatment for
attention-deficit/hyperactivity disorder: United States, 2003.
MMWR Morb Mortal Wkly Rep. 2005;54:842847.
2. Froehlich TE, Lanphear BP, Epstein JN, et al. Prevalence, recognition, and treatment of attention-deficit/hyperactivity disorder in a national sample of US children. Arch Pediatr Adolesc Med. 2007;161:
857864.
3. MTA Cooperative Group. A 14-month randomized clinical trial of
treatment strategies for attention-deficit/hyperactivity disorder. Arch
Gen Psychiatry. 1999;56:10731086.
4. MTA Cooperative Group. Moderators and mediators of treatment
response for children with attention-deficit/hyperactivity disorder.
Arch Gen Psychiatry. 1999;56:10881096.
5. Biederman J, Monuteaux MC, Spencer T, et al. Do stimulants protect
against psychiatric disorders in youth with ADHD? A 10-year followup study. Pediatrics. 2009;124:7178.
6. Epstein JN, Langberg JM, Lichtenstein PK, et al. Attention-deficit/
hyperactivity disorder outcomes for children treated in communitybased pediatric settings. Arch Pediatr Adolesc Med. 2010;164:
160165.
7. Sayal K, Owen V, White K, et al. Impact of early school-based
screening and intervention programs for ADHD on childrens outcomes and access to services: follow-up of a school-based trial at
age 10 years. Arch Pediatr Adolesc Med. 2010;164:462469.
8. American Academy of Pediatrics, Committee on Quality Improvement, Subcommittee on Attention-Deficit/Hyperactivity Disorder.
Clinical practice guideline: diagnosis and evaluation of the child
with attention-deficit/hyperactivity disorder. Pediatrics. 2000;105:
11581170.
9. American Academy of Pediatrics, Committee on Quality Improvement, Subcommittee on Attention-Deficit/Hyperactivity Disorder.
Clinical practice guideline: treatment of the school-aged child with
attention-deficit/hyperactivity disorder. Pediatrics. 2000;108:
10331044.
10. American Academy of Pediatrics, Subcommittee on AttentionDeficit/Hyperactivity Disorder, Steering Committee on Quality
Improvement and Management. ADHD: clinical practice guideline
for the diagnosis, evaluation, and treatment of attention-deficit/
hyperactivity disorder in children and adolescents. Pediatrics. 2011;
128:10071022.
11. Wolraich ML, Lambert EW, Worley KA, et al. Psychometric properties of the Vanderbilt ADHD diagnostic parent rating scale in a
referred population. J Pediatr Psychol. 2003;28:559568.
12. National Institute for Childrens Health Quality. Child health topics:
ADHD. Available at: http://www.nichq.org/areas_of_focus/adhd_
topic.html. Accessed June 24, 2011.
13. Rushton JL, Fant KE, Clark SJ. Use of practice guidelines in the primary care of children with attention-deficit/hyperactivity disorder.
Pediatrics. 2004;114:e23e28.

296

GELTMAN ET AL

14. Ray GT, Levine P, Croen LA, et al. Attention-deficit/hyperactivity

disorder in children: excess costs before and after initial diagnosis
and treatment cost differences by ethnicity. Arch Pediatr Adolesc
Med. 2006;160:10631069.
15. Larson K, Russ SA, Kahn RS, et al. Patterns of comorbidity, functioning, and service use for US children with ADHD, 2007. Pediatrics. 2011;127:462470.
16. Leslie LK, Weckerly J, Plammons D, et al. Implementing the American Academy of Pediatrics attention-deficit/hyperactivity disorder
diagnostic guidelines in primary care settings. Pediatrics. 2004;114:
129140.
17. Wagner EH. Chronic disease management: what will it take to
improve care for chronic illness? Eff Clin Pract. 1998;1:24.
18. Lavigne JV, Dulcan MK, Lebailly SA, et al. Computer-assisted
management of attention-deficit/hyperactivity disorder. Pediatrics.
2011;128:e1e8.
19. Co JPT, Johnson SA, Poon EG, et al. Electronic health record decision
support and quality of care for children with ADHD. Pediatrics. 2010;
126:239246.
20. National Committee for Quality Assurance. Follow-up care for children prescribed ADHD medication. Available at: http://www.ncqa.
org/LinkClick.aspx. Accessed July 5, 2011.
21. Hacker KA, Penfold R, Arsenault L, et al. Screening for behavioral
health issues in children enrolled in Massachusetts Medicaid. Pediatrics. 2014;133:4654.
22. Boaden RJ, Harvey G, Moxham C, et al. Quality Improvement:
Theory and Practice in Healthcare. Coventry, UK: NHS Institute
for Innovation and Improvement; 2008.

ACADEMIC PEDIATRICS
23. Lavigne JV, Dulcan MK, LeBailly SA, et al. Computer-assisted management of attention-deficit/hyperactivity disorder. Pediatrics. 2011;
128:e46e53.
24. Carroll AE, Bauer NS, Dugan TM, et al. Use of a computerized
decision aid for ADHD diagnosis: a randomized controlled trial.
Pediatrics. 2013;132:e623e629.
25. Brinkman WB, Hartl Majcher J, Pling LM, et al. Shared decisionmaking to improve attention-deficit hyperactivity disorder. Patient
Educ Couns. 2013;93:95101.
26. Olfson M, Garmeroff MJ, Marcus SC, Jensen PS. National trends in
the treatment of attention deficit hyperactivity disorder. Am J Psychiatry. 2003;160:10711077.
27. Stevens J, Harman JS, Kelleher KJ. Race/ethnicity and insurance
status as factors associated with ADHD treatment patterns. J Child
Adolesc Psychopharmacol. 2005;15:8896.
28. Rowland AS, Umback DM, Sallone L, et al. Prevalence of medication
treatment for attention deficit-hyperactivity disorder among elementary school children in Johnston County, North Carolina. Am J Public
Health. 2002;92:231234.
29. Hoagwood K, Jensen PS, Feil M, et al. Medication management of
stimulants in pediatric practice settings: a national perspective. J
Dev Behav Pediatr. 2000;21:322331.
30. Fiks AG, Localio AR, Alessandrini EA, et al. Shared decision-making
in pediatrics: a national perspective. Pediatrics. 2010;126:306314.
31. Fiks AG, Hughes CC, Gafen A, et al. Contrasting parents and
pediatricians perspectives on shared decision-making in ADHD.
Pediatrics. 2011;127:e188e196.