We conducted the study at Liverpool Health Service, a tertiary teaching institution of 565 beds, and Macarthur Health Service, a district hospital of 254 beds, with the approval of the South Western Sydney Area Health Service Research Ethics Committee. All patients aged 30-80 years who attended the hospital emergency department or were admitted to the hospitals with chronic obstructive pulmonary disease between September 1999 and July 2000 were identified from their records and invited to participate in the study. Those who agreed were provided with written information about the study and gave written consent. Patients were excluded if they resided outside the region, had insufficient English speaking skills, were resident in a nursing home, or were confused or demented.
The recruited patients were randomised to receive the intervention or usual care. We had intended to use randomised permuted blocks with a block size of four at both sites, but, because of the smaller number of cases at Macarthur Health Service, we used a simple randomisation at that site.
The intervention comprised two home visits by a community nurse. The first, within a week of a patient's discharge from hospital, included a detailed assessment of the patient's health status and respiratory function. The nurses provided verbal and written education on the disease and advised on stopping smoking (if applicable), management of activities of daily living and energy conservation, exercise, understanding and use of drugs, health maintenance, and early recognition of signs that require medical intervention. The nurses also identified problem areas and, if indicated, referred patients to other services, such as home care. After the visit a care plan documenting problem areas, education provided, and referral to other services was posted to each patient's general practitioner, and, if appropriate, the general practitioner was contacted by telephone. At the second home visit, one month later, the nurses reviewed patients' progress and need for further follow up. Patients were encouraged to continue to refer to the education booklet for guidance and to keep in contact with their general practitioner.
Usual care comprised discharge to general practitioner care with or without specialist follow up. Discharge did not include routine nurse or other community follow up.
Evaluation comprised patient interviews at recruitment (baseline) and at three month follow up, conducted either face to face or by telephone by the project officer (OH). At the baseline interview he collected demographic information including country of birth, names of the patient's general practitioner and principal carer, number of people living in the household, main source of income, occupation, and level of education. At follow up, OH sought information on patient satisfaction with care, any readmissions or presentations to the hospital emergency department during the three months, current treatment, frequency of visits to general practitioner, contact with community nurse, smoking habits, immunisation history, knowledge and understanding of the medical condition, help seeking, and self rated health. OH also administered the St George's respiratory questionnaire, a 76 item questionnaire for measuring disease specific quality of life over the previous four weeks,11
at baseline and follow up. It is scored, with the use of empirically derived weights, on a scale of 0-100, and higher scores represent worse impairment of quality life. As well as the total score, three subsidiary scores—symptoms, activity, and impact—can be derived from the questionnaire.12
OH also telephoned each patient's general practitioner at one and three months after hospital discharge and asked about the patient's frequency of consultations and contact with the nurse and the general practitioner's satisfaction with the care provided by the nurse and arrangements for patient follow up. Information on patients' hospital admissions and presentations at the emergency department during the three months after the index admission was obtained from hospital records.
We assessed quality of patient care from the number of general practitioner consultations recorded and the care provided (such as immunisation), number of community nurse visits recorded by patients and the care provided, and patients' and doctors' satisfaction with care. Measures of patient outcome included frequency of presentation to hospital during the three months after the index presentation and quality of life as measured by the St George's respiratory questionnaire at follow up. Intermediate impact measures included patients' knowledge of chronic obstructive pulmonary disease and its management and satisfaction with care at three months after discharge.
We analysed data using the statistical packages EPI INFO 6 and SPSS version 9. Patients' responses to the St George's respiratory questionnaire were summarised as the three subscores (activity, impact, and symptoms) as well as a total score. We used univariate statistical tests to compare the two groups with significance at P<0.05.
We calculated mean scores (with standard deviation) and tested differences between the intervention and control groups using Student's t test. We summarised categorical data as proportions (with 95% confidence intervals) and examined differences between intervention and control groups using contingency tables and the χ2 test.
Before starting the study, we performed a sample size calculation. Based on the assumption that the rate of presentation to hospital over the follow up period would be 30% and that a clinically significant change would halve this rate to 15% or less, we calculated that 120 patients in each group were required to provide a power of 80% to detect a difference of this size at a significance of 5%. As we were unable to recruit sufficient patients, we revised the power of the study and estimated the power to detect a reduction by half was 47.6%. We also estimated that the revised power to detect a 10% change in the total score for the St George's respiratory questionnaire was 50%.