Perspectives · The practice of medicine

Guidelines, automation, and personalised medicine

When I started out, guidelines were not yet a thing — and all medicine was personalised.

By Dr Cindy de Villiers · Originally published c.2020 · Reviewed and edited July 2026 · General information, not medical advice

I recall being a young general practitioner setting out to make my way in the world, before guidelines were a thing. We had not heard of health automation, and all medicine was personalised. I had bought an established, full-time practice in a small provincial town, and was juggling my time between seeing patients, keeping on top of the paperwork, and balancing the books. I was not new to general practice — you may be relieved to hear — but I was new to the business, and to the government funding processes.

This was before the days of a multitude of self-help guides beseeching me to look deep and see what it was I really wanted. I was a doctor, and I instinctively knew I wanted to be in general practice. In practice, my priority was finding the best option for those who trusted me enough to help guide them through the vagaries of medicine. I quickly picked up on subtle cues — the confidence in the walk from the waiting room, the furrowed lines, the florid skin, the direct gaze, the odour, the interaction between parent and child. These cues helped guide me towards a plan that supported healing. That is not to say I did not write the patient's name on the prescription pad as they were sitting down — a complaint often heard from the disgruntled. I did, and it bought me an extra minute with them. On the other side of the interaction: if I did not prescribe a pharmaceutical, the patient would often feel hard done by, with one or two wanting their money back.

Guidelines, and their unintended consequences

Then came the medical guidelines. We were reassured that these were, in fact, just that — guidelines. We were not meant to follow the flowcharts and algorithms at all costs. As clinicians, they were there to help us make the correct decisions on diagnosis and treatment, but ultimately the decisions and the responsibility remained ours. In reality, this was the start of health automation, and the decline of personalised medicine.

Fast-forward twenty-plus years. Guidelines are now not only instilled as “best practice”; there are penalties for clinicians who do not follow them, ranging from the disapproving tut-tuts of colleagues, through financial consequences, to formal proceedings. The process of formulating and instituting guidelines is somewhat murky — I discuss it in my piece on citizen science — and industry most certainly plays a part.

Junior doctors are surprised to see that I write my notes free-style, without a template of checklists. That is how I was taught: observe, listen, examine, recognise a pattern that matches a clinical situation, and formulate a plan. I find the act of writing helps this process. This is not to say checklists and guidelines have not helped me — they have, a great deal. I was simply lucky enough to start in medicine before guidelines became rules, when the art of medicine and the medical ritual were understood to be part of supporting healing.

It is worth noting that alongside the enforcement of guidelines has come a marked reduction in the independence with which a doctor can run a practice. Most general practices now involve a never-ending cycle of programmes, certifications, reporting, and funding rounds with an ever-expanding health bureaucracy.

Have the guidelines helped? It is generally accepted that fewer medical mistakes are made since their introduction, and the health authorities take pride in providing a service where the outcome is the same no matter which doctor is seen. Yet it is my experience that the public are more disgruntled with medicine, often choosing instead to trust influencers and bio-hackers, with predictably mixed results. General practitioners are overworked and caught between one-dimensional protocols and patients who want a magic pill. The guidelines seem to have levelled the playing field — lifting “poor” doctors up, and levelling “good” doctors down.

Meanwhile, in New Zealand the incidence of chronic disease continues to rise, while the medical system consumes a large share of GDP. Telehealth, which became normalised during the pandemic, has genuine benefits in acute medicine, but it is limited in the medicine of life — general practice. There are fewer subtle cues, the conversation does not flow as freely, and the paperwork grows. And while today's flowcharts merely suggest a diagnosis for the clinician to confirm, the advent of personal wearables and health AI may increasingly shape that process. Automation looks likely to reinforce the guidelines. But the arrival of consumer-held data is less predictable, and it creates genuine openings for more personal medicine.

Can automation deliver personalised medicine?

What can a doctor with their patients' best interests at heart actually do? The clock cannot be turned back, and perhaps it should not be. Humans are wired to seek simple solutions to complex problems, and we are notably poor at assessing long-term risk. Ideally, a medical interaction would combine the subtle and subconscious cues, the rigour of checklists, and genuine empowerment for both clinician and patient. Automation will certainly become commonplace, driven by consumer-held data. But automating a sub-optimal process only multiplies its outcomes — there is no point speeding up something that does not work. The guidelines may have had unintended consequences, and more of the same, faster, is not the answer.

Personally, after almost giving up medicine, I found my own way. I run a non-funded private clinic, which allows a modicum of freedom. I encourage patients to be an active part of the medical ritual. I use checklists, but my notes are still typed free-style. For me — and, I hope, for patients — the process has meaning, provides value, and supports long-term health. The trade-off is that I can see only a fraction of the number of patients I once did.

Can automation help? There are proxies for the overall state of an individual that can be measured and tracked, such as heart rate variability, and these may supply some of the subconscious signals a clinician would once have had to sense. More data may, equally, serve only to overwhelm both patient and clinician. The task — the interesting task — is to fold that information back into a relationship in which someone actually knows you. That, in the end, is what a small independent practice is for.