What Is Root Cause Analysis in Clinical Practice?

“Root cause” is one of the most used phrases in functional medicine and one of the least examined. Used loosely, it implies that every chronic illness has a single hidden origin that the right test will reveal. Used carefully, it describes something more modest and more useful: a disciplined way of reasoning from a patient’s story to the factors that are keeping them unwell. The clearest way to show the difference is to follow one illustrative case, built from presentations that are common in practice, from the first appointment to the second follow up.

The presentation

There are 2 tempting responses here, and both are traps. The first is to reassure: the tests are normal, so there is nothing to find. The second is to order a large panel of specialised tests in the hope that something abnormal appears. The first ignores a real and persistent change in how she feels. The second substitutes testing for thinking, and in a panel of dozens of results, some will fall outside the range by chance.

Root cause reasoning starts somewhere else: with the question of when she last felt well, and what changed.

Building the timeline

The history takes most of the first appointment. Some of the most important findings emerge only because the practitioner asks about periods of life rather than symptoms.

Already the story looks different. Nothing here is exotic, but a sequence has appeared: a change in work, followed by shorter sleep, later eating, less activity and more caffeine and sugar, followed by fatigue, weight gain and cognitive complaints.

Mapping the systems

Placed on the functional medicine matrix, her findings cluster in 3 areas. In communication: disrupted sleep, irregular cycles and afternoon cognitive symptoms. In energy: fatigue and poor recovery. In transport and metabolism: central weight gain with a family history of diabetes and hypertension. Digestion, in assimilation, shows milder, meal related symptoms. The lifestyle factors around the matrix are where the strongest signals sit: sleep, movement, nutrition and stress are all disrupted.

This clustering matters because it suggests a small number of shared mechanisms rather than several independent problems.

From findings to hypotheses

A good root cause analysis produces ranked hypotheses, not a single answer. For Meera, the practitioner writes down 4.

1. Chronic sleep restriction and circadian disruption. Short sleep and late eating are known to impair glucose tolerance and alter appetite regulation; experimental studies have shown measurable effects within days. This could plausibly contribute to her fatigue, cognitive symptoms and weight gain.
2. Emerging insulin resistance. Central weight gain, family history, inactivity and afternoon energy crashes after sugary snacks make this likely, even with a normal fasting glucose.
3. A hormonal contribution. Irregular cycles at 42 may reflect early perimenopause; her history of irregular cycles in her twenties also raises the question of polycystic ovary syndrome, which is closely linked with insulin resistance.
4. Less likely but important to exclude. Iron deficiency, vitamin B12 deficiency, vitamin D deficiency, sleep apnoea given weight gain and fatigue, and depression. Some of these have been partly excluded by earlier tests; others have not.

2 points about this list are worth making. The first 3 hypotheses are not competitors; they may all be true and may reinforce each other. And the fourth is there deliberately, because root cause reasoning must never become an excuse to overlook conventional diagnoses.

Choosing what to test, and why

Each test is selected because its result could change the plan.

What is not ordered is as important. There is no salivary cortisol panel, because “adrenal fatigue” is not a recognised diagnosis and such a result would not change the first steps. There is no broad food sensitivity panel, because her bloating is clearly related to meal size and timing. Tests that will not change the plan add cost and noise.

What the results showed

Her HbA1c is 5.6 percent, her fasting insulin is at the upper end of the laboratory range, triglycerides are raised and HDL cholesterol is low. Ferritin is low normal, vitamin D is low, and B12 is normal. Her sleep apnoea screen is low risk and her mood screen does not suggest depression. Her physician arranges further assessment of her cycle changes.

The picture supports the first 2 hypotheses strongly and the third provisionally. It also identifies low vitamin D and borderline iron stores as smaller, treatable contributors.

A plan built around mediators

The plan does not try to change everything at once. It targets the mediators most likely to influence the others, in an order Meera can follow.

• Sleep and timing first. A protected wind down period, a consistent wake time, and negotiating with her employer to move evening calls to 2 fixed nights a week. Dinner moves earlier on the other nights.
• Meal composition. Protein and fibre at breakfast and lunch to reduce the afternoon energy crash, replacing the sugary snacks rather than simply removing them.
• Caffeine. Reducing to 2 cups before midday, because later caffeine is likely to be undermining her sleep.
• Movement. A 10 minute walk after lunch and dinner to begin with, since post meal activity improves glucose handling, building towards the exercise she used to enjoy.
• Nutrients. Vitamin D replacement at a dose agreed with her physician, and dietary iron, with ferritin to be rechecked.
• Coordination. A short letter to her physician summarising the findings, the plan and the request for hormonal assessment.

6 weeks later: testing the hypotheses

At follow up Meera’s sleep has increased to around 7 hours on most nights. Her afternoon cognitive symptoms have improved considerably and her bloating has largely resolved with earlier, smaller dinners. Her energy is better but not normal, and her weight has fallen slightly.

This is the point at which root cause reasoning earns its name. The response to intervention is itself evidence. The rapid improvement in cognition and bloating supports the sleep and meal timing hypotheses. The partial improvement in energy suggests something else is still contributing. The practitioner reviews the plan rather than simply continuing it, and learns that the evening calls have crept back to 4 nights a week, and that her physician’s assessment has confirmed early perimenopausal change. The plan is adjusted accordingly, with the physician managing the hormonal aspect.

At 3 months her HbA1c and triglycerides are repeated to judge the metabolic trajectory. The measure of success is not a single number but a pattern: sleep, energy, cognition, waist circumference and metabolic markers moving in the same direction.

What the case teaches about the method

Where the term comes from

Root cause analysis did not originate in medicine. It was developed in engineering and manufacturing, where techniques such as repeatedly asking “why?” were used to trace failures to their origins, and it was later adopted in patient safety to investigate adverse events. In those settings there is often a single chain of causation. Chronic illness is different. Its causes form a web, with several contributing factors, feedback loops and individual susceptibility. Clinical root cause analysis is therefore less about finding the root than about mapping the network of contributing causes and deciding where to intervene.

The biases to watch for

Diagnostic reasoning research has catalogued the cognitive errors that affect clinicians. 3 are especially relevant here. Anchoring is holding on to the first explanation, such as the patient’s own suggestion of a thyroid problem. Premature closure is stopping once one plausible cause is found, which could have meant missing the perimenopausal contribution. Confirmation bias is seeking only evidence that supports a favoured theory. The ranked hypothesis list, deliberate inclusion of conditions to exclude, and honest review at follow up are all practical defences against them.

Correlation on a timeline is not proof

A timeline shows sequence, not cause. Symptoms that begin after an event may be caused by it, may share a common cause with it, or may simply follow it by coincidence. The timeline generates hypotheses; the response to intervention, and where needed further investigation, tests them.

When not to go looking for a root cause

There are situations in which root cause reasoning should wait. New neurological symptoms, chest pain, unexplained weight loss, rectal bleeding, a new lump, severe or rapidly worsening symptoms and any suggestion of self harm all require prompt conventional assessment first. A structured history can continue afterwards, but it must never delay diagnosis. For practitioners who are not medically qualified, knowing these red flags and having a clear referral route is part of practising the method safely.

A method, not a promise

Meera’s case is ordinary, and that is the point. Most root cause analysis in practice uncovers familiar mechanisms in an unfamiliar combination, unique to one person’s life. The skill lies in listening long enough to see the sequence, mapping it honestly, choosing tests that answer real questions, and letting the patient’s response guide the next step. Done well, it offers patients something that neither reassurance nor exhaustive testing can: an explanation that makes sense of their experience, and a plan they can act on.