Time as a diagnostic test: watchful waiting

A young woman comes to primary care with mild abdominal pain for one day, no fever, no vomiting, and an examination with no warning signs. The doctor considers the possibilities, explains that the picture looks benign, says exactly what would make the pain warrant a same-day reassessment, and schedules a return visit in two days. She did not order tests. She also did not close off the reasoning. She decided to use time as an instrument.

This strategy has a name in the primary care literature: the test of time, or watchful waiting. This text explains where the concept comes from, why the evolution of a picture carries diagnostic information, when waiting is reasonable and when it delays the diagnosis, and how to make waiting a decision with a plan, communicated to the person and recorded in the medical record.

Key points

• The test of time is a recognized diagnostic strategy in primary care, used when the picture does not yet define the diagnosis.
• Many common symptoms improve on their own, and those that persist or change character become more informative.
• Waiting is only safe with a safety net: what to expect, how to know something has changed, and what to do then.
• Some situations call for extra care with waiting, such as diseases that worsen quickly and people at high risk of complications.
• Diagnostic delays in primary care usually start in the consultation itself and in the follow-up of results and symptoms.
• The clinician's sense of alarm has diagnostic value and should weigh against the decision to simply observe.
• Watchful waiting must be explained clearly and recorded in the medical record.

Summary

• What it means to use time as a test
• Why time informs
• When waiting is reasonable
• When waiting gets in the way
• The safety net: waiting with a plan
• Intuition as a signal not to wait
• How to communicate the wait
• How to document the decision
• Frequently asked questions
• Final checklist

What it means to use time as a test

Heneghan and colleagues observed 300 consultations with six family physicians and described the strategies they used to reach a diagnosis (Heneghan, 2009). The reasoning was divided into three stages: the formation of the first hypotheses, the refinement of those hypotheses, and the definition of the final diagnosis. In the last stage, fewer than half of the cases reached a known diagnosis without the need for further tests (Heneghan, 2009).

For the others, the physicians used other strategies: ordering tests, testing a treatment, referring, or using time. The authors define the test of time as the use of the expected evolution of the disease to predict when the person should improve or worsen, in a "wait and see" strategy that allows the diagnosis to become more evident (Heneghan, 2009). The example given is that of an apparently viral gastroenteritis, with no warning signs, in which most physicians would wait one or two weeks before considering other diseases or ordering tests.

The test of time is not synonymous with doing nothing. It appears alongside other refinement strategies described in the same paper, such as the restricted exclusion process, in which the clinician knows the most likely cause of a complaint and a short list of serious diagnoses that must be ruled out (Heneghan, 2009). Time comes after that screening, not in its place.

When no label can be assigned, the authors describe four possible exits: calling the person in for a new assessment, ordering an exploratory test, sharing the uncertainty with the patient, or referring for a second opinion (Heneghan, 2009). Watchful waiting combines the first and the third, and works best when the person takes part in the decision.

Why time informs

The justification for waiting lies in the natural history of symptoms. A study that followed 500 patients who sought outpatient care with a physical complaint found improvement in about four out of five people after five years, and complete resolution in a little more than half (Jackson, 2005). In the same study, lack of resolution at three months was associated with a lower chance of improvement in the long term (Jackson, 2005).

This means that evolution separates the groups. A symptom that disappears within a few days rarely needed investigation. A symptom that persists beyond what is expected, worsens or gains new characteristics carries more weight, and the clinician can act with more information than was available at the first visit.

Time also reveals what has not yet appeared. Almond and colleagues recall that no exam or decision rule in primary care is perfectly sensitive, largely because people arrive at different stages of the disease's evolution (Almond, 2009). The authors note that half of the cases of meningococcal disease in children are not recognized by the family physician at the first visit, often because the typical features have not yet appeared (Almond, 2009).

This example shows both sides of time. It makes the diagnosis clearer, but in diseases that worsen quickly, the interval between the first visit and the appearance of signs can be short. That is why waiting must be vigilant: time only works as a test if someone is watching the result.

The same reasoning appears whenever a single measurement says little and its evolution says a lot. In suspected ectopic pregnancy, for example, a single value of chorionic gonadotropin rarely settles the picture, and it is the curve over days, together with ultrasound, that guides management, as shown in the text on beta-hCG from a single value to the curve. The logic is that of the test of time: repeat the observation at agreed intervals and compare the result with what was expected.

The clinical history is the main instrument for this reading, because the change over days usually appears first in the person's account.

When waiting is reasonable

Vigilant waiting tends to be a good choice when three conditions come together: the initial assessment shows no signs of severity, the most likely diagnosis has a benign and known course, and there is a clear and accessible way for the person to return if something changes. In these situations, immediate tests rarely change management, as discussed in the post on when the exam does not change management.

Acute respiratory infections offer the most studied example. A systematic review with 12 studies and 3,968 participants compared immediate prescription, delayed prescription and no antimicrobial prescription (Spurling, 2023). For many clinical outcomes there was no difference between the strategies, and there was no difference in complications. Delayed prescription reduced antimicrobial use to about three in ten people, against more than nine in ten with immediate prescription, with similar satisfaction (Spurling, 2023).

The authors conclude that, when the clinician considers it safe not to prescribe immediately, not prescribing and advising return if symptoms do not improve tends to produce the lowest antimicrobial use with similar clinical results (Spurling, 2023). It is a concrete form of vigilant waiting: the decision to treat depends on the evolution.

The waiting period depends on the hypothesis. In the example of gastroenteritis without warning signs, most of the physicians observed would wait one or two weeks before considering other causes (Heneghan, 2009). For nonspecific physical symptoms, lack of resolution within three months is already associated with a worse long-term course (Jackson, 2005). In conditions that can worsen quickly, the interval is measured in hours, and reassessment on the same day or the next day is part of the plan. The period should be told to the person and written in the medical record, because it defines when the result of the test of time will be read.

Vigilant waiting is usually considered in some typical situations: acute self-limited conditions without signs of severity, in people without important risk factors; recent musculoskeletal complaints, without relevant trauma and without neurological or systemic signs; nonspecific symptoms of recent onset in a previously well person, with a normal exam; subtle findings on exams whose meaning depends on repetition or evolution; and situations in which the available exam is more likely to generate a false positive than to help.

Waiting also protects against excess. Investigating every symptom on the first day increases the chance of finding unimportant findings and of starting cascades of tests, one of the paths to overdiagnosis.

When waiting gets in the way

The same strategy can delay important diagnoses. Almond and colleagues describe three high-risk situations in which an effective safety net is mandatory: when the diagnosis is uncertain and includes serious diseases that can progress quickly, when the diagnosis is known but carries a risk of serious complication, and when the person has increased risk due to age or other diseases (Almond, 2009). In these scenarios, waiting without a plan is dangerous.

Diagnostic error studies show where waiting usually fails. An analysis of 190 diagnostic errors detected by unexpected returns after primary care visits found 68 different diagnoses, most of them common conditions (Singh, 2013). About four in five failures occurred during the visit itself, especially in the history, the examination and the decision to order tests, and a relevant part involved follow-up and tracking of diagnostic information (Singh, 2013).

A more recent cohort study reviewed records of patients with lung and colorectal cancer diagnosed at an advanced stage, all with at least one primary care visit in the previous two years (Zimolzak, 2025). In more than half of the reviewed cases there were missed opportunities for earlier diagnosis, and the median time between the first diagnostic sign and the completion of the investigation ranged from one to 20 months (Zimolzak, 2025).

Some signs indicate that waiting has stopped being a strategy and become delay: the symptom persists beyond the expected time for the most likely benign cause, a new sign appears that does not fit the initial hypothesis, the person returns several times with the same complaint without the reasoning being reviewed, an abnormal result goes unanswered or without a marked return, or the person belongs to a higher-risk group and the investigation keeps being postponed.

The report on diagnostic error includes delay in its definition, and not only the wrong diagnosis. Waiting too long, therefore, is also a way of erring.

The safety net: waiting with a plan

The term safety net was introduced in general practice by Roger Neighbour, who summed it up in three questions: if I am right, what do I expect to happen; how will I know if I am wrong; and what will I do then (Almond, 2009). The questions apply to any watchful waiting, because they give observation reading criteria, as in any test.

Almond and colleagues recall that up to half of patients leave a general practice visit without a firm diagnosis, and that the need to rule out serious disease remains present in these cases (Almond, 2009). In a consensus with family physicians and hospital specialists, the authors reached five recommendations on the content of safety net advice (Almond, 2009). The table below organizes these elements in a practical way.

Element	Question it answers	How it appears in the visit	What to record
Uncertainty	Is the diagnosis settled?	State clearly that the hypothesis is probable, but not confirmed	Main hypothesis and serious diagnoses considered
What to watch for	How do I know I am wrong?	List specific signs of worsening, not a generic "come back if it gets worse"	Warning signs explained to the person
How to seek help	What to do if it changes?	Indicate where and how to seek care, including outside office hours	Agreed access route
Expected course	What should happen if I am right?	Inform how long improvement usually takes to appear	Expected time frame for improvement
Return	Who will look at the result of the test of time?	Schedule a reassessment or agree on active contact	Date or criterion for return

The authors also describe three ways in which the safety net fails: it is not offered, especially in high-risk situations; it is offered, but not heard or understood; or it is too vague about what to watch for and what to do (Almond, 2009). Each failure corresponds to a row of the table.

The planned reassessment is the point where uncertainty becomes a script: the clinician decides in advance what will count as improvement, what will count as worsening and what to do in each case.

Intuition as a signal not to wait

Not all information from the consultation fits into a list of warning signs. Family physicians describe two types of intuitive feeling: the alarm feeling, which leads to worrying about the person even without a specific finding, and the feeling of reassurance, that everything fits (Oliva-Fanlo, 2022). These feelings have been studied as part of reasoning, and not as something opposed to it.

A prospective study with 155 family physicians and 1,487 patients with a new reason for consultation found some intuitive feeling in almost every consultation, with reassurance in about three in four and alarm in about one in five (Oliva-Fanlo, 2022). After two months, the alarm feeling had sensitivity of about six in ten and specificity of about eight in ten for cancer and other serious diseases, with a positive predictive value of about one in eight (Oliva-Fanlo, 2022).

Another study assessed the alarm feeling in the face of shortness of breath or chest pain, with 25 physicians and 235 questionnaires filled out right after the consultation (Barais, 2020). The positive likelihood ratio was 2.12 and the negative was 0.55, which means that, when the physician felt alarm, the chance of life-threatening disease nearly doubled compared with the initial estimate (Barais, 2020).

The same numbers show the limits of intuition. With a low positive predictive value, most alarm feelings do not end in serious disease, and almost all the people who did not trigger alarm in fact had no serious disease in the following months (Oliva-Fanlo, 2022). Intuition is more useful for raising suspicion than for confirming it.

In practice, the alarm feeling works as one more finding. It confirms nothing on its own, but it weighs against the decision to just observe and in favor of examining again, ordering a test or shortening the return interval. The text about the two types of reasoning discusses how intuition and analysis complement each other, and the post about time, speed and accuracy shows why confidence does not always accompany precision.

How to communicate the wait

To the person, waiting can feel like neglect. A review on managing diagnostic uncertainty in primary care gathered ten studies and organized the topic into three domains: cognitive, emotional and ethical (Alam, 2017). The ethical domain revolved around communicating uncertainty to patients, and the emotional one included accepting uncertainty and working with provisional diagnoses. The authors point out that there is little empirical evidence on how to do this well.

The format of the guidance also matters. A network meta-analysis on safety netting advice for children with acute illness gathered 30 studies and found that paper-based guidance, especially combined with verbal guidance, may reduce antimicrobial prescribing and return visits, while video, websites and online modules may improve parents' knowledge (Burvenich, 2025). The certainty of evidence was low or very low for most comparisons.

Some precautions help communicate watchful waiting:

• Explain what was considered and why the most likely hypothesis is benign, without promising certainty.
• Say that waiting is a diagnostic choice and not a lack of attention.
• Name the specific signs that justify seeking help before the return.
• Agree on where and how to seek care if these signs appear.
• Check whether the person understood, asking them to repeat the plan in their own words.
• Offer the guidance in writing as well whenever possible.

The language matters as much as the content. Phrases like "it's nothing" close the door to coming back, while "most likely this is something that improves on its own, and we'll confirm that together" make it clear that the assessment continues. When the person understands that the wait has a deadline and criteria, they become part of the monitoring, and not just someone who waits.

Shared decision-making applies here: some people prefer to investigate right away, and that preference is part of the decision.

How to document the decision

The last recommendation of the consensus by Almond and colleagues is simple: safety net advice should be recorded in the medical record (Almond, 2009). The record protects the continuity of the reasoning, lets another professional understand what was agreed and leaves the wait documented as a decision.

Interviews with 25 family physicians showed that each one used personal safety net strategies, often learned from past mistakes and without knowing how colleagues did it (Evans, 2018). Under time pressure, they did selective active follow-up of patients seen as higher risk, which left people with low but not zero risk symptoms with less protection (Evans, 2018). The authors argue for systems in the service that allow this group to be followed without relying only on the clinician's memory.

A good record of watchful waiting includes:

• The main hypothesis and the reasoning that supports it.
• The serious diagnoses considered and why they seem unlikely now.
• The warning signs explained to the person, with the words used.
• The agreed route to get help if these signs appear.
• The expected time to improvement and the date or criterion for reassessment.

The post on how to write the assessment shows how to make the reasoning visible, and not just the findings.

At the follow-up, the record serves as a baseline. The clinician compares what happened with what had been predicted and decides whether the hypothesis was confirmed, whether tests are needed or whether a referral should be made. Without this comparison, the test of time has no result.

Frequently asked questions

Is watchful waiting the same as doing nothing?

No. It is a diagnostic strategy with a plan: a defined hypothesis, agreed warning signs and a scheduled reassessment. Family physicians use it to let the course of events clarify the diagnosis (Heneghan, 2009).

When should you not wait?

When there are signs of severity, when the hypothesis includes a serious disease that can worsen quickly, when there is a high risk of complication or when the person belongs to a group at increased risk (Almond, 2009). The clinician's sense of alarm also weighs against waiting.

What should good safety net advice include?

The existence of uncertainty, the specific signs to watch for, how to seek help, the expected course and the record in the medical chart (Almond, 2009). A generic "come back if it gets worse" does not fulfill this role.

Does the physician's intuition have diagnostic value?

It has some. The sense of alarm was associated with a higher chance of serious disease in prospective studies in primary care (Oliva-Fanlo, 2022). It works as one more finding, and not as proof.

Can waiting delay a cancer diagnosis?

It can, when persistent symptoms or altered results go unanswered. Reviews of advanced cases found frequent missed opportunities in the months before diagnosis (Zimolzak, 2025).

Does this text guide individual conduct?

No. It explains a diagnostic strategy in primary care and the evidence about it, and does not replace the clinical assessment of each patient.

Final checklist

• Before waiting, rule out the serious diagnoses on the short list for the complaint.
• Confirm that there are no signs of severity or risk factor that changes the picture.
• Answer the three questions: what am I waiting for, how will I know I was wrong, what will I do then.
• Explain the uncertainty and the specific warning signs to the person.
• Agree on how and where to seek help before the follow-up.
• Schedule the reassessment or agree on active contact.
• Take your sense of alarm seriously.
• Record the hypothesis, warning signs, expected time and return date in the medical chart.

References

1. Alam R, Cheraghi-Sohi S, Panagioti M, Esmail A, Campbell S, Panagopoulou E. Managing diagnostic uncertainty in primary care: a systematic critical review. BMC Fam Pract. 2017. PMID 28784088. DOI 10.1186/s12875-017-0650-0.
2. Almond S, Mant D, Thompson M. Diagnostic safety-netting. Br J Gen Pract. 2009. PMID 19861036. DOI 10.3399/bjgp09X472971.
3. Barais M, Fossard E, Dany A, Montier T, Stolper E, Van Royen P. Accuracy of the general practitioner's sense of alarm when confronted with dyspnoea and/or chest pain: a prospective observational study. BMJ Open. 2020. PMID 32075841. DOI 10.1136/bmjopen-2019-034348.
4. Burvenich R, Bos DA, Lowie L, et al. Effectiveness of safety-netting approaches for acutely ill children: a network meta-analysis. Br J Gen Pract. 2025. PMID 39117428. DOI 10.3399/BJGP.2024.0141.
5. Evans J, Ziebland S, MacArtney JI, Bankhead CR, Rose PW, Nicholson BD. GPs' understanding and practice of safety netting for potential cancer presentations: a qualitative study in primary care. Br J Gen Pract. 2018. PMID 29739779. DOI 10.3399/bjgp18X696233.
6. Heneghan C, Glasziou P, Thompson M, et al. Diagnostic strategies used in primary care. BMJ. 2009. PMID 19380414. DOI 10.1136/bmj.b946.
7. Jackson JL, Passamonti M. The outcomes among patients presenting in primary care with a physical symptom at 5 years. J Gen Intern Med. 2005. PMID 16307629. DOI 10.1111/j.1525-1497.2005.0241.x.
8. Oliva-Fanlo B, March S, Gadea-Ruiz C, Stolper E, Esteva M; CORap group. Prospective observational study on the prevalence and diagnostic value of general practitioners' gut feelings for cancer and serious diseases. J Gen Intern Med. 2022. PMID 35088202. DOI 10.1007/s11606-021-07352-w.
9. Singh H, Giardina TD, Meyer AN, Forjuoh SN, Reis MD, Thomas EJ. Types and origins of diagnostic errors in primary care settings. JAMA Intern Med. 2013. PMID 23440149. DOI 10.1001/jamainternmed.2013.2777.
10. Spurling GK, Dooley L, Clark J, Askew DA. Immediate versus delayed versus no antibiotics for respiratory infections. Cochrane Database Syst Rev. 2023. PMID 37791590. DOI 10.1002/14651858.CD004417.pub6.
11. Zimolzak AJ, Kapadia P, Upadhyay DK, et al. Frequent missed opportunities for earlier diagnosis of advanced-stage colorectal or lung cancer. JAMA Intern Med. 2025. PMID 40690229. DOI 10.1001/jamainternmed.2025.2875.