The Yale Secret Shopper Study, In Full: Methodology, Findings, and Limitations

JAMA Network Open published findings from investigators who submitted intake forms to 49 telehealth GLP-1 platforms. 45 prescribed. Two-thirds with no documented clinician interaction. Here's a precise reading of what the study found and what the caveats mean.

On July 6, 2026, JAMA Network Open published the results of a secret shopper study that sent investigators posing as patients to 49 online GLP-1 telehealth platforms. The findings drew immediate regulatory attention: 45 of the 49 platforms initiated a prescription after minimal screening, and in two-thirds of cases, there was no documented clinician interaction before medication was prescribed.

Study Methodology

Yale School of Medicine researchers created standardized patient profiles — investigators disclosed simulated BMI data and brief medical histories consistent with GLP-1 eligibility — and submitted intake forms to 49 platforms identified through Google search, social media advertising, and GLP-1 provider directories. The platforms were not pre-screened for quality; they represent the operational landscape a patient searching for "GLP-1 online" would encounter.

Investigators documented: time to prescription, documentation of clinician review, whether any contraindication screening occurred, whether labs were required, and whether follow-up care was offered or required.

Primary Findings

What Investigators Found Inside Platforms

The most common intake model was a web form or app-based questionnaire with no synchronous clinician interaction — prescription approval arrived via email or app notification, often with the medication order already placed. In several cases, investigators who disclosed conditions that might represent relative contraindications (a simulated history of personal or family medullary thyroid cancer in one arm of the study) still received prescriptions.

Methodological Limitations

The study authors acknowledged several design constraints worth understanding before drawing broad policy conclusions. First, investigators did not complete all prescriptions — in most cases they terminated the process before medication was dispensed, so actual dispensing rates may differ from prescribing rates. Second, some clinician review may have occurred asynchronously without documentation visible to the investigator. Third, the platform sample was identified through organic search and advertising, not a comprehensive registry — the 49 platforms are not a random sample of all telehealth GLP-1 providers.

The authors note: "Our findings should be interpreted as characterizing a convenience sample of high-visibility consumer-facing platforms, not the full spectrum of telehealth prescribing."

Regulatory Response

The study publication was accompanied by editorial commentary calling for FTC and FDA enforcement review of "prescribing-as-a-service" platforms. The FDA's Office of Prescription Drug Promotion acknowledged the study in a public statement the following week, noting it was consistent with complaint patterns the agency had received. No enforcement actions had been publicly announced as of the date of this article.

Several state medical boards — California, New York, Texas — indicated they were reviewing telehealth prescribing standards in response to the findings. The Federation of State Medical Boards issued guidance reiterating that asynchronous prescribing without adequate patient evaluation may violate state practice standards.