Article Text

Download PDFPDF

Epidemiology and policy
P1-73 Can in-hospital fatality rates following HIP fractures be explained? A meta: regression analysis
  1. S Alves1,2,
  2. M F Pina1,3
  1. 1Instituto de Engenharia Biomédica, Porto, Portugal
  2. 2Escola Superior de Tecnologia da Saúde do Porto—ESTSP, Porto, Portugal
  3. 3Serviço de Higiene e Epidemiologia, Faculdade de Medicina da Universidade do Porto, Porto, Portugal
  4. 4Instituto de Saúde Pública da Universidade do Porto, Porto, Portugal


Introduction The risk of death increases following a hip fracture.

Objective To explain mortality rates of patients hospitalised due to hip fracture, according to multiple co-variables.

Methods A systematic review on Medline was conducted and studies were included if data for in-hospital fatality rates, following a hip fracture admission (ICD- 10 S72.0—S72.2 or ICD- 9- CM 820), was available for patients older than 50 years. Prospective cohorts were considered when appropriate data were available; experimental, review and case studies were excluded as well as studies comparing different treatments. Studies involving specific populations such as cancer or patients with kidney problems were also excluded. Studies published between 2010 and 2000 were considered. Economic, social, health and demographic data were retrieved from OECD—Organisation for Economic Co-operation and Development. A meta-regression was conducted.

Results Preliminary results lead to 21 studies selected, 15 analysed, from 11 different countries, comprising a total of 710 886 cases of hip fractures. Sample sizes differ greatly between studies: 155 to 574 482 Most data refers to no earlier than 1996. Data available presented heterogeneity regarding age groups, availability of information by sex and period of collection. Case fatality rates range from 0.7% in Formosa (2001) to 14% in England (2002–2005).

Conclusions Heterogeneity observed in fatality rates could be explained by a number of variables including allocation of medical resources. Meta regression will allow knowledge incorporation, accounting for sample size and explanation of several covariates.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.