Individualised Versus Conventional Medical Follow-up for Women After Primary Treatment for Ovarian Cancer.

There is recent evidence that routine cancer follow-up is ineffective (Kew, 2011). Knowledge of how best to deliver follow-up remains inadequate and randomised studies are lacking. This trial compares individualised follow-up led by a cancer trained nurse to conventional medical follow-up.

Aim:

The investigators shall conduct the trial using experienced clinical nurse specialists (CNSs) to deliver self-management focused individual follow-up and evaluate whether this brings about greater improvement in quality of life, is acceptable and cost saving when compared to the standard (conventional) model.

Design:

Two arm randomised controlled trial.

Setting:

Specialist gynaecological cancer outpatient services at three cancer centres, one inner city and two urban.

Sample size:

A sample size of 100 patients, randomised equally to the 2 treatment groups, has been determined for this study. See statistical analysis section.

Recruitment:

Based on previous experience of recruiting in this population approximately 50% of women approached will agree to take part. The investigators need 1 year to recruit the sample.

Randomisation:

Participants will be subdivided into two groups those recruited at the inner city centre and those recruited at the suburban. Randomisation will be performed independently for both groups with participants randomly allocated to either the conventional or individualised follow-up in a 1:1 ratio. Allocation of the first participant from a pair to one of the 2 follow-up strategies will be made using randomness derived from atmospheric noise (http://www.random.org); the second participant from each pair will then be allocated to the other group.

The intervention:

Trial follow-up is 2 years from baseline. Conventional follow-up/treatment as usual This will remain unchanged and involve: one post treatment appointment then appointments every 3 months with a doctor. Routinely a medical history is taken and investigations to monitor disease progression including serum cancer antigen 125 (CA125) tumour marker. A physical examination may be performed. The appointment will sometimes involve CNS input and patients may contact the non-study CNS on an ad hoc basis.

Individualised follow-up Patients will meet the study CNS in clinic when they attend for their 4-6 week post treatment appointment. The nurse will negotiate with the patient to agree follow-up arrangements best suited to their needs. E.g. patient-initiated telephone or face-to-face appointments. The nurse will assess the patient at appointments using an assessment proforma.

Data collection:

Quantitative data Baseline qol questionnaires will be provided by the Research Assistant at the time of written consent and completed and returned prior to disclosure of randomisation. Subsequent questionnaires will be posted to participants with a reply paid envelope.

Data will be entered on a 'Patient events' data base and then extracted for all patients for different types of service use during the two-year follow-up period. The investigators will also record primary care contacts and the reasons for these with a questionnaire to General Practitioners.

Qualitative interview data Selected patients will take part in a 45-60 minute one-to-one interview at a location convenient to them. The interview will be audio-recorded following consent and anonymised to maintain confidentiality. A semi-structured interview guide will be used to ensure all important topic areas are covered.

CNSs will be interviewed in a quiet room away from their clinical area as above.

Data Analysis:

Quantitative data Differences in mean patient qol and satisfaction and between the two groups will be assessed by 2 sample t-tests. Assuming a two-sided significance level of 5% with 40 in each sample, to achieve 80% power any differences will have to be quite large to be detectable. For the satisfaction scale (0-100), where the estimated common standard deviation = 14.7 (de Bock et al (2004)), a difference between the two groups of 9.3 would be detectable with 80% power. For the anxiety (0-13) and depression (0-14) measures common standard deviation = 2.75 in both cases (de Bock et al 2004), such a difference would need to be 1.74 in both cases. The investigators may be able to detect smaller differences with the planned sample size when taking into account adjustment for baseline values in the sample size calculation comparing the randomised groups (analysis of covariance).

For the qol scales there are various sub-scales (0-100) and the evidence from literature sources suggest that for most scales the standard deviation is approximately 20-25 (Greimel et al 2003). Using such a value then, a difference in means, detectable with 80% power, would be between 12.7 and 15.9. Even if the ensuing results are not able to detect clinically important differences between the two groups, the scale scores will provide useful information and may justify further inquiry.

The total cost of follow-up for each patient will be calculated by multiplying service use by unit costs obtained from relevant National Health Service (NHS) Reference Costs and summing across all types of use. Unadjusted service use and total costs will be compared between each group using Mann-Whitney two-sample statistics.

Qualitative data Interviews will be fully transcribed and analysed according to principles of grounded theory (Strauss and Corbin, 1991).