Background: Disability scores, such as the Functional Independence Measure (FIM) and Barthel Index, have been shown to correlate with care needs but cannot be used to assess them directly, as they do not indicate the number of people required to help with a task, nor the time taken. The Northwick Park Dependency Score (NPDS) is an ordinal scale that can be used to assess impact on nursing time. It takes 3-5 minutes to complete. Together with a short set of additional questions, it may be used directly to assess care needs in the community and to facilitate discharge planning.
Aims: To develop and evaluate the NPDS for use in a rehabilitation setting.
Methods: (1) DEVELOPMENT: Following a survey of existing instruments, tasks were selected on the basis of their impact on nursing time and divided into Basic Care Needs (BCN) and Special Nursing Needs (SNN). Cut-off points were devised to reflect the number of helpers needed and time taken. Following evaluation of the NPDS version 5, minor changes were made to produce version 6 which was re-evaluated on a smaller scale. (2) EVALUATION: Inter-rater and intra-rater reliability were tested in a cohort of 23 inpatients using five senior nurses. Analysis included assessment of degree of association, significant differences, absolute agreement, and agreement +1 level. Although there is no gold standard, the BCN section should correlate inversely with independently assessed Barthel scores. Re-evaluation of version 6 was undertaken using the same method of analysis in a cohort of 21 patients using three senior nurses.
Results: On initial evaluation inter-rater reliability testing showed an excellent level of association in total composite score between each pair of nurses (rho = 0.73-0.92, p <0.01) and agreement +1 level for individual items ranged from 73 to 100%. Significant disagreements were in six items. On re-evaluation following minor modification, high levels of association were still seen for total BCN, SNN and composite scores both between and within raters, with very satisfactory levels of agreement for individual items. The BCN section of the NPDS showed good inverse correlation with Barthel scores (rho = 0.91, p <0.01).
Conclusion: The NPDS is simple and practical to use in a busy setting. It is shown to be reliable and valid in its assessment of nursing dependency on the ward. Its translation into a directly costable measure of continuing care needs in the community now requires evaluation.