Outcomes Research in Review

Effect of Substituting Nurses for Doctors in Primary Care

Martínez-González NA, Djalali S, Tandjung R, et al. Substitution of physicians by nurses in primary care: a systematic review and meta-analysis. BMC Health Serv Res 2014;14:214.



Study Overview

Objective. To investigate the clinical effectiveness and costs of nurses working as substitutes for physicians in primary care.

Design. Systematic review and meta-analysis of published randomized controlled trials (RCTs) and 2 economic studies that compared nurse-led care with care by primary care physicians on numerous variables, including satisfaction, hospital admission, mortality, and costs of health care.

Settings and participants . The 24 RCTs were drawn from 5 different countries (UK, Netherlands, USA, Russia, and South Africa). In total, there were 38, 974 participants. Eleven of the studies had less than 200 participants and 13 studies had more than 200 (median, 1624). Mean age was reported in 20 trials and ranged from 10 to 83 years.

Analysis. The authors assessed risk of bias in the studies, calculated the study-specific and pooled relative risks (RR) or standardized mean differences (SMD), and performed fixed-effects meta-analyses.

Main results. Nurse-led care was effective at reducing the overall risk of hospital admission (RR 0.76, 95% CI 0.64–0.91) and mortality (RR 0.89, 95% CI 0.84–0.96) in RCTs of ongoing or non-urgent care, longer (at least 12 months) follow-up episodes, and in larger ( n > 200) RCTs. Pooled analysis showed higher overall scores of patient satisfaction with nurse led care (SMD 0.18, 95% Cl 0.13–0.23). Higher-quality RCTs (with better allocation concealment and less attrition) showed higher rates of hospital admissions and mortality with nurse-led care, but the difference was not significant. Subgroup analysis showed that RNs had a stronger effect than nurse practitioners (NPs) on patient satisfaction. The results of cost-effectiveness and improved quality of care analysis with nurses were inconclusive.

Conclusion. Nurse-led care appears to have a positive effect on patient care and outcomes but more rigorous research is needed to confirm these findings.


As the backbone of health care systems around the world, primary care is facing numerous challenges threatening patient access to care. Aging populations, economically strapped governments, and an increasing non-communicable disease burden in developing countries are pushing global health systems to their capacity. In addition, the World Health Organization has highlighted the increasing health worker shortage which further limits the capabilities of health systems [1,2]. One proposed solution to addressing physician shortages is using NPs. Recent studies have shown patient satisfaction, physical, emotional, and social function, and other outcomes associated with nurse-led care to be similar to if not better than those associated achieved by physicians [3–5].

The current meta-analysis has some weaknesses. For example, 13 of the 24 studies had attrition rates of at least 20% and only 10 trials had a sufficient sample size to achieve adequate power in at least 1 outcome, making it more difficult to identify true differences between control and intervention groups. The sample of RCTs were heterogeneous in terms of settings, tasks, and reporting of outcomes. Also, study heterogeneity increased the difficulty of data synthesis and limited the amount of information on cost-effective nursing care and quality of care of patients.

In many of the studies, quality of life among patients was measured inconsistently, using various disease specific and generic scales, making it difficult to compare and provide comprehensive results. Additionally, less than 50% of the patient satisfaction scales used validated questionnaires.

Results should be interpreted with caution as the studies were compiled from 5 different countries. The scope of nursing practice differs in each country and the different cadres of nurses (RN vs NP vs licensed practical nurse [LPN]) also have varying responsibilities. Cross comparisons between RN/LPN, NP/physician, and RN/NP need to consider the country context, regulating bodies, and government policies that dictate the capabilities and practice of each of these licensed professionals.

There was a dearth of economic information. Generally, direct costs such as consultations and cases involving patients less than 65 year of age were lower with nurse-led care, but in other studies costs of nurse-led and physician-led care were not significantly different.

Applications for Clinical Practice

As the health worker shortage continues, health care facilities will have to decide on the appropriate skill mix to provide the best patient outcomes while maximizing cost benefit. While this systematic review and meta-analysis is promising in supporting nursing-led primary care, more research is needed, including longer-term studies with larger sample sizes and more extensive assessment of cost and quality of life. The use of validated and standardized instruments to measure patient satisfaction and quality of care will increase study quality and rigor.

—Melissa T. Martelly, MA, BSN, RN, PCCN, and Allison Squires, PhD, New York University College of Nursing

Next Article: