Dunja Dreesens

29 An iterative approach was used when searching these websites. The hits were checked for relevant descriptions and information by the first author. The information was copied and pasted into a crude data file, including web address and the time and date of the search. When there were more than 100 hits, only the first 100 were checked. In cases where the hits redirected to other websites (see appendix A) these links were followed, as were links to reports and articles (‘snowballing’ method). When reports or publications were unavailable on the website, the organisation was contacted in order to obtain a copy of the report or article. If it could not provide the document, it was obtained through the National Library of the Netherlands. In the case of unclear or unavailable descriptions, Google was used to find more information on the term in combination with the organisation. If Google yielded no or insufficient information, the organisation was contacted by email to ask for the missing information. If no clarification could be provided by email, a phone call or physical appointment was undertaken to retrieve the missing information. Both the emails and the notes of the telephone conversations were added to the crude data file. If the email or follow-up was not successful after two attempts, the researchers limited themselves to the available data from the websites. Data analysis Statements describing or defining the knowledge tool types formed our primary interest. Each tool type found was entered into a table together with the full or partial descriptions identified per data source. All the tools identified were subsequently scored along the inclusion criteria (see figure 1) by the first author. To decide if a tool type had followed an evidence-based approach we checked if it was developed in a transparent and systematic way, and searched and synthesized evidence according to the rules of evidence-based practice (26). Figure 1: Inclusion flow chart for the final tool selection The scoring by the first author was verified by the authors TvdW and LK, by systematically scoring every tenth tool listed, with TvdW starting from 5 and LK from 10 (so rows 5, 15, 25, 35, etc. and rows 10, 20, 30, etc.). When scores varied, we discussed these until we reached consensus on the scoring. The first author then entered the final scores for all tools. The tool types preliminarily included were copied into a new crude data table. Information received from additional searches and inquiries with the organisations were added to this table. On the basis of the additional information, we made a decision on final inclusion. The final table was structured by organisation to provide insight into the number of organisations in the Netherlands active in developing and disseminating knowledge tools, and the different names of Chapter 2