Overdiagnosis: Is more always better?

Abstract

General practice is patient-centred and dynamic, requiring doctors to balance the benefits of early detection with the risks of over-investigation. GPs navigate screening results, incidental findings and shifting disease thresholds. While additional tests and prescriptions may appear helpful, they can contribute to overdiagnosis: correctly identifying abnormalities that would never have caused symptoms or shortened life if left undiscovered. Unlike misdiagnosis, overdiagnosis involves accurate tests and true pathology, but harm arises from the consequences of the label – patient anxiety, unnecessary interventions and diversion of limited NHS resources. Overdiagnosis is rarely visible in an individual consultation, yet its cumulative effects are substantial. This article explores the concept of overdiagnosis, its mechanisms and drivers, and describes its clinical, psychological and economic consequences. It offers practical strategies to help GPs recognise, explain and manage it in day-to-day practice.

Clinical case scenario

Mrs A, a 52-year-old female, attends for a routine health check. Her blood pressure is 132/84 mmHg, and her cholesterol is borderline high. She is concerned about heart disease after her sister’s myocardial infarction and asks whether she should start a statin and have further investigations. Her records show no prior cardiovascular events and her overall QRISK3 (https://qrisk.org) score indicates only a modest risk. She appears anxious and expects early intervention is always better.

Understanding overdiagnosis

For GPs, who work at the interface of population health and individual patient care, understanding overdiagnosis is essential. The term is used variably, but broadly refers to correctly identifying a condition that would not have gone on to cause harm. Unlike misdiagnosis (an incorrect label) or a false positive (where disease is absent), overdiagnosis involves accurate testing and true pathology; the problem lies in the consequences of knowing about it (Brodersen et al., 2018).

The challenge of definition

Defining overdiagnosis is difficult because it requires imagining an unknowable alternative. A patient might be diagnosed with a small, slow-growing cancer but later die from an unrelated cause. With or without the cancer diagnosis, their outcome is the same: the patient dies with the cancer, not from it. However, identifying such cases in real time is impossible; clinicians cannot predict the future with certainty (Brodersen et al., 2018; Hofmann et al., 2021; Walker and Rogers, 2017).

Disease itself is also a shifting concept. Conditions such as hypertension or thyroid nodules exist on a spectrum, and diagnostic thresholds are often revised as evidence evolves (Li et al., 2020; Walker and Rogers, 2017). In multimorbidity, it is rarely possible to attribute outcomes to one single diagnosis (Hofmann et al., 2021). Overdiagnosis is therefore partly a function of how we define and measure ‘disease’, not just underlying biology.

Despite these challenges, ignoring overdiagnosis is not an option. Clinicians and patients need to recognise its existence, understand its consequences, and develop ways to manage it in everyday consultations (Kale and Korenstein, 2018). Some degree of overdiagnosis is unavoidable in modern healthcare because early detection and prevention are central to practice. The more we seek to minimise missed diagnoses (underdiagnosis), the more we risk identifying conditions that might never have caused harm (Kale and Korenstein, 2018). The task for GPs is to balance the two.

Mechanisms of overdiagnosis

Three main mechanisms underlie overdiagnosis:

Biological non-progression. Some conditions, such as certain thyroid or prostate cancers, will never cause symptoms or shorten life expectancy. Detecting and treating them only exposes patients to unnecessary risks (Kale and Korenstein, 2018; Li et al., 2020; Welch et al., 2019).

Competing mortality. A genuine condition may be diagnosed but never become clinically relevant because another illness is more likely to determine outcome, particularly in older, multimorbid populations (Divo et al., 2014).

Risk reclassification. Lowering diagnostic thresholds for conditions such as hypertension or diabetes instantly creates large numbers of new ‘patients’, some of whom will never experience harm at their previous levels (Bell et al., 2022; Kale and Korenstein, 2018; Walker and Rogers, 2017).

These mechanisms highlight that overdiagnosis is not just about biology but also about how medicine defines and interprets disease.

Drivers of overdiagnosis

While biological and diagnostic processes explain how overdiagnosis occurs, systemic and behavioural factors explain why it has become increasingly common in modern medicine.

Technological advancements

Medical technology has transformed our ability to detect disease. Imaging, genetic testing, and artificial intelligence now identify ever-smaller abnormalities (Bohr and Memarzadeh, 2020; Capurro et al., 2022), but increased detection does not always mean improved outcomes.

The widespread use of computed tomography (CT) pulmonary angiography has led to a significant rise in pulmonary embolism diagnoses without a corresponding fall in mortality (Wiener et al., 2013). Abdominal scans and genetic tests also generate numerous ‘incidentalomas’, many of which prompt further investigation but never cause harm (Berland et al., 2010; Diamandis and Li, 2016). For GPs, the challenge is helping patients understand that ‘finding more’ is not always synonymous with better care.

Financial incentives

Financial structures strongly influence how much testing and diagnosing takes place. In the UK, the Quality and Outcomes Framework (QOF) has rewarded practices for achieving biomedical targets such as lowering blood pressure or recording HbA1c levels (Forbes et al., 2017). While this has improved chronic disease management, it has also encouraged the medicalisation of borderline states such as ‘pre-diabetes’ or mild hypertension, and incentivised repeat testing where the benefit is uncertain. This creates tension: meeting targets may secure practice income but will not always align with the best interests of individual patients.

Industry pressures compound this. Pharmaceutical and device companies expand markets by sponsoring awareness campaigns, educational events and sometimes influencing guideline development (Moynihan et al., 2012). These forces can subtly lower intervention thresholds, generating larger ‘at-risk’ populations. For example, changing lipid or osteoporosis criteria has periodically widened eligibility for treatment without robust outcome benefits (Gupta et al., 2020).

Although the NHS buffers against the most aggressive commercial practices, UK clinicians are not immune. Media coverage and patient advocacy groups, sometimes shaped by industry sponsorship, can create public demand for wider screening or treatment, as seen in lobbying for atrial fibrillation screening despite lack of UK National Screening Committee endorsement (McCartney et al., 2023).

For GPs, recognising these pressures helps refocus consultations on patient-centred outcomes rather than external metrics.

Psychology of regret and medical culture

Doctors are more likely to be criticised for missing a diagnosis than for overdiagnosing, and this asymmetry drives defensive practice (Moynihan et al., 2012). Ordering ‘just one more test’ often feels safer than explaining why a test is unnecessary, even when restraint would prevent harm.

Patients, meanwhile, frequently assume that ‘more is better’. Public surveys show many believe early detection always saves lives, while few are aware of overdiagnosis (Kale and Korenstein, 2018). This cultural mindset, reinforced by media, industry and performance targets, encourages both patients and clinicians to equate action with quality care (Pathirana et al., 2017).

The result is a cycle where reassurance is sought through testing and prescribing, but this very process fuels unnecessary diagnoses. Breaking the cycle requires acknowledging uncertainty, resisting defensive instincts and refocusing consultations on proportionate care.

The profile trap

While cancer screening dominates the literature, overdiagnosis in general practice often begins with the ‘routine’ blood test. The digitisation of pathology requesting has made it easy to order extensive testing profiles. A patient presenting with ‘Tired All The Time’ (TATT) may trigger a full blood count, inflammatory markers, thyroid function and a ‘liver screen’ with a single click.

Because laboratory reference ranges typically include the central 95% of a healthy population, by definition, 5% of healthy individuals will have a result outside the reference range on any given test (Doles et al., 2025). Bulk testing magnifies this: if a clinician orders 20 independent tests (a typical profile), the probability of a healthy patient having at least one abnormal result rises to approximately 1 – 0.95²⁰ ≈ 64%. This frequently triggers a cascade of investigation. A borderline liver enzyme result in an asymptomatic patient may lead to repeat tests, then an ultrasound, which identifies an incidental benign lesion, followed by ongoing surveillance. The initial aim was reassurance, but the outcome is medicalisation of a healthy individual.

Lack of continuity of care

This tendency to over-investigate is exacerbated by erosion of continuity in primary care (Baker and Jeffers, 2016). Relational continuity allows a GP to tolerate uncertainty. A doctor who knows a patient well may feel confident in ‘watchful waiting’ for vague symptoms. In fragmented systems where patients see multiple clinicians, the threshold for investigation drops (Pilvar and Watt, 2025). A clinician unfamiliar with the patient lacks the insight of previous consultations and the trust required for shared decision-making. Faced with marginally abnormal results or ill-defined complaints, the safer, defensive option is often to order more tests rather than to risk missing a diagnosis. Thus, the lack of continuity becomes a structural driver for overdiagnosis.

Limited consultation time

The structural constraints of UK general practice are another potent driver. With standard consultation times limited to 10 minutes, GPs must rapidly process undifferentiated symptoms, manage risk and maintain the therapeutic relationship. In this high-pressure environment, investigation can become a coping strategy. Surveys indicate that around 37% of physicians cite inadequate consultation time as a primary driver of overutilisation (Lyu et al., 2017). Qualitative research describes a ‘penalty’ for taking time to educate patients about low-value testing; namely, running behind schedule and compromising care for subsequent patients (Nguyen et al., 2024). To avoid this, clinicians admit to ordering tests or referrals they otherwise would not have chosen, effectively allowing patient care to ‘overflow’ into secondary services to manage the consultation clock (Pilvar and Watt, 2025).

Explaining diagnostic uncertainty is also cognitively demanding. It requires detailed safety-netting and reassurance to manage patient anxiety. By contrast, applying a definitive label (e.g. ‘depression’ for low mood or ‘sinusitis’ for a viral upper respiratory tract infection) offers immediate ‘cognitive closure’. The label validates the patient’s suffering and provides a clear endpoint to the consultation (Jutel, 2009). The system therefore incentivises the quick fix of a diagnosis over the slower, but often safer, strategy of watchful waiting.

Overdiagnosis is therefore not an inevitable by-product of medical progress, but the result of an interplay between detection technologies, professional practice, health system design and societal attitudes. Understanding both mechanisms and drivers is essential to developing effective strategies for reducing harm.

Population-level signatures of overdiagnosis

Overdiagnosis is not directly measurable in individuals; it becomes visible when zooming out to population trends (Adamson et al., 2024; Dunn et al., 2022; Kale and Korenstein, 2018). Rising incidence without a corresponding fall in disease-specific mortality, especially with stable rates of advanced-stage disease, suggests that many additional cases represent overdiagnosis rather than true reductions in lethal disease (Gupta et al., 2020; Welch et al., 2019).

Thyroid cancer is a classic example: widespread ultrasound use has greatly increased detection of small, slow-growing tumours, yet mortality remains stable, implying many are indolent (Dunn et al., 2022; Li et al., 2020). A similar discordant pattern is seen in kidney cancer and melanoma (Adamson et al., 2024; Dunn et al., 2022; Gupta et al., 2020; Welch et al., 2019).The absence of an increase in metastatic disease at diagnosis strengthens the interpretation that enhanced detection has expanded the net to catch harmless disease rather than reducing the burden of lethal cancer (Gupta et al., 2020).

Breast cancer trends illustrate a more complex ‘mixed signal’. With the introduction of mammography, incidence rose while mortality declined, suggesting both benefit and harm (Welch et al., 2019). Stable rates of advanced disease and trial data, such as the UK Age Trial, indicate that much of the additional detection reflects indolent tumours, while reduced mortality appears largely driven by improved treatments rather than screening alone (Duffy et al., 2020; Dunn et al., 2022).

Prostate cancer incidence has fluctuated markedly with prostate-specific antigen (PSA) testing, reflecting changes in screening rather than underlying disease prevalence (Welch et al., 2019). PSA testing is highly prone to detecting indolent cancers, and many men undergo radical treatment, with risks of incontinence and erectile dysfunction, for tumours unlikely to shorten life, reinforcing the need for shared decision‑making around testing (Dunn et al., 2022; Gupta et al., 2020).

In primary care, similar patterns arise not from tumours but from expanding definitions of cardiovascular risk. Lowering thresholds for hypertension, hypercholesterolaemia and ‘prediabetes’, and increasing use of risk calculators such as QRISK, dramatically expand the number labelled as ‘at risk’, many of whom would never have experienced harm at previous levels (Kale and Korenstein, 2018; NICE, 2023).

Unlike cancer, where overdiagnosis can be inferred from rising incidence without a fall in late-stage disease, the ‘signature’ in cardiovascular risk factors is less direct. Here, overdiagnosis emerges as a gradual expansion of disease prevalence as diagnostic boundaries shift – a different but equally important form of overdiagnosis, driven by policy and rooted in preventive medicine.

Systemic incentives and symptomatic conditions

While traditional definitions of overdiagnosis focus on asymptomatic screening (e.g. cancer or hypertension), primary care is increasingly grappling with diagnostic expansion in symptomatic conditions, particularly neurodevelopmental disorders such as attention-deficit/hyperactivity disorder (ADHD) and autism. Unlike cancer screening, where the harm is often physical, the drivers here are frequently societal.

In the current UK context, a medical diagnosis often holds significant ‘social utility’. Access to educational support, welfare benefits and housing assistance is frequently gatekept by the requirement of a formal diagnostic label (Bell et al., 2022; Jutel, 2009). This creates a benevolent driver for overdiagnosis: clinicians may feel pressured to apply a label to a patient with borderline symptoms not because of strict medical necessity, but to ensure the patient receives essential non-medical support. This phenomenon sparks fierce debate. Critics argue that expanding diagnostic criteria pathologises normal human variation and diverts resources from those with severe impairment (Bodfield and Culshaw, 2023; Chiolero et al., 2015). For the GP, this presents a distinct ethical dilemma: withholding a label to prevent ‘overdiagnosis’ may inadvertently deny a patient access to the resources they need to function in society. Thus, the diagnosis becomes a currency for care, forcing doctors to balance the risk of medicalising normality against the risk of social exclusion.

Consequences of overdiagnosis

The consequences of overdiagnosis are wide-ranging, affecting individual patients, healthcare systems and society. While some effects are theoretical and difficult to quantify, others are well documented and carry significant clinical, psychological and economic implications.

Overtreatment

The most direct consequence of overdiagnosis is overtreatment. When indolent or clinically insignificant conditions are detected, interventions such as surgery, chemotherapy or long-term medication may be initiated despite offering no real benefit (Walker and Rogers, 2017). These interventions expose patients to avoidable risks, including surgical complications, infections and adverse drug effects (Gupta et al., 2020; Kale and Korenstein, 2018).

In cancer care, this is particularly evident. PSA screening detects many prostate cancers that may never become symptomatic, yet many men undergo surgery or radiotherapy, risking urinary and sexual dysfunction (Slawomirski et al., 2025). Similarly, thyroid cancer overdiagnosis often leads to thyroidectomy and lifelong hormone replacement without improving outcomes (Slawomirski et al., 2025).

Overtreatment also strains healthcare systems. Unnecessary interventions consume resources that could otherwise benefit patients with symptomatic or life-threatening disease, raising equity concerns in resource-limited settings (Gupta et al., 2020).

Psychological harms and disease labelling

Overdiagnosis can significantly affect mental wellbeing, partly through the assignment of diagnostic labels. Being told one has a ‘disease’ may trigger anxiety, uncertainty and altered life decisions, even when the condition is unlikely to cause harm (Kale and Korenstein, 2018). Women with ductal carcinoma in situ often experience persistent fear of recurrence despite excellent long-term prognosis, and men with small abdominal aortic aneurysms describe feeling as though they carry a ‘ticking bomb’ (Johansson et al., 2015).

Labels also influence behaviour, identity and perceptions of health. Benign heart murmurs in children have led to unnecessary activity restrictions, and incidental imaging findings for mild back pain are associated with increased disability claims and time off work (Kale and Korenstein, 2018)., Expanding diagnostic criteria for conditions such as ADHD and polcystic ovary syndrome (PCOS) can pathologise normal variation, leading to unnecessary treatment and social implications (Slawomirski et al., 2025). Disease labelling may have downstream consequences for employment, insurance and future interactions with healthcare services, highlighting broader societal ripple effects.

The terminology used to describe this problem also carries risk. The term ‘overdiagnosis’ itself can be counter-intuitive or distressing for patients, implying that genuine suffering is being dismissed or that previous care was negligent. Studies of patients who later learn they may have been overdiagnosed, for example, through screening, describe psychosocial distress, regret and a loss of trust in clinicians and health systems (Davies, 2024; McCaffery et al., 2019). Some studies suggest reserving ‘overdiagnosis’ for system-level and policy discussions, while using alternative framings such as ‘unnecessary’ or ‘low‑value’ care, or focusing on the overall balance of benefit versus harm and the goal of optimal rather than simply earlier diagnosis in patient‑facing communication (Wallis et al., 2025)

These psychological and social consequences highlight an ethical challenge for GPs: early detection may not always benefit patients if it undermines mental wellbeing or limits life opportunities. Careful communication, shared decision-making and judicious use of diagnostic terminology are essential to minimise harm.

Financial burden

Overdiagnosis also carries substantial financial consequences for both individuals and healthcare systems. In the USA, breast cancer overdiagnosis in women aged 40–59 years is estimated to cost $1.2 billion annually (Kale and Korenstein, 2018). Thyroid cancer overdiagnosis contributes to significant global expenditure on surgery and long-term hormone therapy (Slawomirski et al., 2025), and, in Australia, up to one in five cancers may represent overdiagnosis (Capurro et al., 2022).

These costs represent inefficiencies and divert limited resources away from interventions with proven population health benefits. At a societal level, rising expenditures may increase insurance premiums or reduce access to essential care, while patients may face harms from unnecessary interventions that reduce quality of life without clinical benefit (Gupta et al., 2020).

Addressing overdiagnosis

Overdiagnosis cannot be eradicated, but it can be anticipated, reduced and managed responsibly. Because most testing and screening decisions originate in primary care, strategies are particularly important here. To be effective, interventions must operate across four distinct levels: the individual clinician, medical education, the health system and public health.

The individual clinician

Shared decision-making

Many patients and clinicians assume that more testing is always better, and that early detection inevitably saves lives. These assumptions can drive unnecessary investigations when net benefit is uncertain. GPs can address this by discussing the possibility of overdiagnosis, particularly around screening decisions (Kale and Korenstein, 2018). Decision aids presenting absolute benefits and harms help patients understand trade-offs. Careful language also matters: describing a borderline blood pressure as ‘slightly raised’ rather than ‘hypertension’ may prevent premature labelling (Dunn et al., 2022).

Time as a diagnostic tool

In an environment where the ‘quick fix’ of a label is incentivised, time is often the most powerful diagnostic tool. For undifferentiated presentations, such as non-specific fatigue or transient lymphadenopathy, a ‘diagnostic pause’ or ‘test of time’ allows benign, self-limiting conditions to resolve while significant pathology declares itself. This is not passive inaction, but an active decision that requires robust safety-netting so patients feel safe to return.

Continuity and de-labelling

Perhaps the most potent tool in managing uncertainty is relational continuity. A strong doctor–patient relationship buffers anxiety and supports ‘watchful waiting’ through planned, active monitoring rather than immediate investigation. Prioritising continuity is therefore not just a logistical preference, but a structural necessity for managing risk without over-investigation. GPs can also protect patients by reviewing existing labels. For example, re-evaluating historical ‘asthma’ diagnoses or removing inaccurate drug allergy records, thereby preventing future cascades of unnecessary care.

Medical education

Normalising uncertainty

Overdiagnosis is fuelled by fear of litigation, the psychology of regret, and an ethos that frame missing disease as the greatest failure. These pressures drive clinicians to act ‘just in case’, even when the probability of benefit is marginal (Chiolero et al., 2015; Moynihan et al., 2012). Training programmes should normalise uncertainty, teach risk communication without defaulting to action, and emphasise that “watchful waiting” and safety netting are core elements of responsible care.

Research and evidence

Incorporating overdiagnosis into medical curricula and continuing professional development can recalibrate expectations (Pathirana et al., 2017). Priorities include clarifying the natural history of borderline abnormalities and developing transparent ways of quantifying overdiagnosis in trials.

Additionally, practical point-of-care tools (e.g. Cochrane, GP Evidence) can help clinicians present balanced, information, without increasing consultation burden.

The health system

Reforming incentives and guidelines

Financial incentives often encourage overdiagnosis by rewarding volume of activity rather than value of care. Remuneration and quality schemes should instead prioritise quality of life and safety, and guidelines should include explicit thresholds for when not to test or treat and clearly describe potential harms (Pathirana et al., 2017). At policy level, interventions that cause net harm should be phased out. South Korea’s reduction in thyroid cancer overdiagnosis after national screening changes illustrates what is possible (Li et al., 2020).

Optimising diagnostic criteria

Many conditions have seen diagnostic thresholds lowered over time, expanding the pool of people labelled as ‘diseased’ without clear evidence of benefit (Pathirana et al., 2017). Optimising criteria to reflect clinically meaningful disease, rather than biochemical variation or subclinical findings, can reduce unnecessary treatment and medicalisation. While awaiting updated guidelines, GPs can emphasise risk stratification and careful monitoring for borderline findings.

System safeguards

In the UK, the National Screening Committee and NICE already weigh harms as well as benefits when considering screening programmes and new technologies, providing important safeguards against overdiagnosis (NICE, 2022; UK National Screening Committee (UK NSC), 2019, 2020).

Public health

Public campaigns such as Choosing Wisely (https://www.choosingwisely.org/) and BMJ’s Too Much Medicine (https://www.bmj.com/too-much-medicine) are essential for reinforcing a cultural shift. They challenge the assumption that ‘more’ or ‘newer’ is always better. By raising public awareness about the risks of over-investigation, they can empower patients to ask critical questions about the necessity of tests and treatments, creating a more receptive environment for shared decision-making in the consultation room.

Mitigating overdiagnosis therefore requires more than individual restraint; it demands alignment between consultation practice, professional culture and the structures that determine what counts as disease in the first place.

Conclusions

Overdiagnosis is not simply an inevitable by-product of technological progress and broadened disease definitions; it emerges from the interaction of detection technologies, professional culture, health system design and societal expectations. Its psychological, social, financial and iatrogenic harms undermine both patient wellbeing and the sustainability of healthcare. While complete eradication is impossible, the evidence shows that overdiagnosis can be anticipated, reduced and managed more responsibly.

System-level safeguards in the UK, such as the UK National Screening Committee and NICE, already provide important checks, but they must evolve. Progress will require more systematic monitoring of overdiagnosis, transparent reporting of harms alongside benefits in trials, and routine incorporation of overdiagnosis metrics into health technology assessments, alongside activede-implementation of interventions that cause net harm, as seen in South Korea’s reversal of thyroid cancer screening.

Policy change alone is insufficient. Embedding the principles of proportionate, evidence-based care into clinical training, public communication, incentive structures and everyday practice is essential. This includes recognising the limits of labels, acknowledging the potential harms of the term ‘overdiagnosis’ in patient communication, and reframing early diagnosis in terms of optimal rather than simply earlier detection.

GPs are uniquely positioned at the interface of population strategies and individual care. Through relational continuity, judicious testing, thoughtful use of time as a diagnostic tool and clear, balanced communication, GPs can help patients navigate uncertainty while avoiding unnecessary interventions. For GPs, the goal is clear: diagnose wisely, treat responsibly and prioritise patient wellbeing over the mere accumulation of labels.

Key points

Overdiagnosis is a correct diagnosis of a condition that would not have caused harm, distinct from misdiagnosis and false positives

Population-level trends can reveal overdiagnosis in cancer and chronic disease

Overtreatment, psychological distress, harmful labelling, and financial burden are major consequences for patients and healthcare systems

Technological sensitivity, bulk testing, lowered diagnostic thresholds, limited consultation time, loss of continuity, defensive practice and societal incentives for labels all contribute to overdiagnosis in UK general practice

‘Overdiagnosis’ may be best reserved for system-level discourse, while patient-facing conversations focus on unnecessary or low-value care and the balance of benefit versus harm

Key strategies include shared decision-making, using time as a diagnostic tool, robust safety-netting, continuity and de-labelling, supported by education, evidence-based guidelines, cautious disease definition and UK safeguards

Footnotes

ORCID iD

Chiara Zaffaroni

References

Adamson

Naik

Jones

, et al. (2024) Ecological study estimating melanoma overdiagnosis in the USA using the lifetime risk method. BMJ Evidence-Based Medicine 29(3): 156–161. DOI: 10.1136/bmjebm-2023-112460.

Baker

Jeffers

(2016) Continuity of Care in Modern Day General Practice. London: Royal College of General Practitioners. Available at: www.rcgp.org.uk/getmedia/11f26527-5d11-47f2-a593-1a894c2fff1b/Continuity-of-care-in-modern-day-general-practice1.pdf (accessed 3 February 2026).

Bell

Doust

Sanders

, et al. (2022) A novel methodological framework was described for detecting and quantifying overdiagnosis. Journal of Clinical Epidemiology 148: 146–159. DOI: 10.1016/j.jclinepi.2022.04.022.

Berland

Silverman

Gore

, et al. (2010) Managing incidental findings on abdominal CT: white paper of the ACR Incidental Findings Committee. Journal of the American College of Radiology 7(10): 754–773. DOI: 10.1016/j.jacr.2010.06.013.

Bodfield

Culshaw

(2023) The place for diagnosis in the UK education system? Emotional and Behavioural Difficulties 28(4): 316–328. DOI: 10.1080/13632752.2024.2316398.

Bohr

Memarzadeh

(2020) The rise of artificial intelligence in healthcare applications. In: Bohr

Memarzadeh

(eds) Artificial Intelligence in Healthcare. Academic Press, pp. 25–60. DOI: 10.1016/B978-0-12-818438-7.00002-2.

Brodersen

Schwartz

Heneghan

, et al. (2018) Overdiagnosis: what it is and what it isn’t. BMJ Evidence-Based Medicine 23(1): 1–3. DOI: 10.1136/ebmed-2017-110886.

Capurro

Coghlan

Pires

DEV

(2022) Preventing digital overdiagnosis. JAMA 327(6): 525. DOI: 10.1001/jama.2021.22969.

Chiolero

Paccaud

Aujesky

, et al. (2015) How to prevent overdiagnosis. Swiss Medical Weekly 145: w14060. DOI: 10.4414/smw.2015.14060.

10.

Davies

(2024) Informed by research, transformed by research. BMJ Evidence-Based Medicine 29(1): 62–63. DOI: 10.1136/bmjebm-2023-112646.

11.

Diamandis

(2016) The side effects of translational omics: overtesting, overdiagnosis, overtreatment. Clinical Chemistry and Laboratory Medicine (CCLM) 54(3): 389–396. DOI: 10.1515/cclm-2015-0762.

12.

Divo

Martinez

Mannino

(2014) Ageing and the epidemiology of multimorbidity. European Respiratory Journal 44(4): 1055–1068. DOI: 10.1183/09031936.00059814.

13.

Doles

Ye Mon

Shaikh

, et al. (2025) Interpretating normal values and reference ranges for laboratory tests. The Journal of the American Board of Family Medicine 38(1): 174–179. DOI: 10.3122/jabfm.2024.240224r1.

14.

Duffy

Vulkan

Cuckle

, et al. (2020) Effect of mammographic screening from age 40 years on breast cancer mortality (UK Age trial): final results of a randomised, controlled trial. The Lancet Oncology 21(9): 1165–1172. DOI: 10.1016/s1470-2045(20)30398-3.

15.

Dunn

Woloshin

Xie

, et al. (2022) Cancer overdiagnosis: a challenge in the era of screening. Journal of the Naional Cancer Center 2(4): 235–242. DOI: 10.1016/j.jncc.2022.08.005.

16.

Forbes

Marchand

Doran

, et al. (2017) The role of the Quality and Outcomes Framework in the care of long-term conditions: a systematic review. British Journal of General Practice 67(664): e775–e784. DOI: 10.3399/bjgp17x693077.

17.

Gupta

Koul

(2020) Overdiagnosis and overtreatment; how to deal with too much medicine. Journal of Family Medicine and Primary Care 9(8): 3815–3819. DOI: 10.4103/jfmpc.jfmpc_433_20.

18.

Hofmann

Reid

Carter

, et al. (2021) Overdiagnosis: one concept, three perspectives, and a model. European Journal of Epidemiology 36(4): 361–366. DOI: 10.1007/s10654-020-00706-4.

19.

Johansson

Hansson

Brodersen

(2015) Estimating overdiagnosis in screening for abdominal aortic aneurysm: could a change in smoking habits and lowered aortic diameter tip the balance of screening towards harm? BMJ 350: h825. DOI: 10.1136/bmj.h825.

20.

Jutel

(2009) Sociology of diagnosis: a preliminary review. Sociology of Health & Illness 31(2): 278–299. DOI: 10.1111/j.1467-9566.2008.01152.x.

21.

Kale

Korenstein

(2018) Overdiagnosis in primary care: framing the problem and finding solutions. BMJ 362: k2820. DOI: 10.1136/bmj.k2820.

22.

Maso

Vaccarella

(2020) Global trends in thyroid cancer incidence and the impact of overdiagnosis. The Lancet Diabetes & Endocrinology 8(6): 468–470. DOI: 10.1016/s2213-8587(20)30115-7.

23.

Lyu

Brotman

, et al. (2017) Overtreatment in the United States. PLOS ONE 12(9): e0181970. DOI: 10.1371/journal.pone.0181970.

24.

McCaffery

Nickel

Pickles

, et al. (2019) Resisting recommended treatment for prostate cancer: a qualitative analysis of the lived experience of possible overdiagnosis. BMJ Open 9(5): e026960. DOI: 10.1136/bmjopen-2018-026960.

25.

McCartney

McCutcheon

Cooke

, et al. (2023) Investigation into financial conflicts of interest and screening for atrial fibrillation in the UK: a cross-sectional study. BMJ Evidence-Based Medicine 28(1): 15–20. DOI: 10.1136/bmjebm-2022-112004.

26.

Moynihan

Doust

Henry

(2012) Preventing overdiagnosis: how to stop harming the healthy. BMJ 344: e3502. DOI: 10.1136/bmj.e3502.

27.

Nguyen

M-LT

Honcharov

Ballard

, et al. (2024) Primary care physicians’ experiences with and adaptations to time constraints. JAMA Network Open 7(4): e248827. DOI: 10.1001/jamanetworkopen.2024.8827.

28.

NICE (2022) NICE Health Technology Evaluations: The Manual. Available at: www.nice.org.uk/process/pmg36/chapter/introduction-to-health-technology-evaluation (accessed 8 September 2025).

29.

NICE (2023) Cardiovascular disease: risk assessment and reduction, including lipid modification. Available at: www.nice.org.uk/guidance/ng238/chapter/rationale-and-impact (accessed 8 September 2025).

30.

Pathirana

Clark

Moynihan

(2017) Mapping the drivers of overdiagnosis to potential solutions. BMJ. DOI: 10.1136/bmj.j3879.

31.

Pilvar

Watt

(2025) The effect of workload on primary care doctors on referral rates and prescription patterns: evidence from English NHS. The European Journal of Health Economics 26(5): 817–837. DOI: 10.1007/s10198-024-01742-7.

32.

Slawomirski

Kelly

de Bienassis

, et al. (2025) The economics of diagnostic safety. OECD Health Working Papers. DOI: 10.1787/fc61057a-en.

33.

UK National Screening Committee (2019) Breast Cancer. GOV.UK. Available at: https://view-health-screening-recommendations.service.gov.uk/breast-cancer (accessed 8 September 2025).

34.

UK National Screening Committee (2020) Prostate Cancer. GOV.UK. Available at: https://view-health-screening-recommendations.service.gov.uk/prostate-cancer (accessed 8 September 2025).

35.

Walker

Rogers

(2017) Defining disease in the context of overdiagnosis. Medicine, Health Care and Philosophy 20(2): 269–280. DOI: 10.1007/s11019-016-9748-8.

36.

Wallis

Donnelly

Sullivan

, et al. (2025) From early diagnosis to optimal diagnosis: a conference report on what excellent innovation should look like. BMJ Innovations 11(3): 161–164. DOI: 10.1136/bmjinnov-2024-001304.

37.

Welch

Kramer

Black

(2019) Epidemiologic signatures in cancer. New England Journal of Medicine 381(14): 1378–1386. DOI: 10.1056/nejmsr1905447.

38.

Wiener

Schwartz

Woloshin

(2013) When a test is too good: how CT pulmonary angiograms find pulmonary emboli that do not need to be found. BMJ 347: f3368. DOI: 10.1136/bmj.f3368.