Skip to main content

Archived Comments for: Hip fracture risk in relation to vitamin D supplementation and serum 25-hydroxyvitamin D levels: a systematic review and meta-analysis of randomised controlled trials and observational studies

Back to article

  1. Socioeconomic factors are most important

    Måns Rosén, The Swedish Council on Health Technology Assessment (SBU)

    21 January 2011

    I read the systematic review on hip fracture and vitamin D with great interest. The authors seem puzzled by the fact that the RCTs and observational studies showed somewhat different results. I claim that socioeconomic factors as an uncontrolled confounding is the primary reason for the discrepancy. Reading through the observational studies you also notice that none of the studies controlled for socioeconomic factors. Two often quoted examples of misleading results from observational studies are hormone replacement therapy (HRT) and vitamins. In both these cases the results from observational studies were in line with RCTs when they controlled for socioeconomic factors (1,2). I suggest that you always should adjust for socioeconomic factors when assessing interventions with observational studies (3). Only a minority of observational studies adjust for those important factors today (4).

    Måns Rosén
    Executive Director, SBU
    Adjunct professor in HTA at Karolinska Institutet, Sweden.

    References:

    1. Lawlor DA, Davey Smith G, Ebrahim S. Socioeconomic position and hormone replacement therapy use: explaining the discrepancy in evidence from observational and randomized controlled studies. Am J Public Health 2004;94:2149-54.

    2. Lawlor DA, Davey Smith G, Brucksdorfer KR et al. Those confounded vitamins: what can we learn from the differences between observational versus randomised trial evidence? Lancet 2004;363:1724-7.

    3. Rosén M, Axelsson S, Lindblom J. When can RCTs and observational studies mislead us and what can we do about it? Int J Clin Practise 2009;63:1562-4.

    4.Rosén M. Observational studies versus RCTs: what about socioeconomic factors? Lancet 2009;373:2026.

    Competing interests

    No conflict of interest.

Advertisement