Umeå University's logo

umu.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Comparison of three statistical methods for analysis of fall predictors in people with dementia: negative binomial regression (NBR), regression tree (RT), and partial least squares regression (PLSR)
Umeå University, Faculty of Medicine, Department of Community Medicine and Rehabilitation, Physiotherapy. Umeå University, Faculty of Medicine, Department of Community Medicine and Rehabilitation, Geriatric Medicine.
Umeå University, Faculty of Science and Technology, Department of Mathematics and Mathematical Statistics.ORCID iD: 0000-0003-1524-0851
Umeå University, Faculty of Medicine, Department of Community Medicine and Rehabilitation, Geriatric Medicine.
Umeå University, Faculty of Medicine, Department of Community Medicine and Rehabilitation, Physiotherapy.
2009 (English)In: Archives of gerontology and geriatrics (Print), ISSN 0167-4943, E-ISSN 1872-6976, Vol. 49, no 3, p. 383-389Article in journal (Refereed) Published
Abstract [en]

Searching for background factors associated with falls in people with dementia is difficult because the population is heterogeneous. The aim of this study was to compare the efficacies of three statistical methods for analysis of fall predictors in people with dementia. NBR, RT and PLSR analyses were compared. Data used for the comparison were from a prospective cohort study of 192 patients at a psychogeriatric ward, specializing in patients with cognitive impairment and related behavioral and psychological symptoms. Seventy-eight of these patients fell a total of 238 times. PLSR and RT analyses are directed at finding patterns among predictor variables related to outcome, whereas an NBR model is directed at finding predictor variables that, independent of other variables, are related to the outcome. The NBR analysis explained an additional 10–15% variation compared with the PLSR and RT analyses. The results of PLSR and RT show a similar plausible pattern of predictor variables. However, none of these techniques appears to be sufficient in itself. In order to gain patterns of explanatory variables, RT would be a good complement to NBR for analysis of fall predictors.

Place, publisher, year, edition, pages
2009. Vol. 49, no 3, p. 383-389
Keywords [en]
partial least squares regression, regression tree, negative binomial regression, accidental falls, fall predictors, dementia and falls
Identifiers
URN: urn:nbn:se:umu:diva-2821DOI: 10.1016/j.archger.2008.12.004PubMedID: 19168232Scopus ID: 2-s2.0-69349092812OAI: oai:DiVA.org:umu-2821DiVA, id: diva2:141105
Available from: 2007-11-23 Created: 2007-11-23 Last updated: 2023-03-24Bibliographically approved
In thesis
1. Falls in people with dementia
Open this publication in new window or tab >>Falls in people with dementia
2007 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]

Falls and concomitant injuries are common problems among large groups of the elderly population, leading to immobility and mortality. These problems are even more pronounced among people suffering from dementia. This thesis targets fall risk factors for people with dementia in institutions. The overall aim of this thesis was to investigate risk factors for falls, predisposing as well as related to circumstances surrounding falls, and to do this as efficiently as possible.

In a prospective cohort study including residents of residential care facilities with and without dementia, the fall rate was higher for those with dementia, the crude incidence rate ratio (IRR) was 2.55 (95% CI 1.60–4.08) and the adjusted IRR was 3.79 (95% CI 1.95–7.36). In the group of people suffering from dementia, including 103 residents, a total of 197 falls resulted in 11 fractures during the 6-months follow-up period. From the same baseline measurements 26% and 55%, respectively, of the variation in falls could be explained in the group of residents with and without dementia. Fall predictors significantly and independently associated with an increased risk of falls in the group of people suffering from dementia were the category “man walking with an aid” and the use of more than four drugs.

In a prospective cohort study, including 204 patients in a psychogeriatric ward, a total of 244 falls resulted in 14 fractures. Fall predictors significantly and independently associated with an increased risk of falls were male sex, failure to copy a design, use of clomethiazole, and walking difficulties. Treatment with statins was associated with a reduced risk of falls. With these fall predictors in the negative binomial regression (Nbreg) model, 48% of the variation in falls was explained.

The data from the psychogeriatric ward were also analysed with the use of partial least squares regression (PLS) and regression tree to be compared with the results of the Nbreg analysis. PLS and regression tree are techniques based on combinations of variables. They both showed similar patterns, that a combination of a more severe level of dementia, behavioral complications and medication related to these complications is associated with an increased fall rate. Thirty-two percent and 38%, respectively, of the variation in fall rate were explained in the PLS and regression tree analysis.

The circumstances surrounding the falls in the psychogeriatric ward were analysed. It was found that the fall rate was equally high during the night and the day. A large proportion of the falls was sustained in the patients’ own room and a small proportion of the falls was witnessed by the staff. This pattern was even more pronounced during the night. The proportion of diurnal rhythm disturbances and activity disturbances was higher for falls at night than for falls during the day. Circumstances associated with an increased risk of falls, as shown by a short time to first fall, were anxiety, darkness, not wearing any shoes and, for women, urinary tract infection. The proportion of urinary tract infection was also higher in connection to falls sustained by women than to falls sustained by men.

This thesis confirms that people suffering from dementia are prone to fall. Walking difficulties, male sex and impaired visual perception are factors that should be considered in the work of reducing falls among people suffering from dementia. Furthermore, falls at night, behavioral complications and medication related to these complications should also be considered in this work, especially as the dementia disease progresses. A larger portion of the variation of the outcome variable was explained by the Nbreg model than the regression tree and PLS. However, these statistical methods, based on combinations of variables, gave a complementary perspective on how the fall predictors were related to falls.

Place, publisher, year, edition, pages
Umeå: Sjukgymnastik, 2007. p. 106
Series
Umeå University medical dissertations, ISSN 0346-6612 ; 1135
Keywords
Dementia, behavioural disturbances, accidental falls, risk factors, circumstances, analysis of fall predictors, male sex, walking difficulties, visual perception
National Category
Geriatrics
Identifiers
urn:nbn:se:umu:diva-1449 (URN)978-91-7264-433-5 (ISBN)
Public defence
2007-12-15, B 101, Vårdvetarhuset, 901 87, Umeå, 09:00 (English)
Opponent
Supervisors
Available from: 2007-11-23 Created: 2007-11-23 Last updated: 2009-05-18Bibliographically approved

Open Access in DiVA

No full text in DiVA

Other links

Publisher's full textPubMedScopus

Authority records

Eriksson, StaffanLundquist, AndersGustafson, YngveLundin-Olsson, Lillemor

Search in DiVA

By author/editor
Eriksson, StaffanLundquist, AndersGustafson, YngveLundin-Olsson, Lillemor
By organisation
PhysiotherapyGeriatric MedicineDepartment of Mathematics and Mathematical Statistics
In the same journal
Archives of gerontology and geriatrics (Print)

Search outside of DiVA

GoogleGoogle Scholar

doi
pubmed
urn-nbn

Altmetric score

doi
pubmed
urn-nbn
Total: 584 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf