Care Keys Data and Statistical Methods

  • Ene-Margit Tiit
  • Kai Saks
  • Marja Vaarama

In order to elaborate the instruments, test the Care Keys methodology and build models, it was necessary to have different databases, covering all relevant client groups in all participant countries and containing measurements of all variables that might influence the crQoL of clients. The first databases were developed to test and pilot the various instruments for final data collection (see Chapter 2). After the instrumentation was finalised, national data collection took place during November 2004–March 2005, using common tools and a common data collection procedure. The national data sets were then combined to create a pooled database for empirical research that covered all the areas necessary for checking hypotheses and model building.

The Care Keys pooled database (CKPD) contained data on about 1,500 clients from 5 European countries, with an average more than 500 measured variables per case. The structure of data was quite complicated, with the data being drawn from interviews with clients, caregivers (in some cases also relatives), extracted from the care documentation and from managers of services. The overall response rate was moderate, mainly because of the health of clients and lack of information within care documentation. The following data processing tasks were required to create the pooled database:
  1. 1.

    Merging national data sets, cleaning, handling missing values, making imputations (where necessary and possible).

  2. 2.

    Exploratory analysis of data (on national and integrated, conditionally European level) to find the leading tendencies and estimate the distributions of variables, checking working hypotheses.

  3. 3.

    Compressing the data, calculating new variables (indexes).

  4. 4.

    Finding key-indicators with the aim of optimising the list of variables that would be measured in future.

  5. 5.

    Measuring dependencies and building models describing the factors influencing quality of life (QoL).

  6. 6.

    Creating meta-models to follow the multilevel structure of factors influencing the QoL of clients via quality of care (QoC) and other factors.


Mainly classical multivariate statistical methods were used in the data analysis, using the SPSS and SAS packages.


Home Care Descriptive Power Home Care Client Care Documentation Pool Database 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Brancato, G. (2006). Recommended practises for questionnaire design and testing in the European Statistical System. <>.
  2. Folstein, M., Folstein, S., & McHugh, P. (1975). Mini mental state, a practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12, 187–198.CrossRefGoogle Scholar
  3. Joreskog, K. G. (1969). A general approach to confirmatory maximum likelihood factor analysis. Psychometrika, 34, 183–202.CrossRefGoogle Scholar
  4. Little, R. J. A., & Rubin, D. B. (1987). Statistical analysis with missing data. New York: John Wiley.Google Scholar
  5. Maruyama, G. M. (1998). Basics of structural equation modeling. Thousand Oaks, CA: SAGE Publications.Google Scholar
  6. Revised Long-Term Care Resident Assessment Instrument User’s Manual. Version 2.0. (2002). Centers For Medicare & Medicaide Services.Google Scholar
  7. Saks, K., Tiit, E.-M., & Vähi, M. (2003). Measuring and modelling the quality of life of older Estonians. The 7th Tartu Conference on Multivariate Statistics. Abstracts.Google Scholar
  8. Toothaker, L. E. (1991). Multiple comparisons for researchers. Newbury Park, CA: Sage Publications, Inc.Google Scholar

Copyright information

© Springer Science+Business Media, LLC 2008

Authors and Affiliations

  • Ene-Margit Tiit
  • Kai Saks
  • Marja Vaarama
    • 1
  1. 1.Department of Social WorkUniversity of LaplandFinland

Personalised recommendations