Safety netting in consultation: what to tell the patient

A middle-aged man comes to primary care with a cough for three weeks, no fever and a normal lung exam. The doctor concludes it must be the end of a viral infection and closes with a familiar phrase: "if it doesn't get better, come back". The patient leaves reassured. Two weeks later, the cough is still there, but he doesn't know whether that already counts as "not getting better", or whether he should wait longer, call or book an appointment. He ends up waiting.

The closing phrase was not wrong. It was incomplete. The primary care literature calls the set of guidance and arrangements that protect the person when the diagnosis is still uncertain a safety net. This text explains what the safety net includes, what studies with recorded consultations show about practice, why specific guidance works better than vague guidance, how to share responsibility for follow-up and how to record what was agreed.

Key points

• A safety net means agreeing with the person what to expect, how to notice that something has changed and what to do in that case, including when and where to seek help.
• Studies with recorded consultations show that guidance is frequent, but it is often vague and does not always reach the medical record.
• Patients prefer specific, planned guidance; generic phrases can sound like neglect.
• Checking understanding by asking the person to repeat it in their own words is one of the strategies proposed by patients and professionals.
• Follow-up of tests and referrals needs a defined person in charge, and the double safety net involves both the service and the patient.
• Telephone consultations, multimorbidity, lower continuity and mental health are situations in which the net tends to be weaker.

Summary

• What a safety net is
• What recorded consultations show
• Specific or vague guidance
• How to know whether the person understood
• Who is responsible for follow-up
• Pending results and referrals
• How to record the safety net
• When the net becomes weaker
• Frequently asked questions
• Final checklist

What a safety net is

The term was popularized in family medicine to describe what the clinician does when ending a consultation without diagnostic certainty. Almond and colleagues summed up the idea in three questions the professional should answer with the person: what is expected to happen, how to know whether the course is not as expected and what to do in that case (Almond, 2009). The safety net, in this sense, is a tool for managing uncertainty, not a farewell phrase.

A literature review that gathered 47 studies detailed what is usually part of this conversation (Jones, 2019). The components described were:

• communicating diagnostic uncertainty honestly;
• explaining the warning signs that justify a new assessment;
• stating the expected course over time, so the person knows when the delay becomes abnormal;
• saying how and when to seek help, and in what type of service;
• agreeing on follow-up, such as a booked return visit or active contact;
• tracking tests ordered and referrals;
• organizing the service so that all of this actually happens.

The last item is the least remembered. The safety net depends not only on what is said in the room, but also on systems that ensure a result reaches someone, that a call is returned and that the person can book the agreed return visit (Jones, 2019).

The same review points to groups in which the safety net matters more: children, people with acute illness that can worsen quickly, people with several chronic diseases and people with mental health problems (Jones, 2019). In the infant with fever, for example, the margin for waiting is smaller, as discussed in the text on what changes for the clinician in the febrile baby.

The safety net complements other strategies for managing uncertainty. When the return visit is planned with clear hypotheses and criteria, uncertainty stops being a loose problem and gains a script, an idea developed in the text on planned reassessment.

What recorded consultations show

Much of what is known about actual practice comes from video-recorded consultations, with consent. Edwards and colleagues analyzed 318 routine primary care consultations and found safety netting advice in 205 of them, about two in three consultations (Edwards, 2019). When the analysis moved to the problem level, the advice appeared for 257 of the 555 problems discussed, less than half (Edwards, 2019).

The content also varied. Specific advice, which says exactly what to watch for and what to do, appeared in almost half of the occasions, and the rest was generic (Edwards, 2019). Safety netting was more likely for acute problems, for the first problem discussed in the consultation and among younger doctors (Edwards, 2019). In other words, the second or third topic of a long consultation tends to be left without safety netting.

A study of recorded consultations in out-of-hours primary care services, with 77 videos of 19 doctors treating respiratory infections, found a similar pattern: the advice existed, but it was generally vague (Colliers, 2022). The authors observed that the advice was more frequent when an antimicrobial was prescribed, as if the prescription pulled the conversation toward what to do next (Colliers, 2022).

These findings do not point to individual carelessness. They show that safety netting competes for space with other tasks at the end of the consultation, when time is short and attention is already on the next patient. That is why making the advice a structured habit, and not an improvisation, is one of the most repeated recommendations in the literature.

Specific or vague advice

The difference between "come back if it gets worse" and specific advice seems small to the person speaking, but it is large to the person listening. A study of interviews with patients and professionals concluded that the advice needs to be adapted to the clinical context and, in particular, to the encounter between that person and that professional (Fernholm, 2023). In practice, this means making clear what to watch for and saying when and how to come back, components described in Jones's review (Jones, 2019).

The patients' perspective reinforces this idea. In interviews with 20 people who had sought care with respiratory symptoms that could indicate lung cancer, participants preferred active, planned advice, such as an already scheduled return visit or a follow-up test, to open-ended instructions like "come back if you need to" (Black, 2022). Vague advice was perceived by some as neglect, and the feeling was worse when the consultation had been by phone (Black, 2022).

The table below summarizes the practical differences between the two types of advice.

Element	Vague advice	Specific advice
Expected course	"It should get better"	"The cough usually improves within a few weeks; if it continues beyond that, I want to see you again"
Warning signs	"If it gets worse"	Concrete, observable signs, such as shortness of breath, blood in the sputum or a fever that returns
Deadline	Absent	A date or interval for reassessment, even if everything is fine
Where to seek help	"Seek help"	Which service, at what hours and what to do outside those hours
Responsible person	Implied	Who will look at the test and how the person will be notified
Record	Absent or generic	Signs, deadline and plan written in the medical record

Specificity is not the same as a long list of symptoms. Advice with ten warning signs can be as unhelpful as none, because the person cannot retain everything. Participants in the interview study argued for adapting the advice to the clinical context and to the consultation at hand (Fernholm, 2023). In children, advice for parents and caregivers should include the signs of deterioration that most often go unnoticed, the topic of the text on why the child who gets worse misleads.

Written guidance can help, but it does not replace conversation. In a co-design study, patients and professionals suggested combining written and verbal information to avoid misunderstandings and help the person remember, and the professionals said they adapt the medium to each patient (Wannheden, 2024). Jones's review found several sources recommending written guidance, but none said what it should contain or who else would benefit from it (Jones, 2019).

How to know whether the person understood

Saying no does not guarantee that the person heard, and hearing does not guarantee that they understood. In an interview study with nine doctors and eight patients, participants described the conditions for the safety net to work: the person needs to be in a position to receive the information, the guidance must make sense within what they already know, and both sides need to leave the consultation with the same idea of what can happen and what to do (Fernholm, 2023). The authors called this a shared mental model.

A co-design study with patients, caregivers and professionals organized the strategies into three groups: conveying the information, making sure it was understood, and supporting the person to act after the consultation (Wannheden, 2024). To make sure of understanding, the proposed strategy was to ask the person to explain in their own words what they will watch for and what they will do (Wannheden, 2024). To support action, measures included making the contact channel clear and making it easier to come back.

Asking the person to repeat is not a memory test. A respectful way is to take responsibility for clarity: "So I can be sure I explained it well, what will you watch for in the coming days, and when will you come to us?". The answer shows gaps that the professional can correct before the person leaves.

The check also reveals what the person fears. Often, the concern that brought the person to the consultation was not addressed, and the safety net ends up disconnected from what they will watch at home. Paying attention to the cues the patient offers during the consultation helps avoid this disconnection, as discussed in the text about cues the patient offers.

When there is a language barrier, checking understanding becomes even more important. The presence of a professional interpreter and the way the conversation is conducted make a difference, as shown in the text about how to conduct an interpreted consultation.

Who is responsible for follow-up

Agreeing on what to watch for is half of the net. The other half is deciding who does what after the consultation. Evans and colleagues interviewed 25 family doctors and 23 patients about the follow-up of symptoms that could indicate cancer (Evans, 2019). The authors described two ways of dividing responsibility: a transactional one, in which the doctor hands over the guidance and responsibility passes to the patient, and an interdependent one, in which doctor and patient keep active and complementary roles over time (Evans, 2019).

In the interviews, the risk situations recurred: guidance with no deadline, information insufficient for the patient to know what to watch for, excessive reassurance that discouraged return, and patient concerns that were not addressed (Evans, 2019). In these situations, responsibility was transferred to the patient without the patient having the means to exercise it.

A survey of 2,879 primary care professionals from 11 jurisdictions showed that the practice of keeping responsibility for follow-up varies widely between health systems (Nicholson, 2018). In some places, fewer than one in five professionals said they kept this responsibility when the patient was referred; in others, almost all did (Nicholson, 2018). The authors argue for a double safety net, in which the service actively tracks tests and referrals and the patient also knows what to expect and when to follow up (Nicholson, 2018).

The double safety net protects against failures that neither side notices alone. If the result is lost in the service, the patient who knows they should have received news within a week can call. If the patient forgets, the service that tracks the list of pending items can call them. Participants in the codesign study recognized that it is not always possible to maintain continuity with the same professional and, for this reason, emphasized that the person needs to feel welcome to come back and have easy access to follow-up (Wannheden, 2024).

Result follow-up and closing the diagnostic loop appear among the recommendations of major reports on diagnostic safety, as summarized in the text on the eight recommendation axes. When the failure happens, the way of talking with the person about it also matters, as discussed in the text on what patients expect to hear.

Pending results and referrals

Tests ordered and referrals made are open promises. Each of them needs three answers: who will look at the result, how long until the person should have news, and what happens if the news does not come. Jones's review includes the tracking of tests and referrals among the components of the safety net, alongside the guidance given to the patient (Jones, 2019).

In practice, the rule "no news is good news" is one of the most fragile forms of net. It assumes that every abnormal result will be seen and communicated, and it makes silence a form of reassurance. In Evans's interviews, the lack of a deadline and excessive reassurance were among the situations that left the patient not knowing whether they should ask for an answer (Evans, 2019). One alternative is to agree that the person will receive news in any scenario, or to say how long they should wait before contacting the service if they have received nothing.

Planned long-term follow-up is a particular case. Findings that require reassessment in months, such as an incidental pulmonary nodule, depend on a record that survives changes of professional and of service. The text on from recommendation to practice shows how this type of follow-up usually fails.

A few simple precautions help close the loop on pending results:

• record in the medical chart which tests were ordered, why, and who will check the result;
• agree with the person on a deadline for receiving the news and on what to do if it does not arrive;
• keep a team list of pending items, reviewed at defined intervals;
• for referrals, tell the person what to expect from the other service and how long it will take;
• when the professional is away, leave the pending items with a named colleague.

None of these measures depends on sophisticated technology. They depend on someone feeling ownership of the pending item, which is easier when the team has clear roles.

How to record the safety net

The record protects the person and the professional. It lets the next professional know what was agreed, helps the team pick up pending items again, and shows the reasoning behind the decision. In Edwards's study with recorded consultations, the safety net given verbally was recorded in the medical chart for 109 of 242 problems, less than half, while follow-up plans were recorded more often, in 295 of 354 problems (Edwards, 2019).

A retrospective cohort study of 1,886 consultations at out-of-hours primary care services found safety net guidance recorded in about four out of five consultations, but a deadline for seeking care again appeared in only about one in five (Edwards, 2025). Recording was more frequent in in-person consultations than in those done by telephone, and less frequent in consultations about mental health (Edwards, 2025).

Phrases such as "advised to return if necessary" record that something was said, but not what. A useful record includes the main hypothesis, the warning signs that were explained, the deadline for reassessment, and who was responsible for tests and referrals. This fits in a few lines and fits into the part of the medical chart dedicated to assessment and plan, as described in the text on how to write the assessment.

Recording also helps avoid automatic repetition. Ready-made templates that always insert the same guidance sentence can give the impression that the safety net was provided when the conversation never happened. Adapting the text to that day's problem is part of the work.

When the net becomes more fragile

Some situations weaken the safety net in predictable ways. Jones's review considers the safety net particularly important in emergency services, out-of-hours care and telephone consultations, and includes people with several chronic conditions among those who need it most (Jones, 2019). In each of these situations, a part of the net that would normally work on its own stops working.

In a telephone or video consultation, the professional loses part of the examination and of nonverbal reading, and the person may find it harder to retain the guidance. The studies described above show less frequent recording of the safety net in telephone consultations (Edwards, 2025) and a greater perception of neglect when vague guidance came by telephone (Black, 2022). The text on history, warning signs and plan in telemedicine discusses how to compensate for these losses.

In multimorbidity, the symptoms of one disease can mask those of another, and the person receives guidance from several professionals who do not always talk to each other. A safety net for each problem, without integration, can become a list that is impossible to follow. Prioritizing the signs that really require quick action and gathering the guidance in one place helps.

In mental health, recording of the safety net was less frequent (Edwards, 2025), although Jones's review includes this group among those who need it most (Jones, 2019). Agreeing with the person what to do if symptoms worsen, whom to contact and how to make contact out of hours is part of care, not a bureaucratic extra.

There are also access limits. Guidance to "come back in two days" is no use if the person cannot book an appointment in two days. The safety net must be consistent with how the service actually works, which includes knowing how the person will be seen if they come back outside the schedule, and with the person's conditions, such as the ability to understand and decide, distance and work. This brings the topic back to the organizational component described by Jones: the best verbal guidance fails if the service does not offer the agreed path (Jones, 2019).

Frequently asked questions

Is a safety net the same thing as watchful waiting?

No. Watchful waiting is the decision to use the course of the condition as diagnostic information. The safety net is the set of guidance and agreements that protect the person during that wait and in any other situation of uncertainty, including after ordering tests or making a referral (Almond, 2009).

How many warning signs should I explain?

There is no fixed number. Patients and professionals argue for adapting the guidance to the clinical context and to the consultation in question (Fernholm, 2023), which in practice means prioritizing the signs most relevant to the hypothesis instead of listing every possible symptom.

Do a leaflet or a written message replace verbal guidance?

Not necessarily. Patients and professionals suggest combining written and verbal information, to avoid misunderstandings and help people remember, and the medium can be adapted to each person (Wannheden, 2024).

How can I check understanding without embarrassing the person?

Take responsibility for clarity and ask the person to explain in their own words what they will watch for and when they will seek help. This strategy was proposed by patients, caregivers and professionals in a codesign study (Wannheden, 2024).

Who is responsible for a test I ordered?

Whoever ordered the test must ensure that someone will check the result and tell the person. The recommendation of a double safety net also includes informing the patient when to expect news and what to do if it does not arrive (Nicholson, 2018).

What should be written in the medical record?

The main hypothesis, the warning signs explained, the timeframe for reassessment and who was responsible for tests and referrals. In a study of nearly 1,900 consultations, the timeframe for seeking care again appeared in only about one in five (Edwards, 2025).

Final checklist

• Say what is expected to happen and in how much time.
• Choose a few concrete warning signs for that hypothesis and explain them in plain language.
• Define where and how the person should seek help, including outside regular hours.
• Ask the person to explain in their own words what they will watch for and what they will do.
• For each test or referral, define who checks the result and in how much time the person will hear back.
• Record the hypothesis, the signs explained, the timeframe and who is responsible.
• In remote consultations, with multiple conditions or in mental health, reinforce the guidance and the path back.

References

1. Almond S, Mant D, Thompson M. Diagnostic safety-netting. Br J Gen Pract. 2009. PMID 19861036. DOI 10.3399/bjgp09X472971.
2. Black GB, van Os S, Renzi C, Walter FM, Hamilton W, Whitaker KL. How does safety netting for lung cancer symptoms help patients to reconsult appropriately? A qualitative study. BMC Prim Care. 2022. PMID 35858826. DOI 10.1186/s12875-022-01791-y.
3. Colliers A, Philips H, Bombeke K, Remmen R, Coenen S, Anthierens S. Safety netting advice for respiratory tract infections in out-of-hours primary care: A qualitative analysis of consultation videos. Eur J Gen Pract. 2022. PMID 35535690. DOI 10.1080/13814788.2022.2064448.
4. Edwards PJ, Finnikin S, Wilson F, et al. Safety-netting advice documentation in out-of-hours primary care: a retrospective cohort from 2013 to 2020. Br J Gen Pract. 2025. PMID 38950945. DOI 10.3399/BJGP.2024.0057.
5. Edwards PJ, Ridd MJ, Sanderson E, Barnes RK. Safety netting in routine primary care consultations: an observational study using video-recorded UK consultations. Br J Gen Pract. 2019. PMID 31740458. DOI 10.3399/bjgp19X706601.
6. Evans J, Macartney JI, Bankhead C, et al. How do GPs and patients share the responsibility for cancer safety netting follow-up actions? A qualitative interview study of GPs and patients in Oxfordshire, UK. BMJ Open. 2019. PMID 31515421. DOI 10.1136/bmjopen-2019-029316.
7. Fernholm R, Wannheden C, Trygg Lycke S, Riggare S, Pukk Harenstam K. Patients' and clinicians' views on the appropriate use of safety-netting advice in consultations-an interview study from Sweden. BMJ Open. 2023. PMID 37798020. DOI 10.1136/bmjopen-2023-077938.
8. Jones D, Dunn L, Watt I, Macleod U. Safety netting for primary care: evidence from a literature review. Br J Gen Pract. 2019. PMID 30510099. DOI 10.3399/bjgp18X700193.
9. Nicholson BD, Goyder CR, Bankhead CR, et al. Responsibility for follow-up during the diagnostic process in primary care: a secondary analysis of International Cancer Benchmarking Partnership data. Br J Gen Pract. 2018. PMID 29686134. DOI 10.3399/bjgp18X695813.
10. Wannheden C, Hagman J, Riggare S, Pukk Härenstam K, Fernholm R. Safety-netting strategies for primary and emergency care: a codesign study with patients, carers and clinicians in Sweden. BMJ Open. 2024. PMID 39107019. DOI 10.1136/bmjopen-2024-089224.