Impact of payment model on the behaviour of specialist physicians: A systematic review

Author: Quinn AE, Trachtenberg AJ, McBrien KA, Ogundeji Y, Souri S, Manns L, Rennert-May E, Ronksley P, Au F, Arora N, Hemmelgarn B, Tonelli M, Manns BJ.

Geographical coverage: USA and Canada

Sector: Impact/Economic Evaluation

Sub–sector: Costs

Equity focus: Not reported

Study population: Specialist physicians and their patients

Review type: Effectiveness review

Quantitative synthesis method: Narrative synthesis

Qualitative synthesis method: Not applicable

Background: Physician payment models are key levers for shaping healthcare delivery; they aim to improve quality, access, efficiency and value. Although their effects are well documented in primary care, much less is known about how specialists respond. Fee‑for‑service (FFS) remains the dominant model worldwide and incentivises high service volumes, whereas salary and capitation may encourage more coordinated and cost‑conscious care. Because specialists differ from generalists in patient complexity, referral pathways and procedural mix, their behavioural response to financial incentives may also differ. Emerging evidence therefore suggests that specialists do not always react to payment reforms as economic theory would predict.

Objective: To synthesise evidence on how specialist physician payment models affect healthcare quality, clinical outcomes, service utilisation, access, costs, and patient as well as physician satisfaction.

Main findings: Eleven studies examining seven specialist payment reforms—mostly from the USA and Canada—met the inclusion criteria. Sample sizes ranged from 290 surgical cases to 343,844 patients. Eight studies used a controlled before‑after design and three an interrupted time‑series design; the latter generally had a lower risk of bias.

• FFS models were associated with greater service utilisation, for example higher outpatient‑visit rates for haemodialysis patients, and improved access to emergency care.
• Moving from FFS to capitation or salary reduced certain elective procedures (e.g. cataract surgery and tubal ligation), while outpatient‑visit volumes often remained unchanged. Episode‑based payment achieved cost savings mainly by shifting the location of care rather than reducing procedure numbers.
• Blended models (capitation plus FFS) achieved some intended goals—such as increasing specialist visits for dialysis patients—but also led to unintended consequences, including lower uptake of preferred home‑dialysis modalities.
• Salary‑based reforms produced mixed results: they reduced elective‑surgery volumes but did not consistently lower overall utilisation and, in some instances, increased physician income.

Overall, payment models influenced specialist behaviour, but the magnitude of change was modest and highly context dependent.

Methodology:  MEDLINE, Embase, the Cochrane Central Register of Controlled Trials, CDSR, DARE, NHS EED, EconLit and Web of Science were searched from inception to October 2018. Eligible study designs included randomised controlled trials, controlled clinical trials, controlled before‑after studies and interrupted time‑series analyses evaluating specialist payment reforms. Only English‑language publications were considered. Two reviewers independently screened records, resolved disagreements by consensus, extracted data and assessed risk of bias using the Cochrane Effective Practice and Organisation of Care criteria. Findings were synthesised narratively.

Applicability/external validity: As all included studies were conducted in the USA or Canada, applicability to other health systems is uncertain. Methodological heterogeneity, speciality differences and reliance on non‑randomised designs further limit external validity. Robust, context‑specific studies are needed before firm policy recommendations can be made.

Geographic focus: No geographical limits were specified, but every included study originated from North America (USA and Canada).

Summary of quality assessment: Confidence in the findings is medium. Searches were comprehensive, inclusion criteria explicit, dual‑reviewer screening employed and study quality appraised with validated tools. Nevertheless, the language restriction, absence of an exclusion list and residual bias in the primary studies temper the conclusions.