Keywords attention-deficit/hyperactivity disorder - information dissemination - surveys and
questionnaires - communications - other clinical informatics applications
Background and Significance
Background and Significance
Attention-deficit/hyperactivity disorder (ADHD) is the most common inheritable chronic
childhood neurobehavioral disorder, affecting approximately 6.5% of school-aged children.[1 ]
[2 ]
[3 ] The effects of ADHD manifest in multiple environments, but most commonly lead to
issues at both home and school.[2 ] As a result, effective management of children with ADHD involves collaboration between
parents, clinicians, and teachers. ADHD rating scales (e.g., Vanderbilt scales) are
the standard method for clinicians to collect data and monitor disease severity during
evaluation and treatment of ADHD.[2 ]
[4 ] In clinical practice, rating scales represent a one-way, parallel communication
from parent and teacher to clinician; there is a missed opportunity to foster communication
between parents and teachers, key members of a child's ADHD management team.
Recognizing the need for collaboration, the 2011 American Academy of Pediatrics Guidelines
for the care of children with ADHD stress the need for shared decision making (SDM).[2 ] SDM involves the identification of possible treatment options, sharing of preferences
and goals for intervention, and implementation of the preferred treatment plan.[5 ] Although research has begun to demonstrate benefits of SDM in ADHD care,[6 ]
[7 ] communication challenges between families, clinicians, and schools remain. Standard
use of data collection instruments does not support SDM as parents and teachers do
not have access to data collected by clinicians. The role of technology in facilitating
this information sharing is under investigation.[8 ]
Despite the need for teamwork across settings, fragmentation of interventions between
schools and the health care system often undermines ADHD care and limits SDM.[9 ]
[10 ]
[11 ]
[12 ] Parents and teachers may wish to collaboratively set behavior and education goals
informed by areas of concern identified during ADHD evaluations. In addition, through
sharing parents may become aware of issues that present exclusively at school that
should be addressed in treatment plans. When families experience difficulty communicating
their preferences and goals for treatment to providers and educators involved in their
child's care, the quality and adherence to treatment plans may be adversely impacted.[13 ] Despite ubiquitous mobile devices transforming interpersonal communication,[14 ] sharing of ADHD information between interested parties continues to be inconsistent.[15 ]
[16 ] Multiple barriers, including school policies, can lead to decreased communication
between parents, teachers, and clinicians.[17 ]
[18 ]
[19 ]
[20 ] While validated rating scales have helped clinicians diagnose and monitor ADHD,[4 ]
[21 ]
[22 ] less work has been done to address barriers to communication and fragmentation of
care, which can result in suboptimal outcomes.[15 ]
[16 ]
[17 ] In addition, poor communication and coordination may result in duplication of services
and inadequate monitoring and tailoring of treatment to align with a child's needs
and his or her family's preferences and goals.[23 ]
[24 ] Quality of care and uptake of evidence-based therapies have been reported to be
low in poorly integrated systems.[9 ]
[20 ]
[25 ] Informatics-based interventions that can facilitate cross-systems collaboration
and support ongoing communication are urgently needed.[16 ]
Building on prior success in using an internally developed tool for ADHD monitoring
by clinicians,[26 ]
[27 ]
[28 ]
[29 ]
[30 ] we sought to develop a system to improve communication between parents and teachers
to decrease fragmentation of care. We targeted communication between parents and teachers
due to recognition that this data exchange channel was unsupported by our electronic
health record (EHR), and hypothesized that our current clinical decision support (CDS)
could be modified to support this information sharing. During this study, we evaluated
patient characteristics that may have influenced parents and teachers to use this
system to share and view patient ADHD information. Recently, it has been demonstrated
that Web-based portals help improve coordination among parents and providers for the
management of pediatric chronic medical conditions.[31 ]
[32 ] Web-based interventions have been developed for ADHD as well, and have been shown
to improve rating scale completion and patient satisfaction and patient outcomes.[26 ]
[33 ]
[34 ]
[35 ]
[36 ]
[37 ] Although portals have been developed for pediatricians to collect ADHD information
from parents and teachers and can support communication, no previous interventions
support automated sharing of ADHD survey responses between parents and teachers. Based
on studies of children with other chronic medical conditions,[31 ]
[32 ] we hypothesized that parents of more severely affected children would be more likely
to share information with teachers.
Objective
We sought to adapt our institution's EHR-linked system for ADHD symptom monitoring
to support communication between parents and teachers. Upon completion, we then evaluated
child characteristics associated with sharing of ADHD information to determine the
relationship between socioeconomic and other clinical factors on willingness to share
information.
Methods
Study Setting and Population
We conducted this study within the Children's Hospital of Philadelphia (CHOP) ambulatory
care network. We undertook a retrospective cohort study to evaluate the use of a system,
the ADHD Care Assistant, to facilitate sharing of information on ADHD symptoms, performance
impairments, medication side effects, comorbid symptoms, and treatment goals between
parents and teachers. The original software was made available on November 1, 2014.
The study period began on January 25, 2017 when the updates to the software were implemented
to allow for sharing between parents and teachers and ended on June 16, 2017 to align
with the end of the school year. In our analysis, we included all patients whose parent/guardian
used the ADHD Care Assistant at least once during the study period (N = 590). A study flow diagram ([Fig. 1 ]) demonstrates how the study sample was derived. The ADHD Care Assistant was displayed
to clinicians for all patients with a diagnosis of ADHD or a chief complaint of “behavior problem ” within the reason for visit select box. Use of this system was optional and offering
it to families was solely at the discretion of the primary clinician. The institutional
review board at CHOP approved this project and granted a waiver of consent.
Fig. 1 Study flow diagram. Approximately 8.7% of patients seen since the introduction of
the attention-deficit/hyperactivity disorder (ADHD) Care Assistant have ADHD or had
a concern raised for ADHD. Of these, 14% were registered by clinicians to use the
intervention. Nearly half (45%) of parents who registered to use the intervention
completed at least one survey. Of parents who used the system, 44% completed at least
one survey during the study period.
The ADHD Care Assistant
The ADHD Care Assistant ([Fig. 2 ]) is an internally developed CDS module used across the CHOP's network of 31 primary
care sites to improve data collection from parents and teachers of school-aged children
with ADHD.[26 ] Through the ADHD Care Assistant, a clinician can register a parent and teacher to
receive scheduled ADHD symptom surveys, which are delivered directly to the parent's
personal email address. A form given to the teacher provides instructions on how to
register and receive surveys, which are then delivered directly to the teacher's email
address. Parents and teachers completed surveys by clicking the survey-specific link
and assessing information on symptoms, performance impairments, comorbid symptoms,
and medication side effects based on the Vanderbilt ADHD Rating Scales.[4 ]
[21 ] In addition, parents provided information on their ADHD treatment preferences and
goals using the ADHD Preference and Goal Instrument (PGI).[6 ]
[38 ] Both the Vanderbilt and PGI are validated survey instruments for children with ADHD.
The initial parent survey instrument contained 104 questions while the initial teacher
survey and follow-up survey instruments each contained approximately 45 questions.
Prior to initializing emailed surveys, a clinician obtains a Health Insurance Portability
and Accountability Act (HIPAA) release in the office and gave the parent a Family
Educational Rights and Privacy Act (FERPA) release form so that the school could document
parental permission for the teacher to use the system.
Fig. 2 The Attention-Deficit/Hyperactivity Disorder (ADHD) Care Assistant. The ADHD Care Assistant has these functions: (1) email-based support for gathering
information from parents and teachers including ADHD symptoms, performance impairments,
and treatment preferences and goals, (2) automatic delivery of information to clinicians
through the electronic health record (EHR), and (3) links to educational materials.
Our institution uses this system for children with behavior concerns as a screening
tool and to track symptom severity, performance, treatment goals, and, when applicable,
medication side effects for patients diagnosed with ADHD. No particular script was
provided to help clinicians introduce the system. Use of the system was voluntary
and activation of the system required a primary care clinical encounter. In this system,
parents annually updated their child's teacher. Consistent with standard practice,
we initially developed this system with a unidirectional data flow from parents and
teachers directly into the EHR for clinician review and did not facilitate parent–teacher
communication. To inform development of the system to support parent–teacher communication,
we solicited input from 8 parents, 11 pediatricians, and 8 educators in a series of
stakeholder meetings between September 2015 and September 2016. Based on stakeholder
input, we enhanced the system to support parents selectively sharing portions from
each submitted survey with a teacher, a key feature not present in similar software
systems, and one that had yet to be studied. Secure viewing of submitted surveys was
achieved using Cisco Registered Envelope Service (Cisco System, Inc., San Jose, California,
United States), which requires users to register prior to accessing secured messages.[39 ] The teacher–parent information sharing functionality is shown in [Fig. 3 ].
Fig. 3 Messages handling upon survey completion. Upon survey submission, the clinician is
alerted, the raw data are stored in the electronic health record (EHR), and a copy
is sent to the survey submitter. Parent survey data can be selectively shared with
teachers and teacher survey data are automatically shared with parents.
Endpoints
We selected as the primary study endpoint parent/guardian sharing of any component
of the parent ADHD survey at first opportunity during the study period. When sharing
occurred, we further analyzed the type of information shared (symptoms, performance,
goals, and/or medication side effects). For patients with multiple opportunities to
share, we conducted a secondary analysis to assess patterns in sharing over the study
period. For this secondary analysis, we restricted the sample to children whose parent/guardian
shared at the first opportunity and completed at least two surveys during the study
period. Finally, we investigated the frequency with which teachers and parents viewed
shared information. For this analysis, we included all surveys completed during the
study period for our study population.
Child Characteristics Associated with Parent Sharing
We assessed child characteristics that were potential drivers of parent/guardian sharing.
Characteristics included ADHD symptom and performance scores calculated using standard
Vanderbilt Rating Scale scoring practices.[4 ] Specific characteristics used in the analysis included: (1) ADHD symptoms (categorized
as no symptoms, inattentiveness only, hyperactivity only, or combined symptoms), (2)
performance impairment (categorized as no impairments, academic impairment only, interpersonal
impairment only, or both), (3) symptoms of comorbidities (none, any, and symptoms
for oppositional defiance disorder, conduct disorder, or anxiety/depression), (4)
prior diagnosis or problem list entry of ADHD, (5) medication side effects (categorized
as none, mild, or moderate/severe side effects), (6) survey type (initial or follow-up),
and (7) completion of a teacher survey during the study period. In our primary analysis,
we included each family's first survey during the study period. That first survey
could be either an initial or follow-up ADHD rating scale survey depending on whether
it was the parent's first time using the ADHD Care Assistant rating scale collection
system (initial survey) or whether they were already using the system at the start
of the study (follow-up survey). While the initial and follow-up survey both have
sections addressing symptom and performance impairment, it is important to note that
only the initial survey contained the treatment goal section and only the follow-up
survey contained the medication side effect section. Parent, clinician, and teacher
demographics were not obtained as part of this study as we relied on data collected
during the provision of routine medical care.
Covariates
Covariates included patient age (3–12 vs. 13–18 years), sex, race (white, African
American, or other race), Hispanic/Latino ethnicity, insurance status (Medicaid vs.
private), practice setting (urban vs. suburban), ADHD medication orders, and parent
report of ADHD medication use. We selected these age ranges to evaluate differences
in sharing in primary school versus high school. Ten parents of preschool-aged children
used the system to submit ADHD information from parents and preschool educators. We
performed a subgroup analysis on the 9- to 12-year-old cohort to assess if this population
was significantly different from the younger primary school group. If either a parent
reported ADHD medication use on the survey or the EHR contained an order for ADHD
medications, this patient was classified as a medication user.
Statistical Analysis
We examined the distribution of ADHD characteristics and patient demographics in the
study population. For our primary analysis, we examined bivariate associations between
these characteristics and whether parents decided to share information at the first
opportunity using chi-squared tests. After identifying patient ADHD characteristics
associated with sharing (at the p < 0.1 level), we used multivariable logistic regression to estimate associations
of these characteristics with sharing at the first opportunity. Characteristics examined
in the full model included ADHD symptoms, performance impairments, survey type, and
teacher survey completion. We controlled for patient covariates that were associated
with sharing in bivariate analyses, including age, race, insurance status, and ADHD
medication use based on documentation in either the EHR or parent report through the
ADHD Care Assistant.
We then examined viewing patterns of survey results by parents and teachers. We calculated
the proportions of shared parent surveys that were viewed by the teacher and teacher
surveys that were viewed by the parent (all teacher surveys are shared automatically).
We looked for associations between patient characteristics and teacher viewing of
surveys. Finally, we examined patterns of sharing over time. We calculated the proportion
of children whose parent/guardian shared at least two surveys, and examined the distribution
of total number of surveys shared by each parent. We examined whether there were differences
in parent and teacher viewing of survey results between parents who chose to share
a subsequent survey and those who did not.
Results
Study Population
Between January 25, 2017 and June 16, 2017, 590 parents/guardians completed at least
one survey using the ADHD Care Assistant for their child; 232 of these children had
a teacher survey completed during the study period (39.3%). A total of 1,006 parent
surveys and 493 teacher surveys were completed overall (range: 1–10 parent surveys
per child and 0–10 teacher surveys per child).
[Table 1 ] presents the distribution of demographic characteristics of the study population.
Most children (88%) in the study had a diagnosis of ADHD either before or during the
study period. Overall, 85.9% of patients were aged 3 to 12 years, 70.2% were male.
Note that 48.3% were white, 40.0% were African American, and 5.1% were Hispanic/Latino.
A total of 63.6% were privately insured, and 61.5% were from suburban practices. ADHD
medication use was documented and/or reported for 87% of children.
Table 1
Demographic and clinical characteristics of study population and bivariate associations
with whether parents chose to share ADHD survey information with teachers on the first
survey in the study period
Demographic characteristic
N (%) with characteristic (column %)
Shared at first opportunity[a ]–N (row %)
p -Value[c ]
Total, N
590
380 (64.4)
Age (y)
< 0.001
3–12
507 (85.9)
345 (68.1)
13–18
83 (14.1)
35 (42.2)
Sex
0.8
Male
414 (70.2)
268 (64.7)
Female
176 (29.8)
112 (63.6)
Race
0.03
White
285 (48.3)
169 (59.3)
African American
236 (40.0)
160 (67.8)
Other race
69 (11.7)
51 (73.9)
Ethnicity
0.3
Hispanic/Latino
30 (5.1)
22 (73.3)
Not Hispanic/Latino
560 (94.9)
358 (63.9)
Insurance status
0.06
Medicaid
215 (36.4)
149 (69.3)
Private
375 (63.6)
231 (61.6)
Primary care practice type
0.1
Urban
227 (38.5)
155 (68.3)
Suburban
363 (61.5)
225 (62.0)
ADHD medication use[b ]
< 0.001
Yes
513 (86.9)
314 (61.2)
No
77 (13.1)
66 (85.7)
Abbreviations: ADHD, attention-deficit/hyperactivity disorder; EHR, electronic health
record.
a Parent and teacher surveys were administered at regular intervals set by the child's
primary care clinician. In our primary analysis, we included each family's first survey
during the study period and determined whether the parent chose to share any component
of the survey with their child's teacher.
b Based on EHR documentation and parent report. Of children with medications listed
on medication recorded in the EHR, 92.0% were on a stimulant only, 4.7% were on a
stimulant and either an α agonist or atomoxetine, 1% were on an α agonist only, and < 1%
were on other combinations.
c
p -Values calculated using chi-squared tests to compare the proportion of parents in
each category that chose to share information with teachers at the first opportunity.
Value of Sharing to Parent and Teacher Stakeholders
During the preimplementation stakeholder meetings, parents and teachers both indicated
desire for access to the information they personally submitted and to the information
that was submitted by their counterparts. Stakeholder parents expressed concern that
some parents may wish to control which information was shared with the teacher, using
examples from the Vanderbilt Rating Scale of answers to particular questions that
people may not like to share (e.g., #32: Has stolen things that have value).[21 ] To address this acceptability concern, we developed our system to allow selective
sharing at the section level. The shareable sections were ADHD symptoms, performance
impairments, treatment goals (initial survey only), and medication side effects (follow-up
surveys only).
Parents and teachers both requested to see “what the clinician sees” in terms of presentation
of results. To support this function, we provided both a longitudinal view and raw
data within the reports sent to parents and teachers. We used the same color-coding
of results that was provided to clinicians—green for good, yellow for borderline,
and red for areas of concern. We did not provide scores or cutoffs for diagnosis,
as this was felt to be beyond the reach of the CDS application and would not facilitate
communication between parents and teachers.
Parent and teacher stakeholders reviewed prototype reports and provided feedback during
the iterative development process. The stakeholders indicated that the layout and
content of the result reports in their final form would be beneficial in supporting
SDM surrounding ADHD diagnosis and management ([Fig. 4 ]). Overall, stakeholders valued the possibility of an electronic tool to address
known difficulties in communication of ADHD information.
Fig. 4 Example of emailed report sheet. This example of a patient-directed report sheet
generated by the Attention-Deficit/Hyperactivity Disorder (ADHD) Care Assistant includes
longitudinal data summarizing all scores submitted since the patient was registered
and includes the questions/answers submitted on the most recent ADHD survey response.
Parent and teachers reports are formatted the same.
Sharing at the First Opportunity
Out of 590 patients whose parent completed a survey during the study period, 380 parents
chose to share results with their child's teacher at the first opportunity (64.4%,
[Table 1 ]). Most parents (304, 80.0%) shared all sharable survey components with their child's
teacher. In bivariate analysis, parents of children aged 3 to 12 and of children who
were publically insured were more likely to share than parents of adolescents or parents
of privately insured children ([Table 1 ]). In a subgroup analysis, we noted no difference between the youngest (3–8 years)
and middle aged (9–12 years) patient groups. Parents of white children were less likely
to share, as were parents of children using medication for ADHD. Practice location
(urban vs. suburban) did not affect the likelihood of decision to share.
In bivariate analyses ([Table 2 ]), parents were more likely to share an initial compared with a follow-up survey
(p < 0.001). In addition, parents whose child was manifesting elevations in ADHD symptoms,
performance impairments, or symptoms of any comorbidity (scoring as per Vanderbilt
Rating Scale instructions) were significantly more likely to share information with
teachers (p < 0.05 for all comparisons). Having no prior diagnosis of ADHD was associated with
an increased likelihood of a parent selecting to share the results with the teacher.
Medication side effects were not associated with sharing. Parents of children whose
teacher completed a survey during the follow-up period were also more likely to share
information with teachers (p = 0.001). In a secondary analysis, we looked at the relationship between parental
preferences and goals and information sharing. We found that these were not associated
(p ≥ 0.05 for all preference and goal categories).
Table 2
Bivariate associations of ADHD characteristics with parents' decision to share information
with teachers at the first opportunity
ADHD characteristics: Parent surveys[a ]
N (%) with characteristic (column %)
Shared at first opportunity –
N (row %)
p -Value[c ]
Total, N
590
380 (64.4)
Survey type
< 0.001
Initial
180 (30.5)
141 (78.3)
Follow-up
410 (69.5)
239 (58.3)
Teacher completed survey during study
0.001
Yes
232 (39.3)
168 (72.4)
No
358 (60.7)
212 (59.2)
ADHD diagnosis or problem list entry
< 0.001
Yes
520 (88.1)
325 (62.5)
No
70 (11.9)
55 (78.6)
ADHD symptoms reported[b ]
0.02
No ADHD symptom criteria met
302 (51.2)
176 (58.3)
Inattentive symptom criteria met
124 (21.0)
89 (71.8)
Hyperactive symptom criteria met
31 (5.3)
21 (67.7)
Both inattentive and hyperactive symptom criteria met
133 (22.5)
94 (70.7)
Performance impairments[b ]
< 0.001
No impairment criteria met
184 (31.2)
95 (51.6)
Academic only
163 (27.6)
115 (70.6)
Interpersonal only
69 (11.7)
47 (68.1)
Both academic and interpersonal
174 (29.5)
123 (70.7)
Symptoms of comorbidities[d ] (N = 180)
0.03
No criteria for comorbidities met[b ]
102 (56.7)
74 (72.6)
Criteria met for 1 or more comorbidity
78 (43.3)
67 (85.9)
Criteria met for specific comorbidities
Oppositional defiance
66 (36.7)
55 (83.3)
0.2
Conduct disorder
12 (6.7)
10 (83.3)
0.7
Anxiety or depression
37 (20.6)
34 (91.9)
0.03
Medication side effects (N = 410)
0.8
No side effects
46 (11.2)
25 (54.4)
Mild side effects
119 (29.0)
69 (58.0)
Moderate/severe side effects
245 (59.8)
145 (59.2)
Abbreviation: ADHD, attention-deficit/hyperactivity disorder.
a ADHD symptoms and performance items were asked on both initial and follow-up surveys
(N = 590 parent surveys and 232 teacher surveys). Comorbidities were only assessed on
initial surveys (N = 180 parent surveys and 104 teacher surveys), while side effects were only assessed
on follow-up surveys (N = 410 parent surveys and 128 teacher surveys).
b Scoring performed according to instructions provided in the Vanderbilt survey handout.
No symptoms/evidence in each category indicates that a parent or teacher did not indicate
a positive in any subsection of the scale. Inattentive and hyperactive symptoms require
6/9 positive responses. Other sections of the Vanderbilt had varying criterion for
positives.
c
p -Values calculated using chi-squared tests to compare the proportion of parents in
each category that chose to share information with teachers at the first opportunity.
d This reflects symptoms of comorbidities reported by parent at time of survey completion—not
diagnosed comorbidities. The comparator for each comorbidity group was no comorbidity.
Many children screened positive for more than one comorbidity.
In multivariable models controlling for patient age, race, insurance status, and ADHD
medication use, parents of children experiencing ADHD-related performance impairments
in academics, interpersonal relationships, or both were significantly more likely
to share relative to those with no impairments (odds ratios [ORs] of 1.95 [95% confidence
interval [CI]: 1.21, 3.15], 2.03 [1.09, 3.76], and 2.09 [1.26, 3.44], respectively,
[Table 3 ]). In addition, parents of children whose teacher had completed a survey during the
study period had 1.5 times greater odds of sharing (OR, 1.52 [95% CI: 1.03, 2.23].
Finally, parents tended to be more likely to share an initial compared with a follow-up
survey (OR, 1.61 [95% CI: 0.97, 2.66], p = 0.06). In these models, prior diagnosis of ADHD and ADHD symptom scores were not
associated with information sharing.
Table 3
Multivariable associations of ADHD characteristics with parents' decision to share
ADHD survey information with teachers at the first opportunity
Child's ADHD characteristics
Parent shared with teacher at first opportunity, odds ratio (95% CI)[a ]
Survey type – Initial (vs. follow-up)
1.61 (0.97, 2.66)
Teacher completed a survey during study (vs. not)
1.52 (1.03, 2.23)
ADHD diagnosis or problem list entry
0.79 (0.40, 1.59)
ADHD symptoms at time of survey completion
No symptoms
Reference
Inattentiveness only
1.12 (0.67, 1.88)
Hyperactivity only
0.92 (0.40, 2.14)
Combined symptoms
0.88 (0.52, 1.50)
ADHD-related performance impairments
None
Reference
Academic impairment only
1.95 (1.21, 3.15)
Interpersonal impairment only
2.03 (1.09, 3.76)
Both academic and interpersonal impairments
2.09 (1.26, 3.44)
Abbreviations: ADHD, attention-deficit/hyperactivity disorder; CI, confidence interval;
EHR, electronic health record; OR, odds ratio.
a Results are from multivariable logistic regression of the characteristics above and
sharing at the first opportunity. ADHD characteristics that were included were bivariately
associated with sharing (p < 0.1 as cutoff). Models additionally adjusted for patient demographic and clinical
characteristics that were bivariately associated with sharing, including child age,
race, insurance payer, and ADHD medication use (documented in the EHR or based on
parent report).
Viewing of Surveys
Of the 667 shared parent surveys, 107 (16.0%) were viewed by the child's teacher,
whereas 149 of the 493 teacher surveys were viewed by parents (30.2%). Parents and
teachers viewed their own survey results at similar rates: 71 out of 1,006 total parent
surveys (7.1%) and 38 out of 493 total teacher surveys (7.7%) (chi-square, p = 0.2). Parents and teachers were both more likely to view information provided by
the other rather than viewing their own results. There was no relationship between
ADHD diagnosis and either parent or teacher viewing email results. Teachers of children
in the suburban setting and of children with private insurance were significantly
more likely to view shared surveys (8.8% vs. 20.3% [p < 0.001] and 19.6% vs. 19.9% [p = 0.001], respectively).
Subsequent Sharing
A total of 152 parents/guardians shared survey results with teachers on the first
opportunity and completed at least two surveys during the study period (25.8% of the
total sample). Among those with the opportunity to share again, 135 (88.8%) shared
results with teachers on a second occasion. There was no difference in the percent
of parents viewing teacher surveys between those who did and did not share a later
survey (17.0% vs. 17.7%). In addition, 22 (30.9%) parents/guardians who elected not
to share at the first opportunity did choose to share at a subsequent opportunity.
Discussion
Fragmentation of services may lead to care inefficiencies and suboptimal care.[15 ] Technology can help promote SDM and decrease fragmentation but has not been widely
integrated in ADHD management.[40 ] SDM in ADHD depends upon information exchange among parents, clinicians, and teachers,
yet multiple barriers to this process remain. Our results indicate that the majority
of families were willing to share relevant health information with teachers using
a technology-assisted approach, which can promote SDM. Performance impairments (OR,
2.09) were the strongest patient characteristic associated with sharing, with younger
patient age also noted to be an important predictor, consistent with other literature
on ADHD help-seeking adolescents.[41 ]
[42 ] Younger patient age was also noted to be an important predictor which may be reflective
of a trend for parents to become less involved in their children's schooling when
they become adolescents.[43 ]
[44 ] This is consistent with other studies addressing barriers to ADHD treatment.[20 ] Despite a large quantity of sharing, only a minority of shared surveys were viewed;
parents and teachers viewed 16 and 30% of shared surveys, respectively.
Our results indicate that the majority of families are willing to share health information
with teachers when that shared information is essential to effective treatment. Despite
good reasons for having privacy laws (HIPAA for health care and FERPA for scholastic
areas), these regulatory burdens can impair the sharing of information, even when
all stakeholders agree that sharing is desirable. Our ADHD information sharing system,
which adhered to all regulatory constraints, demonstrated that a technological approach
to sharing is acceptable to families. Despite needing to opt in for sharing, most
parents (64.4%) decided to share at least part of their survey responses with their
child's teacher and 80% who elected to share, shared everything. We also found that
parents who started using the system after the sharing features were activated (i.e.,
families completing initial Vanderbilt surveys with the new system) tended to share
more than established users of the system (i.e., parents completing follow-up surveys
when sharing features became available). This could reflect that the novelty of the
system was a driving force for sharing.[45 ]
[46 ] Sharing rates for initial surveys may also be higher for other reasons, including
potentially the ability to share the treatment goal section, which is unique to the
initial survey. Those parents who elected to share on the initial survey were very
likely to share subsequent surveys suggesting a perceived benefit for parents from
sharing, given the need to opt in at each opportunity.[47 ] Other factors that could have influenced sharing include parental trust in the medical
system and the influence of the primary care physician on the parent decision-making
process, even outside of the office visit.
We hypothesized that increased ADHD severity would be associated with increased sharing.
Interestingly, we found performance impairments were associated with sharing after
controlling for confounding, but simply having symptoms was not. This finding likely
reflects the perceived importance of performance (e.g., academic problems) as opposed
to symptoms in seeking to enlist a teacher's help by sharing information.[41 ]
[42 ]
[48 ]
[49 ] Parents also could be more interested in sharing if they believed that there was
something that needed to be fixed.[31 ]
[32 ]
[41 ]
[42 ] Our results suggest the need to especially encourage system use among families who
most need support to address these impairments, as they are those most likely to utilize
the tools.
Although sharing was common, our results indicate a need for developing improved strategies
to support viewing of shared results. Despite only 16% of parents and 30% of teachers
cross-reviewing submitted surveys, we were encouraged by these rates of viewing given
numerous barriers to accessing messages through the secure messaging software.[39 ] These findings suggest that some families and teachers see more value in reviewing
the information than others. There is known variation in patient engagement and adherence
to recommendations in families of children with ADHD.[20 ] We could potentially improve sharing and viewing by highlighting these features
during patient registration, providing support for clinicians, onboarding families,
and facilitating teacher enrollment.
Within our system, viewing surveys required parents and teachers to access information
through a secure email delivery service, which required its own logon and password.
This process was thought to be more difficult than simply electing to share via a
survey question. Simplifying access to this system using our institution's personal
health record, providing logon support, the creation of a mobile-friendly portal,
and/or allowing teachers restricted access to a patient's EHR portal could improve
viewing of shared information. Evidence-based strategies could also be employed to
encourage the viewing of shared information including sending electronic reminders
or the use of incentives.[50 ]
[51 ] Teachers of children in suburban practices and on private insurance were more likely
to view information, potentially indicating that socioeconomic status of the child
or resources of the school system play a role in teachers viewing information. From
stakeholders, we learned that urban teachers may be responsible for large numbers
of students, with diverse educational needs, and time constraints alone may have limited
viewing. Additional supports may be needed to help teachers in the urban setting and/or
who provide education for children with public insurance in accessing and viewing
the shared surveys. Finally, the emails only provided the data and data trends without
guidance for parents or teachers on what to do next. Parents and teachers may have
been less likely to open emails if they did not have a clear action to take based
on the information provided. Future revisions to the tool may best support families
if suggested actions are highlighted. It is also possible that the parents elected
to share information with teachers without informing the teachers. In these circumstances,
it would be entirely likely that a teacher would not view the email.
Many children with ADHD are managed by external providers, and clinicians caring for
these patients would be unlikely to utilize the ADHD Care Assistant for these families,
which helps to explain why only 14% of identified patients were registered for the
system. In addition, certain clinicians and practices had existing established workflows
for ADHD and were reluctant to change their processes by using the electronic system.
This study had several limitations. The patients in this study represented a limited
subset of all patients seen in our practices with ADHD and for whom screening for
ADHD was triggered by the EHR. It may be difficult to draw generalizations to the
larger population, especially those with ADHD managed outside of primary care or whose
parents are less comfortable using technology to manage health concerns. The focus
of the evaluation was selection of sharing for parents who already opted in to use
the system. Parents who elected to use this system may be different from parents who
opted not to use this system and why parents and teachers elected to complete or not
complete a survey was beyond the scope of this study. An evaluation of the factors
that influence use of this type of system will require future study. From a design
standpoint, the use of observational data limits our ability to establish causation,
and therefore it is difficult to identify precisely those factors that can improve
sharing. We had limited control over when and why clinicians chose to use the system,
resulting in activation of the system for children less than 5 years old. Given the
small number of preschoolers included, it would be difficult to draw conclusions from
this study for this population. We also lacked information regarding clinician, parent,
and teacher demographics so we cannot draw conclusions related to nonpatient demographics.
In addition, our reliance on technology may limit the generalizability of the approaches
we used in health systems with older or more constrained EHRs, which are common in
many practice settings. Nonetheless, Web services like Health Level-7's Fast Healthcare
Interoperability Resources and our own Care Assistant framework are increasingly common
with many EHR vendors and are likely to increasingly support the implementation of
novel informatics approach to care, such as the partnership between families and teachers
described in this article. As with many CDS interventions, sustainability and dissemination
are concerns. We have transitioned ownership of this project from a research team
to our operational team and hope to generate lessons learned about this process and
to support dissemination beyond our organizational walls. Future studies involving
this intervention will address how sharing of information impacts ADHD-related outcomes
(i.e., symptoms, performance impairment, and medication use), the value of sharing
ADHD scores between parents and teachers, whether we can decrease barriers to sharing
of information, and whether this approach can be applied to other disorders affecting
health and school performance.
Conclusion
Electronic tools to support communication can decrease fragmentation of care and support
SDM for children with ADHD. Most parents were willing to share ADHD-related information
with teachers, and those with more severely affected children were most likely to
share information. Despite barriers to accessing information, parents and teachers
successfully viewed the shared surveys at modest rates. System enhancements to facilitate
the viewing of shared surveys are technologically feasible and should be explored
in future studies. Future work is needed to determine if sharing information impacts
patient care outcomes relevant to ADHD.
Clinical Relevance Statement
Clinical Relevance Statement
Treatment for children with ADHD is often fragmented leading to suboptimal care. Electronic
tools may foster communication but have not yet been applied for improving communication
among parents, teachers, and clinicians during the provision of ADHD management. We
demonstrate the potential of using electronic systems provided by the medical home
to promote parental sharing of information about their child's ADHD with teachers.
Multiple Choice Questions
Multiple Choice Questions
Sharing of results and viewing of results were the two new functions developed and
evaluated as part of the update to this intervention. Which of the following correctly
match a clinical or patient characteristic with increased utilization of the new function?
Survey Type AND Sharing of Surveys by Parents.
Survey Type AND Viewing of Surveys by Teachers.
Suburban Office Setting AND Sharing of Surveys by Parents.
Suburban Office Setting AND Viewing of Surveys by Teachers.
Performance Impairments AND Sharing of Survey by Teachers.
Correct Answer: The correct answer is option d. This question addresses the evaluation of results
presented using bivariate and covariate analysis. In bivariate analyses, potential
relationships between patient characteristics may appear significant, but upon a covariate
analysis, after controlling for patient demographics, these relationships may be found
to no longer be significant. Answer a is only an association identified during the
bivariate analyses. When controlling for patient characteristics, it was not found
to be significant. Additionally, the question asks for a clinical or patient characteristic
and this is an ADHD survey characteristic. Answer b is incorrect as viewing of surveys
by teachers was not investigated for different survey types (insufficient power to
perform this analysis). Answer c is incorrect as suburban office setting was found
to be unrelated to sharing in the bivariate analysis. Answer d reports a correct association
of a patient characteristic and a significant finding. Answer e is incorrect because
it is sharing by parents and not sharing by teachers that was investigated.
Multiple potential barriers were identified to parent and teacher viewing of survey
results. Which follow-up study could be performed to evaluate a potential solution
to a technologic barrier of viewing emailed surveys?
Repeating the study with an increase in sample size to overcome barriers to viewing.
Randomizing parents to receive secure versus nonsecure emails with ADHD survey results.
Switching from an email-based system to a Web portal-based system for accessing records.
Compensating families and teachers each time they viewed a new survey result.
Allowing multiple teachers to have access to the system to increase the likelihood
of a tech-savvy participant.
Correct Answer: The correct answer is option b. This question addresses different ways future studies
could be performed to address technical limitations of this study. Answer b directly
addresses the proposed technical limitation that SEND SECURE messages deter people
from opening for results. Answer a would not address the technical barrier, but could
be useful in determining if patient characteristics are related to viewing by increasing
the study power. Answer c would potentially increase views but would not directly
target the technical barrier of needing to have a login and password to obtain the
results. Compensating families (answer d) might improve rates but would not address
the technical barriers. Answer e would also potentially increase views but would not
address the technical barriers.
