Healthcare is increasingly fragmented. Patients, particularly those with chronic diseases, navigate complex healthcare systems with many service providers but often no-one “captaining the ship”. Payers mainly finance discrete services and episodes of care rather than wholistic care and clinical outcomes. The cost of healthcare continues to increase in real terms and as a proportion of gross domestic product (GDP) in Western countries (1).

There are many reasons for healthcare fragmentation. Some of these include the following. First, medical generalists, including family physicians in the United States (US) and general practitioners (GPs) elsewhere, have been devalued to a medical triage role in many well-serviced and high-resourced locations e.g., metropolitan centres. This is encouraged by governments, the medical hierarchy, and medical indemnity providers, but, ultimately, increases the number of high paid medical specialists. Second, evidence-based medicine demands specialist care if available. Indeed, patients with multiple comorbidities attend numerous medical specialists who apply clinical practice guidelines (CPGs) within their own speciality, sometimes to the determent of overall care. An example is the prescribing of a set list of drugs after myocardial infarction without consideration of potential interactions with existing drugs for pre-existing comorbidities, resulting in drug-drug interactions and adverse effects (2, 3). Third, allied health providers are working in areas traditionally serviced by family doctors, such as chronic disease management e.g., pharmacist and nurse prescribing from restricted drug formularies (4, 5). Fourth, specialized novel services that are expensive are increasingly added to healthcare budgets e.g., genetic testing (6). Finally, there are sometimes unrealistic societal expectations to provide treatments for all, independent of disease trajectory, age, and social circumstances (7) e.g., major surgery for small bowel obstruction in a patient with advanced dementia and very short life expectancy.

Given the decentralized models of patient care and the unsustainable costs of healthcare (>10% of GDP in Western countries), strategies to reduce healthcare fragmentation are attractive from payer and patient perspectives (8). This commentary introduces Virtual Twin for Healthcare Management (VTHM), a conceptual patient-centred smart phone application that disrupts the healthcare sector by addressing fragmentation and transferring control of health information from multiple sources to the patient. The system is a combined personal- and electronic-health record (PHR/EHR). Importantly, VTHM is designed to deliver precision medicine initiatives as they roll-out in more affluent economies, such as access to molecular diagnostic results, pharmacogenomic (PGx)-guided prescribing, and model-informed precision dosing (MIPD) (9–11).

During their interactions with healthcare, patients tell their story repeatedly, relating medical history, allergies, current and past medications etc., time and time again to multiple people. Repetition is often exhaustive for patients with multiple chronic medical conditions and for their families and carers. Electronic-health records (EHRs) have been implemented to better capture this information, with the aims of improving continuity of care, patient safety and patient satisfaction (12). However, such e-health systems are often locally oriented, rarely communicate well with each other, and are difficult to access, either by healthcare providers without access, or by patients themselves.

Although some e-health systems are superior at linking patient data between various EHRs, for example Surescripts® is a US e-prescribing platform that covers about 90% of prescriptions in the community setting, there is considerable redundancy in this area. Information technology providers promise novel solutions but deliver little innovation to improve functionality, workflow and, critically, interoperability. Basic information that is essential for medical care, such as the patient's medical history, is collected repeatedly and stored in numerous e-health systems, whilst detailed information, such as a complete medical assessment with treatment plan from an experienced medical specialist, remains buried in a specific system because of access restrictions. This leads to expensive service repetition. A common example is the re-ordering of blood tests by different healthcare providers based on the pathology service they use.

The only common factor linking all the fragments of healthcare together—i.e., multiple healthcare providers, multiple healthcare records (paper, personal, electronic), multiple medication management systems, multiple pathology services, multiple medical imaging services etc.—is the patient. To date, attempts to centralize health information with national EHRs have had limited success. For example, the Australian Government started “My Health Record” in 2010, but participation by healthcare providers is relatively low and patients are resistant over privacy issues (13, 14). Personal health records have also been implemented, but again, these are locally oriented and focused on single chronic diseases, such as diabetes and heart failure, rather than providing more universal platforms for healthcare management (15).

With the advent of ubiquitous smart phone use, this device is the ideal platform for patient control of their health information. Virtual Twin for Healthcare Management (VTHM) is a multi-functional patient-centred smart phone application using a human model at the hub (“virtual twin”—see below) and EHR modules at the spokes. The application is in the design and proof-of-concept phases but is presented here showing how VTHM would ideally operate in the future. It is important not to confuse the aspirational nature of VTHM with current actual capabilities. Indeed, this perspective describes the “finished product”, and many challenges must be overcome on the path from concept to reality (see later section). With this in mind, the organizational layout of VTHM is shown in the Figure 1. Data is stored on a secure server, allowing healthcare providers to access the VTs (virtual twins) of their patients from any device with an internet connection.

A VT is a physiologically based (PB) human model constructed from a base model that is “virtualized” with each individual's unique genetic, molecular, physiological, and disease characteristics (16). It is the hub of VTHM and allows the user to visualise the body and its health data. An avatar of the patient overlays organ systems (“systems data”) arranged and separated with bloods flows, and tissue perfusion is modelled with differential equations based on physiology. The software that powers VT is Simcyp (Certara, USA), which is used to simulate drug pharmacokinetics (±pharmacodynamics) in the pharmaceutical industry (17, 18) and is increasingly being studied as a MIPD tool in clinical practice (19–22).

The systems data required to make a VT comes from sources that already inform medical decision making, including the prescribing of drugs. Data include basic demographics (age, weight, height, sex, ethnicity), pathology results (liver function tests, creatinine clearance, inflammatory markers etc.), biomarkers of disease and therefore potential treatment monitoring (real-time monitoring of blood glucose, prostate specific antigen, etc.), medical imaging (liver size, kidney size, ECHO cardiac output data etc.), and any values for physiological/molecular structure or function considered necessary for disease and treatment modelling and simulation. A Monte Carlo approach is employed to generate realistic values for unknown parameters of the system based on population covariate relationships whenever actual values are unknown e.g., the liver size of a Japanese male aged 50 years or the glomerular filtration rate of a 90 yo Caucasian female weighing 45 kg (17).

Virtual twin is a dynamic PB model that changes as the patient ages (23). The VT also reflects medical changes. When a patient is diagnosed with a new disease, the pathophysiology changes that occur in the condition, if known, are updated in VT (24). When a woman becomes pregnant, the physiological system changes that occur during pregnancy are automatically activated (25). When information is known about drug metabolizing enzyme and transporter abundances and/or activities (26, 27), or a patient takes a PGx test to determine the genotype of a drug metabolizing cytochrome P450 enzyme (CYP) (28), this information is added to their VT to increase the level of individualisation (referred to as “virtualisation”) (20). When drug changes are made, drug-drug interactions, drug-gene interactions and other variables are considered as VT uses MIPD (see below) to predict drug pharmacokinetics (±pharmacodynamics).

A series of theoretical case studies based on real scenarios are presented here to show how VTHM could potentially improve patient care and clinical efficiency.

As with aeronautical engineering, VT can be used to simulate real scenarios that are difficult or impossible to study during drug development or post-marketing. The pharmacokinetics (±pharmacodynamics) of multiple drugs used together in patients with several comorbidities (such as Case 5) is an example where various drug regimens can be tested in silico prior to selecting the best one for the individual patient. Outcomes of non-pharmacological interventions can be incorporated in cases where this is well-understood e.g., the potential effects of a low-calorie ketogenic diet on glucose tolerance and body weight (53). As confidence with the models increases, so will confidence in the outcomes and subsequent recommendations.

As VTHM becomes more sophisticated, and with increasing coverage of certain patient populations, it is envisaged that the data curated from VTHM could be used for health-economic analyses, which, in turn, could inform healthcare policy (54). There would be extensive data mining to help answer pharmacoepidemiology questions (55). Future generations could extract benefits from these data and establish “lessons learnt” regarding healthcare patterns and barriers to success. For example, artificial intelligence and machine learning would rationalise drug and dose selection from various guidelines together with real-world data on the outcomes from similar patients with regard to ethnic status, socioeconomic status, genetic makeup, medical conditions and co-medications (56, 57). Powerful “Precision Medicine 3.0” algorithms (58) could then select one treatment regimen over another based on novel “fingerprints” of disease and treatment biomarkers, thus making high-level recommendations to policymakers. A patient's VT could indeed “live” on after death, providing valuable data on a human life for future generations.

The functionalities on the spokes of VTHM (Figure 1) are well-established in healthcare to improve clinical outcomes e.g., e-prescribing (12). The VT hub is new in healthcare, but routine in the pharmaceutical industry (59), and there are many scientific, technical, clinical and logistical challenges to overcome prior to acceptance as a CDSS and precision dosing tool in the clinic and at the bedside (16). Perspectives from regulators, payers, clinicians, and patients about such tools are addressed in the previous reviews on MIPD (11, 60). The major challenge of VTHM is to unite the individual components of the system into one smart phone application (Figure 1). This is a formidable undertaking not underestimated by the developers. The technology is disruptive, which understandably brings opposition from those with affected business interests. Therefore, the business model for VTHM is complex and involves partnerships with other e-health system companies and healthcare service providers. The legal and regulatory frameworks for some components of VTHM are evolving and will require novel solutions, but there is already strong regulatory support for precision dosing tools under medical device guidance (11, 60). Issues of data privacy and data security must be addressed (61). Monitoring the clinical utility of VTHM is also critical, particularly when technical changes are implemented e.g., the addition of new drugs to the VT hub. A collaborative effort across the pharmaceutical, regulatory, academic, and healthcare sectors may be required for this large job. Finally, some patients will be unable to use VTHM, such as the young, elderly and the critically unwell. Caregivers would help bridge this gap.

Escalating healthcare costs and adverse clinical outcomes have not been improved adequately by e-health platforms that are locally oriented with restricted access. In this commentary, a conceptual patient-centred smart phone application called VTHM is proposed that works as a combined personal- and electronic-healthcare record. The platform is revolutionary because it transfers the control of health information from service providers to patients. The overachieving aims of VTHM are simple but ambitious—to lower healthcare costs and improve patient care by decreasing healthcare fragmentation. With big ambitions come big challenges, and those for VTHM are formidable and not underestimated. But the potential benefits of VTHM would be worth the effort. Notably, the VTHM platform is designed to operationalize current and future precision medicine initiatives, such as access to molecular diagnostic results, PGx-guided prescribing, and MIPD.