Prognosis is the clinical estimation of the likely future course, outcome or consequences of a disease, injury or health condition. It can encompass expected recovery, progression, complications, recurrence, functional outcome, survival or response to treatment, depending on the condition and context. Prognosis is not the same as certainty about an individual's future. It is an estimate based on available evidence, clinical characteristics, disease severity, treatment options and factors known to influence outcomes. Clinicians may communicate prognosis qualitatively, such as describing an expected recovery pattern, or quantitatively using measures such as estimated survival, probability of recurrence or likelihood of functional improvement when sufficiently reliable evidence exists.
Prognosis is derived from prognostic factors and evidence about populations with similar conditions. Factors can include disease stage, severity, physiological status, age, comorbidities, functional capacity, biological characteristics, treatment response and timing of intervention. Prognostic models may combine multiple variables to estimate an outcome, but their accuracy depends on the population in which they were developed and validated. A statistically important association does not necessarily provide a sufficiently accurate individual prediction. Prognosis should therefore be understood as probability-based reasoning rather than deterministic forecasting. Even when a factor is associated with poorer outcomes at population level, substantial variation can exist among individuals.
Prognosis differs from diagnosis because the two answer different questions. Diagnosis asks what condition is present or most likely responsible for the presentation, whereas Prognosis asks what is likely to happen subsequently. Prognosis also differs from prediction in some contexts: prediction is a broad term for anticipating an outcome, while prognosis is generally used specifically in relation to the expected course or outcome of a disease or health condition. Risk refers to the probability of an event occurring, and prognostic risk concerns future outcomes after a particular clinical state or diagnosis. These concepts overlap, but they should not be treated as interchangeable because their reference populations and time horizons can differ.
Prognostic information can concern several timescales. Short-term prognosis may address immediate deterioration, response to initial treatment or discharge readiness; intermediate prognosis may concern recovery, complications or functional improvement; long-term prognosis may address recurrence, chronic disability, disease progression or survival. The appropriate endpoint must therefore be specified. For example, a treatment can have a favourable effect on short-term symptoms without changing long-term disease progression, while another intervention may have little immediate effect but substantially alter long-term risk. Prognosis is consequently multidimensional rather than reducible to a single statement such as "good" or "poor".
Communicating Prognosis is clinically and ethically important. Patients may need information about likely outcomes to make informed decisions, plan care and understand the consequences of treatment options. Prognostic communication should distinguish what is known from what remains uncertain and should avoid false precision. Population statistics can inform discussions but do not necessarily describe exactly what will happen to an individual. The relevance of a statistic depends on whether the study population, treatment era, disease characteristics and outcome definition resemble the patient's situation. Clinicians may need to explain ranges, uncertainties and alternative trajectories rather than present one numerical outcome as inevitable.
Prognosis can change as new evidence emerges. A patient's response to treatment, new test results, complications, disease progression or improved understanding of the underlying condition can alter the expected course. Treatments themselves can modify prognosis, meaning that historical outcome statistics may not accurately represent outcomes under current care. Prognostic estimates can also be affected by lead-time and selection effects in studies, differences in diagnostic criteria and changes in treatment standards. For these reasons, prognostic evidence should be interpreted according to its date, population, intervention and endpoint rather than assuming that any published percentage applies universally.
Prognostic factors can be favourable, unfavourable or neutral depending on the outcome being considered, but the relationship is rarely absolute. Some factors influence several outcomes simultaneously, while others predict one endpoint but not another. Prognostic models can support clinical decision-making, resource planning and research, but they require calibration and validation and can perform differently across demographic and healthcare populations. Models may also reproduce biases present in their underlying data. A prognostic estimate should therefore complement rather than replace clinical judgement, patient preferences and consideration of treatment goals.
Common misconceptions include treating Prognosis as a guarantee, assuming that one person's outcome determines another's, or interpreting a population survival statistic as an individual's fixed lifespan. Prognostic uncertainty is inherent in medicine, especially when diseases have heterogeneous courses or evidence is limited. Prognosis also does not necessarily determine what should be done: two treatments can have similar expected survival but different effects on symptoms, function, side effects or quality of life, making patient priorities relevant. Proper prognostic reasoning integrates clinical evidence, statistical uncertainty, disease biology, treatment effects and individual circumstances while clearly distinguishing likelihood from certainty.
