What "Hypothesis-Generating" Means: How to Read a Small Study

Last updated: September 17, 2026

Written by: Jay Hastings, CEO of PlexusDx, and Jayden Lee, PharmD, EMBA, PlexusDx Medical Director

Jay Hastings is the CEO of PlexusDx, a precision health company focused on genetic testing, blood biomarker insights, and personalized wellness recommendations. He has more than 20 years of experience across healthcare innovation, genomics, laboratory operations, healthcare investing, and strategic finance.

Medically reviewed by: Jayden Lee, PharmD, EMBA

Jayden Lee, PharmD, EMBA, is the PlexusDx Medical Director, with a PharmD and MBA specializing in pharmacogenomics and clinical product development, with a proven ability to bridge the gap between genomic research and practical patient outcomes. Dr. Lee has more than 10 years of professional experience in clinical pharmacy, academia, and research.

This article is part of the PlexusDx Education Hub — your resource for evidence-based guidance on Peptides & GLP-1. Browse all Peptides & GLP-1 education

When scientists call a result “hypothesis-generating,” they mean this: the result is interesting enough to test properly, and not strong enough to act on by itself. The PlexusDx microdose tirzepatide study uses that phrase about itself on purpose. This article explains the six ideas behind it, each in two plain sentences, and then shows how each one applies to our study. Microdosing is not a medical term. In the PlexusDx study it meant a total weekly dose of 1 or 2 mg of injectable tirzepatide, below the usual 2.5 mg starting dose.

Sample size (the N)

N is how many people a number is based on. A number from 8 people wobbles more than a number from 800.

In our study, N changes by window: 57, then 19, then 8. When you see a percentage, look for its N right next to it.

Time window

A window is the stretch of days a measurement had to fall inside to count. A weight from day 20 and a weight from day 70 tell you different things, so they go in different windows.

Our windows were 14–35, 36–59 and 60–89 days after the first dose. No one reached 90 days, so nothing is claimed past that.

Comparison group

A comparison group is a set of similar people who did not get the treatment. Without one, you cannot tell how much change would have happened anyway.

Our study had no comparison group. So it can describe what happened, but it cannot say the medicine caused it.

Self-report

Self-report means the patient told us the number instead of a clinic measuring it. People are mostly honest, but home scales differ, and memory is imperfect.

Every weight in our study was self-reported at a routine check-in. That is a real limit, and the study says so.

Completer bias

Completer bias happens when the people who stick around are different from the people who leave. If those who felt it was not working quit early, the later numbers look better than the truth.

Our later windows have far fewer people: 19 and then 8. The study did not fill in the missing ones, and it warns that the later windows are not “the same group getting better.”

Confidence interval

A confidence interval is the range of values that would fit the data reasonably well. A wide interval means less certainty; a narrow one means more.

Every mean in our study comes with a 95% confidence interval. When the interval is wide, treat the mean as a rough estimate, not a fact.

Putting it together

Here is the authors’ own conclusion: “Weight loss was observed within the 14–35-, 36–59-, and 60–89-day observation windows among evaluable adults initiating compounded injectable tirzepatide at 1–2 mg/week. These hypothesis-generating observations support prospective comparative study, not conclusions about causality, equivalent dosing, or long-term safety.” Read that last sentence slowly. It does not say the medicine works at a low dose. It says something was observed, and that it is worth testing the right way. That is what “hypothesis-generating” means.

How to use a small study well

Use it to ask better questions. Bring it to your prescriber and ask what they make of it. Do not use it to set your own dose, to skip a clinician, or to expect a specific number on your scale. And hold every other page you read about “microdose results” to the same six checks: N, window, comparison group, how weights were measured, who dropped out, and how wide the interval is. The results themselves are in microdose tirzepatide results at 3, 7 and 9 weeks, and the record checks are in how we checked the data.

About the PlexusDx study

A retrospective, single-organization observational study of adults documented as new to GLP-1 medicines with BMI ≥27 kg/m²; no comparison group; self-reported weights; not a clinical trial; preprint not yet peer reviewed. The records came from the commercial care program that PlexusDx operates. Care was delivered by an independent network of U.S.-licensed clinicians through a state-licensed telehealth platform. The medication was compounded injectable tirzepatide with vitamin B12 at a total weekly dose of 1 to 2 mg, prepared by state-licensed compounding pharmacies operating in FDA-registered facilities. The study was written by Jay Hastings (PlexusDx CEO) and Jayden Lee, PharmD, EMBA (PlexusDx Medical Director). See the PlexusDx microdose tirzepatide study (60 patients, 2026).

The authors’ conclusion, in their words: “These hypothesis-generating observations support prospective comparative study, not conclusions about causality, equivalent dosing, or long-term safety.” The full preprint is on medRxiv: Weight Loss With Microdose Tirzepatide: Real-World Outcomes in Patients Initiating Treatment Below 2.5 mg Weekly (doi:10.64898/2026.09.15.26363172; posted September 17, 2026; CC BY 4.0). It has not yet been peer reviewed.

Where PlexusDx fits

PlexusDx offers compounded tirzepatide through its Weight Management Protocols, including the entry-tier Microdose GLP-1 Protocol. You fill out a short intake. A licensed clinician reviews it and decides what, if anything, fits you. The study described here looked back at records. It is not a promise of any result.

This article is educational. PlexusDx offers compounded tirzepatide through its Weight Management Protocols. The GLP-1 & Peptide Pathways Genetic Test analyzes how your genes influence peptide-related biological pathways. It does not recommend, prescribe, or determine which peptides you should use. Consult a qualified healthcare provider before beginning any peptide protocol. Compounded tirzepatide is not an FDA-approved drug product. It is prepared by state-licensed compounding pharmacies under federal compounding rules and is not the same as Zepbound® or Mounjaro®. Zepbound® and Mounjaro® are registered trademarks of Eli Lilly and Company. PlexusDx is not affiliated with or endorsed by Eli Lilly and Company.

Frequently Asked Questions

What does hypothesis-generating mean?

It means a finding is interesting enough to test properly but not strong enough to act on alone. The PlexusDx study uses the term because it is retrospective, has no comparison group, uses self-reported weights, and has small groups in later windows. It supports a prospective study, not conclusions.

Why does a small study need a comparison group?

Without a group of similar people who did not get the treatment, you cannot separate the treatment’s effect from what would have happened anyway. The PlexusDx study had no comparison group, so it describes what was observed and does not claim the medicine caused it.

What is completer bias?

It is the distortion that appears when people who stay in a study differ from those who leave. If people who saw no change stopped early, later averages look better than reality. The PlexusDx study’s later windows shrink to 19 and 8 patients, so this limit applies.

What does a 95% confidence interval tell me?

It is the range of values that fit the data reasonably well. A wide interval means the true group average is less certain. Every mean in the PlexusDx study is reported with its interval, and small groups, like 8 patients, produce wider ones than large groups do.

Should I change my dose based on a small study?

No. A hypothesis-generating study is not a basis for changing your own treatment. Bring it to a licensed prescriber and ask questions. Dose decisions belong to your clinician, who knows your history and can weigh a small study against everything else they know.

This article is part of the PlexusDx Education Hub. Browse all Peptides & GLP-1 education

Medical and Editorial Standards

Medical review process: This article was reviewed for medical accuracy, scientific clarity, evidence alignment, and appropriate discussion of genetics, medications, supplements, biomarkers, and health-related claims.

Sources and evidence: PlexusDx educational content is developed using peer-reviewed research, clinical literature, reputable medical references, and, where applicable, public health or regulatory guidance.

Commercial transparency: PlexusDx offers genetic testing, blood biomarker testing, personalized supplement recommendations, and related precision wellness services. Product mentions are intended to help readers understand available options and should not be interpreted as medical advice.

Important disclaimer: PlexusDx educational content is for informational purposes only and should not be used as a substitute for professional medical advice, diagnosis, or treatment. Always consult a qualified healthcare provider before making decisions about medications, supplements, genetic testing, lab testing, or health-related care.