The Impact of Primary Care Practice Structural Capabilities on Nurse Practitioner Burnout, Job Satisfaction, and Intent to Leave

Abstract

Background:

Lack of structure for care delivery (i.e., structural capabilities) has been linked to lower quality of care and negative patient outcomes. However, little research examines the relationship between practice structural capabilities and nurse practitioner (NP) job outcomes.

Objectives:

We investigated the association between structural capabilities and primary care NP job outcomes (i.e., burnout, job dissatisfaction, and intent to leave).

Research Design:

Secondary analysis of 2018–2019 cross-sectional data.

Subjects:

1,110 NPs across 1,002 primary care practices in six states.

Measures:

We estimated linear probability models to assess the association between structural capabilities and NP job outcomes, controlling for NP work environment, demographics, and practice features.

Results:

The average structural capabilities score (measured on a scale of 0–1) across practices was 0.6 (higher score indicates more structural capabilities). After controlling for potential confounders, we found that a 10-percentage point increase in the structural capabilities score was associated with a 3-percentage point decrease in burnout (p<.001), a 2-percentage point decrease in job dissatisfaction (p<.001), and a 3-percentage point decrease in intent to leave (p<.001).

Conclusions:

Primary care NPs report lower burnout, job dissatisfaction, and intent to leave when working in practices with greater structural capabilities for care delivery. These findings suggest that efforts to improve structural capabilities not only facilitate effective care delivery and benefit patients, but they also support NPs and strengthen their workforce participation. Practice leaders should further invest in structural capabilities to improve primary care provider job outcomes.

INTRODUCTION

Access to high-quality primary care is critical to improving population health,¹ yet persistent workforce shortages in the United States (US) pose substantive barriers to meeting care demands.² In addition to a scarce workforce, the need for primary care services has increased considerably with Medicaid expansion under the Affordable Care Act,³ the growing needs of the aging US population,⁴ and the increasing prevalence of chronic diseases.⁵ Nurse practitioners (NPs) serve as an evidence-based solution to the demand for primary care. NPs are the fastest growing segment of the primary care workforce; their number doubled from 91,000 to 190,000 between 2010 and 2017,⁶ with additional and rapid expansion projected in the future.⁷ Furthermore, NPs maintain a holistic, biopsychosocial education model, which focuses on caring for the “whole person”, including addressing medical complaints, behavioral health, the social determinants of health, and managing complex family caregiver dynamics.⁸ In addition to this whole health focus, NPs have key clinical privileges (e.g., diagnosing and treating health conditions; ordering, performing, and interpreting diagnostic tests; prescribing medications; patient/family counseling and education), making them well-suited to meet the growing demand for primary care.⁸

However, NPs are vulnerable to negative working conditions in primary care, including high patient volumes, lack of resources, insufficient salaries, and poor work environments.^9–11 These factors contribute directly to negative job outcomes like burnout, a work-related condition characterized by emotional exhaustion,¹² and are also significant predictors of workforce turnover.¹³ Primary care provider burnout may be as high as 60%,¹⁴ underscoring the critical need to better understand how embedded structures of primary care can be leveraged to improve NP retention.

Structural capabilities in primary care practices– such as disease registries, care reminders, shared systems for communication with patients, care management, community health referral systems, electronic health records (EHRs), performance feedback, and after-hours care– may improve NP job outcomes through greater resource support. Prior research shows that these structural capabilities are important avenues for improving care delivery processes and lead to higher quality of care.^15–20 However, limited research has assessed the impact of structural capabilities on clinician job outcomes, let alone NP job outcomes.

Investments in practice infrastructure may help NPs improve efficiency and provide greater support for care delivery. For example, having access to a registry with a list of patients with overdue chronic disease services (e.g., hemoglobin A1c testing for diabetes; cholesterol testing for coronary artery disease), may support NPs in identifying patients with time-sensitive care needs. Additionally, a system of reminders, such as a flowsheet or checklist, may support NPs in following recommended guidelines for patients with chronic conditions (e.g., asthma, hypertension, diabetes). With the continued growth of patient-centered medical homes (PCMHs), the pressure to adopt structural capabilities has been heightened, making it paramount to understand how they affect clinician job outcomes.²¹ Greater details of structural capabilities, including definitions and their impact on clinician job outcomes can be found in Table 1.

Table 1.

The Impact of Structural Capabilities on Clinician Job Outcomes

Electronic Health Records (EHRs)

• Provide a platform for care delivery, documentation, and information exchange through computerized systems for various clinical tasks and have been considered by The National Academy of Medicine as a key component of the professional practice environment.25   • While EHRs are critical to care delivery, their use has been associated with higher burnout,26 although some recent findings indicate these systems may help improve workflow and provider experience.27

Shared Communication Systems

• Shared system of communication are structures that organizations use to facilitate communication between healthcare providers or between providers and patients and are often integrated within the EHR.28   • Patients reported greater satisfaction with computerized notification systems,29 suggesting that these may reduce administrative burden and potentially improve provider working conditions without compromising patient satisfaction.

Care Reminders

• Care reminders are also often integrated within the EHR and aid providers in remembering patient care responsibilities (e.g., prompts, clinical decision support)30 and are associated with delivery of care31 and process adherence.30   • Little is known about how these reminders ultimately influence clinician job outcomes.

Community Health Referral Systems

• Community health referral systems connect patients, especially those with chronic conditions, to relevant community health resources.32   • Examples include senior support services, mental health and addiction services, and chronic disease prevention and management.   • Despite the important potential benefits of equipping providers to address social risk factors, it is unclear whether these systems impact provider job outcomes.

Disease Registries

• Disease registries are deployed in primary care settings to track and manage particular patient populations, health outcomes, care lapses, and set care management goals.33   • Disease registries are often used to manage patients with one or more chronic conditions and are helpful in centralizing efforts to gather, track, sort, and report data from multiple sources at the patient- and population-level in automated ways, helping providers identify gaps in care and make appropriate follow-ups.33

Care Management

• Care management involves the use of a dedicated clinician team (or a dedicated clinician) to help patients improve their functional health status, minimize the need for acute care services, and improve care engagement.34   • Primary care providers using care management teams report higher job satisfaction, retention,35 and lower stress levels.36,37

Performance Feedback

• Performance feedback is necessary for providers to improve care delivery and assess their performance across professional standards.38   • A lack of feedback on performance is associated with over a 2-fold increase in burnout among primary care workers.39   • Performance feedback may also help advance medical trainees through their residency and increase confidence in their performance.40   • These findings indicate that clinicians benefit from feedback and performance evaluations, enabling them to better achieve their professional goals.

After-Hours Care

• After-hours care is a measure of a facility’s extended practice hours, where regular ambulatory care services are available during evenings and weekends.41   • The infrastructure for after-hours care may provide greater staff coverage, though one study found it can be an additional stressor.41   • NPs conducting after-hours services may experience greater opportunities for career progression and autonomous practice,42 as well as increased job satisfaction.41

However, most research on NP job outcomes has focused on the work environment. Specifically, this research shows that primary care practices with better work environments—where NPs have support for independent scope of practice, experience positive collaborative relationships with physicians and administrators, and work in an environment where their roles and competencies are clearly understood and utilized^10,11—have lower levels of negative job outcomes like burnout and intent to leave, as well as greater job satisfaction.^22–24 These findings are consistent with the broader literature concluding that organizational characteristics like autonomy, resource support, professional relationships, and leadership are greater determinants of burnout than personal attributes or a lack of individual resilience.²⁵ Yet, this body of research focuses on the work environment in isolation, rather than also considering important practice structural capabilities, and little is known about how structural mechanisms intended to support patient care delivery impact NP job outcomes. Thus, this study aims to assess the relationship between primary care practice structural capabilities and NP burnout, job dissatisfaction, and intent to leave.

MATERIALS & METHODS

Design and Data

We conducted a secondary analysis of cross-sectional survey data collected in 2018 and 2019 from NPs working in primary care practices. As part of a parent study, data was collected from NPs working in six geographically diverse states to mirror national variation in NP scope of practice at the time of the survey.⁴³ California (CA) and Texas (TX) represent restricted scope of practice states, where NPs must work under the supervision of another healthcare provider. New Jersey (NJ) and Pennsylvania (PA) represent reduced scope of practice states, where NPs are required to have a collaborative agreement with another provider. Arizona (AZ) and Washington (WA) represent full scope of practice states, where NPs are able to practice independently without supervision or collaborative agreements.

NPs were identified from the IQVIA OneKey database, which includes facility and physician specialty data.⁴³ Primary care practices were defined if more than half of physicians working in the practice had specialties of pediatrics, preventative medicine, internal medicine, geriatrics, general practice, or family practice⁴³. NPs from these practices were recruited for the study. Further details on the parent study’s survey approach have been published,⁴³ but in brief, a modified Dillman method was used to send surveys to 10,237 NPs. Three surveys and two postcard reminders were sent, and three phone call reminders were made to non-respondents. There was a 21.9% response rate, and the final sample included 1,244 NPs across 1,109 primary care practices.

Sample

To ensure consistency in our practice sample across statistical models, we required that NPs respond to survey questions regarding structural capabilities, demographics, and practice features. Based on this inclusion criteria, the final study sample included 1,110 NPs working across 1,002 primary care practices.

Measures

STRUCTURAL CAPABILITIES were derived from the structural capabilities index (SCI), a validated tool to measure the presence of structural capabilities in primary care and has been used to link structural capabilities to quality of care measures.^15–20 For example, a 2014 study found that disease registries used to help providers identify patients overdue for chronic disease services were associated with higher levels of diabetic screenings.¹⁵ Another study found that disease registries and care management led to better performance on diabetes and cancer care measures and lower levels of hospitalization and emergency department utilization.¹⁶ Along these lines, other studies have found promising results regarding structural capabilities’ impact on care quality.^17–20 The SCI includes 27 questions grouped into eight themes: disease registries, care reminders, shared systems for communication with patients, care management, community health referral systems, electronic health records, performance feedback, and after-hours care.

To create a practice-level SCI measure, we computed an individual NP’s SCI score by averaging their response to all SCI items and then averaging across all NPs in the practice; this created a continuous practice-level SCI score ranging from 0–1. Scores closer to 1 indicate that a greater number of structural capabilities are available in the practice. Prior work has demonstrated the SCI’s content validity for being capable of measuring practice structural capabilities.^15–20 As it is difficult to establish content validity in a secondary data analysis, we assessed the internal consistency reliability by computing Cronbach’s alpha of all SCI items. Our Cronbach’s alpha was 0.88, indicating there is good internal consistency among the individual SCI items, and we can aggregate to build an overall practice-level SCI measure.

JOB OUTCOMES were derived from three single-item questions in the NP survey. To determine burnout status, NPs were asked how they would rate their overall level of burnout on an ordinal scale, ranging from 1–5 (i.e., Level 1 indicating no symptoms of burnout to Level 5 indicating completely burned out). This burnout measure has been validated for use against the Maslach Burnout Inventory⁴⁴ and has been used in prior studies of primary care NPs.^24,27 The single item burnout question was used instead of the full Maslach Burnout Inventory to reduce survey burden and improve the response rate. We dichotomized the measure so that scores ranging from 3–5 indicate burnout, and scores 1–2 indicate no burnout—an approach that is consistent with prior research.^24,27 Job dissatisfaction was determined using a single-item measure with a 1–4 ordinal scale, which has been shown to be comparable to measures using several items⁴⁵ and has been used in previous studies of primary care NPs.^22,23 We dichotomized this measure so that NPs were either satisfied or dissatisfied with work. Intent to leave was measured using a single item asking NPs if they intended to leave their jobs within the next year (i.e., yes/no were possible answers), which has also been used in prior studies.^22,23

COVARIATES. In our adjusted analyses, we controlled for the following NP characteristics: age, sex, race, ethnicity, marital status, education, years of experience, average hours worked per week, and whether NPs managed their own patient panel. We selected these specific characteristics because prior research has identified them as important mediating factors that affect job outcomes.⁴⁶ We also controlled for the following practice features: practice type, state scope of practice, and the number of NPs working in the practice. We further adjusted for the NP work environment, because prior research has found that work environment features are important determinants of job outcomes.^22,24 Our work environment measure was derived from the Nurse Practitioner Primary Care Organizational Climate Questionnaire (NP-PCOCQ), which includes 29 items grouped into four subscales: NP-Physician Relations, Professional Visibility, NP-Administration Relations, and Independent Practice and Support. NPs reported on the degree to which work environment features were present in their practice on a 1–4 Likert scale. The NP-PCOCQ has been validated in prior studies.^10,11 We created a practice-level summary measure by aggregating individual NP responses by subscale to the practice-level if NPs responded to at least 70% of the items. We then averaged the 4 subscales together to create one continuous work environment measure. The work environment measure has a Cronbach’s alpha ranging from 0.87 to 0.96, has strong psychometric properties, and has been widely used in prior research of NPs.^10,11

Data Analysis

We summarize the characteristics of our NP and practice sample using means and standard deviations to describe continuous variables, and frequencies and percentages to describe categorical variables. We describe the variation in NP job outcomes across practices with different levels of structural capabilities using chi-squared tests. The differing levels of structural capabilities were determined by dividing practices into four groups based on the aggregate structural capabilities score. Low-scoring practices include those in the first quartile (SCI score range 0–0.39); medium-scoring practices include those in the second and third quartiles (SCI score range 0.41–0.73); and high-scoring practices include those in the fourth quartile (SCI score range 0.75–1.0).

We conducted a series of stepwise regressions to estimate the association between the SCI score and NP job outcomes. Model 1 tests the bivariate association, Model 2 adds NP and practice characteristics, and Model 3 adds the work environment. Model 4 adds an interaction term between the SCI and the work environment. To adjust for the clustering of NPs within practices, Huber-White Sandwich estimators were used.^47,48 Variance Inflation Factors (VIF) were used to assess the risk of multicollinearity, which we found to be very low, as the average VIF was 1.33 (<2.5 considered low risk of multicollinearity).⁴⁹ Significance was assessed using a 0.05 threshold in STATA 15.1 (StataCorp LLC, College Station, TX, USA).

Our primary specification is a linear probability model, which measures the impact of the SCI score on the probability of the outcome = 1 (i.e., probability of reporting burnout, job dissatisfaction, or intent to leave). The model coefficient is constructed to measure the impact of a 1-unit increase in the SCI score (i.e., going from 0 to 1, or 0% to 100% of structural capabilities = yes). To facilitate interpretation, we rescale to measure the impact of a 0.1-unit increase in the SCI score, which corresponds to a 10-percentage point increase (e.g., going from 0.5 or 50% = yes to 0.6 or 60% = yes). To do this, we divide the coefficient by 10 (to change from a 1-unit increase to a 0.1-unit increase) and then multiply by 100 to convert to a percentage point change in the probability. We take this approach because it is unlikely that practices would go from having 0% to 100% of all structural capabilities, but would rather experience incremental increases.

RESULTS

We summarize available NP characteristics in Table 2. On average, NPs were 49 years old and 87% identified as female. In terms of racial composition, 80% of our sample identified as White. Over half of the NPs worked full-time (65.8%) and had greater than 4 years of experience (77.5%). Most NPs earned at least a master’s degree (83.8%). Over half co-managed their patient panel (55.9%).

Table 2.

Descriptive characteristics of NP sample

	All NPs (n=1,110)
Age (years), m (SD)
Age	48.6 (11.9)
Sex, n (%)
Women	966 (87.0)
Men	144 (13.0)
Race, n (%)
White	889 (80.1)
Black	41 (3.7)
Asian	110 (9.9)
Other	70 (6.3)
Ethnicity
Not Hispanic	1,017 (91.6)
Hispanic	93 (8.4)
Marital status, n (%)
Married	832 (75.0)
Not Married	278 (25.1)
Education, n (%)
ADN, ASN, BSN, Other	41 (3.7)
MSN	930 (83.8)
DNP, PHD	139 (12.5)
Years of experience, n (%)
≤3 years	250 (22.5)
4–9 years	383 (34.5)
≥10 years	477 (43.0)
Hours per week, n (%)
Less than 20 hours	53 (4.8)
20–40 hours	327 (29.5)
40+ hours	730 (65.8)
Panel management, n (%)
Co-managed panel	620 (55.9)
Managed own patient panel	490 (44.1)

Abbreviations: m = mean, SD = standard deviation, n = number, NP = nurse practitioner

Table 3 summarizes the characteristics of the primary care practices employing the sampled NPs. Almost half of the practices were physician-owned (46.9%).Additionally, over half of practices employed between 2–6 NPs (53.2%). Over 75% of practices were in reduced or restricted scope of practice states.

Table 3.

Descriptive characteristics of primary care practices

	All Practices (n= 1,002)
Practice Type, n (%)
Community Health Center (1)	210 (21.0)
Hospital Based Clinic (2)	92 (9.2)
Physician-Owned Practices (3)	470 (46.9)
All Other (0)	230 (23.0)
Number of NPs in Practice, n (%)
≥7 NPs	66 (6.6)
2–6 NPs	533 (53.2)
1 NP	403 (40.2)
State Scope of Practice, n (%)
Full	245 (24.5)
Reduced/Restricted	757 (75.6)
Work Environment Score, m (SD)	3.2 (0.51)
Structural Capabilities Score, m (SD)	0.57 (0.22)

Abbreviations: m = mean, SD = standard deviation, n = number, NP = nurse practitioner. Notes: The work environment score range is between 1.41 and 4. The structural capabilities score range is between 0 and 1.

Table 4 reports the NP job outcomes observed across practices with different SCI scores, which we categorized as low (0 to 0.39), medium (0.41 to 0.73), and high (0.75 to 1.0). Over the entire sample, 26.7% of NPs reported being burned out, over 10.2% were dissatisfied with their jobs, and over 20.2% planned to leave their jobs within the next year. NP job outcomes varied significantly across the structural capabilities categories, with NPs reporting consistently higher rates of negative job outcomes in practices with low SCI scores compared to practices with high SCI scores. For example, 33.6% of NPs reported intent to leave in practices with low SCI scores, compared to just 12.8% of NPs in practices with high SCI scores (p<.001). We found similar trends for burnout (low SCI: burnout=31.4% vs. high SCI: burnout=19.3%; p=.002) and job dissatisfaction (low SCI: job dissatisfaction=35.4% vs. high SCI job dissatisfaction=13.3%; p=.002).

Table 4.

NP job outcomes across practices with low, medium, and high structural capability scores

		Structural Capability Index Score
	All NPs (n=1,110)	Low (n=280)	Medium (n=549)	High (n=281)	P-value
Burnout status, n (%)					.002
Burned out	296 (26.7)	93 (31.4)	146 (49.3)	57 (19.3)
Job satisfaction, n (%)					.002
Dissatisfied	113 (10.2)	40 (35.4)	58 (51.3)	15 (13.3)
Intent to leave, n (%)					<0.001
Likely to leave	235 (21.2)	79 (33.6)	126 (53.6)	30 (12.8)

Abbreviations: m = mean, SD = standard deviation, n = number, NP = nurse practitioner.Notes: P-values generated from chi-squared tests. Structural capability index score ranges are based on dividing practices into four groups based on the aggregate structural capabilities score. The low-scoring practices include those in the first quartile, with scores ranging between 0 and 0.39. The medium-scoring practices include those in the second and third quartiles, with scores between 0.41 and 0.73. The high-scoring practices include those in the fourth quartile, with scores ranging between 0.75 and 1.0.

Table 5 presents the regression results. We report the coefficient for SCI, as well as the percentage point change in the probability that the outcome = 1 corresponding to a 0.1-unit (or 10-percentage point) increase in the SCI score (e.g., going from 0.5 or 50% to 0.6 or 60%). Focusing on the results from Model 2 (which controls for NP and practice characteristics), we find consistently significant relationships between the SCI score and job outcomes. Specifically, a 10-percentage point increase in the SCI score is associated with a 3.01-percentage point, 1.76-percentage point, and 3.32-percentage point decrease in the probability that an NP reports burnout (p<.001), job dissatisfaction (p<.001), and intent to leave (p<.001), respectively.

Table 5.

Effect of the Structural Capability Index on NP Burnout, Job Dissatisfaction, and Intent to Leave

Burnout
	Model 1: Unadjusted			Model 2: NP and Practice Characteristics			Model 3: Work Environment			Model 4: Interaction
	β (95% CI)	Percentage point change	P	β (95% CI)	Percentage point change	P	β (95% CI)	Percentage point change	P	β (95% CI)	Percentage point change	P
Structural Capability Index (SCI)	−0.27 (−0.39, −0.15)	−2.67	<.001	−0.30 (−0.42, −0.18)	−3.01	<.001	−0.11 (−0.23, 0.01)	−1.07	.082	−0.80 (−1.56, −0.04)	−8.00	.040
Work Environment	--	--		--	--		-0.28 (−0.34, −0.23)	--	<.001	−.40 (−.53, −.27)	--	<.001
SCI x Work Environment	--	--		--	--		--	--		.21 (−0.01, 0.43)	--	.058
Job Dissatisfaction
	Model 1: Unadjusted			Model 2: NP and Practice Characteristics			Model 3: Work Environment			Model 4: Interaction
	β (95% CI)	Percentage point change	P	β (95% CI)	Percentage point change	P	β (95% CI)	Percentage point change	P	β (95% CI)	Percentage point change	P
Structural Capability Index (SCI)	−0.17 (−0.25, −0.09)	−1.69	<.001	−0.18 (−0.26, −0.09)	−1.76	<.001	−0.04 (−0.12, 0.04)	−0.04	.284	−0.88 (−1.52, −0.24)	−8.79	.008
Work Environment	--	--		--	--		-0.19 (−0.24, −0.15)	--	<.001	−0.33 (−0.45, −0.22)	--	<.001
SCI x Work Environment	--	--		--	--		--	--		0.26 (0.07, 0.44)	--	.006
Intent to Leave
	Model 1: Unadjusted			Model 2: NP and Practice Characteristics			Model 3: Work Environment			Model 4: Interaction
	β (95% CI)	Percentage point change	P	β (95% CI)	Percentage point change	P	β (95% CI)	Percentage point change	P	β (95% CI)	Percentage point change	P
Structural Capability Index (SCI)	-0.29 (−0.40, −0.18)	-2.87	<.001	-0.33 (−0.44, −0.22)	-3.32	<.001	-0.15 (−0.26, −0.05)	-1.55	.006	-0.81 (−1.55, −0.07)	-8.12	.032
Work Environment	--	--		--	--		-0.26 (−0.31, −0.21)	--	<.001	-0.37 (−0.50, −0.24)	--	<.001
SCI x Work Environment	--	--		--	--		--	--		0.20 (−0.01, 0.42)	--	.065

Abbreviations: SCI = Structural Capability Index, β = Beta Coefficient, CI = Confidence Interval, P = p value.Notes: Coefficients and p values generated from linear probability models. Model 1: Unadjusted relationship between SCI and NP burnout/job dissatisfaction/intent to leave. Model 2: Add NP characteristics (i.e., age, sex, race, ethnicity, marital status, education, years of experience, hours/week, managed own panel) and primary care practice characteristics (i.e., practice type, scope of practice, number of nurse practitioners) to Model 1. Model 3: Add work environment to Model 2. Model 4: Add interaction of SCI x Work Environment to Model 3. To interpret the findings produced by the linear probability models, we applied the transformations described in theData Analysis section. We report the percentage point change in the probability of the outcome = 1 corresponding to a 0.1-unit increase in SCI score.

In Model 3, we additionally control for the work environment. While burnout and job dissatisfaction are no longer significantly associated with the SCI, intent to leave retains a significant relationship: A 10-percentage point increase in the SCI score is associated with a 1.55-percentage point decrease in the probability that an NP reports intent to leave (p=.006).

In Model 4, we additionally include an interaction term to assess if the SCI and work environment interact to jointly affect job outcomes. Notably, for job dissatisfaction, we find that there is a significant interaction between the SCI and work environment (p=.006), which slightly moderates the effects that the SCI and work environment individually have on the probability of reporting job dissatisfaction.

DISCUSSION

Using a large sample of NPs across 1,002 primary care practices in six states, we evaluated whether structural capabilities are associated with NP job outcomes. After controlling for NP and practice characteristics, we found that improvements in structural capabilities are significantly associated with reductions in burnout, job dissatisfaction, and turnover. Our findings indicate promising results, supporting the hypothesis that investing in infrastructure (e.g., EHRs, care reminders, care coordination) to support care delivery efforts can generate beneficial effects on NP labor outcomes.

We also considered if structural capabilities have an effect beyond the benefits of a good work environment. Prior studies have found that good work environments – professional settings where NPs have the resources and support to do their work well, professional visibility commensurate with physicians, and positive working relationships with colleagues – are associated with better outcomes, such as lower burnout and turnover and higher job satisfaction.^22,24 We found that after controlling for the work environment, structural capabilities maintained an independent and significant effect for intent to leave, but not for burnout or job dissatisfaction. These findings indicate that investments in structural capabilities can confer additional benefits beyond work environment improvements alone.

PCMHs have been viewed as one potential solution to many primary care challenges, and the adoption of structural capabilities are central to this effort.²¹ However, adoption of structural capabilities within primary care, particularly by small- and medium-sized practices,²¹ has not reached its full potential. To encourage uptake of structural capabilities in primary care, efforts should be made to improve both the work environment and primary care practice infrastructure simultaneously, as there is a synergistic relationship between the two. A 2018 study of community health centers found that practices with more favorable workplace climates were able to successfully incorporate structural capabilities into their practices.¹⁹ This is likely because integrating structural capabilities into primary care practices requires significant investment, not only in the structure itself, but also in the human capital required to support it. Our results for job dissatisfaction reinforce this point, as we find that structural capabilities and the work environment significantly interact to impact NP job dissatisfaction.

Primary care practices already face staffing shortages and a lack of resources;⁹ adding new care delivery infrastructure without the requisite support (e.g., staff, training, protected time to learn during work hours) can add additional burden to healthcare providers. A critical part of the work environment is the relationships between NPs and practice administrators.¹⁰ To ensure successful adoption and integration of structural capabilities, practice leaders should make concerted efforts to involve NPs and other clinicians in workflow redesign and implementation to support integration upfront. We show that successful integration of structural capabilities associated with PCMHs are promising not only for patients, but also for providers like NPs.

Limitations and Future Research

This study has several limitations. We utilize cross-sectional data, which restricts causal inference. We rely on NPs to self-report their working conditions and outcomes, which may be subject to reporting bias. It is also possible that selection bias influenced the results, as surveyed NPs could be unobservably different from not-surveyed NPs.

To determine the relationship between structural capabilities and NP job outcomes, we used a summary measure based on eight structures. However, this is not an exhaustive list, and there are other important structural capabilities in primary care, such as on-site language interpreters and quality improvement initiatives (i.e., frequent meetings on quality performance, presence of practice leader for quality improvement), that should be studied in future research. While we are unable to establish content validity in this study because it is a secondary data analysis, we demonstrated that SCI has good internal consistency reliability with Cronbach’s alpha of ~0.88.

Our sample of NPs was also limited to those working within six states, and while geographically diverse, they are not representative of national variability. Our sample of NPs was also limited in terms of racial, ethnic, and gender diversity, which reflects the lack of diversity in the national NP workforce.⁵⁰ Future research should investigate how to improve job outcomes among minority NPs, in order to retain a more diverse NP workforce.

Our study analyzes existing survey data, so our included covariates are limited to the variables available in the survey. For example, additional survey questions on NPs’ family life, commuting time, and plans to move or retire could help better understand what factors influence their work-life balance and job outcomes. Thus, there may be unobservable confounders affecting our analytic findings. Our analysis is also limited by the options NPs had to respond to survey questions, particularly around sex, race, and marital status.

Conclusions

We investigated the relationship between structural capabilities and NP job outcomes in primary care practices. We found that NPs in practices with higher levels of structural capabilities are less likely to be burned out, dissatisfied, or planning to leave their position. While improving the work environment remains an important way that practice leaders can increase NP retention, administrators should take action to integrate structural capabilities into the care delivery process, which will better support NPs in their work and, ultimately, improve the quality of primary healthcare services delivered to patients, families, and communities.