Inclusion and exclusion criteria

All relevant published and unpublished systematic reviews reported from January 1, 2010 up to December 2, 2022, with no restrictions on jurisdiction or language, were included. To be retained, authors needed to provide sufficient information so the reviewer could identify all the components of a research question (i.e., PICOS), inclusion and exclusion criteria, and use methods to identify relevant published and unpublished evidence to reduce the risk of bias [16]. We excluded reviews if the PHCNP impact was not reported separately from other types of nurses or team members or if no indicators could be identified. We also excluded reviews that addressed broad research questions (e.g., scoping reviews). Nurse midwives were excluded because these roles are not consistently identified as advanced practice nursing roles internationally, and regulatory bodies in different countries do not all require these roles to be filled by nurses.

We considered reviews that included all types of study designs to capture the impact of a complex intervention like the addition of PHCNPs in healthcare teams because different types of information are needed to inform decisions about role effectiveness [19]. Participants included patients of any age, groups or communities receiving PHC care in all types (e.g., teaching and non-teaching), sizes (e.g., small, large) and locations (e.g., rural, urban) of community or care agencies (e.g., primary care, home care, long-term care).

Search strategy

We limited the search from January 1, 2010 up to December 2, 2022 to capture the most up-to-date trends. Pieper et al. [20] argue that evidence in about half of published reviews is outdated after five years. Small changes were made in the databases as proposed in the protocol as some were no longer available at the time data were collected [18]. The following electronic databases were searched: Cumulative Index to Nursing and Allied Health Literature (CINAHL), Cochrane Library Database of Systematic Reviews and Controlled Trials Register, Database of Abstracts of Reviews of Effects (DARE), Embase, Global Health, Joanna Briggs Institute (JBI) EBP, Ovid Healthstar, Ovid MEDLINE, PubMed, and Web of Science Core Collection. The search strategy combined subject headings and keywords related to primary healthcare, advanced practice nursing, and different types of outcomes. They were subsequently combined with a search filter we derived from the Canadian Agency for Drugs and Technologies in Health (CADTH)’s systematic reviews and meta-analyses search filter [21] and that designed by Lunny et al. for reviews of systematic reviews [22]. The search strategy was adapted for each database, and reviewed by an academic librarian (See S1 Appendix). In addition, the reference lists of relevant reviews were hand-searched.

The grey literature was searched for the same period, and included: CADTH Information Services, relevant websites from CADTH’s Grey Matters tool, Health Evidence, Heath Systems Evidence, OpenGrey Repository, Organization for Economic Co-operation and Development (OECD), PDQ-Evidence, ProQuest Dissertation and Theses, and World Health Organization. The PROSPERO International Prospective Register of Systematic Reviews was searched to identify registered review protocols. For each website, the content was searched using the same search terms (e.g., indicator AND (primary care) AND (nurse practitioner)) as those used for the published literature. If there was not an inherent search function on the website, a search was conducted of all webpages and weblinks. The search strategy for the unpublished literature is included in S2 Appendix.

Study selection

All reviewers (eight in total, grouped into four pairs) were trained to use the screening instrument and the inclusion/exclusion criteria. Regular follow-up sessions were held to answer any questions. The retained studies were uploaded into the EndNote software and the RAYYAN web platform [23]. Duplicates were removed. Two reviewers independently screened titles and abstracts using the prespecified inclusion/exclusion criteria. Discrepancies were discussed among the reviewers. A third reviewer acted as a tie-breaker if the first reviewers could not come to a consensus. If no abstract was available, a full text review was completed. A full text review was completed for all the reviews included after the initial screening. The reviewers reached an agreement level of 84.3% using the abovementioned inclusion/exclusion criteria, corresponding to a fair inter-rater agreement (Cohen’s kappa: 0.394) [24]. Exclusions related to “wrong role” included systematic reviews with no PHCNP role or with a PHCNP role not meeting our pre-specified definition, focusing on the wrong patient population/setting (i.e., not PHC), or whose topic was irrelevant to answer our research question. The most important challenge for reviewers was to identify consistent PHCNP role definitions.

Data extraction

For the included papers, all data were extracted by one coder and reviewed by a second coder. A structured tool was developed for the study and pilot-tested by the investigators [25]. The extracted data included the aim of the review; review characteristics (e.g., year); number of databases searched; countries; number of studies, date range and designs; population characteristics, intervention, comparator, and synthesis method. Specifications of patient, provider and health system outcomes, how the outcomes were measured, number of patients and providers in the intervention and control groups, effect sizes, p values, and funding source were recorded. The results of the meta-analyses, if conducted, were included. If the data were not available in the review, extractors indicated ‘not reported’ in the data extraction form.

Risk of bias assessment

Each review’s methodological quality was assessed using AMSTAR 2 criteria [26]. Training sessions were conducted with the reviewers who conducted the assessments to enhance rater understanding of the items. The reviewers independently rated each review, and inter-rater agreement was estimated using the kappa statistic [24, 26]. Any disagreements were discussed among the reviewers until consensus was reached. A summary table with the AMSTAR 2 ratings can be found in Table 1. Inter-rater agreement was 91.1% and substantial for all the items using Cohen’s kappa (κ) (κ = 0,791). Estimates using Cohen’s kappa was assessed as fair to complete agreement with κ values ranging from 0.377 to 0.952 and perfect agreement. Disagreements between reviewers were noted for some elements of the AMSTAR 2 assessment tool (see Discussion). Nineteen studies were assessed to be at low risk of bias with four or fewer criteria that were not met and no critical domains missing [11, 27–45]. Four studies were deemed to be at high risk of bias [46–49]. The studies at a high risk of bias have a limited impact on review findings as Fraser [46] was one of 17 studies reporting on cost outcomes; Mileski [47] reported on how frequently a theme was reported in their review; Ness [48] was one of 15 studies reporting on prescribing; and Smith [49] was one of seven studies reporting on illness prevention. Funding sources were often not reported in the included reviews.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Risk of bias assessment for included studies.

https://doi.org/10.1371/journal.pone.0290977.t001

Outcomes

The primary outcomes of the review of reviews are those that assess patient, provider and health system outcomes directly evaluating PHCNP roles. According to Laurant and colleagues [13], those working predominantly in complementary roles provide additional services that are intended to complement or extend existing services, and those working in predominantly alternative roles provide similar services to those for whom they are substituting (usually physicians). We used these distinctions to ascertain if care was in favour of the comparator, equal to or superior to usual care.

Analysis

Narrative synthesis was used to summarize the findings. As proposed by Olry de Labry et al. [50] and Ramalho et al. [4], outcomes were categorized at the patient (e.g., blood pressure), provider (e.g., knowledge) and health system (e.g., cost) levels. Given the large number of indicators, an iterative process was used to code indicators into categories to synthesize the findings. The indicators included in each category are presented below. Summary tables were developed to describe review characteristics (e.g., year of publication), outcomes, quality assessment, and findings. A record was kept of all review-related decisions (See S3 Appendix). As proposed by Smith et al. [51], no additional quantitative analyses were planned because of the potential overlap in studies included in more than one review.

Patient indicator categories

Activities of daily living were reported in seven studies [27, 32, 34–37, 52, 53], and included measures of health status and functional abilities in older adults and children with special needs. Donald [27] and Osakwe [37] noted significant improvements in ADLs (p = 0.04 and p = 0.02, respectively). Parents of children with special needs missed less work (pre vs post 26.3% vs 14.1%, p < .05) and children with special needs attended school significantly more often (missed school >20 days (pre vs post 10.4% vs 11.7%, p >.05) following the provision of NP home care visits [53]. Lovink [32], Morilla-Herrera (2016) [34] and Newhouse/Stanik Hutt (2013) [35, 36] found mixed results, with some studies showing statistically significant improvements in ADLs (e.g., basic ADLs), while other studies in their reviews found no difference (e.g., social functioning) between the intervention and control groups.

Adaptation to health conditions was reported in five studies [34, 52, 54–56]. This indicator was examined using 12 different measures. Significant improvements in the NP group were noted in adaptation-related goals [54], reduced disempowerment in treatment [55], life review therapy for homebound elderly patients with depression [55], reduced uncertainty for depressed women with cancer (p = 0.018) [55], reduced distress in women with cancer (p < 0.0001) and in adolescents (p < 0.0001) in the first 24 months (no difference at 36 months) [55], empowerment for patients with arthritis at 12 months (no difference at 24 months) [56], care planning (p = 0.005) [34], and reduced caregiver distress (p = 0.05) [34]. Non significant between-group differences were noted in overall goal-attainment for long-term care residents, and advance directive measures [54] and caregiver burden [52].

Clinical conditions represented the largest number of indicators to summarize. These indicators were categorized into health, cardiovascular, cancer care, diabetes, mental health, renal, respiratory, rheumatoid arthritis outcomes. Each one is detailed below:

1. Clinical/Health outcomes were found in eight studies [12, 35, 36, 41, 42, 47, 53, 57, 58] and included health risk reduction, health status, self-reported perceived health, SF 36 physical composite score, physiologic measures, and clinical outcomes. Milesky et al. [47] reported that NPs improved health outcomes in almost one fifth of theme occurrences. No significant between-group differences were noted in three reviews [12, 35, 36, 57]. Improvements in health outcomes were noted in the NP group in three other reviews [41, 53, 58], but p values were not reported in two studies [41, 58]. Van Vliet et al. highlighted that little is known about the influence of NP roles in ambulance care [42].
2. Clinical/Cardiovascular outcomes were identified in 10 studies [11, 32, 33, 35, 36, 39, 57, 59–62] and included blood pressure, lipids, heart failure, ejection fraction, N-terminal pro brain natriuretic peptide levels (marker of heart failure), functional exercise capacity (cardiac function), left-ventricular end-diastolic volume index, undetected hypertension, and vascular risk reduction. Significant improvements in favour of the NP group were found for blood pressure outcomes in four reviews [11, 33, 39, 60]. Trends towards improvement of metabolic outcomes and reduced vascular risk favouring the NP group were noted in studies by Carranza [59] and Smigorowsky [57] (no p value reported). Non significant results were noted in four studies [33, 39, 60, 61]. Rates of previously undetected high blood pressure ranged from 14% to 63% in one review (no p value reported) [62]. Lipids were measured using cholesterol, mean level of total cholesterol, reduction in high density lipoprotein and triglycerides. Trends in reduction of total cholesterol and high density lipoprotein favouring the NP group were found in studies by Schadewaldt [39] and Newhouse/Stanik-Hutt [35, 36], while no significant differences were noted in three studies [33, 39, 60]. Schadewaldt et al. [39] identified fluctuations in lipid control at 3, 6, 12 and 18 months. Mixed results were observed in two reviews, with equal to superior results for the NP group [35, 36, 39]. Statistically significant improvements were noted in the ejection fraction and composite score for patients with heart failure in favour of the NP group [32].
3. Clinical/Cancer care was identified in two reviews [31, 49]. Measures included clinical examination of skin lesions, accuracy of skin lesion examinations, and post-treatment survivorship care. Loescher et al. [31] indicated a trend towards a greater number of examinations and a decrease in the number of referrals to a dermatologist and in the number of biopsies (no p value reported). Smith and colleagues [49] highlighted they were unable to identify studies involving NPs in their assessment of post-treatment survivorship care.
4. Clinical/Diabetes outcomes were highlighted in seven studies [32, 33, 35, 36, 53, 59–61]. Indicators included HbA1C, HbA1C at 3 months, two or more HbA1C tests, glucose, diabetic control, foot examination, retinal exams, and annual eye exams. No significant differences between groups in HbA1C were noted in three reviews [32, 35, 36, 60], including a meta-analysis that included four studies (n = 589 patients) [33]. Mixed results were reported for HbA1C in one study [61]. Improvements in HbA1C were found in two studies [53, 59], but no p value was indicated in one review [59]. Glucose control was identified in two studies (no p value reported) [35, 36, 59], with a trend towards improvement in the NP group noted by Carranza (2021) [59]. Diabetes control was indicated in two studies [60, 61], with significant improvements favouring the intervention group in one study [61] and no differences found in the other study [60]. Retinal exams and foot exams were conducted more frequently in the NP group in one review [60] and no significant differences were noted in requests for annual exams by ophthalmologist in the second study [61].
5. Clinical/Mental health outcomes were reported in eight reviews [12, 27, 34, 39, 44, 52, 55, 57]. Significant reductions were noted in the NP group in anxiety rehabilitation (p < 0.0001) [27], psychological stress [55], uncertainty for depressed women with cancer [55], depression (p = 0.0005) [55], and illicit drug use (p = 0.029, p = 0.046) [44]. NP-delivered motivational interviewing led to a significant increase in patient depression self managemen in one review (p ˂ 0.001) [44]. Trends towards reductions in anxiety and in cognitive impairment and severe depression at 12 months were noted but p values were reported inconsistently [44, 52]. There were no significant differences in: depression and anxiety in long-term care [34]; primary care at 1 and at 4 years [39]; the SF 36 mental composite score for patients with cardiovascular disease [57]; subjective health status [12]; and alcohol use disorder identification test scores [44]. Schadewaldt et al. [39] cautioned against a risk of a cross-over effect in the control group at 4 year with decreased odds for depression (p = 0.001).
6. Clinical/Renal outcomes were identified in two studies [33, 60]. Significant differences favouring the NP group were noted in urinalysis results in one review (p < 0.01) [60], whereas no significant differences between groups were noted in mean urine sodium excretion and serum creatinine at six months in another review [33]. There was a trend towards higher levels of urinary albumin excretion favouring the control group, but no p value was reported.
7. Clinical/Respiratory outcomes were outlined in six reviews [27, 29, 33, 35, 36, 53, 60]. No significant differences were noted between the intervention and control groups for change in maximal peak flow in asthma, emergency nebulization, frequency of exacerbation, 6-month follow-up, asthma severity, asthma symptoms, absence from school or work due to asthma, forced expiratory volume at one second (FEV1), peak flow rate, airway hyper-reactivity using PD/PC20 methacholine/histamine, as well as lung function at 12 and 24 months. Mixed results noted for asthma control. No significant difference noted by Donald et al. (2015). Significant reductions in nightime symptoms (t test: 3.966 df 19 p < .05), oral steroids (t test: 3.750 df 18, p < .05), better adherence to therapy (z score -3.272, p < .05), and level of control (z score -4.132, p < .001) noted in one review [53].
8. Clinical/Rheumatoid arthritis outcomes were noted in two studies [56, 59]. Disease activity in rheumatoid arthritis was assessed in one review [56] with significant reductions in the NP group across all included studies (p < 0.05). Disease progression and severity were assessed in one review, with trends towards improvements noted [59]. Morning stiffness was examined in one review, with equal to significant improvements reported in the NP group (p = 0.01) [56].

Diagnosis was identified in three studies [59, 63, 64] and included diagnostic accuracy, most common diagnoses, and nurse-initiated X-Rays and treatments. The most common diagnoses in the Emergency department included soft tissue injuries and fractures. Improvements were noted in adenoma detection (p < 0.001), diagnosis and treatment (p < 0.001) with the NP group.

Education-Patient included seven studies [11, 34, 41, 59–61, 64], all of which noted positive trends with the NP group for this indicator category. Measures included patient information, health education, pharmaceutical treatment, symptom relief, discharge information, providing written documentation as a strategy, increased knowledge, information on who to contact if needed, dietary and activity recommendations including sodium reduction, moderation in alcohol consumption, weight control and reduction. Statistical significance in favour of the NP group was reached in four reviews [11, 34, 60, 61].

Mortality was examined in eleven studies [10, 27, 28, 30, 32, 34–36, 39, 52, 56, 59]. Seven reviews [27, 30, 32, 34–36, 52, 56] found no significant differences between groups in mortality, mortality at 90 days and 24 months. Significant reductions were noted for the NP group in three reviews [28, 52, 59] for mortality and mortality at 12 months. Mixed results were identified in two reviews, where one review found that 1 out of 4 included studies identified a lower risk of mortality in the control group [10]; the other review [39] found statistically significant reductions in total mortality for the NP group at the 4-year follow-up (p = 0.038). This trend became non significant at the 10-year follow-up because of the risk of a cross-over effect among participants [39].

Patient Adherence was examined in three studies [12, 27, 29]. Results favoured NP care [27] with increased patient adherence with beta-blockers, statins, and angiotensin-converting enzyme inhibitors, and attendance at cardiac rehabilitation. These estimates did not reach statistical significance. Swan and colleagues reported that patient adherence to follow-up was higher in the NP group, but did not report a p value [12].

Quality of Life (QOL) was identified in 12 reviews [27, 29, 32, 34, 37, 39, 41, 55, 57, 59–61]. Statistically significant improvements favouring NPs were highlighted in three reviews for physical and mental QOL over time (p < 0.001), diabetes [60], and QOL and general health perception [39]. No differences between groups were noted in asthma or respiratory QOL [27, 29], mental QOL [60], QOL at 18 months and four years [39], and quality-adjusted life years, as well as health-related QOL after cardiac surgery [57]. Kuethe [29] undertook a meta-analysis of two studies and found no between-group difference in QOL. Equal to superior findings on QOL was found in three reviews [34, 41, 61]. One review identified significant deterioration in physical QOL in the NP group (p = 0.4) [60].

Satisfaction-Patient and Family was examined in 18 reviews [11, 12, 27, 34–36, 38, 40, 41, 45, 52–54, 59, 60, 64–67]. Equal to improved patient and family satisfaction were noted consistently for all reviews. Statistical significance and p values reported inconsistently. Significant improvements favouring the NP group were specified in four reviews [11, 27, 40, 67]. Patel et al. [38] noted contradictory findings for patient satisfaction with usual care (control group) in jurisdictions in the United States with the least restrictive vs the most restrictive NP scope of practice policies.

Signs and Symptoms included nine reviews [27, 29, 34, 39, 40, 56, 59, 61, 68], and comprised: symptom management, symptom improvement, urinary tract infections, pain, fatigue, arthritis impact, symptom free days, symptoms of dementia, cognitive and behavioural changes, angina symptoms, continence after treatment, urinary symptoms (i.e., frequency, nocturia, urgency and urinary incontinence) at three and six months, daytime and night-time leakage, leakage severity at three months, total number of symptoms at three months, overall symptoms at six months, and management of asymptomatic urinary tact infection. Significant differences favouring the NP group were noted for symptom management [59], managing symptoms of dementia [34], all urinary symptoms, total number of urinary symptoms and daytime severity of leakage at three months [40] and managing asymptomatic urinary tract infactions with increased fluids (p < 0.001), frequent toileting (p < 0.001), and cranberry juice (p < 0.05). Equal to significant results favouring the NP group for pain and fatigue were reported [56]. All other findings showed no significant differences between the groups.

Provider indicator categories

Adherence to best practice-Providers was identified in six studies [12, 27, 32, 39, 44, 61]. NPs were reported to improve adherence to best practice guidelines for medications including beta-blockers, statins, angiotensin-converting enzyme inhibitors, and aspirin intake at 1-year follow-up (p < 0.001) [27, 32, 39, 61]. No significant differences were noted for medications including aspirin and clopidogrel [27]. No significant differences were noted for NPs providing mental health guideline-recommended care for medications, counselling, cognitive behavioral and problem-solving therapy to older adults, patient monitoring and motivational interviewing in one review [44]. Mixed findings were observed for one review in long-term care [32] and another in primary care [12]. In long-term care, Lovink et al. [32] found no significant differences in adherence and compliance with guidelines including annual mandatory histories and physical examinations. Significant findings favouring the NP group were noted for care of vulnerable elders (p < 0.001), secondary prevention, patients with dementia, incontinence, and adherence to care and co-management [32, 61]. In primary care, Swan and colleagues [12] found that results favoured NPs in providing disease-appropriate care, informing patient on the cause of illness, symptom relief, and chances of illness recurrence in two out of three studies, with no significant differences in the third study.

Education-Provider included five studies [31, 32, 64, 69, 70]. Provider knowledge was assessed for the use of the Australasian Triage Scale, patient counselling on obesity, skin cancer detection and training, early detection of skin cancer, and orientation in long-term care. Two reviews [31, 70] reported that knowledge of clinical examination of skin lesions improved with training but no p value was indicated. Stratton et al. [70] identified that didactic sessions for providers in dermatology lasted 14 minutes and clinical apprenticeships were up to six months but no indication of timing or frequency was provided. The modes of delivery of the didactic interventions were face to face or using observations by experts.

Illness Prevention was identified in seven studies [27, 39, 49, 55, 56, 59, 60]. Illness prevention indicators included health behaviours (physical activity, diet/nutrition, smoking status, cessation and cessation up to 1 year), weight/body mass index and weight loss, screening women for depression, health assessment, screening for cancer (cervical, breast, colorectal) and human papillomavirus (HPV) vaccination. Significant differences favouring the NP group were noted in four reviews [49, 55, 56, 60]. Results favoured the NPs for weight loss in one study (no p value reported) [59]. Another review noted equal to significant findings for smoking cessation up to 1 year [39]. Smith and colleagues [49] highlighted that no studies examined HPV vaccination performed by NPs, and that two out of four studies showed that physicians reported more colorectal cancer screening than advanced practice nurses and physician assistants [49].

Interprofessional Team Functioning was examined in four reviews [32, 44, 52, 69]. Providers expressed positive perceptions of high functioning interprofessional teams [69] and improved communication (McParland, 2022) (no p value reported) [52]. In their review on long-term care, Lovink et al. stated that no provider outcomes were identified [32]. In mental health, NP-led collaborative care led to significant improvements in patients’ depressive symptoms [44]. In addition, a collaborative practice model between NPs and NP specialized in mental health increased the number of case reviews per month from 5 to 15, and reduced the number of referrals to the specialized NP in mental health from 19 to 5 (no p value reported) [44].

Prescribing included 15 reviews [10–12, 28, 29, 32, 34, 43, 44, 48, 56, 59, 60, 68, 69]. Significant results favouring NPs were found for the initiation, up-titration and target doses of beta-blockers (p < 0.001) [28], as well as for the prescription of cholinesterase inhibitors and antidepressants [34]. Prescription of anti-microbial agents favoured NPs in one review (no p value reported) [48]. No significant differences in prescribing were noted in six reviews [11, 12, 29, 32, 59, 68]. A meta-analysis comprising 26 studies by Zhang et al. [43] examined awareness of pre-exposure prophylaxis (PrEP) implementation cascade by NPs and physicians and found lower scores in the NP group (p < 0.05), no difference in willingness to prescribe PrEP, and higher odds of prescribing PrEP for NPs (p < 0.05). Equal to statistically significant results favouring the NP group were noted in the appropriateness of medication prescribing in two reviews [10, 60], and prescribing diagnostic tests [12]. Under-prescribing of weight-loss pharmaceuticals was noted by Hyer [69]. Mixed results were noted in one review [44] where NPs prescribed equal to significantly more anxiolytics (p ˂ 0.001) and significantly more anti-depressants. No differences were noted in the odds of being prescribed pharmacothrapy for alcohol use disorder [44].

Health system indicator categories

Access to care was found in five reviews [47, 58, 59, 65, 70]. Improved access to care was noted using open access scheduling for the NP group (no p value reported) [65]. The addition of pediatric NPs increased the proportion of urgent appointments (p < 0.001) and access between visits (no p value reported) [59]. The implementation of an NP-led surveillance clinic increased access to treatment for precancerous lesions with an increase of 18.3% (n = 828) and 6.2% were treated for non melanoma skin cancer [70]. In their respective reviews, Yang [58] reported equal to significant improvements in primary care access (no p values reported), and Milesky [47] noted that the theme of increased access to health care occurred in 10.3% of theme occurrences.

Consultations were identified in 11 studies [11, 12, 27, 29, 31, 32, 42, 56, 58–60]. Results included significant reductions in the number and duration of consultation calls in rehabilitation (p = 0.05), and the total number and duration of consultations calls (p < 0.05) [27] for the NP group. Significantly higher number of visits and referrals to physicians were identified when there were restrictions on scope of practice [56, 58]. Referrals for echocardiographs among patients with presumed congestive heart failure were signifiantly higher in the NP group (p < 0.001). Consultation times were longer for NPs in several reviews [11, 56, 60]. A meta-analysis that included 2500 patients estimated an increase of 4.1 minutes in the intervention group (p < 0.0001) [11]. For patients with chronic conditions (diabetes and hypertension), one review [60] reported an average increase of 11 minutes in consultation time (p <0.001). No comparison in consultation times was presented in one review but consultation times decreased between the first and third visit [29], and one review did not report measures of consultation times [59]. No significant between-group differences were noted in the number of referrals in three reviews [11, 12, 42]. Unnecessary referrals to dermatologists and biopsies decreased [31] (no p value reported), and unplanned consultations for acute conditions in long-term care increased significantly in the NP group (p < 0.0001) [32].

Costs were reported in 17 studies [11, 12, 27, 29, 30, 32, 34, 41, 46, 47, 52–54, 56, 58, 59, 67]. Reductions in costs were noted in five studies for the NP group [11, 32, 34, 46, 47]. Martin-Misener [11] also completed a meta-analysis of two studies that reported costs (n = 2689 patients) and found a mean difference of -6.41 euros (2006 euros). No significant differences in costs were noted in five reviews [27, 52, 54, 56, 67]. In addition, Fraser et al. [46] noted that broader scope of practice was projected to have an important impact on state-level economic activity. Similarly, Yang and colleagues [58] found that fewer restrictions on scope of practice was associated with lower costs, and increased hourly earnings of NPs was associated with a decrease in physician earning between 2005–2010. Equal to lower costs were observed in three reviews for outpatient visits [29], laboratory services [12], and home-based primary care [53]. Costs were higher for NPs vs. physicians in one review of patients with lung disease (p < .001) [59]. Mixed results were noted in two reviews with both increased and decreased costs for patients requiring long-term care and end of life care [30] and a wide range of chronic conditions [41].

Emergency Room Visits were included in 14 studies [11, 28, 30, 32, 35–37, 42, 45, 53, 55, 60, 63, 66, 67]. Reductions in Emergency Room visits were noted in half of the reviews [30, 32, 37, 45, 60, 66, 67]. Statistical significance not always specified. Increases in the number of Emergency Room visits were identified in two reviews [32, 55], and the number of Emergency Room visits remained unchanged in studies reported in three reviews [11, 28, 35, 36]. One review reported equal to statistically significant reductions in Emergency Room visits [53]. Jeyaraman et al. noted an increase in the number of Emergency Room visits in the NP-led care model to improve patient flow through the Emergency Room compared with the control group (no p value reported) [45]. No data were available on non-conveyance rates (ambulance transport to the Emergency Room) for NPs [42].

Healthcare Service Delivery was outlined in 17 reviews [11, 12, 27, 30, 32, 34, 37, 41–43, 55, 56, 58–60, 65, 67], and was examined using: no show rates, continuity of care, additional interventions, transitional model of care for patients with schizophrenia, screening women with depression, home care, number of primary care visits, acceptability of nurse-led care, models of care, end-of-life care, hospital days, and number of outpatients contacts. Significant results favouring the NP group were noted in five reviews [30, 37, 42, 55, 59], and equal to superior care was noted in another eight reviews [12, 32, 34, 41, 56, 58, 65, 67]. Zhang et al. [43] described key barriers and ideal location to providing PrEP. A meta-analysis of three studies (n = 2562 patients) led by Martin-Misener [11] found that NPs were more likely to ask patients to return than physicians within two weeks of the index visit (p < 0.001), but return visits for any reason were equal at 1 year.

Hospitalization were included in 15 reviews [10, 11, 27–30, 32, 34–37, 47, 53, 57, 59, 60]. Reviews measured unplanned hospital transfers, hospital admission rates for patients with lung disease, rehospitalizations, index re-hospitalization up to 42 days, over 90 days and 180 days, any re-hospitalization up to 180 days, hospitalizations for congestive heart failure, hospital admissions, hospital days, hospital admission in long-term care, number of hospital days in long-term care, number of hospital admissions from primary care, at least one hospitalization, institutionalization for residential care, readmissions, 30-day readmission rates for heart failure. No significant differences were noted in five reviews [11, 28, 29, 37, 57]. No differences were found following a meta-analysis of index rehospitalizations up to 42 days and any rehospitalizations up to 180 days [27]. Equal to statistically significant outcomes favouring NPs were noted in nine reviews [10, 27, 30, 32, 34–37, 53, 60]. Reductions in nursing home admissions were noted in intervention groups in one review but no p value was reported [53]. One study included in one review found a significant reduction for hospital admissions rates for patients with lung disease in favour of the control group (66 patients in the NP group vs. 42 in the physician group, relative rate: 1.52, p = 0.03) [59].

Length of Stay was identified in seven studies [30, 34–36, 45, 60, 63, 64], and was assessed using time spent in the Emergency Room and for patients who were admitted to the hospital. Equal to statistically significant reductions in ED length of stay noted in one review (median: -28.50 minutes) [45]. Trends towards a decrease in the time to triage and a greater number of patient discharges from the ED at 60 minutes, 90 minutes, 4 hours and under 6 hours favour the NP group but level of significance was not indicated [45]. No significant between-group differences were noted in three reviews [35, 36, 60, 63]. Decreased length of stay was noted in the reviews led by Leduc [30] (no p value), and by Morilla-Herrera [34] (p < 0.001) in long-term care. Galiana-Camacho et al. [64] did not provide a comparison of the average length of stay, but the authors noted that 78.5% of patients were seen in less than four hours.

Patient Safety was reported in 11 reviews [27, 28, 31, 34, 52, 56, 58–60, 64, 67], and was examined using: treatment complications, adverse effects, falls, adverse events, unplanned readmissions, adequacy of diagnostic tests, medication interactions, unplanned visits to family physicians, hospitalization, risk factor management, inappropriate management, missed injuries, malpractice, unnecessary referrals, and unnecessary biopsies. Five reviews [27, 31, 34, 60, 64] favoured the NP group and five reviews [28, 52, 56, 59, 67] showed no significant differences between the intervention and control groups. Further, Yang and colleagues [58] highlighted that 31% lower malpractice payment occurred over a 13-year span per 1,000 physicians for states with full NP practice authority compared to those with a restricted practice authority.

Quality of Care was examined in six studies [29, 32, 41, 47, 58, 67]. Significant improvements in quality of care were noted for patients receiving medical care from Emergency Room NPs (p <0.02), asthma-specific care for children (p < 0.05), vulnerable elders (p < .001), and positive health outcomes [58]. Equal to statistically significant improvements were found in the NP group by Kuethe et al. [29] and Yang [58]. Improvements were highlighted in the Tsiachristas [41] review but no p value was reported. The Milesky review [47] indicated that NPs improved quality in almost one fifth of the theme occurrences identified in their sample.

Scope of Practice was identified in six reviews [38, 47, 58, 63, 64, 69]. Milesky and colleagues [47] reported that unrestricted or least restrictive scope of practice was mentioned in 7.35% of theme occurences in their sample. Positive trends and positive impacts reaching statistical significance were noted when less restrictive scope of practice regulations, as well as increased or full practice autonomy, were found in the reviews.

Wait Times were reported in seven studies [31, 42, 45, 59, 63, 65, 67]. Four studies reported reductions in wait times for appointments [45, 59, 63, 65] and a fifth [31] reported an increased number of patients seen in a timely manner with the NP group. Equal to statistically significant reductions in wait times in the Emergency Room were found for the NP group in two reviews [63, 67]. No comparison for on-scene treatment time were available [42]. Equal to statistically significant reductions in time to initial provider assessment (mean time difference: -3.00 min, 95%CI [-3.47, -2.43] to -50 min, 95%CI [-53.63, -46.37]) noted in one review [45]. Equal to statistically significant reductions in the number of patients who left the Emergency Room without being seen in one review (no p value reported) [45]. The number of patient who left against medical advice was unchanged in one review (no p value reported) [45].