Measuring emergency department crowding in an inner city hospital in The Netherlands

Overcrowding in the emergency department (ED) is an increasing problem worldwide. In The Netherlands overcrowding is not a major issue, although some urban hospitals struggle with increased throughput. In 2004, Weiss et al. created the NEDOCS tool (National Emergency Department Over Crowding Study), a web-based instrument to measure objective overcrowding with scores between 0 (not busy at all) to above 181 (disaster). In this study we tried to validate the accuracy of the NEDOCS tool by comparing this with the subjective feelings of the ED nurse and emergency physician (EP) in an inner city hospital in The Netherlands.

In a 4-week period, data of a total of 147 time samplings were collected. The subjective feelings of being overcrowded and feeling rushed by the ED nurse and EP were scored on a survey using a 6-point Likert scale on answering the question of how busy they would say the ED is right now. NEDOCS tool scores were calculated, and these were compared with the subjective feelings using the kappa statistic assessing linear weights according to Cohen's method.

Of all the time samplings, approximately 80% of the surveys were completed. The ED was rated as overcrowded 9% of the time by the ED nurses and 11% of the time by the EPs. The median NEDOCS score was 37 (0 to 120) and scored as overcrowded in 3%. There was a good intrarater agreement for the ED nurse and EP for the feeling of overcrowding and feeling of being rushed (κ = 0.79 and 0.73, respectively); the interrater agreement was moderate (κ = 0.53 and 0.43, respectively). The agreement between the NEDOCS and the subjective variables was moderate (κ = 0.50 and 0.53, respectively). A composite variable was created as the average of both the scores of feeling overcrowded of the nurse and the EP and the score of the EP of feeling rushed. The agreement between this and the NEDOCS was κ = 0.53.

The NEDOCS tool is a reasonably good tool to quantify the subjective feelings of overcrowding. When overcrowding is encountered and immediately recognised, specific measures can be taken to guarantee the timely provision of necessary medical care to the patients in the ED at that time. However, possibly more accurate agreements could be obtained as approximately 20% of the surveys were not completed because of perceived crowdedness. An important limitation is that only 3% of the NEDOCS is scored as overcrowded, so no conclusions can be drawn about the agreement for higher categories of overcrowding. It is suggested to repeat the study in a busier period. As the triage category was not taken into account in the formula, a high workload with only a few patients giving high scores in subjective overcrowding in spite of a low NEDOCS score could have led to lower agreements. Incorporating the triage category in the NEDOCS tool possibly will lead to better agreement, but further research is needed to assess this idea.