When does a nutrient become a worthwhile health offering?

01

Vitamin D: population and outcome belong together

A meta-analysis pooled 4,190 people with prediabetes across three placebo-controlled trials: two of vitamin D3, one of an active analogue. The adjusted three-year absolute diabetes risk reduction was 3.3 percentage points (95% confidence interval 0.6–6.0). Individual trials were nonsignificant; rare adverse effects remained insufficiently assessed. [1]

A fracture trial involving 25,871 generally healthy middle-aged and older adults found no significant vitamin D3 benefit: 5.9% experienced fractures versus 6.0% with placebo over a median 5.3 years; hazard ratio 0.98 (95% confidence interval 0.89–1.08). Recruitment was explicitly not based on vitamin D deficiency or osteoporosis. [2]

For a development decision, we would therefore require the intended population and proposed benefit to be documented together. A project team should place its planned application alongside the study protocol: do the starting conditions, outcome and observation period match? Each departure should become an explicit development question. This would provide an assessable rationale for taking a particular project forward. Even an appealing ingredient would have to pass that test.

02

Define the need precisely

A 12-week placebo-controlled RCT studied 198 menstruating women with fatigue, ferritin below 50 µg/L and no anaemia. Iron produced an additional 3.46-point improvement on a subjective 40-point fatigue scale (95% confidence interval 0.3–6.7), without significant quality-of-life improvement. The small sample and possible unblinding limit interpretation. [3]

The randomised MRC trial enrolled 1,817 women after a pregnancy affected by a neural tube defect. Among 1,195 assessable subsequent pregnancies, defects occurred in 6/593 with periconceptional folic acid versus 21/602 without it: 1.0% versus 3.5%; relative risk 0.28 (95% confidence interval 0.12–0.71). Early stopping and the high-risk population limit interpretation. [4]

Our proposed development approach would begin with a carefully designed assessment of need. Symptoms, relevant history and existing findings would receive professional consideration. A digital questionnaire could prepare that discussion, with explicit responsibility for decisions requiring a professional. The process should also provide a clear explanation when someone does not meet the offering's criteria. That outcome deserves as much attention in the service design as accepting someone into a programme.

03

Develop the supplement in context

A meta-analysis of 49 RCTs involving 1,863 healthy adults found that added protein produced an average additional 2.49 kg gain in the maximum weight lifted once (95% confidence interval 0.64–4.33). Every trial included at least six weeks of resistance training. Prevention of care dependency was not established. [5]

For Human Nexus, this could inform a specific development hypothesis: design the supplement and the accompanying daily routine as one coherent process. Before a pilot begins, the team should establish what support participants actually receive, who answers questions and how interruptions are handled. That requires clear operational responsibilities. In this proposed model, an information page alone would not constitute support; accessible contacts and a defined sequence would be essential.

04

Give the digital service a specific purpose

We propose three jobs for the digital service: make information understandable, organise agreed steps and route feedback to the responsible professional. The interface should explain why it asks a question and what happens after an answer. Reminders would follow an agreed process. When health questions remain unresolved, the route to personal support should be clear. A pilot could then investigate whether people understand the offering and can follow its processes.

For investment decisions, we would also require separate assessments of health outcomes, actual use and the resources needed to deliver the service. A completed questionnaire would count as a usage event. A professionally relevant change would need assessment against a measure defined in advance. Consultation time, follow-up questions and withdrawals would belong in the operational costing alongside those findings. This would help Human Nexus identify where a concept needs improvement before extending it to additional populations.