Asking a follow-up that narrows an answer

A broad question gets a broad answer. Research tools can summarize many study types, so their first answer often mixes human data, animal data, and lab data in one place. That is not wrong. It is what the broad question asked for.

The second question is where you decide what you actually want. You can narrow by who was studied, what was measured, or how the study was designed. Each move removes a set of studies you did not mean. The result is an answer you can judge against the question in your head, instead of a general summary you have to sort on the fly.

The second question in each row is not longer for its own sake. It names the part of the research that should stay in view.

The next three sections walk through each filter, one at a time.

Say the group you are asking about first. If you leave the group out, the answer may still draw from trials in older adults, younger adults, animals, or lab cells. ChatPEP states the evidence level, but narrowing by population keeps the answer on the group you mean. A broad question can produce a broad stack of sources, and the reader is the one who has to separate them.

"Adults" narrows to adult participants. "Postmenopausal women" narrows further. "Human trials only" removes animal and cell studies entirely. A population detail is a filter. It does not make a study better by itself, but it makes the answer match the group you are trying to understand.

When a question includes the population, the tool has less room to combine study types. A cartilage repair question may return laboratory cell work unless you ask for human trials. Including "in adults" keeps the answer from drifting into pediatric research when you did not ask for it. The narrower question is not more technical. It is more honest about whose evidence you want.

Name the outcome you care about. If you ask only about a compound, the answer may report effects on many different measures. Some of those may not be the one you meant. A compound can be studied for skin thickness, tendon strength, cartilage volume, or reported pain, and the first answer may not know which one matters to you.

An outcome is the thing a study measured. When you name the outcome, the follow-up tells the tool to prefer studies that recorded that measure. It also gives you a way to check the answer. If the answer talks about a different measure, that is a sign to narrow again.

This step matters because one compound can be studied for different outcomes. A question can pull in studies that measured wound healing, skin thickness, or something else entirely. Saying "for skin" keeps the answer on skin rather than the other directions. You are not asking the tool to prove anything new. You are asking it to sort the research it already found.

Ask for a study design. Naming the study design you want makes the request clearer; ChatPEP follows the conversation as described in ¹. Study design is a filter on the method, not on the result.

A lab study can show a mechanism. A randomized controlled trial can show how a treatment performed in people under controlled conditions. They answer different questions. When you do not name the design, the answer may lay them side by side or leave you to sort the mixture. That makes it harder to spot the difference between a promising observation and a tested result.

The phrase "human trials only" is one of the most useful filters. It removes animal studies and cell studies while leaving human evidence. A more specific phrase like "randomized controlled trials" goes further, because it favors studies with a comparison group and an assigned treatment. If you want the strongest human evidence, say so in the follow-up instead of hoping the tool guesses.

In a research tool that keeps the thread, the follow-up can be short. ChatPEP follows the conversation, so a short follow-up picks up where the last answer left off.¹ You do not need to repeat the compound every time, but naming the filter helps.

When you ask a narrower follow-up, the answer should change in two visible ways. It should state the evidence level more clearly, and it should draw from a smaller set of sources. ChatPEP says whether a finding comes from human trials or from animal and lab studies.² Open the sources under the answer to see whether the narrower answer actually cites the kind of studies you asked for.³

Most answers end with a follow-up question about a related part of the research. You can reply to it or ask your own.⁴ A reply counts as a message, like any other.⁴ If the suggested question is close but too broad, add your own filter. For example, reply with "in human trials only" to trim the answer into the evidence you meant.

A good narrowed answer still needs checking. Select a citation to open the study. If the answer still mixes designs, ask the tool to separate them: "List the human trials first, then note any animal studies separately." That follow-up continues to narrow. The point is not to demand a different answer. The point is to make the tool show its work where you can see it.

• People who want a protocol to follow. ChatPEP never tells you what to take, how much, or how.⁵
• People who want to save every message. Each follow-up you send counts toward your daily message limit, and the limit uses a rolling 24-hour window.⁶
• People who want a source-free summary. ChatPEP cites its sources, but you are meant to open and check them.³
• People who want medical advice. The line under the message box says for research only, and ChatPEP is not a substitute for a qualified provider.²