Referrals are a power hurdle; who decides which doctor you can see?

After being rejected by the first clinic, Nora had a referral document in her hand.
The document was marked "Urgent." It was issued by her family doctor, stating that her condition required immediate intervention by a gynecologist. It was legal, correctly formatted, and clearly written.
It solved nothing.
The existence of a referral document does not equate to a referral actually taking place. There is a gap between the document and the actual medical visit. Whether that gap is opened depends not entirely on the patient's needs, nor entirely on the doctor's judgment, but rather on a more complex set of institutional arrangements.
This arrangement is what we will be dissecting in this article: the referral mechanism and the power logic behind it.

(Image caption) The existence of a referral document does not equate to a referral actually taking place. There remains an institutional hurdle between the document and the actual medical treatment.

I. The Surface Function and Real Function of Referral In medical education, referral is described as a collaborative mechanism: when a general practitioner determines that a patient needs a specialist evaluation, he or she refers the patient to a specialist with the appropriate capabilities to ensure that the patient receives the most appropriate treatment.
The description is correct, but it is incomplete.
Referrals also serve as a cost control mechanism. In most gatekeeper-designed healthcare systems, patients are required to be referred by a general practitioner before receiving specialist services. One of the core purposes of this is to prevent patients from directly flooding into the more expensive specialist and hospital systems. This is an efficiency logic, not a patient service logic.
Referrals also serve as an information filtering mechanism. Ideally, general practitioners, acting as gatekeepers, should be able to distinguish which symptoms require specialist evaluation and which can be addressed at the primary care level. However, the quality of this filtering heavily depends on the general practitioner's competence, time, and available information, and these three factors vary significantly across different regions, clinics, and physicians.
Referral is also a power distribution mechanism. Who is eligible for referral, where to refer, and within what time frame—these decisions constitute a set of control logics about who can move within the healthcare system, at what speed, and to where.
When we combine these three functions, referral becomes more than just a medical procedure; it becomes an institutionalized power checkpoint.

II. The United States: The Design Dilemma of a Market-Based Gatekeeper The United States’ healthcare system has created a complex and unique structure in its referral design.
Health maintenance organizations (HMOs) require patients to select a primary care physician (PCP), and all specialist visits must first be authorized by a PCP's referral. Unauthorized specialist visits are typically not covered by insurance or have very low coverage. This is a typical gatekeeper design: the patient's movement is strictly controlled within the PCP's judgment.
Preferred Provider Organizations (PPOs) offer greater flexibility: patients can see in-network specialists directly without needing referral authorization, although seeing out-of-network doctors is more expensive. This design gives patients more autonomy in their treatment path, but at the cost of higher premiums.
The problem with this design is that it shifts the determinant of healthcare accessibility from "what patients need" to "what insurance patients can afford." A patient with HMO insurance has more stringent referral controls than a patient with PPO insurance, and the root of this difference is premiums, not medical needs.
Nora's case presents another layer of complexity: even with referral documents, her experience at the first clinic demonstrates that referral documents alone cannot guarantee the quality of a referral. Documents are procedural, while the quality of care is substantive, and there is virtually no institutional mechanism to bridge the gap between the two.

(Image caption) In most healthcare systems, general practitioners are not only providers of medical care, but also "gatekeepers" for access to specialist medical resources.

A deeper issue is prior authorization. In the United States, many insurance companies require physicians to obtain prior authorization from the company before prescribing certain tests, medications, or treatments. While ostensibly designed to prevent unnecessary medical expenses, this mechanism, in practice, gives insurance companies substantial influence over medical decisions.
According to a survey by the American Medical Association (AMA), American physicians spend an average of more than 14 hours per week on administrative work related to pre-authorization, and about one-third of physicians said that delays in pre-authorization have led to serious health consequences for patients.
This is a question of sovereignty in medical decision-making: when an insurance company's administrative procedures can delay or overturn a doctor's clinical judgment, to whom does medical sovereignty belong at that specific moment?

III. The UK NHS: Waiting Time as an Implicit Gatekeeper Mechanism. The UK NHS referral system is built upon the general practitioner (GP) system: each NHS patient is registered under a GP, and specialist referrals typically require going through a GP. The advantage of this design is its systemic nature: it establishes a clear primary healthcare layer, theoretically allowing problems to be resolved at the grassroots level in many cases, reducing unnecessary consumption of specialist resources.
However, the NHS gatekeeping mechanism has a significant problem in practice: waiting time.
The NHS mandates that referred patients receive specialist care within 18 weeks. However, this goal has been consistently missed in recent years, with actual wait times for some specialties far exceeding this standard. In some regions, wait times for certain specialties can exceed one year.
Waiting time is an implicit gatekeeping mechanism. Unlike the explicit approval process required by HMOs, it effectively limits patients' access to certain healthcare resources through time costs.
For a patient with acute symptoms—like Nora—waiting time can be a healthcare safety issue. For a patient with chronic symptoms, a long wait could worsen the problem while waiting, ultimately increasing the final healthcare costs.

(Image caption) The pathways in the healthcare system are not entirely determined by the patient, but rather shaped by the system, insurance, resources, and time. Controlling the pathways is essentially a matter of power distribution.

IV. Germany: The Illusion and Reality of Free Choice Germany's healthcare system is relatively flexible in its referral design: patients with statutory health insurance (GKV) can see their family doctor, pediatrician, or gynecologist directly without a referral. For other specialties, theoretically, direct consultation is also possible, although some insurance plans encourage referrals through family doctors to obtain lower out-of-pocket expenses.
This design superficially gives patients greater autonomy in their treatment pathways. However, it faces the problem that when everyone can directly see a specialist, specialists will face considerable scheduling pressure.
In Germany, there is a significant difference in waiting times between patients with private insurance and those with statutory insurance. Patients with private insurance typically get specialist appointments much faster, while those with statutory insurance sometimes have to wait for months.
This is a hidden dual-track system: on the surface, it is the same free choice, but in reality, the ability to pay determines the waiting time, and the waiting time determines the actual accessibility.

V. China: The pull of top-tier hospitals and the hollowing out of primary healthcare. China's healthcare system faces a structural dilemma in its referral design: patients are highly concentrated in top-tier hospitals, while primary healthcare institutions generally lack trust and use.
There are deep-seated reasons for this phenomenon. For a long time, primary healthcare institutions have been relatively disadvantaged in terms of resource allocation, physician training, and salary incentives. Patients' choice to go directly to tertiary hospitals is a rational response to the current distribution of resources.
In recent years, China has implemented a tiered healthcare system reform, attempting to guide patients to receive treatment at the primary care level first, and then transfer to higher-level hospitals through a referral mechanism. However, this reform faces the challenge that, before the actual capacity of primary healthcare institutions is significantly improved, mandatory referrals only increase procedural friction for patients without improving the quality of care they ultimately receive.
China's referral difficulties illustrate a more general problem: the design of a tiered medical system must be based on the actual development of medical capabilities at each level; otherwise, it is merely an administrative procedure, not a mechanism for improving healthcare.

(Image caption) When insurance companies' pre-authorization mechanisms intervene in the medical process, some of the decision-making power of clinical judgment is transferred to the administrative and payment systems.

VI. Singapore and Japan: Relatively Mature Examples of Tiered Healthcare Systems Singapore's tiered healthcare system is relatively mature within the Asian context. Patients initially consult with community clinics (GPs or multiple clinics), and are referred to general hospitals or specialist centers by their GPs when specialist evaluation is required. This design operates relatively smoothly overall, but the cost structure—the cost gap between the public and private systems—still creates a stratification effect in resource accessibility.
Japan's referral system allows patients to directly seek treatment at large hospitals, but in practice, direct treatment at large hospitals usually requires paying an additional "no referral fee," which essentially creates a financial gatekeeping mechanism.
Both cases illustrate that even in a relatively well-designed system, the referral mechanism is still inevitably intertwined with cost structure, waiting time, and resource distribution, resulting in varying levels of actual accessibility for different patients.

VII. The essence of referral: Control of pathways is the allocation of sovereignty. Putting these comparisons together, a conclusion becomes clear:
Referrals are never just a medical procedure. They are a control mechanism about who can move within the healthcare system, at what speed, and to where. The design of this mechanism determines how healthcare resources are actually allocated among different population groups.

(Image caption) Waiting time is not a neutral cost, but rather a hidden filtering mechanism. The longer the wait, the lower the actual accessibility.

For a patient with sufficient knowledge, time, and financial resources, referral mechanisms are usually manageable. However, for a patient with language barriers, insufficient information, time constraints, or limited financial resources, referral mechanisms can become a real barrier.
Nora's experience illustrates that even with referral documents, even if those documents are marked "Urgent," you can still be turned away. The guard at that door isn't a rule, but a person not adequately bound by rules, in a scenario without an immediate error-correction mechanism.
This is the crux of the referral problem: it's not just about efficiency or cost, but a sovereign issue about who has the power to decide on healthcare pathways.
When this decision-making power is not in the hands of the patient, when it is dispersed among the judgment of general practitioners, the prior authorization of insurance companies, the administrative logic of waiting lists, and the service willingness of institutions, medical sovereignty has already been partially transferred.
Understanding this shift is one of the most important prerequisites for redesigning the healthcare system.
In the next article, we will shift our focus from institutional design to industrial structure, and examine how AI, insurance capital, pharmaceutical companies, and platform-based healthcare are redistributing control over healthcare.

Next article: When Healthcare Meets Capital

This comparative study is based on publicly available healthcare policy documents and academic research from various countries, including the American Medical Association's pre-authorization survey report, UK NHS performance statistics, and publicly available data from various national healthcare systems. It does not constitute any medical or legal advice.