When a self-management experiment goes wrong in software, a release slips. When it goes wrong in a hospital, someone can die. That is why nursing is the toughest test of the claim that teams run better without a layer of managers above them, and why its evidence deserves to be read carefully rather than cheered.
An organization of nurses, almost no managers
The reference case is Buurtzorg, the Dutch home-care organization. Frank Martela and Sharda Nandram (2024) document that it operated more than 900 self-managing teams of up to 12 nurses each, around 14,700 staff as of January 2021, with just two directors, roughly 50 back-office people, and no middle managers or supervisors. Coordination across some 950 teams ran through about 21 internal coaches, who held no formal authority and no responsibility for team productivity. A KPMG study around 2015 found administrative overhead of about 8 percent against a Dutch home-care sector average near 25 percent. The authors are careful about boundary conditions: the model works best where tasks have low interdependence, which is a limit, not a slogan.
The outcome data points the same way, with the caveats attached. A scoping review by de Bruin and colleagues (2022), covering 27 studies with Buurtzorg the most examined, reported that job satisfaction was high because of greater autonomy and closer patient relationships, that turnover fell by roughly half, and that continuity of care improved with small stable teams, with patient satisfaction reaching around 98 percent in one Scottish study. The same review is honest about the ledger. Administrative time ran about 30 percent lower, yet the cost per hour of care was higher, near 54.74 euros against 48.74, and most of the underlying studies were qualitative. The cost case for the wider health system rests on a KPMG analysis reporting up to 50 percent fewer care hours per patient and roughly 40 percent savings, savings that were not explained by an easier patient mix (Commonwealth Fund, 2015).
The mechanism generalizes beyond one Dutch pioneer
Skeptics can dismiss a single admired case. They have a harder time with the hospital-wide data. McHugh and colleagues (2013) compared 56 Magnet hospitals, recognized for strong nursing work environments and high clinical autonomy, with 508 non-Magnet hospitals across more than 56,000 surgical discharges. Patients in Magnet hospitals had 14 percent lower odds of 30-day mortality and 12 percent lower odds of failure-to-rescue. Nurse autonomy was part of what distinguished those environments, though the authors are clear it explains part of the advantage, not all of it.
The meta-analytic evidence closes the loop. Reviewing decades of studies, Ryu, Neubert and Gonzalez-Mule (2022) found team autonomy positively related to team effectiveness overall, while mapping the moderators that decide when the effect is strongest. Stewart (2006), synthesizing 93 studies of team design, likewise found greater team autonomy associated with higher performance, with the size of the gain depending on the type of task. Autonomy is a real lever, not a universal guarantee, which is precisely the shape of the nursing evidence.
Autonomy is not the same as absence
The caveats are not fine print, they are the design brief. Self-managing teams do not perform because a manager was subtracted. They perform when the inputs are present. Reviews of self-managing teams stress that results depend on the right composition, task design, and organizational support, and that members need genuine competence to run without a boss (Magpili and Pazos, 2018; Doblinger, 2022). Buurtzorg supplies exactly that scaffolding: small teams, a digital platform, and coaches who develop rather than command. Strip the scaffolding and keep only the slogan of no managers, and the outcomes reverse.
The practical read
The self-managing nurse team is not a feel-good story. It is a claim with mortality figures, retention numbers, and a cost ledger that includes an inconvenient line: a higher hourly rate offset by far fewer hours and much lower overhead (de Bruin et al., 2022; Commonwealth Fund, 2015). Read with its caveats intact, the evidence still lands on the same side. Give expert teams real autonomy, plus the composition, competence, and support that autonomy requires, and the outcomes that matter, mortality, continuity, retention, tend to improve (McHugh et al., 2013; Stewart, 2006; Ryu, Neubert and Gonzalez-Mule, 2022). The lesson for any leader is not to copy Buurtzorg's chart. It is to notice that in the setting with the highest stakes of all, fewer managers and more autonomy was not the reckless option. It was the one the numbers supported.
Sources
- Martela and Nandram (2024). Buurtzorg: scaling up an organization with hundreds of self-managing teams but no middle managers. Documents 900-plus nurse teams and about 14,700 staff run with two directors and no middle managers, at roughly 8 percent overhead. https://link.springer.com/article/10.1007/s41469-024-00184-y
- de Bruin, Doodkorte, Sinervo and Clemens (2022). The implementation and outcomes of self-managing teams in elderly care: A scoping review. Reports higher satisfaction, turnover cut by half, and lower admin time, while flagging higher cost per hour and mostly qualitative evidence. https://onlinelibrary.wiley.com/doi/abs/10.1111/jonm.13836
- Commonwealth Fund (2015). Home Care by Self-Governing Nursing Teams: The Netherlands' Buurtzorg Model. Summarizes a KPMG study finding up to 50 percent fewer care hours and roughly 40 percent system savings, not explained by patient mix. https://www.commonwealthfund.org/publications/case-study/2015/may/home-care-self-governing-nursing-teams-netherlands-buurtzorg-model
- McHugh, Kelly, Smith, Wu, Vanak and Aiken (2013). Lower Mortality in Magnet Hospitals. Magnet hospitals, marked by high nurse autonomy, had 14 percent lower odds of 30-day mortality across 56,000-plus discharges. https://pubmed.ncbi.nlm.nih.gov/23047129/
- Ryu, Neubert and Gonzalez-Mule (2022). Putting the team in the driver's seat: a meta-analysis on team autonomy. Finds team autonomy positively related to effectiveness, with moderators that set when the effect is strongest. https://onlinelibrary.wiley.com/doi/10.1111/peps.12468
- Stewart (2006). A Meta-Analytic Review of Relationships Between Team Design Features and Team Performance. Across 93 studies, greater team autonomy is associated with higher performance, varying by task. https://journals.sagepub.com/doi/10.1177/0149206305277792
- Magpili and Pazos (2018). Self-Managing Team Performance: A Systematic Review of Multilevel Input Factors. Self-managing team results depend on composition, task design, and organizational support. https://doi.org/10.1177/1046496417710500