Description
Please read the article and answer the questions
Unformatted Attachment Preview
Nembhard et al. BMC Health Services Research
https://doi.org/10.1186/s12913-020-4986-0
(2020) 20:137
RESEARCH ARTICLE
Open Access
A quasi-experiment assessing the sixmonths effects of a nurse care coordination
program on patient care experiences and
clinician teamwork in community health
centers
Ingrid M. Nembhard1* , Eugenia Buta2, Yuna S. H. Lee3, Daren Anderson4, Ianita Zlateva4 and Paul D. Cleary5
Abstract
Background: Recognition that coordination among healthcare providers is associated with better quality of care
and lower costs has increased interest in interventions designed to improve care coordination. One intervention is
to add care coordination to nurses’ role in a formal way. Little is known about effects of this approach, which tends
to be pursued by small organizations and those in lower-resource settings. We assessed effects of this approach on
care experiences of high-risk patients (those most in need of care coordination) and clinician teamwork during the
first 6 months of use.
Methods: We conducted a quasi-experimental study using a clustered, controlled pre-post design. Changes in staff
and patient experiences at six community health center practice locations that introduced the added-role approach
for high-risk patients were compared to changes in six locations without the program in the same health system. In
the pre-period (6 months before intervention training) and post-period (about 6 months after intervention launch,
following 3 months of training), we surveyed clinical staff (N = 171) and program-qualifying patients (3007 preperiod; 2101 post-period, including 113 who were enrolled during the program’s first 6 months). Difference-indifferences models examined study outcomes: patient reports about care experiences and clinician-reported
teamwork. We assessed frequency of patient office visits to validate access and implementation, and contextual
factors (training, resources, and compatibility with other work) that might explain results.
Results: Patient care experiences across all high-risk patients did not improve significantly (p > 0.05). They improved
somewhat for program enrollees, 5% above baseline reports (p = 0.07). Staff-perceived teamwork did not change
significantly (p = 0.12). Office visits increased significantly for enrolled patients (p < 0.001), affirming program
implementation (greater accessing of care). Contextual factors were not reported as problematic, except that 41%
of nurses reported incompatibility between care coordination and other job demands. Over 75% of nurses reported
adequate training and resources.
Conclusions: There were some positive effects of adding care coordination to nurses’ role within 6 months of
implementation, suggesting value in this improvement strategy. Addressing compatibility between coordination
and other job demands is important when implementing this approach to coordination.
Keywords: Nurse care coordination, Patient care experience, Office visit frequency, Teamwork
* Correspondence: [email protected]
1
The Wharton School, University of Pennsylvania, Health Care Management
Department, 3641 Locust Walk, 207 Colonial Penn Center, Philadelphia, PA
19104, USA
Full list of author information is available at the end of the article
© The Author(s). 2020 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Nembhard et al. BMC Health Services Research
(2020) 20:137
Background
Efforts to improve care coordination have increased in
recent years due to the recognition that coordination is
a central part of high quality care yet is often less than
optimal in healthcare [1–4]. Care coordination refers to
“the deliberate organization of patient care activities between two or more participants (including the patient)
involved in a patient’s care to facilitate the appropriate
delivery of healthcare services” ([5], p., 5]). In the United
States (U.S.), 35% of patients with serious illnesses or
chronic conditions report having experienced a coordination failure [6]. Such failures have resulted in medical
complications, preventable hospitalizations, duplicative
testing, and morbidity increases [4] estimated to cost the
U.S. healthcare system $25 to 45 billion in 2011 [7, 8].
Other countries (e.g. Canada, France, Norway, Sweden,
and Switzerland) struggle with care coordination as well,
where 30% or more of patients report experiencing coordination failures [6].
To improve care coordination in several countries [4,
9–11], many health systems and organizations have implemented or are implementing nurse care coordination
initiatives in which nurses provide additional care and
support to patients with coordination needs such as
those with a chronic illness, transitioning from hospital
to home, or with multiple medical and behavioral health
issues [12–14]. In this approach, nurses work closely
with designated patients and providers to coordinate
multi-specialty care and help patients manage their illnesses. Core responsibilities in this role include monitoring patient health and facilitating development,
communication, and delivery of care plans with other
care team members [15, 16]. Nurses in many organizations perform these activities, which are within their
scope of practice [17, 18]. New programs have structured these activities, clarifying authority, tasks, options,
and responsibility, to enhance nurses’ visibility, effectiveness, and efficiency as coordinators. These programs address calls from professional and scientific groups for
nurse coordinator roles to be more explicit, developed,
and designed deliberately into training and delivery organizations [15, 18–20]. These programs should lead to
better experiences for patients and clinicians because assignment of responsibility to one person and coordination improve the logic, continuity, and efficiency of
care [5, 12, 18, 21–23].
Currently, two approaches to nurse care coordination
are common. In the first, adopted mainly in large health
systems and medical groups, a nurse serves exclusively
as care coordinator for a panel of patients. This “exclusive-role approach,” has been used, for example, by participants in the U.S. Medicare Care Coordination
Demonstration program [24]. In the second “added-role
approach”, a nurse performs care coordination in
Page 2 of 14
addition to existing responsibilities. Although reviews of
research on the first approach indicate mixed results
[25–27], increasing evidence from controlled trials indicates that patients with these coordinators experience
better technical quality of care, lower hospital readmissions, lower costs, and better care experiences (e.g.,
provider-patient communication) than patients who do
not have a coordinator [12, 17, 28–37].
Little is known about the experiences of patients
served by nurses in the added-role approach, which may
be pursued more by smaller organizations or those in
lower-resource settings, which are many of the settings
across the world. There is also limited information, particularly in primary care settings about the effectiveness
of this approach, even though these settings are increasingly expected to coordinate care with patients and other
providers [38]. In primary care settings such as federally
qualified health centers (FQHCs) in the U.S., a type of
community health center that serves disproportionately
more complex patients with multiple co-morbidities and
socioeconomic disadvantages than do private practices
and health systems [39], the imperative for coordination
is especially great, but there is little evidence about the
effects of adding care coordination to the nursing role.
Nurses dedicating any increased attention to these tasks
may be positive for patients in need and clinicians. On
the other hand, the potential positive effects of the
added-role approach may not be realized because of the
inability to focus exclusively on coordination tasks.
In this manuscript, we examine the early (six-month)
effects of a nurse care coordination program in FQHC
practices that use the added-role approach for high-risk
patients using two measures: care experiences of these
patients and clinician-reported teamwork. High-risk patients have complex and/or multiple medical and psychosocial problems, which may require them to see as
many as 16 physicians per year, making them most in
need of care coordination, most at risk for coordination
failures, and most likely to benefit from care coordination, [1, 40] although recent studies suggest that benefits may extend beyond this group [41]. We also
examine an indicator of implementation effectiveness,
the frequency of patient office visits, and contextual factors because they can influence implementation, and
thus program outcomes [42].
We focus on effects in the program’s first 6 months
because early experiences with a program are often consequential for long-term success [43–46]. Also, departure from past patterns is often salient to participants
early, before they become accustomed to new patterns
and adjust expectations, [47] making early assessments a
window into program functioning. Currently, there is
limited investigation of the early effects of nurse care coordination programs, leaving organizations with little
Nembhard et al. BMC Health Services Research
(2020) 20:137
Page 3 of 14
knowledge about what to expect. Research on other
patient-nurse and coordination interventions in other
settings (e.g., skilled nursing visits in home health care
[48–50]) suggests that positive effects can materialize in
6 months.
Methods
Study setting and design
This study was conducted in a statewide, multi-site
FQHC with 12 sites that provide comprehensive primary
medical, dental, and behavioral healthcare services to
over 140,000 patients a year. The center serves patients
with all types of primary care needs and emphasizes
serving the uninsured, underinsured, and special populations such as patients with HIV/AIDS, diabetes, and
chronic mental health issues. The FQHC has been recognized as a Primary Care Medical Home by the Joint
Commission [51] and a level 3 Patient-Centered Medical
Home by the National Commission on Quality Assurance [52]. Thus, each site has demonstrated commitment to patient-centered care, comprehensive care,
coordinated care, access to care, and a systems-based approach to quality and safety.
We conducted a cluster quasi-experiment in which
pre-post intervention changes in clinician and patient
experiences in six sites (clusters) that introduced a nurse
care coordination program for high-risk patients using
the added-role approach (“intervention group”) were
compared to changes in experiences in six sites without
the program at the time of our study (“comparison
group”). Sites in the comparison group implemented the
program after our data collection. The FQHC used a sequential roll-out plan (all locations (3) in one county
every 3 months) as it does for certain large-scale initiatives for operational reasons (e.g., maintaining crosscoverage between providers in county and having sufficient resources for implementation). When deciding
about comparison sites, the FQHC's leadership selected
pairs of sites that were relatively similar based on number of patients, patient population profile, and the
organization of sites. Sites were allocated to the intervention group if the intervention could begin sooner
there than at pair site, given staff work and training
schedules, etc. The selected intervention and comparison
sites were similar at baseline and follow-up on all but
two characteristics for which we could obtain data
Table 1 Comparison of Intervention and Comparison Groups’ Characteristics at Baseline and Follow-up
Baseline (Median[range])
Characteristics
Intervention
Centers
(N = 6)
Follow-up (Median[range])
Comparison Total
PIntervention
value* Centers
(N = 6)
Comparison
Centers
(N = 6)
Total
Centers
(N = 6)
Centers
(N = 12)
501 [333;
511]
443 [301;
528]
450 [301;
528]
0.20
918 [631;
1114]
841 [611;
1000]
841 [611;
1114]
0.42
Medicaid patients (%)
71 [59; 74]
64 [55; 81]
68 [55;
81]
0.75
72 [60; 74]
65 [56; 81]
68 [56;
81]
0.94
Medicare patients (%)
8 [4; 10]
13 [9; 17]
10 [4; 17]
0.01
8 [5; 10]
12 [8; 15]
9 [5; 15]
0.02
Private insurance patients (%)
10 [9; 12]
12 [7; 20]
11 [7; 20]
0.42
11 [8; 12]
12 [7; 18]
11 [7; 18]
0.26
Uninsured patients (%)
9 [3; 23]
6 [2; 18]
6 [2; 23]
0.26
9 [3; 22]
6 [2; 16]
7 [2; 22]
0.29
White patients (%)
28 [13; 66]
41 [27; 65]
32 [13;
66]
0.34
27 [13; 64]
38 [25; 63]
31 [13;
64]
0.38
Black patients (%)
8 [7; 20]
15 [2; 22]
12 [2; 22]
0.52
8 [5; 19]
12 [2; 18]
10 [2; 19]
0.87
Hispanic patients (%)
49 [12; 64]
34 [18; 58]
43 [12;
64]
0.26
50 [13; 64]
34 [18; 59]
43 [13;
64]
0.30
Other race patients (%)
4 [2; 5]
5 [4; 12]
4 [2; 12]
0.05
4 [2; 7]
6 [3; 22]
4 [2; 22]
0.20
Race unknown patients (%)
Number of patient visits per full-time employee in
6-month period
Centers
(N = 12)
Pvalue*
Patient insurance status
Patient race
4 [2; 10]
2 [1; 11]
4 [1; 11]
0.19
5 [3; 10]
4 [2; 12]
4 [2; 12]
0.51
Patients eligible for care coordination^
330 [114;
1396]
745 [162;
1538]
410 [114;
1538]
0.63
.
.
.
.
Productivity indicator
1.21 [1.04;
1.37]
1.00 [0.96;
1.19]
1.12 [0.96; 0.06
1.37]
1.23 [1.02;
1.28]
1.05 [0.81;
1.16]
1.06 [0.81; 0.07
1.28]
Supervisor support for staff, indicative of work
climate (staff reported, 1–4 scale)~
3.60 [2.96;
3.80]
3.67 [3.49;
3.84]
3.66 [2.96; 0.57
3.84]
3.37 [3.23;
3.88]
3.57 [3.23;
3.93]
3.56 [3.23; 0.57
3.93]
^ Baseline values apply to follow-up period as well because the starting sample of eligible patients remained the focus throughout the study. *p-value from
Wilcoxon rank-sum tests comparing intervention to comparison centers. ~Supervisor support measured by 5 items from the FQHC’s staff survey
Nembhard et al. BMC Health Services Research
(2020) 20:137
(Table 1). Wilcoxon rank-sum tests indicated that the
groups differed significantly with respect to percent of
patients with Medicare as their health insurer (p = 0.02
and p = 0.01 at baseline and follow-up, respectively) and
percent of patients with “other race” (p = 0.05 at baseline). We adjust for these differences in our analyses.
Our primary study outcomes were two indicators of
program effectiveness: patient reports about their care
experiences and clinician reports of teamwork in their
centers. If care coordination programs function as
intended, patient experiences, as reflected in responses
to questions about care coordination, timeliness of care,
and support for self-management should improve, as
should clinician teamwork.
Because degree and fidelity of program implementation
are critical determinants of program effectiveness, we collected the implementation information that we could,
given resource limitations and concerns about staff burden. We obtained information about numbers of telephone calls to patients, but those data turned out to be
inconsistent and of poor quality and so are not presented.
The other measure of program implementation that we
have is the number of patient office visits, which is a proxy
measure of accessibility of care, engagement with patients,
monitoring, and follow-up to achieve care plan goals (e.g.,
condition controlled, no preventable hospitalization). If
the care coordination program was implemented as
intended, there should be an increase in patient office
visits in the early months of the program to address outstanding patients’ care needs and self-management training. Research on programs that incorporate the exclusiverole approach has found that primary care office visits increase with coordination programs in the first 2 years,
while emergency department visits decline for high utilizers [53]. Over a longer period, not covered by this study,
office visits should decline due to better patient health and
self-management. Because implementation and effectiveness are often influenced by resources, training, and compatibility with current work, [42, 54, 55] we also assessed
these contextual factors via nurse surveys, because these
factors may help explain our results. Other non-program
specific contextual factors (e.g., employee workload, patient population profile, and supervisor support for
workers, which shapes work climate) were examined as
well (Table 1).
Intervention
In intervention sites, every nurse’s role was expanded to
include care coordination for adult patients who were
expected to benefit most from this effort. These were
defined by the organization as patients who were 18
years of age or older, had two or more visits with a primary care provider (PCP) in the past 12 months, and
had been identified as “high risk.” Patients were
Page 4 of 14
classified as high risk if they had: 1) two or more emergency room visits in the past 12 months; 2) one or more
hospitalizations in the past 12 months; 3) a Type 2 diabetes diagnosis on their problem list and a hemoglobin
A1C test in the past 12 months greater than 9%; 4) a
diagnosis of persistent asthma diagnosis on problem list
and two or more asthma control test scores < 19 in the
past 12 months; or 5) four or more of specified chronic
illnesses on their active problem list, including Type 2
diabetes, chronic obstructive pulmonary disease, hypertension, asthma, coronary artery disease, or behavioral
health diagnosis. A subset of the eligible patients (those
with greatest immediate need as perceived by staff) was
enrolled in the program at the outset due to time and
resource constraints. Other patients were also enrolled
when a PCP or nurse identified the patient as needing
care coordination (e.g., newly discharged from a
hospital).
As part of the new program, nurses were expected to
work with enrolled patients to help them navigate their
healthcare and lead a weekly panel management meeting
with enrollees’ PCP and behavioral health provider. The
sessions were to be used to review patient progress, identify
additional patients who needed coordination, and plan coordinated care. To implement the program, the
organization introduced the nurse care coordinator role to
all staff via meetings and other communications (e.g., newsletters). It also provided three resources to nurses to support their effectiveness as coordinators: training, a
“playbook”, and an electronic dashboard. All nurses in the
intervention sites received 23 h of training over a period of
2 to 3 months from experts within the organization and
outside consultants. The training covered care plan development, panel management, documentation, transition care
support, motivational interviewing, self-management goal
setting, chronic disease management, and behavioral health
disorders — evidence-based components of nurse care coordination [17]. The playbook provided instructions for
each task within the new nurse role, information on additional resources, and measures to evaluate performance.
The electronic dashboard leveraged information in the organization’s electronic health record system, which aided
nurse tracking of patients and activities. No other group
was assigned care coordination responsibilities. The
organization reinforced its commitment to the role change
by monitoring nurse performance and providing feedback
reports to nurses. It was expected that the program would
lead to more coordinated and timely care, greater patient
support for self-management, and care for mental health.
Study outcomes
Patient care experiences
We collected patient surveys that asked about care experiences during two periods at each center. The first
Nembhard et al. BMC Health Services Research
(2020) 20:137
(baseline) period covered the 6 months prior to start of
nurse training in the intervention centers, and was before nurses were told about the intervention and patients
who would be in the program were known. In intervention and comparison centers, we invited a random sample of the high-risk (i.e., program-eligible) patients
described earlier that had visited the center in the preceding 6 months (N = 5525) to complete the Consumer
Assessment of Healthcare Providers and Systems Clinician & Group (CG-CAHPS) survey [56, 57] and PatientCentered Medical Home (PCMH) Supplemental Item
Set [57, 58]. These surveys assess multiple aspects of patient care experiences, [57, 59] and have been used in
other studies of care coordination [60, 61]. The sites
already administered these surveys for performance
monitoring. With funding provided by the CAHPS Program, we supplemented sites’ surveying to capture the
patients in this study.
We assessed the program’s impact using patients’ responses to questions about four aspects of care targeted
by the program and therefore expected to be affected by
experiencing the program: timeliness of care, coordination of care, support for patient self-management, and
care for mental health. Timeliness of care was hypothesized to increase because patients in the program would
have priority access to care; their nurse care coordinators would try to be highly responsive. Coordination of
care for program enrollees was to improve because
nurses would focus on ensuring that enrollees’ needs
were met as seamlessly as possible. Support for selfmanagement and care for mental health were additional
program foci and areas of training for nurses; therefore,
we expected that nurse efforts in these areas would be
reflected in patient reports of their experiences. We focused on these four standard measures of patient care
experience, rather than care coordination alone, recognizing that nurse care coordination efforts should manifest in multiple ways [19, 20]. Table 2, Part A lists the
items used from the CG-CAHPS survey to measure
these aspects of care, response options, and the reliability of the scales in our sample. Patients indicated
whether they experienced the action described in each
question using a four-point scale (1 = never to 4 = always) or No (=1)/Yes (=4) response. We averaged responses for the items in each composite to arrive at a
score for each aspect of their experience. The four composite scores are highly correlated (p-values < 0.001), so
to simplify analyses and presentation, we averaged them
to arrive at an overall patient care experience score for
each person.
After the program had been in effect for 6 months following nurse training, we again invited a random sample
of program-eligible patients that had visited the center
in the preceding 6 months (N = 4661) to complete the
Page 5 of 14
Table 2 Study Measures
A. Patient-reported care experience (4 components)
Timeliness of care (Cronbach’s alpha (α) = 0.89)
▪ When you phoned this provider’s office to get an appointment for care you needed right
away, how often did you get an appointment as soon as you needed?
▪ When you made an appointment for a check-up or routine care with this provider, how often
did you get an appointment as soon as you needed?
▪ How often were you able to get the care you needed from this provider’s office during
evenings, weekends, or holidays?
▪ When you phoned this provider’s office during regular office hours, how often did you get
an answer to your medical question that same day?
▪ When you phoned this provider’s office after regular office hours, how often did you get an
answer to your medical question as soon as you needed?
▪ How often did you see this provider within 15 min of your appointment time?
Care coordination (Cronbach’s alpha (α) = 0.73)
▪ How often did this provider seem to know the important information about your medical
history?
▪ When this provider ordered a blood test, x-ray, or other test for you, how often did someone
from this provider’s office follow up to give you those results?
▪ Did you get the help you needed from this provider’s office to manage these different
providers and services?
▪ How often did the provider named seem informed and up-to-date about the care you got
from specialists?
▪ How often did you and anyone in this provider’s office talk about all the prescription
medicines you were taking?
Support for patient self-management (Cronbach’s alpha (α) = 0.65)
▪ In the last 6 months, did anyone in this provider’s office talk with you about specific goals for
your health?
▪ In the last 6 months, did anyone in this provider’s office ask you if there are things that make
it hard for you to take care of your health?
Care for mental health (Cronbach’s alpha (α) = 0.78)
▪ In the last 6 months, did anyone in this provider’s office ask you if there was a period of time
when you felt sad, empty or depressed?
▪ In the last 6 months, did you and anyone in this provider’s office talk about things in your life
that worry you or cause you stress?
▪ In the last 6 months, did you and anyone in this provider’s office talk about a personal
problem, family problem, alcohol use, drug use, or a mental or emotional illness?
B. Clinician-reported teamwork (2 components)
Interprofessional Collaboration (Cronbach’s alpha (α) = 0.77)
▪ Nurses and physicians plan together to make decisions about care for complex patients.
▪ Open communication between care providers takes place as decisions are made for complex
patients.
▪ Decision-making about patient care for complex patients is well-coordinated.
▪ The input of ancillary staff is regularly sought when developing care plans.
Relational coordination (Cronbach’s alpha (α) = 0.75)
▪ The people on this team share my goals for the care of patients.
▪ The people on this team know about the work I do with patients.
▪ The people on this team respect me and the work I do with patients.
▪ The people on this team communicate with me in a timely way about the status of patients.
Note: Cronbach’s alpha (α) above 0.70 indicates satisfactory reliability of a measure and
between 0.50 and 0.70 indicates moderate reliability. The reported alphas are based on
baseline data. For the first two aspects of care, patients indicated whether they experienced
the action described in each question using a four-point scale (1 = never to 4 = always),
except for the third item in the care coordination scale for which they replied No (=1) or
Yes (=4). For the third and fourth aspects of care, they replied No (=1) or Yes (=4). For staffreported teamwork, staff responded using a four-point response scale (1 = strongly disagree
to 4 = strongly agree)
CG-CAHPS survey with additional items. All 145 program enrollees received an invitation by design. Followup at 6 months allowed us to avoid contamination of the
comparison group: per the organization’s fixed roll-out
plan, the program (training) was scheduled to begin in
Nembhard et al. BMC Health Services Research
(2020) 20:137
the first set of comparison centers at this time. This
planned endpoint also aligned with our study objective
to assess early effects of the added-role approach.
In both the baseline and follow-up periods, we mailed
a copy of the survey in English and Spanish to each patient in the sample. Approximately 2 weeks after the first
mailing, members of the sample were sent a thank you/
reminder postcard. Approximately 2 weeks after that,
another survey package was mailed to those who had
not responded. If no response was received after two to
three more weeks, we called the patients. A minimum of
six calls per person were made on different days and at
different times of the week.
In the baseline period, 3209 patients of the 5525 contacted (58%) answered the survey; of those, 3007 (94%)
confirmed having visited the center in the prior 6
months (intervention group = 934; comparison group =
2073). In the follow-up period, 2306 patients of the 4661
contacted (49%) answered the survey; of those, 2101
(91%) confirmed having visited the center in the prior 6
months (intervention group sample size = 774; comparison group sample size = 1327). In total, 943 patients answered the survey in both periods (643 in control group;
300 in intervention group), and 113 program enrollees
responded (78% of the 145 enrolled).
Teamwork
During the month in which we began both the baseline
and follow-up patient surveys, we administered an
“organizational assessment survey” via the internet or
paper to all primary care team members (PCP, nurses,
medical assistants, and behavioral health providers). We
recruited team members to participate via informational
presentations during lunchtime staff meetings and email,
and confirmed willingness to participate via signed consent forms. The survey consisted of validated survey
scales for assessing core aspects of teamwork i.e., relational coordination and interprofessional collaboration
[62–64]. Interprofessional collaboration refers to the degree of cooperation among individuals with different disciplinary backgrounds [65], while relational coordination
refers to the presence of high-quality communication
and relationships characterized by shared goals, shared
knowledge, and mutual respect needed for task integration [66]. Each scale included four items (Table 2, Part
B). Team members indicated their level of agreement
with each item using a four-point response scale (1 =
strongly disagree to 4 = strongly agree). Because scores
for the two scales were highly correlated (p < 0.001), we
averaged them to arrive at a summary teamwork score
reported by each respondent.
At baseline, 96 of 190 (51%) team members completed
the survey (intervention group = 43; control group = 53).
At follow-up, 135 of 188 (72%) members completed the
Page 6 of 14
survey (intervention group = 57; control group = 78).
Sixty members participated at both baseline and followup. We used their responses in our analyses to assess
program effect based on the experiences of a stable
population and minimize the possible confounding effect
of respondents new to the centers. This longitudinal
sample was 39% PCPs, 22% nurses, 24% medical assistants, and 15% behavioral health providers. The majority
were female (71%), full-time staff (89%), and with the
organization more than 2 years (82%). Except for the
percentage with more than 2 years with the organization
(63%), this sample was demographically like the full
sample consisting of 33% PCPs, 23% nurses, 28% medical assistants, 18% behavioral health providers, 83% female, and 88% full-time staff.
Implementation measures
Office visit frequency
We obtained information about patients’ number of office visits via response to a question in the CG-CAHPS
survey: “In the last 6 months, how many times did you
visit this provider to get care for yourself?” Seven response options were offered: none, 1 time (coded as 1),
2 (coded as 2), 3 (coded as 3), 4 (coded as 4), 5 to 9
(coded as 7, the midpoint), and 10 or more times (coded
as 10). Patients who did not recall any visits were excluded from study (N = 202 (7%) at baseline and 205
(9%) at follow-up).
Contextual factors: training, resources, and compatibility
with other job demands
The organizational survey administered to primary care
team members during the follow-up period included
additional questions for nurses about program training,
resources, and their new role’s compatibility with other
job demands, which we used to assess whether these factors posed a challenge to implementation and effectiveness. Four items were adapted from Venkatesh et al.’s
[67] facilitating attributes scale: “I have the resources necessary to coordinate care for complex patients,” “I have
the knowledge necessary to coordinate care for complex
patients,” “Coordinating care for complex patients is not
compatible with other tasks that I’m required to perform,” and “It is easy for me to coordinate care for complex patients.” A fifth resource-related item drew from
the FHQC’s employee survey: “I have adequate authority
to carry out my work.” We asked nurses at intervention
centers to report their level of agreement with each
statement (1 = strongly disagree to 4 = strongly agree).
Other non-program specific contextual factors that can
affect implementation (e.g., supervisor support for
workers and workload) and could be assessed for intervention and comparison groups at baseline and followup were evaluated for potential inclusion as covariates.
Nembhard et al. BMC Health Services Research
(2020) 20:137
Covariates
In models assessing patient care experience