DNP Preparation of Primary Care Nurse Practitioners and Clinical Outcomes for Patients with Chronic Conditions

Abstract

Many professional nursing organizations have proposed that the Doctor of Nursing Practice (DNP) is the most appropriate entry-level degree for nurse practitioners (NPs). There have been no studies to date examining the impact of DNP preparation on quality of care or patient outcomes. In this study, we use survey data from over 1,000 primary care NPs in 6 states linked to Medicare claims data to examine differences in emergency department utilization and hospitalizations among patients with chronic conditions cared for by Master of Science in Nursing (MSN)- and DNP-prepared primary care NPs. We use regression models to control for various patient, NP, and practice characteristics that might confound the relationship. We find that patient outcomes are not statistically different between patients attributed to MSN- and DNP-prepared primary care NPs. These findings have important implications for NP education. Further empirical analysis related to the clinical outcomes other than health care utulization of the DNP degree is warranted. Future studies might consider examining 1) NPs in settings other than primary care, 2) practice-wide or system-wide outcomes, 3) other measures of care quality, and 4) impact of DNP program content.

Introduction

In 1968, Lorretta Ford, a nurse, and Henry Silver, a pediatrician, wrote the seminal paper first articulating the role of the nurse practitioner (NP) (Silver et al., 1968). The role was envisioned to allow registered nurses to “provide comprehensive well-child care to well children and identify, appraise, and temporarily manage certain acute and chronic conditions of the sick child.” Ms. Ford became the first NP, and the role of NPs became one of the greatest successes in nursing education and practice. As of 2022, there are more than 355,000 NPs across the United States working in a variety of roles in hospitals, primary care practices, specialty practices, home care, urgent care, and other settings (American Association of Nurse Practitioners, 2022). Many studies have shown that NPs provide equivalent or even better care than their physician colleagues (Barnett et al., 2022; Newhouse et al., 2011; Swan et al., 2015). NPs are an important workforce to ensure access to high-quality, equitable care in the United States (Buerhaus, 2018; Shalala et al., 2011).

The educational requirements for NPs have changed significantly over the last 50 years. In 1973, approximately 65 NP programs existed in the United States, the majority of which were short certificate programs located in medical schools (American Association of Nurse Practitioners, 2019b). From its inception, NP education was often controversial within nursing. Because many of the programs were housed in medical schools, they were often seen as too closely connected to organized medicine (Pulcini et al., 2019). Over time, many clinicians and educators felt that NP education must be offered at the master’s level to legitimize the educational track within nursing schools (Pulcini et al., 2019). Although the first master’s-level NP program was established at Boston College in 1967, it was only in the 1980s that a significant portion of certificate programs transitioned into master’s and post-master’s programs (Fairman, 2009; Pulcini et al., 2019). The National Organization of Nurse Practitioner Faculties wrote a position paper in 1984 stating that NP programs should be graduate-level and offered support for the transition (Fairman, 2009; Pulcini et al., 2019). By 1989, 90% of NP programs were either master’s-level or a post-master’s degree (Pulcini et al., 2019).

In 2004, the American Association of Colleges of Nursing adopted a position statement supporting the Doctor of Nursing Practice (DNP) as the most appropriate educational level for NP entry to practice (American Association of Colleges of Nursing, 2004). The DNP is a clinical doctorate like the clinical doctorates for medicine, physical therapy, occupational therapy, and audiology. The statement proposed a process to phase out the Master of Science in Nursing (MSN) degree and replace it with the DNP by 2015. There were many reasons cited by AACN to justify the position. Many MSN-level NP programs already included enough credits to count as doctoral education. Moving from MSN-DNP simply recognized the level of education that already existed. Other clinical roles, such as physical therapy and occupational therapy, were moving toward doctoral preparation, and the DNP would allow NPs to maintain degree parity with other professions including physicians. The DNP was also meant to respond to the increasing complexity of healthcare both at the clinical level and the system level. The degrees would offer advanced clinical education that would improve direct patient care as well as content, such as quality improvement and public policy, that would allow NPs to help lead system-level change.

This decision has not been without controversy. At the time of announcement, many nursing leaders argued that the current educational system for NPs was not broken (Nelson, 2005). Others argued that requiring the DNP, given that the degree would take longer to complete than the MSN to complete, might lead fewer nurses to prepare at an advanced practice level, negatively affecting cost, quality, and access to care (Chase & Pruitt, 2006; Dracup et al., 2005). Despite these concerns, many nursing schools have developed DNP programs for NPs. As of 2022, 384 nursing schools across the country offer DNP education, and there are 35,755 students enrolled in these programs (American Association of Colleges of Nursing, 2022). Some of these schools have made a full transition to the DNP and have closed their MSN programs. However, many schools have adopted the DNP as a post-MSN completion and have retained their MSN programs, while still others have maintained the MSN-only track for NPs. Programs that have not fully transitioned to the DNP have cited several concerns, including perceived student and employer demand, issues concerning accreditation and certification, and resource constraints (Martsolf et al., 2015).

Therfore, as of today, NPs have multiple educational options that can prepare them to receive liscensure and certification as NPs. There are important differences between the MSN and DNP degree which may lead to improved patient outcomes. First, the DNP degree was originally designed to provide NPs with advanced clinical training (Mundinger & Carter, 2019). Thus, the DNP-prepared NP would ideally be better-suited to care for patients, especially those with complex care needs, leading to improvements in quality and patient outcomes. Second, DNP programs include a number of courses related to health policy, quality improvement, and practice management, above and beyond the MSN curriculum. In this way, we might expect to see differences in outcomes on two levels. First, we might observe differences in outcomes in patients attributed to DNP-prepared versus MSN-prepared NPs driven by better clinical skills among DNP-prepared NPs. Second, we might see that practices with more DNP-prepared NPs report better overall outcomes, due to DNPs’ deeper knowledge of quality improvement and practice management.

Despite the promise of the DNP degree in improving quality of care and patient outcomes, there have been no studies to date that systematically examine the extent to which the DNP has impacted NPs’ clinical practice. A recent opinion piece by prominent nursing school deans argued for such evidence, stating that, “Nursing deans and nurse health care executives must call for data to tangibly demonstrate distinguishing features of DNP-educated APRNs, compared to other provider preparations, and then highlight those metrics for health system stakeholders” (McCauley et al., 2020).

A number of studies have examined the roles, responsibilities, and practice patterns of DNP-prepared NPs. One previous study found that MSN- and DNP-prepared NPs have a few differences in terms of their practice environment, independence in patient care, and their roles in clinical practice, administration, and quality improvement (Martsolf et al., 2021). Specifically, compared to MSN-prepared NPs, DNP-prepared NPs experience no difference in practice independence, have marginally better relationships with physicians, and spend slightly more time in practice leadership and less time in direct clinical care (Martsolf et al., 2021). A large mixed methods study of DNP employers in hospital settings, health systems, public health settings, and primary care found that DNP employers do not prefer the DNP degree over other advanced nursing degrees and report uncertainty regarding the difference between DNP- and MSN-prepared NPs (Beeber et al., 2019). Overall, these studies indicate that the role, responsibilities, and practice of DNP- versus MSN-prepared NPs are relatively similar across multiple settings.

Despite recent interest in examining the impact of DNP preparation on NP practice, no studies to date have compared clinical outcomes between patients cared for by MSN- and DNP-prepared NPs. In this study, we use survey data from over 1,000 primary care NPs in 6 states linked to Medicare claims data to examine differences in emergency department (ED) utilization and hospitalizations between patients with chronic conditions cared for by MSN- and DNP-prepared primary care NPs. Emergency department and hospitalization utilization can serve as a signal of the quality of care delivered in primary care practices.

Methods

Data

Survey sample

We used the IQVIA OneKey database to identify primary care NPs across 6 states: Arizona (AZ), New Jersey (NJ), and Washington (WA), Pennsylvania (PA), California (CA) and Texas (TX). These states were chosen because they were diverse and have variation in state scope of practice regulations at the time of data collection. IQVIA is a proprietary dataset that tracks all providers in nearly the entire universe of ambulatory practices across the United States. To do this, we first identified every ambulatory primary care practice in the United States as any practice for which least 50% of providers had a primary care specialty. For each of the practices, we obtained from IQVIA the practice address and phone number. Of those practices, we then selected all that employed at least one NP. We also called each of the practices to confirm that they were open, were primary care, and that the NP we identified still worked at the practice. The final sample consisted of 5,689 eligible NPs.

We conducted the survey from 2018 and October 2019 working with the Survey Research Institute at Cornell University. maximize the response rate, we followed a Dillman approach for mixed-mode surveys (Dillman et al., 2014). In total, 1,244 NPs completed and returned the surveys, for a response rate of 21.9%. The survey data was collected in 2018–2019. More details about the survey methodology and non-response bias is published elsewhere (Harrison et al., 2021).

Medicare claims files and practice attribution

We obtained Medicare claims for all Medicare beneficiaries (n=1,123,861) who had at least one visit to the practices which had at least one NP in our study. We used a common attribution approach to attribute patients to practices where they received care (Mehrotra et al., 2010). We first attributed patients to clinicians by NPI. To do this, we calculated the proportion of primary care evaluation and management (E&M) paid amounts provided to a given beneficiary by each clinician submitting at least one claim for that beneficiary in the target year (2019). Next, the beneficiary was assigned to the clinician who provided the highest proportion of E&M paid amounts (plurality) as long as that provider accounted for at least 30% of the E&M paid amount (Mehrotra et al., 2010). In the rare cases (<1%) of ties, one primary care clinician was randomly selected. We then retained all patients with age ≥ 65 and who had at least one of the following six chronic conditions: asthma, chronic obstructive pulmonary disease (COPD), hypertension, congestive heart failure (CHF), coronary heart disease (CVD), and diabetes. These conditions were chosen because they are the most common illnesses among Medicare beneficiaries. The conditions were identified in CMS Chronic Condition Warehouse (Chronic Conditions Data Warehouse, 2022). Primary and secondary International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) diagnosis from both outpatient and inpatient claims files were used to define the chronic conditions.

Variables

Dependent variables

We had four dependent variables in this study, including: 1) hospitalizations, 2) ambulatory care sensitive (ACS) hospitalizations, 3) ED visits, and 4) ACS ED visits. ACS can be conceptualized as any hospitalization or ED visit that could reasonably be prevented through adequate ambulatory primary care (Agency for Healthcare Research and Quality, 2001). We define ACS for each of the measures below.

Hospitalizations are identified as any record in the CMS inpatient claims file with length of stay more than one day during the study period. ACS hospitalizations are a subset of hospitalizations for which one or more of the 10 prevention quality indicators for being ambulatory care-sensitive, according to the Agency for Healthcare Research and Quality, was identified in the claims (Agency for Healthcare Research and Quality, 2001).

ED visits are any visits with Part B claims for Healthcare Common Procedure Coding System (HCPCS) codes 99281, 99282, 99283, 99284, and 99285 (Venkatesh et al., 2017). We identified ACS ED visits using the “NYU ED Algorithm” (Ballard et al., 2010; Gandhi & Sabik, 2014; NYU ED Algorithm).

Independent Variable

The independent variable for this study was highest educational attainment by the NPs completing the survey. Each NP was coded as having a DNP or MSN as their highest degree. Because the current policy and educational issue of interest is in comparing DNP and MSN experiences, we excluded NPs with only an associate’s or bachelor’s degree and those with a PhD.

Covariates

We included in the regression models a series of clinician, practice, and patient characteristics. Clinician-level covariates included years in the current practice setting, age, gender, and number of years since receiving initial NP license.

Practice-level covariates included practice type (e.g., physician practice, community health center, or hospital-based clinic) and practice size (i.e., the total count of NPs, physician assistants, and physicians). We also calculated a Structural Capability Index (SCI), which measures the extent to which practices have adopted capabilities that are consistent with the delivery of high-quality care (Martsolf, Ashwood, et al., 2018; Martsolf, Kandrack, et al., 2018; Martsolf et al., 2016). We controlled for whether the state in which the NP practiced had full (i.e., AZ and WA), reduced (i.e., NJ and PA), or restricted (i.e., TX and CA) scope of practice (American Association of Nurse Practitioners, 2019a). We further included the NP work environment scale from the Nurse Practitioner-Primary Care Organizational Climate Questionnaire (NP-PCOCQ). This scale contains 29 items that ask NPs to rate their practice environment based on NP-Administration Relations (NP-AR), NP-Physician Relations (NP-PR), Independent Practice and Support (IPS), and Professional Visibility (PV) domains. This is described extensively in other studies (Poghosyan et al., 2017; Poghosyan et al., 2019; Poghosyan et al., 2013).

Finally, we controlled for a number patient-level characteristics such as age, sex, race, and ethnicity. We also included in our models 27 comorbidities that we accessed via the Medicare Chronic Condition Warehouse. In the descriptive statistics table, we report on the average count of these chronic conditions for each beneficiary. But, in the regression models, these comorbidities were included as 27 different binary variables indicating the presence or absence of each chronic condition for each beneficiary. The specific chronic conditions are listed in Appendix Table 1. We also included a measure of the socioeconomic status of patients’ communities. We calculated this using the Area Deprivation Index (ADI) (University of Wisconsin School of Medicine and Public Health, 2022). We used 2019 data from the University of Wisconsin Neighborhood Atlas ADI to characterize the socioeconomic disadvantage of the community in which practices are located. The ADI measure uses American Community Survey Five Year Estimates in its construction. The 2019 ADI uses the American Community Survey data for 2019, which is a 5-year average of ACS data obtained from 2015 to 2019. It includes factors from the domains of income, education, employment, and housing quality (Kind & Buckingham, 2018; University of Wisconsin School of Medicine and Public Health, 2022).

Data Analysis

We calculated bivariate associations between each covariate and the DNP-MSN variable to assess differences in the NP, practice, and patient characteristics between MSN- and DNP-prepared NPs. For categorical variables, we used chi-square to compare differences. For continuous variables, we used t-tests. We then estimated multivariable generalized linear models (with a logistic link function) to measure the association between educational attainment of NPs and the dependent variables. We first estimated unadjusted odds ratios comparing the odds of hospitalizations or ED visits for patients cared for by MSN- versus DNP-prepared NPs. We then estimated adjusted odds ratios that included all clinician, practice, and patient covariates. We estimated robust standard errors using a sandwich estimator to account for the clustering of patients within NPs (Wooldridge, 2015).

Results

Our final sample size is 48,182 patients attributed to 752 NPs at 626 primary care practices. In Table 1, we present NP demographic characteristics. Of the 752 NPs, 677 reported holding an MSN degree, while the remaining 75 held a DNP degree. There were no significant differences between MSN- and DNP-prepared NPs in age, gender, or years of experience since receiving NP license. In Table 2, we show patient demographic characteristics. MSN- and DNP-prepared NPs cared for similar patients, though patients cared for by MSN-prepared NPs were more likely to be female (64.53% of MSN patients vs. 60.9% of DNP patients), Black (5.56% vs. 4.64%), and non-Hispanic White (81.48% vs. 80.49%), yet the observed differences in racial composition were very small. In Table 3, we show characteristics of practices employing NPs. There were no significant differences in the types of practices employing MSN- versus DNP-prepared NPs. However, DNP-prepared NPs were significantly more likely to work in full practice authority states (54.67% vs. 28.51%) and less likely to work in restricted states (20% vs. 33.23%). In Table 4, we present unadjusted differences in outcomes between patients cared for by MSN-prepared and DNP-prepared NPs. There were significant differences in the rates of ED visits (MSN: 31.52%; DNP: 33:76%), ACS ED visits (MSN: 21.21%; DNP: 22.86%), hospitalizations (MSN: 20.68%; DNP: 22.56%), and ACS hospitalizations (MSN: 3.08%; DNP: 4.03%). These differences were all significant at a p-value of 0.05. In the unadjusted regression models (Table 5), patients cared for by DNP-prepared NPs had higher odds of ED visits and hospitalizations than patients cared for by MSN-prepared NPs, but those differences were not statistically significant. After controlling for potential confounders, we still found that any observed differences in acute care utilization were not statistically significant (Table 5). In summary, we found no statistically significant differences in ED visit or hospitalizations rates between patients cared for by MSN-prepared and DNP-prepared NPs.

Table 1.

Descriptive statitics of Nurse Practitioner Demographics

	NP characteristics
	Overall	MSN	DNP	P-Value
	N= 752	N= 677	N= 75
Age – n (%)				0.94
<30	20 (2.66)	18 (2.66)	2 (2.67)
31–45	281 (37.37)	250 (36.93)	31 (41.33)
45–55	180 (23.94)	164 (24.22)	16 (21.33)
55–65	186 (24.73)	169 (24.96)	17 (22.67)
65+	85 (11.3)	76 (11.23)	9 (12)
Gender – n (%)				0.69
Male	99 (13.16)	88 (13)	11 (14.67)
Female	653 (86.84)	589 (87)	64 (85.33)
Years since receiving initial NP license - n (%)				0.08
0–2 years	63 (8.38)	58 (8.57)	5 (6.67)
3–8 years	309 (41.09)	268 (39.59)	41 (54.67)
9–19 years	255 (33.91)	237 (35.01)	18 (24)
20–39+	125 (16.62)	114 (16.84)	11 (14.67)

Note. NP = nurse practitioner; MSN = Master of Science in Nursing; DNP = Doctorate in Nursing Practice; n = number

Table 2.

Descriptive Statistics of Patient Demographics

	Patient Characteristics
	Overall	MSN	DNP	P-value
	N= 48182	N= 43867	N=4315
Age – mean (SD)	71.45 (12.03)	71.49 (12.0)	71.16 (12.38)	0.08
Number of chronic conditions – mean (SD)	4.38 (2.85)	4.38 (2.85)	4.39 (2.88)	0.76
ADI National Rank – mean (SD)	46.3 (22.4)	46.25 (22.52)	46.7 (21.21)	0.22
Female sex – n (%)	31,261 (64.2)	28,570 (64.53)	2,691 (60.9)	<0.0001
Race – n (%)				0.0003
American Indian/Alaska Native	381 (0.78)	336 (0.76)	45 (1.02)
Asian	979 (2.01)	891 (2.01)	88 (1.99)
Black	2,665 (5.47)	2,460 (5.56)	205 (4.64)
Hispanic	4,011 (8.24)	3,608 (8.15)	403 (9.12)
Non-Hispanic White	39,635 (81.29)	36,078 (81.48)	3,557 (80.49)
Other	315 (0.65)	273 (0.62)	42 (0.95)
Unknown	710 (1.46)	631 (1.43)	79 (1.79)

Note. NP = nurse practitioner; MSN = Master of Science in Nursing; DNP = Doctorate in Nursing Practice; n = number; SD = standard deviation; ADI = area deprivation index

Table 3.

Descriptive Statistics of Nurse Practitioner Practice Characteristics

	Practice Characteristics
Covariates	Overall	MSN	DNP	P-Value
	N= 626	N= 567	N= 59
Practice type – n (%)				0.28
Physician practice	322 (42.82)	297 (43.87)	25 (33.33)
FQHC	141 (18.75)	123 (18.17)	18 (24)
RUHC	26 (3.46)	25 (3.69)	1 (1.33)
Hospital-based clinic	78 (10.37)	69 (10.19)	9 (12)
Other	185 (24.6)	163 (24.08)	22 (29.33)
NP practice environment – mean (SD)
NP-Physician Relations	3.31 (0.51)	3.30 (0.51)	3.37 (0.52)	0.26
Independent Practice and Support	3.48 (0.44)	3.48 (0.44)	3.50 (0.46)	0.70
Professional Visibility	3.13 (0.65)	3.13 (0.64)	3.11 (0.73)	0.75
NP-Administration Relations	2.87 (0.71)	2.87 (0.71)	2.84 (0.73)	0.72
Practice size – n (%)				0.31
Solo providers	22 (2.93)	21 (3.1)	1 (1.33)
1–2 providers	85 (11.3)	72 (10.64)	13 (17.33)
3–5 providers	297 (39.49)	269 (39.73)	28 (37.33)
>5 providers	348 (46.28)	315 (46.53)	33 (44)
Structural Capability Index – mean (SD)	0.64 (0.21)	0.64 (0.20)	0.63 (0.21)	0.73
Scope of practice – n (%)				<0.0001
Full	234 (31.12)	193 (28.51)	41 (54.67)
Reduced	278 (36.97)	259 (38.26)	19 (25.33)
Restricted	240 (31.91)	225 (33.23)	15 (20)

Note: n = number; MSN = Master of Science in Nursing; DNP = Doctorate in Nursing Practice; NP = nurse practitioner; SD = standard deviation; FQHC = federally qualified health center; RUHC = rural health center

Table 4.

Descriptive Statistics of Patient Outcomes

	Overall	MSN	DNP	P-value
	N= 48,182	N= 43,867	N=4,315
ED visit – n (%)	15,446 (31.72)	13,954 (31.52)	1,492 (33.76)	0.002
ACS ED visit – n (%)	10,399 (21.35)	9,389 (21.21)	1,010 (22.86)	0.01
Hospitalization – n (%)	10,154 (20.85)	9,157 (20.68)	997 (22.56)	0.003
ACS hospitalization – n (%)	1,543 (3.17)	1,365 (3.08)	178 (4.03)	0.001

Note: n = number; MSN = Master of Science in Nursing; DNP = Doctorate in Nursing Practice; ED=Emergency department; ACS=Ambulatory Care Sensitive

Table 5.

Findings from Regression Models Examining Patient Outcomes by MSN and DNP (MSN referent) (n=48,182)

	Unadjusted		Adjusteda
	Odds Ratio	P-Value	Odds Ratio P-Value
ED	1.12	0.22	1.06	0.35
ACS ED	1.10	0.30	1.03	0.72
Hospitalization	1.12	0.44	1.08	0.38
ACS Hospitalization	1.01	0.51	1.05	0.74

Note: MSN = Master of Science in Nursing; DNP = Doctorate in Nursing Practice; ACS = ambulatory care sensitive; ED = emergency department

^a: models adjusted for practice, clinician, and patient characteristics.

Discussion

This is the first study to compare the outcomes of patients cared for by NPs with different educational backgrounds. In particular, we found that there were no significant differences in acute and emergecy care utilization patterns between patients cared for by MSN- versus DNP-prepared NPs. While the raw observed differences in outcomes indicated that patients cared for by DNP-prepared NPs had higher rates of ED visits and hospitalizations, those differences were not statistically significant in regression models that controlled for various NP, practice, and patient characteristics. Our findings suggest that the DNP degree has not led to major and observable improvements in these patient outcomes. Below, we discuss a number of important implications of these findings for NP practice, education policies, and future research.

Our findings show that the training and education of DNPs has not translated into preventing acute and emergency care use among older adults. It may be that the the current confuguration of DNP programs are not, in fact, providing NPs with stronger clinical skills compared to MSN programs. DNP programs have grown significantly over time, arguably the largest growth in programs has been in post-MSN DNP programs (Mundinger & Carter, 2019). These programs often focus on leadership and policy rather than advanced clinical practice, and they often have no clinical component (McCauley et al., 2020; Mundinger & Carter, 2019). Post-BSN DNP programs tend to be clinically-focused APRN programs, but little is known about systematic differences in clinical hours between post-BSN DNP programs and traditional MSN programs. There is some evidence that BSN-DNP programs require more clinical hours compared to MSN programs, but there are likely significant variations across programs (McCauley et al., 2020). Furthermore, it remains unclear if any differences in clinical hours contribute to differences in readiness to practice among graduating NPs (McCauley et al., 2020). The field would be well-served to produce an extensive study that examines differences in the clinical competencies that DNP education provides above and beyond the MSN.

Our findings provide important insights about continued efforts of various professional nursing organizations to promote the DNP for entry-to-practice for NPs. In addition to AACN’s commitment to the position that DNPs should replace MSNs as the entry to APRN practice (American Association of Colleges of Nursing, 2004), the National Organization of Nurse Practitioner Faculty (NONPF) also announced their support for “a seamless, integrated DNP curriculum without a master’s exit point as preparation for entry to the NP role” (The National Organization of Nurse Practitioner Faculties, 2018). The primary justification for the NONPF statement is that doctoral education will help NPs “lead” and “deliver high quality care” (The National Organization of Nurse Practitioner Faculties, 2018). Our findings indicate no difference between MSN and DNP education on the outcomes we have investigated. Health care systems actions suggest that they perceive little benefit from DNP above MSN education. Several previous studies found that health systems are not using NPs with DNPs any differently than NPs with MSNs (Beeber et al., 2019; Martsolf et al., 2021). Evidence shows that MSN programs continue to produce high-quality NPs (Buerhaus, 2018). Our findings suggest that it remains premature to push systematic elimination of the MSN without more evidence that doctoral education improves patient outcomes and meaningfully helps health systems.

Although this paper represents an important contribution to the literature, there are several limitations. We were unable to assess the extent to which “incident-to” billing might impact our results. Incident-to occurs when care provided by NPs is billed by a supervising physician. Incident-to billing would bias the results if MSN- and DNP-prepared NPs billed incident-to at different rates and incident-to was correled with ED utilization or hospitalizations. Previous studies suggest that incident-to billing accounts for up to 40% of all NP-delivered care (Patel et al., 2022). We were unable to account for incident-to billing. Our response rate was relatviely low. However, when we compared responding and non-responding NPs on common observable characteristics, the differences were relatively small (Harrison et al., 2021). We did control for many clinician, practice, and patient characteristcis including extensive control for patient comorbidities. However, it is possible that there remain important unobserved differences between MSN- and DNP-prepared NPs for which we could not control. So, these results should not be interpretted as strictly causal to the extent to which there are important missing unobserved covariates. Finally, the sample size of DNP-prepared NPs was small. This may have impacted our ability to identify statistically significant differences in the outcomes by education status.

We recognize that more evidence is needed to guide the debate regarding the future of NP education. We propose a number of potential studies that should be conducted to better understand the impact of DNP education on patient outcomes. Our study only examined patients that were directly attributed to and cared for by NPs. However, one of the primary justifications for the DNP is that doctoral education prepares NPs to address practice-wide or system-wide outcomes. Future studies should examine how the DNP’s unique skillset might impact population- or system-wide outcomes. Our study was also limited in terms of the outcomes of interest. We focused solely on acute and emergency care utilization, which is an important indicator of clinical quality. Yet, it is possible that DNP preparation might impact other aspects of quality of care, such as consistent delivery of guideline-recommended treatment for chronic conditions. Future studies should focus on comparing process measures between DNP- and MSN-prepared NPs. Additionally, our study focused exclusively on primary care. DNP preparation may be especially important in other settings, such as acute care or population health. Research has shown that more NPs are entering specialty practice settings (Martsolf, Barnes, et al., 2018). Studies should examine the impact of DNP-prepared NPs in acute care, specialty practice, and population health. We had no information about the type of DNP degree that the NPs had completed. We do not know if the degrees were BSN-DNP or MSN-DNP programs. Future studies should gather information on the type of DNP degree that NPs have completed. We also had no information on the DNP programs that the students completed. Future studies might focus on examining the content of DNP programs and the impact of varying DNP program content on outcomes.

Conclusion

This study is the first to compare patient outcomes between MSN- and DNP-prepared NPs. Our study found that patient outcomes were not statistically different between patients attributed to MSN- and DNP-prepared NPs. These findings call for further investiagtion of the impact of DNP edcuation on patient outcomes. We hope that these findings jumpstart further empirical investigations into the clinical impact of the DNP degree.

Highlights.

• Many have proposed the Doctor of Nursing Practice to be the most appropriate entry-level degree for nurse practitioners.

• No studies to date examined the impact of DNP-preparation on patient outcomes.

• We found no differences in hospitalizations or ED visits among patients cared for by DNP-prepared NPs compared to MSN-prepared NPs.

Appendices

Appendix Table 1.

27 Conditions Included in Model

Alzheimer’s Disease

Alzheimer’s Disease and Related Disorders or Senile Dementia

Acute myocardial infarction

Anemia

Asthma

Atrial fibrillation

Breast cancer

Colorectal cancer

Endometrial cancer

Lung cancer

Prostate cancer

Cataracts

Chronic heart failure

Chronic kidney disease

Chronic obstructive pulmonary disease

Depression

Diabetes

Glaucoma

Hip fracture

Hyperlipidemia

Benign Prostatic Hyperplasia

Hypertension

Hypothyroidism

Ischemic heart disease

Osteoporosis

Rheumatoid Arthritis/Osteoarthritis

Stroke/TIA