top of page
Latest Insights
(1) Alzheimer’s disease: how close are we to a cure, and what impact would it have on patients and society?
Alzheimer’s disease (AD), the seventh leading cause of death in the United States in 2022 (Kumar et al., 2024), is a neurodegenerative condition with insidious onset and progressive impairment of behavioural and cognitive functions. As life expectancy rises, Alzheimer’s disease will become increasingly prevalent, with a predicted increase to over 1.4 million by 2024 solely in the UK (Cooper, 2026). The sheer number of individuals diagnosed with AD demonstrates that the search for a cure is ever so urgent. So, how close is humanity to restoring identity, autonomy and memories? Though a cure is highly unlikely in the immediate future, recent scientific advances suggest AD could be reversed to achieve full neurological recovery, revolutionising the modern society by offloading immense strain from health infrastructure, catalysing economic growth, and restoring connections between patients and loved ones.
To understand the challenges in finding a cure for Alzheimer’s disease, it is important to understand its pathology. Amyloid plaques and neurofibrillary tangles are considered hallmarks of AD (Chinthapalli, 2014). Amyloid plaques are made up of degenerating neuronal material surrounding deposits of the beta-amyloid protein, which is derived from the amyloid precursor protein. Neurofibrillary tangles are twisted protein fibres found within nerve cells, consisting of a normal neuronal protein called tau. When misprocessed through hyperphosphorylation, tau molecules clump together to form tangles (Britannica, 1998). These abnormalities interfere with communication between neurons, eventually leading to neuronal death. From this, the basis of the amyloid hypothesis is formed, which proposes that beta-amyloid accumulation, where the molecules generated more quickly than their removal, initiates a signalling sequence of neurodegenerative processes, such as synaptic disruption and mitochondrial dysfunction (Sheppard, 2020). During these pathological processes, patients experience disorientation, mood shifts, and recent memory loss. In theory, the use of anti-amyloid medication should slow the progression of Alzheimer’s disease. However, these drugs only address a small fraction of a far more complex issue. Studies revealed that the hyperphosphorylation of tau was independent of beta-amyloid accumulation, inferring that multiple disease pathways may contribute to neurodegeneration in AD (An et al, 2008). Furthermore, several medications were successful in removing amyloid plaques from the brain, but did not slow cognitive deterioration (Anon, 2023). This suggests that amyloid deposition is a contributing factor to AD rather than an underlying cause. Unsuccessful results from anti-amyloid medication has created criticism of the amyloid hypothesis, encouraging other theories to emerge. This lack of scientific consensus for the disease pathway of AD highlights the incomplete understanding of its mechanisms, hence exemplifying why manufacturing a cure is so difficult. Additionally, individuals who carry autosomal dominant AD mutations can remain symptom-free for decades before the onset clinical manifestations (Chaubey et al., 2025). This means that irreversible brain damage has already occurred by the time cognitive decline becomes noticeable, hindering scientists from understanding the full pathology of AD. Furthermore, the disease is not caused by a single pathway but rather influenced by a combination of multiple genes with lifestyle and environmental factors (NIA, 2023). Consequently, the multifactorial nature of AD must be fully addressed in treatments, which complicates approaches to a universal cure.
The known complexity of the disease has discouraged many researchers from supporting the possibility of neurological reversal in Alzheimer’s disease. However, is such a reversal as easy as adopting a healthier lifestyle? A randomised, controlled clinical trial led by Dr Dean Ornish investigated the effect of a whole-foods diet, moderate aerobic exercise for at least 30 minutes a day, stress management including meditation, and support groups for patients and spouses (News desk, 2024). The results found a significant correlation between lifestyle and both the plasma Aβ42/40 ratio (a biomarker linked to beta-amyloid accumulation) and cognitive function (Ornish et al., 2024), meaning there was reduced amyloid deposition, which may indicate improved synaptic function and overall brain productivity. However, the study excluded patients with severe AD, suggesting their reversal may prove to be more complicated. Moreover, despite patients demonstrating increased cognitive function, a full return of memory was not noted, suggesting lifestyle changes alone cannot reverse AD. Another recent study using mouse models fosters hope towards advanced AD patients by demonstrating the potential reversibility of cognitive decline using the experimental compound P7C3-A20. Unlike traditional immunotherapy drugs that reduce the impact of AD, P7C3-A20 restores nicotinamide adenine dinucleotide (NAD+), reverses tau phosphorylation, blood-brain barrier deterioration, oxidative stress, and DNA damage, resulting in full cognitive recovery (Chaubey et al., 2025). Recovering NAD+ was an especially groundbreaking discovery, as it plays a critical role in energy production and cell survival, suggesting neurons in the brain could be free of degeneration by Alzheimer’s disease. These findings controverted the traditional stance that AD is irreversible and suggested that its treatment should be more focused on stabilising the brain’s metabolic energy levels rather than simply removing plaque buildup that does not address the underlying cause. While a universal cure using P7C3-A20 seems extremely feasible, it is important to note the early experimental stages the study was conducted in, such as mice which differ in brain complexity to humans. Ultimately, recent scientific breakthroughs involving both lifestyle changes and experimental drugs provide hope for a cure, yet further evidence is still necessary to establish a basis for a cure for Alzheimer’s disease.
It is estimated that there are fourteen to fifteen million unpaid Alzheimer’s caregivers devoting seventeen billion hours to looking after a relative or friend with the disease, and that their combined economic loss is greater than $200 billion per year (Petsko, 2026). With a cure potentially on the horizon, the immense societal burden of AD could be alleviated, allowing millions of these caregivers to return to full-time employment and daily life. In principle, global economies would experience a massive surge in labour participation, improving economic productivity. With cognitive function and independence of the elderly maintained, a “longevity society” would dominate, meaning there are even more individuals contributing to financial growth. A cure to AD would also reduce government expenditure on dementia care, significantly relieving financial strain on public healthcare systems such as the NHS. Furthermore, a cure to AD could promote social cohesion between families due to the likely growing prominence of multi-generational housing. However, along with its immediate benefits, creating a cure that affects millions is bound to also carry its own challenges. With AD prevalence doubling every 5 years after age 65 (Morris, 2026), a cure would fundamentally alter the trajectory of global ageing. Demographics would shift toward an exponential rise in the 85+ and 100+ age brackets, potentially increasing demand for healthcare, adult social care, and residential care. Moreover, the systemic issue of unequal access to these treatments could prevent lower-income regions from growing economically, exacerbating current global inequalities. In sum, the impact of a cure would reshape society by alleviating pressure on public systems, whilst simultaneously introducing new social and economic shifts due to a changing age distribution.
The effect of a cure on AD patients would be immensely transformative, restoring autonomy and allowing patients to spend valuable time with their families without the persistent sense of emotional burden. Currently, symptoms of the disease can be physically and emotionally deterring; patients lose memories of loved ones, find difficulty in swallowing, and even lose the ability to speak. As the condition progresses, these symptoms only worsen, meaning dependency on others is more apparent than ever. A cure would preserve dignity by enabling individuals to carry out daily tasks unaided that are otherwise stripped away by the disease, such as shopping for groceries, cooking traditional family recipes, or reading bedtime stories to grandchildren. Overall, the emotional well-being of patients would be significantly enhanced, highlighting the groundbreaking impact a cure could potentially cultivate.
While previous treatments primarily focused on slowing symptom progression of AD, contemporary research has opened pathways to manufacture drugs that target the root causes of the disease, such as how neurons in the brain initially deteriorate. Scientific optimism for a cure is now more realistic than ever, with promising trials investigating both non-pharmacological and pharmacological interventions. However, with a mixture of success and failure, it is evident that further research is needed, and so a cure in the immediate future is unlikely. Hypothetically, the emergence of a cure would relieve the economic and emotional burden placed on social and healthcare services, acting as an economic enabler for society. On a more personal level, a cure would also restore identity, independence, and human connection to millions of people, reshaping patient experience and society as a whole.
Works Cited
An, Yuhui, et al. “Main Hypotheses, Concepts and Theories in the Study of Alzheimer’s Disease.” Life Science Journal, vol. 5, no. 4, 1 Jan. 2008, www.researchgate.net/profile/Chun-Xia-Yao/publication/228483723_Main_hypotheses_concepts_and_theories_in_the_study_of_Alzheimer.
Anon. “The Amyloid Hypothesis of Alzheimer’s: Are We on the Right Track?” Alzheimer’s Research Association, 31 Aug. 2023, www.alzra.org/blog/the-amyloid-hypothesis-of-alzheimers-are-we-on-the-right-track/.
Britannica , The . “Alzheimer Disease | Definition, Causes, Symptoms, & Treatment.” Encyclopedia Britannica, 20 July 1998, www.britannica.com/science/Alzheimer-disease?utm_source.
Chaubey, Kalyani, et al. “Pharmacologic Reversal of Advanced Alzheimer’s Disease in Mice and Identification of Potential Therapeutic Nodes in Human Brain.” Cell Reports Medicine, Dec. 2025, p. 102535, www.cell.com/cell-reports-medicine/fulltext/S2666-3791(25)00608-1?_returnURL=https%3A%2F%2Flinkinghub.elsevier.com%2Fretrieve%2Fpii%2FS2666379125006081%3Fshowall%3Dtrue, https://doi.org/10.1016/j.xcrm.2025.102535. Accessed 31 Dec. 2025.
Chinthapalli, Krishna. “Still a Perplexing Problem.” BMJ: British Medical Journal, vol. 349, 2014. JSTOR, www.jstor.org/stable/26516517, https://doi.org/10.2307/26516517.
Cooper, Emma . Google.com, 2026, www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=www.alzheimersresearchuk.org/news/dementia-is-still-uks-biggest-killer-where-do-we-go-from-here/&ved=2ahUKEwiGx6PTlsGUAxVKTkEAHblNOH4QFnoECBoQAQ&usg=AOvVaw2XG6CMCjDf5RldFukAoKVH. Accessed 10 May 2026.
Kumar, Anil, et al. “Alzheimer Disease.” National Library of Medicine, StatPearls Publishing, 12 Feb. 2024, www.ncbi.nlm.nih.gov/books/NBK499922/.
Morris, John. “Is Alzheimer’s Disease Inevitable with Age?” Google.com, 2026, www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=pmc.ncbi.nlm.nih.gov/articles/PMC409830/&ved=2ahUKEwi_qOC81M-UAxUHT0EAHXfzK6QQFnoECCYQAQ&usg=AOvVaw1ADRxRigmjahITYf_oazFJ. Accessed 23 May 2026.
National Institute On Aging. “Alzheimer’s Disease Genetics Fact Sheet.” National Institute on Aging, 2023, www.nia.nih.gov/health/alzheimers-causes-and-risk-factors/alzheimers-disease-genetics-fact-sheet.
News desk. “Lifestyle Changes Significantly Improve Cognition and Function in Early Alzheimer’s in Study First.” Neuro Rehab Times - the World’s Leading Neurorehabilitation Magazine, 7 June 2024, nrtimes.co.uk/study-lifestyle-changes-significantly-improve-cognition-and-function-in-early-alzheimers-disease-for-the-first-time-in-a-randomized-controlled-trial/. Accessed 22 May 2026.
Ornish, Dean, et al. “Effects of Intensive Lifestyle Changes on the Progression of Mild Cognitive Impairment or Early Dementia due to Alzheimer’s Disease: A Randomized, Controlled Clinical Trial.” Alzheimer’s Research & Therapy, vol. 16, no. 1, 7 June 2024, p. 122, rdcu.be/dK2yn, https://doi.org/10.1186/s13195-024-01482-z.
Petsko, Gregory. “The Coming Epidemic of Neurologic Disorders: What Science Is — and Should Be — Doing about It.” Google.com, 2026, www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=direct.mit.edu/daed/article/141/3/98/27048/The-Coming-Epidemic-of-Neurologic-Disorders-What&ved=2ahUKEwiQg9-QmM-UAxXAd0EAHc8fCA4QFnoECBgQAQ&usg=AOvVaw3teeDjLWuaF-yrfKsJgWzH. Accessed 23 May 2026.
Sheppard, Olivia, and Michael Coleman. “Alzheimer’s Disease: Etiology, Neuropathology and Pathogenesis.” PubMed, Exon Publications, 18 Dec. 2020, www.ncbi.nlm.nih.gov/books/NBK566126/.
(2) The uses of pharmacological and nonpharmacological treatments
Pharmacological treatments, the treatment of disease through the administration of drugs to mitigate or manage pathological processes (Kirker-Head et al., 2014), are often pitted against non-pharmacological interventions (NPIs), therapies and measures that do not involve taking medication (Kooijmans et al., 2024). Over time, medicine has evolved greatly, from willow bark used over 3,500 years ago for pain relief to aspirin used in the 21st century. With a swallow of a pill, symptoms of pain, fever and inflammation are alleviated within 30 minutes, and it is this level of instant efficacy that popularised medicinal drugs. However, while pharmacological treatments have, in many ways, revolutionised modern medicine, the widespread accessibility of medicines has created an over-reliance on drugs. This can initiate a cascade of challenges, including antimicrobial resistance, iatrogenic harm, and adverse side effects. Other NPIs, such as psychological therapies and therapeutic activities, should be emphasised and used in tandem with drug-based solutions to reduce the reliance on drugs. The two intervention techniques address the different dimensions of the disease, thereby supporting a holistic, patient-centred approach to care (Castellano-Tejedor, 2022).
Pharmacological treatments are a cornerstone of modern medicine, underpinning both the management and treatment of disease. Drug treatments act by binding to molecular targets, altering their activity to produce therapeutic effects (Crowley, 2022). This process highlights the precision of pharmacological treatments, as they can target disease mechanisms directly. The specificity of drugs can be shown by the example of SSRIs targeting serotonin reuptake in depression and directly increasing serotonin activity (Chu et al., 2023). Hence, pharmacological treatments enable rapid symptom relief that may not be achieved with more generalised, non-pharmacological approaches. The undeniable efficacy of pharmacological treatment can also surpass that of NPIs in serious diseases that cannot be managed by lifestyle changes. An example is taking insulin for Type 1 Diabetes, an autoimmune disease of the pancreas that causes decreased insulin secretion (Chakrabarti et al., 2002). Without insulin, patients can develop severe hyperglycaemia and ultimately, life-threatening diabetic ketoacidosis (Lucier, 2024). Before the availability of insulin, T1D was fatal within 3 years. Now, life expectancies are close to normal. A Finnish study involving 42,936 persons with T1D estimated that the remaining life expectancy at age 20 was 51.64 years, only 9.88 years lower than that of the general Finnish population (Arffman et al., 2023). Drugs can play a pivotal role in reducing healthcare costs and alleviating strain on medical systems. For instance, prescribing antihypertensive drugs to prevent strokes could avoid a costly toll of $59,000 in the US if hospitalisation and long-term rehab are needed. Overall, the importance of pharmacological treatments is marked by their efficacy, especially in severe cases, their specificity in targeting action to areas, and, in many cases, lasting affordability.
Pharmacology has advanced significantly, particularly since World War II, with the discovery of transformative drugs such as Streptomycin for tuberculosis. Despite these developments offering instant and targeted relief, excessive drug reliance could lead to a plethora of risks. A growing concern is antimicrobial resistance (AMR), the phenomenon in which microorganisms develop the ability to survive in antimicrobial medicines that normally kill them (Klemperer et al., 2024). Repeated drug use exerts continuous selective pressure on bacteria, thus promoting resistance through horizontal gene transfer and genetic changes (Bala K. Nimmana et al., 2026). AMR is a slow-paced epidemic, whereby infections effect more than 2.8 million people a year in the US, resulting in around 35,000 deaths annually (CDC, 2024). In addition to prompting direct mortality and morbidity, AMR jeopardises medical advances in cancer, organ transplants, burns, and joint replacement. (Bristol, 2020). Furthermore, an over-reliance on drugs raises concerns about iatrogenic harm, including opioid dependence. Opioid prescriptions have increased by 300% in the US since 1999, despite there not being a change in the overall pain reported by people (McCarthy, 2015). While opioids aim to provide relief from chronic pain, dependence on them can increase the probability of side effects such as sedation, dizziness, nausea, vomiting, constipation, and respiratory depression. Tolerance to these symptoms rarely develops (Benyamin et al., 2008), resulting in a drastic decline in quality of life. Moreover, an over-reliance on drugs can lead to polypharmacy, defined as the regular use of 5 or more medications at the same time (Masnoon, 2017). Polypharmacy is especially concerning in older adults with multiple chronic health conditions, making them especially susceptible to medication-related harm (Varghese et al., 2024). Supporting this, the Moli-sani study found that hospitalisation and mortality were the main outcomes of polypharmacy in older adults, with a 30% increased risk of mortality and a 61% higher risk of hospitalisation than those not taking multiple medications (Costanzo, 2024).
Non-pharmacological treatments are also utilised in medical care and often serve as long-term measures for managing a condition, reducing symptoms and risk of disease progression. Hypertension ( a high blood pressure disorder prevalent in the obese population) can be intervened earlier with lifestyle changes that reduce cardiac risk. A randomised controlled trial assigned 98 hypertensive patients to 4 groups: yoga, brisk walking, reduced salt intake, and a control. The data indicated a significant BP reduction in all 3 non-pharmacological intervention groups (Fig 1a) (Soudarssanane et al., 2011). Whilst drugs may lower the BP quickly, NPIs offer long-term benefits by addressing underlying lifestyle-related issues and reducing the risk of disease progression. Drug-free interventions are also generally safer and side-effect-free , enhancing or supporting regular physiological processes, without the introduction of exogenous substances. Most NPIs, such as physical therapy, pose minimal threat to a person’s health, aside from the possibility of muscle strain and fatigue. Conversely, pharmacological treatments are sometimes associated with adverse outcomes, such as NSAIDs leading to jaundice, fulminant hepatitis, liver necrosis, and hepatic failure (Anon, 2026). NPIs are also advantageous for their holistic, patient-centred approach, allowing a greater breadth of patient autonomy and independence. Whereas pharmacological treatments focus primarily on the patient’s physiological symptoms, NPIs address the emotional, cognitive, and social factors of health. For instance, Alzheimer’s disease is a neurodegenerative disorder that significantly impairs memory, cognitive function, and the ability to perform daily tasks independently. Research shows that social activities can improve the quality of life of people with Alzheimer’s, prevent Alzheimer’s-related apathy, and reduce the amount of care needed (Anon, 2006). By promoting social engagement, the patient can overcome psychological barriers such as conversational difficulties, thereby retaining independence for a prolonged period. Overall, NPIs are essential in creating long lasting benefits and promoting patient-centred care, all whilst posing minimal side effects.
Whilst both pharmacological and non-pharmacological interventions have benefits, it is important to note that each has limitations when used in isolation. NPIs often target the root causes of the disease, addressing underlying lifestyle issues and emotional factors. However, using them as a sole treatment will not provide sufficient improvements to physiological symptoms of severe conditions such as advanced heart failure. Contrarily, pharmacological treatments can provide instant relief but may only foster passive coping styles and fail to address other factors that promote disease progression (Becker et al., 2017). Combining the two therapies has the potential to increase effectiveness compared to either treatment alone, as they may lead to beneficial impacts across different domains. For instance, ADHD medication is used to target the main symptoms, such as inattention and hyperactivity, and drug-free treatment targets secondary problems and associated, coexisting conditions. Combining the two treatments is also found to have long-lasting impacts through the development of cognitive skills, and instant effects through medication of ADHD symptoms (National Guideline Centre, 2018). In a meta-analysis of 101 trials and 11,910 participants with depression, the results showed 27% were more likely to respond to combination treatment (their symptoms reduced by half) than receiving therapy alone, and 25% more likely than receiving drug treatment alone (National Institute for Health and Care Research, 2020). It is important to recognise that the investment of combined therapy is currently more expensive to deliver than a single treatment. However, the 3E model in diabetes management (Fig 1b) exemplifies how integrated non-pharmacological and pharmacological approaches can achieve significant reductions in costs without compromising care quality (Raghav et al., 2025). Hence, there is future potential for integrated care that is both effective in managing symptoms and economically sustainable for both patients and healthcare systems.
Over the decades, pharmacological treatments have irrefutably transformed the medical field; drugs act as effective, target-specific, and, in many cases, cost-effective interventions against disease. However, in recent years, its widespread popularity has fostered a global over-reliance. This has resulted in polypharmacy, heightening the risk of adverse symptoms (Varghese et al., 2024), and also contributed to antimicrobial resistance, which endangers various medical advances such as organ transplants (Bristol, 2020). However, these challenges can be mitigated by incorporating nonpharmacological interventions, which complement the work done by pharmacological treatments, thereby reducing the dependency on drugs. Not only does this integration reduce drug over-reliance, but it also promotes holistic care, encouraging patients to take an active role in their care and gain a sense of independence.
Reference list
Arffman, M., Hakkarainen, P., Keskimäki, I., Oksanen, T. and Sund, R. (2023). Long-term and Recent Trends in Survival and Life Expectancy for People with Type 1 Diabetes in Finland. Diabetes Research and Clinical Practice, [online] 198, p.110580. doi:https://doi.org/10.1016/j.diabres.2023.110580.
Becker, W.C., Dorflinger, L., Edmond, S.N., Islam, L., Heapy, A.A. and Fraenkel, L. (2017). Barriers and facilitators to use of non-pharmacological treatments in chronic pain. BMC Family Practice, [online] 18(1). doi:https://doi.org/10.1186/s12875-017-0608-2.
Benyamin, R., Trescot, A., Datta, S., Buenaventura, R., Adlaka, R., Sehgal, N., Glaser, S. and Vallejo, R. (2008). Opioid Complications and Side Effects. [online] Pain physician. Available at: https://pubmed.ncbi.nlm.nih.gov/18443635/.
Bristol, N. (2020). The U.S. Government and Antimicrobial Resistance. [online] JSTOR. Available at: https://www.jstor.org/stable/resrep24759.
Castellano-Tejedor, C. (2026). Non-Pharmacological Interventions for the Management of Chronic Health Conditions and Non-Communicable Diseases. [online] Google.com. Available at: https://www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=https://pubmed.ncbi.nlm.nih.gov/35886388/&ved=2ahUKEwi3r6nMh9CTAxX-VPEDHTxIHW8QFnoECAwQAQ&usg=AOvVaw13dKTJDDNh2CV51jJUJR4H [Accessed 2 Apr. 2026].
CDC (2024). Antimicrobial Resistance Facts and Stats. [online] Antimicrobial Resistance. Available at: https://www.cdc.gov/antimicrobial-resistance/data-research/facts-stats/index.html.
Chakrabarti, R. and Rajagopalan, R. (2002). Diabetes and insulin resistance associated disorders: Disease and the therapy. Current Science, [online] 83(12), pp.1533–1538. Available at: https://www.jstor.org/stable/24108177.
Chu, A. and Wadhwa, R. (2023). Selective serotonin reuptake inhibitors. [online] PubMed. Available at: https://www.ncbi.nlm.nih.gov/books/NBK554406/.
Costanzo, S., Augusto Di Castelnuovo, Panzera, T., Curtis, A.D., Falciglia, S., Mariarosaria Persichillo, Chiara Cerletti, Donati, M.B., Gaetano, G. de and Iacoviello, L. (2024). Polypharmacy in Older Adults: The Hazard of Hospitalization and Mortality is Mediated by Potentially Inappropriate Prescriptions, Findings From the Moli-sani Study. International Journal of Public Health, 69. doi:https://doi.org/10.3389/ijph.2024.1607682.
Crowley, R. (2023). How Do Medicines Work? | National Institute of General Medical Sciences. [online] Nih.gov. Available at: https://nigms.nih.gov/biobeat/2023/10/how-do-medicines-work.
D. Nguyen, A. and K. Nimmana, B. (2026). [online] Google.com. Available at: https://www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=https://www.ncbi.nlm.nih.gov/books/NBK513277/&ved=2ahUKEwiqpLLVy8-TAxXSTEEAHfMRMq4QFnoECBgQAQ&usg=AOvVaw0H0tG98V_G9Vba3ROnk1tV [Accessed 2 Apr. 2026].
Google.com. (2006). Alzheimer’s disease: Learn More – Non-drug Interventions for Alzheimer’s Disease. [online] Available at: https://www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=https://www.ncbi.nlm.nih.gov/books/NBK279355/&ved=2ahUKEwiI44if9c-TAxUnZ0EAHYa6ED4QFnoECBgQAQ&usg=AOvVaw1vozeTqVQZZVU_tC54-fXS [Accessed 2 Apr. 2026].
Google.com. (2026). Available at: https://www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=https://www.ncbi.nlm.nih.gov/books/NBK547742/&ved=2ahUKEwjpvtfO9cqTAxU6T0EAHapfCoQQFnoECCoQAQ&usg=AOvVaw0YcEeGhAduzU8CJpC6GnFR [Accessed 31 Mar. 2026].
Kirker-Head, C.A. and Feldmann, H. (2014). Pharmacotherapy of joint and tendon disease. Equine Sports Medicine and Surgery, pp.473–502. doi:https://doi.org/10.1016/b978-0-7020-4771-8.00023-5.
KLEMPERER, K., GUZMAN, J. and MCDONNELL, A. (2024). Stopping Drug Resistance in Its Tracks. [online] Center for Global Development. doi:https://doi.org/10.2307/resrep63546.
Kooijmans, E.C.M., Hoogendijk, E.O., Drapała, N., Antonenko, O, Burchell, G.L., Barańska, I., Pokladníková, J., Szczerbińska, K., Fialová, D., van Hout, H.P.J. and Joling, K.J. (2024). Defining and Categorizing Nonpharmacologic Interventions in the Older Population: A Systematic Review. Journal of the American Medical Directors Association, [online] 26(1), p.105306. doi:https://doi.org/10.1016/j.jamda.2024.105306.
Lucier, J. and Weinstock, R.S. (2024). Diabetes Mellitus Type 1. [online] Nih.gov. Available at: https://www.ncbi.nlm.nih.gov/books/NBK507713/.
Masnoon, N., Shakib, S., Kalisch-Ellett, L. and Caughey, G.E. (2017). What Is polypharmacy? a Systematic Review of Definitions. BMC Geriatrics, [online] 17(1), p.230. Available at: https://bmcgeriatr.biomedcentral.com/articles/10.1186/s12877-017-0621-2.
McCarthy, M. (2015). Opioids should be last resort to treat chronic pain, says draft CDC guideline. BMJ: British Medical Journal, [online] 351. doi:https://doi.org/10.2307/26523799.
Milac Passali (2022). Benefits and Significance in Pharmacology. Journal of Contemporary Medical Education, [online] 12(1). Available at: https://www.jcmedu.org/jcmedu-articles/benefits-and-significance-in-pharmacology-84597.html [Accessed 27 Mar. 2026].
National Guideline Centre (UK) (2018). Evidence review for combined pharmacological and non-pharmacological treatments review: Attention deficit hyperactivity disorder: diagnosis and management: Evidence review F. [online] PubMed. London: National Institute for Health and Care Excellence (NICE). Available at: https://www.ncbi.nlm.nih.gov/books/NBK578099/.
National Institute for Health and Care Research (2020). NIHR Evidence - Combined drug and psychological therapies may be most effective for depression - Informative and accessible health and care research. [online] evidence.nihr.ac.uk. Available at: https://evidence.nihr.ac.uk/alert/combined-drug-and-psychological-therapies-may-be-most-effective-for-depression/.
OECD. (2025). Integrated care: Health at a Glance 2025. [online] Available at: https://www.oecd.org/en/publications/2025/11/health-at-a-glance-2025_a894f72e/full-report/integrated-care_3e34a223.html.
Raghav, S., Kumar, S., Ashraf, H. and Khanna, P. (2025). Cost-effectiveness of the 3E model in diabetes management: a machine learning approach to assess long-term economic impact. Frontiers in Public Health, 13. doi:https://doi.org/10.3389/fpubh.2025.1571546.
RAYMOND, I.H. (1935). HOW DRUGS ACT. Science Progress (1933), [online] 30(118), pp.252–267. doi:https://doi.org/10.2307/43423811.
Soudarssanane, Mb., Thiruselvakumar, D., Sahai, A, Subramanian, H., Jayalakshmy, R., Navasakthi, D. and Saptharishi, L. (2011). Non-pharmacological interventions in hypertension: A community-based cross-over randomized controlled trial. Indian Journal of Community Medicine, [online] 36(3), p.191. doi:https://doi.org/10.4103/0970-0218.86519.
Strilciuc, S., Alecsandra Grad, D., Radu, C., Chira, D., Stan, A., Ungureanu, M., Gheorghe, A. and Muresanu, F.-D. (2021). The economic burden of stroke: a systematic review of cost of illness studies. Journal of Medicine and Life, [online] 14(5), pp.606–619. doi:https://doi.org/10.25122/jml-2021-0361.
Varghese, D., Koya, H.H., Ishida, C. and Patel, P. (2024). Polypharmacy. [online] National Library of Medicine. Available at: https://www.ncbi.nlm.nih.gov/books/NBK532953/.
(3) Will AI become a Doctor?
ChatGPT or other AI tools are used on a daily basis. Whether this be asking what the weather is like tomorrow, or to help write an email, or even what you should cook for dinner. As AI is becoming a huge part of our lives, reliance on it reaching record high levels. But – to what extent is this reliance? Would you trust AI to tell you that you have cancer? Would you trust it to decide your treatment?
The role of AI in medicine is big hot topic now, many say it has the potential to improve the accessibility and efficiency of the healthcare sector using pattern analysis in a way no human ever could, or some worry that it could replace the jobs of young individuals fresh out of university with a painstakingly earned degree.
In this discussion, I will explain how AI is already integrated, how it could be used in the future, assess the both the positives and negatives, including ethical concerns, and ultimately answering the question: Can AI become a doctor?
Before delving into how AI be applied, we first need to know what it is. AI, or artificial intelligence, can mean many different things, as shown in this image. Simply put, AI refers to the science and engineering of making intelligent machines, through algorithms or a set of rules, which the machine follows to mimic human cognitive functions, such as learning and problem solving. AI is not one universal technology, rather it represents several subfields such as machine learning and deep learning, that individually or in combination add intelligence to applications.
Unlike traditional technologies which follow a fixed set of instructions, modern AI systems learn from data. By analysing millions of samples, they identify the patterns that humans might miss and use those patterns to make predictions.
There are 3 ways in which AI can learn: supervised learning, unsupervised learning, and reinforcement learning. In the former, AI is essentially fed data that is already labelled with the correct answer. It compares its guesses to the labels and adjusts it until it gets the right answer. In unsupervised learning, AI looks for hidden patterns or groupings in unlabelled data without human guidance. Lastly, reinforcement learning is where AI learns through trial and error, receiving rewards for correct actions and penalties for mistakes.
Now, we can discuss why medicine might need AI.
There is an ever-increasing demand on the NHS, with 36.2% of patients waiting for more than 4 hours in A&E, and even more shockingly, 10.1% of patients waiting over 12 hours. This is due to a multitude of factors, a prominent one being workforce shortages caused by low pay, work related stress and failure to train enough domestic staff. An ageing population also contributes to this, due to multimorbidity among the elderly, which is having 2 or more long-term health conditions simultaneously. Because these conditions are mostly chronic, it means that’s constant monitoring is required, which takes up a lot more resources than short-term illnesses.
Because of this immense strain on the NHS, it is becoming harder to deliver what is called the “quadruple aim” for healthcare. This is to improve population health, improve the patient’s experience of care, enhance caregiver experience and reduce the rising cost of care. As a result, healthcare systems are turning to artificial intelligence as a potential tool to improve efficiency, reduce clinician’s workload, and support better patient outcomes.
Now we can take a look at how AI is already integrated in healthcare.
AI video tools are being developed in areas such as minimally invasive diagnosis and treatment, for example endoscopy and laparoscopy, which use cameras as part of the procedure to routinely capture and use video footage. During these procedures, AI records a surgical intervention and writes a draft report using natural language processing to summarise the intervention, which staff can then review. Then, it is able to provide feedback and coaching to a surgeon, enabling reduced compilations and reduced length of stay for some patients.
I myself witnessed the integration of AI during a work experience placement I underwent in Lewisham hospital. Here, AI assisted with imaging of fractures. The software analysed X-rays and highlighted the suspected fracture, as well as helping with pre-operative planning by mapping the injury to help guide where to place the implant.
Another striking example is AI actually predicting sepsis. the FDA has recently cleared an AI based sepsis detection system, called the Targeted Real-Time Early Warning System, for approval. Sepsis is a life-threatening medical emergency where the body has an extreme, harmful reaction to an infection -> the most advances stage carries an appallingly high mortality rate of 30% to 50% even with immediate intensive care. It is for this reason why sepsis needs to be detected as early as possible. This new system catches symptoms hours earlier than traditional methods, making patients 20% less likely to die from sepsis. The machine-learning system tracks patients from the time they arrive in the hospital through discharge, ensuring that critical information isn’t overlooked even if staff changes or a patient moves to a different department. While a doctor might notice a fever, a low blood pressure, or a fast rate. AI looks much deeper. Each change alone might not worry a clinician, but together they form a pattern that resembles thousands of previous sepsis cases that the AI has been trained on.
If AI can already diagnose disease, analyse scans and predict sepsis, should we trust every decision it makes?
There are many limitations about the accuracy of AI: the first one being that AI is only as powerful as the data it learns from. AI systems can be trained on biased datasets, meaning they could inadvertently widen health disparities, particularly for underrepresented populations. In these datasets, there could be imbalanced samples sizes for certain patient groups. One illustrative example of class imbalance affecting model performance is the prediction of melanoma from skin images. These AI models are trained on data sets that are heavily composed of light-skinned images from patients in the US, Europe, and Australia. As a result, there is reduced performance for images of lesions in darker skin tones, potentially crossing the line between a patient receiving life-saving treatment or not.
Have you ever asked ChatGPT something, but it gives you a wrong response? This is called a hallucination. In other words, the AI generates information that sounds convincing but is actually false.
A common type of hallucination is fabricated data, which includes inventing non-existent drug dosages, fake medical histories, or incorrect treatment paths in patient notes. Another is phantom references, which is essentially generating completely fake journal citations and author names to back up false claims.
The consequences of hallucinations could be catastrophic. A clinician who unknowingly trusts this information could make decisions based on false evidence, possibly putting a patient’s life at risk.
Even if we could solve every issue with accuracy and bias, would that be enough? what does AI lack that human doctor’s do not?
The most obvious choice that comes into mind is the lack of physical embodiment. The inherent lack of a physical body within AI limits the potential for a genuine patient encounter. The AI would also struggle in observing the non-verbal cues such as body language and facial expression, at least in a much less detailed way compared to a human physician.
Another major quality of a human doctor is empathy. Imagine having to tell a patient they have a terminal illness – it’s not just about delivering the facts, but doing so with compassion, understanding, and emotional support. While AI can generate empathetic responses, mimicking the warmth of human clinician, it is unclear if patients perceive this chatbot empathy as genuine.
Another is leadership. Doctors lead multidisciplinary teams, coordinate patient care, delegating responsibilities to nurses, pharmacists, physiotherapists, and specialists. While AI can provide recommendations and support clinical decision making, it cannot take responsibility or motivate a team.
Medicine is also not always straightforward, and not every patient fits into a textbook diagnosis. Doctors usually have to think creatively when faced with rare diseases, something that is hard for AI to replicate. AI can recognise patterns, but it cannot truly think outside the boundaries of what it has been trained on.
Another thing to note is that currently AI systems are not reasoning engines, meaning they cannot reason the same way as human physicians, who rely on common sense or clinical intuition and experience. Instead, AI is much like a signal translator, translating patterns from datasets. For example, when analysing radiographs, AI can detect subtle abnormalities, but what it cannot do is truly understand what it is seeing.
One of the key principles for the use of AI for health care is protecting human autonomy. The use of AI can lead to situations in which decision-making power could be transferred to machines. If this happens, who is responsible for the death of the patient? Is it the software developers, healthcare organisations, or the doctor in charge.
Another ethical concern includes risk to patient’s privacy from data collected by AI. Deploying medical AI models without protective measures can pose substantial privacy risks to individuals. This data can be used for malicious purposes such as identity theft and targeted cyberattacks. However, there are some ways healthcare organisations can minimise the risks. Strong encryption could be used, meaning if data is intercepted, it cannot be read without the correct decryption key. Names and addresses could also be removed to reduce the risk of individual patients being identified.
Thirdly, advanced AI technologies may only be available in wealthier hospitals or countries, increasing healthcare inequalities. It would have to take a lot of investment to create affordable AI systems. Developing, implementing, and maintaining these systems require significant financial investment, as well as specialised infrastructure and trained staff. As a result, lower-income countries can fall much behind, leaving patients without access to the same tools and treatments.
Lastly, excessive reliance on AI could lead to the dehumanisation of care, eroding a doctor-patient relationship, which is a cornerstone of effective medical practice. The emphasis on data-driven decisions can overshadow the empathy, trust and personalised care traditionally provided by human clinicians. AI has what is known as a “black box” nature, which is a system where the internal workings of AI is a mystery to its users. Users can see the system’s inputs and outputs, but they can’t actually see what happens within the AI tool to produce those outputs. This “black box” nature can exacerbate issues with transparency and patient trust.
Despite these limitations, AI is not going away. In fact, its role in healthcare is only expected to grow. So, how will AI shape the future of medicine?
One rapid area of development is continuous health monitoring. Devices such as smartwatches already detect irregular heart rhythms, but future AI systems may monitor blood pressure, glucose levels, breathing patterns, and other physiological signals continuously. This allows diseases to be detected before symptoms even appear, shifting healthcare from treating illness to preventing it.
Another is transforming idea is drug discovery. Developing a medicine traditionally takes over a decade and costs billions of pounds. AI can analyse enormous chemical databases and identify promising drug candidates far more quickly, helping researchers to decide which compounds are worth testing.
Since we’ve now discussed the benefits of AI, as well as how it could be used in the future, let’s see exactly what a future consultation could look like with the integration of AI.
Before you even walk into the consultation room, the AI system has already analysed your data from a wearable sensor such as a smart watch. Its highlighted possible diagnoses and suggested investigations. The doctor walks in, and they role remains essential, they listen to your concerns, explain the options in a way you understand, consider your personal values, and provide an empathetic ear. AI provides the information, while the doctor provides human judgement and reassurance.
Finally, the answer to the question “Can AI be doctors?”, is that it is highly unlikely. AI lacks human qualities that make a doctor. Patients cannot be expected to immediately trust AI, as it lacks empathy and compassion, leadership, and tangible connection.
It is however much more reasonable to say that AI can act as a physician assistant. By recognising patterns from patient records, streamlining workflows to increase efficiency by scheduling appointments and generating reports, AI can free up the physician’s time to focus on providing more direct patient care.
Ultimately, AI will by heavily integrated into healthcare, transforming the way medicine is practised, hopefully allowing high-quality care to be accessed by a larger population. Again, its greatest potential won’t be from replacing doctors but instead working alongside them.
Medical Research Insights
Medical Insights explores advanced scientific concepts and intricate medical developments. We navigate through these sophisticated subjects using a logical framework to ensure curious minds can track the progression.
Stay Updated
bottom of page