We performed a criteria-based clinical audit as described by the Health Care Quality Improvement Partnership (13). The project process consisted of setting criteria and standards according to research and legal demands, applying a suitable audit tool, measuring baseline documentation quality, implementing tailored interventions based on identified barriers and facilitators, and re-measurement of documentation quality.
Setting
The project took place at two wards in a Norwegian community care centre, which was organised as a residential care home for people with dementia needing 24-hour nursing care. The regular staff included eight nurses (19 %), 14 assistant nurses (35 %), and 20 assistants without a health care degree (47 %). The staff organization was based on a primary care model (14).
The community care centre applied an electronic patient records system that integrated the nursing process in care planning, with freewriting for diagnoses, aims, interventions, evaluation and daily reports. According to local guidelines, nurses and assistant nurses were responsible for the development of nursing care plans, while all staff compiled progress reports in the electronic patient records.
Criteria, standard and audit tool
Criteria and standard were defined in line with the step-wise nursing process and the nursing care plan (6), the N-Catch II audit instrument (15) and Norwegian regulations for patient records (16).
The N-Catch II was culturally adapted and translated into Norwegian (17) from the valid and reliable Dutch D-Catch instrument (11). It is designed to assess electronic written nursing documentation and includes a quantitative (e.g presence of relevant information), and a qualitative (e.g. correctness of language) assessment of documentation. In this audit, we focused on the quantitative assessment of the documentation, assessing the following steps (criteria) in the nursing process; ”Nursing assessment on admission”, “Nursing diagnoses”, “Aims for nursing care”, “Nursing interventions”, and “Evaluation/progress report” (Table 1). Each step was scored on a scale from 0 – 3, with zero indicating inadequate or missing documentation and three complete documentation.
To assess inter-rater reliability for the audit instrument, two nurses individually scored all criteria in seven patient records. In advance, the nurses had met for 45 minutes to reach a mutual understanding of the instrument. Agreement was calculated as the proportion of equal responses on all parts defining a criteria (18), and ranged from 0 % on the criterion “Nursing assessment on admission” to 87 % on “Aims for nursing care” (Table 2). The reasons for discrepancies between raters were different interpretation of items in the audit tool and dissimilar interpretation of information in the nursing records. The raters reached consensus based on a discussion of scores and clarification of audit tool.
Data collection
A power analysis informed that at least 33 patient records were needed to estimate a mean with 95 % confidence and precision equal to 0.2, this when assuming an expected population standard deviation (SD) of 1.5 and a population size of 38. For comparison of mean values at audit and re-audit, a total sample size of 72 (36 in each group) was needed to detect a difference in means of 1 unit and assuming SD equal to 1.5. This to achieve a power of 80 % and a two-sided significance level of 5 %. The software R (19) as applied for sample size calculations using the sample.size.means function in the Samplingbook package (20), and the pwr.t.test function in the pwr package (21).
We included nursing records for residents at the two included wards. If nursing care plans had been initiated before the study period, the records were included if the patients` health problem was still present. We excluded assessments, transfer notes, archived or former edits of nursing care plans, and daily reports written by others than the regular staff at the two wards.
Baseline data was collected retrospectively from electronic patient records during January to March 2018. In this period, 39 patient records were eligible and all, but one fulfilled the inclusion criteria. The re-audit was based on 38 patient records during March to June 2019.
Data analyses
Frequency bar charts were established to compare the distribution of score values at audit and re-audit. We presented mean scores at audit and re-audit, and mean score differences (MD) with 95 % confidence intervals (CI) for each criteria. Changes in mean scores were tested with the independent sample t-test. P-values less than 0.05 indicated statistical significance. The statistical software IBM SPSS Statistics version 25 (22) was used for the statistical analyses.
Implementation strategy
We designed a multifaceted implementation strategy (23), with interventions tailored to feedback (24), printed educational materials (25) and local opinion leaders (26). The first step of our strategy was feedback on baseline results to staff and leaders in small groups. We held one session with all leaders and five staff sessions where 47% attended. At the end of all sessions, we performed a brainstorming where we explored barriers and facilitators for changing documentation practice. Further, we performed a root-cause analysis and described interventions tailored to local barriers.
The identified local barriers were related to lack of knowledge and skills, lack of resources (computers) and insufficient time for writing patient records (Table 3). Local opinion leaders were identified as head nurses, assistant head nurses and experts in electronic nursing documentation software. We applied checklists for the staffs’ training in electronic documentation, developed educational material and cards with documentation guidelines to be available as reminders by computers, supplied the wards with additional computers, and encouraged staff to organize time for documentation. At the same time, the leaders at the community care centre performed an educational session where staff were introduced to updated terminology and criteria for nursing documentation, in line with a national quality project in nursing documentation (27).