Study designs of evaluations included in the review
Inclusion criteria were not specified in terms of study design. Studies which aimed to study the effectiveness of nursing care processes in improving health related patient outcomes for broadly defined populations of patients entering in-hospital settings were eligible. Included studies used an experimental or quasi-experimental design or were descriptive and evaluated a process-link outcome by determining whether performance of a process/processes increased the likelihood of a good outcome.
Specific interventions included in the review
Nursing care processes within the usual scope of registered nurses were eligible. Interventions were carried out in hospitals (including urban teaching and non teaching, teaching, and rural) or in-home health agencies.
Participants included in the review
Adult patients hospitalised on general care wards (neither ICU nor CCU) for medical-surgical conditions in the United States, Canada, the United Kingdom or Scandinavian countries were eligible. Patients included: elderly patients referred to home health care at discharge; patients with peritoneal dialysis; patients receiving one or more of six medications; patients receiving erythropoietin; patients with congestive heart failure, abdominal hysterectomy, HIV and pneumocystis carinii pneumonia, acute myocardial infarction, pneumonia, cerebrovascular accident, hip fracture, percutaneous transluminal coronary angioplasty, community-acquired pneumonia coronary artery bypass graft, valve replacement; and patients on medical, surgical or bone marrow transplant units.
Outcomes assessed in the review
Studies that included at least one patient-related outcome were eligible.
Data on outcomes were obtained using surveys, from medical or administrative records, or through direct observation. Outcomes used to evaluate the assessment process included: medication, knowledge, pain relief, discharge status, length of stay and mortality. Outcomes used to evaluate problem identification processes included: functional status of patients and the proportion of patients discharged to nursing homes. Outcomes used to evaluate management processes included: health-related patient outcomes (patient self-rated physical status, physical condition, HIV-Quality marker score, iron balance, haematocrit, haemoglobin, bleeding, coronary care unit days, time to sheath removal, time on bedrest, discomfort, medication/health knowledge, psychological status, length of stay, mortality, and patient/family involvement); staff-related outcomes (clinical nurse specialist resources, nurse-physician collaboration, staff sense of accountability, staff competency) and economic outcomes (direct costs of service utilization).
How were decisions on the relevance of primary studies made?
Two investigators reviewed titles and abstracts of studies identified from the computerised searches. Nine graduate students reviewed all identified studies using a literature coding form and screened for data based studies. Inter-rater reliability was assessed by having two or more graduate students review approximately 10% of the studies. Reported inter-rater reliability was 100%. Study investigators screened all data based studies for process-outcome studies. The content of articles was analysed by the authors by summarising the following details of articles onto an article abstraction form: study objectives; sample details; findings; variables; reliability; validity; and the presence of a process-outcome link. At least three investigators reviewed the final set of studies using a predesigned abstraction form. Discrepancies were resolved by consensus.