Using the internet for information exchange: design and evaluation of a website for use in follow up of cancer patients at low risk of recurrence

Using Technology to Improve Clinical Care

UKCRN Portfolio Number

4666

Background

For low risk patients cancer recurrence is not usually discovered at review clinic but is recognized either by patients between visits or from standard tests (i.e. blood, mammography). At review staff monitor functional status, identify late effects and provide support to increasing numbers of survivors. There has been a huge rise in use of the Internet for collecting and distributing information. A specifically designed website could provide not only a way forward for undertaking follow up of patients but also information for patients on support, information and hospital contacts. This would reduce the need for hospital follow up, relieve pressure on overstretched services and would provide a systematic method to collect long-term outcomes from the growing number of survivors.

Aim

The objective of this project is to develop and evaluate a web-based two-way information platform for use in a large, geographically diverse cancer network.
Specific aims are:

  • 1. identify the number and characteristics of cancer patients with/without internet accesss
  • 2. test web-site accessibility, feasibility, acceptability, security and usefulness
  • 3. estimate patient compliance with using the website
  • 4. examine score equivalence in patient self-report data collected via the new system compared to traditional paper data methods

Analysis

Comparison of those without (Group 1) and with (Group 2) access to the internet The Chi-square test and t-tests will be used to look at differences between those with and without access to the internet. Regression analysis will be undertaken to explore the impact of age, sex, disease site and deprivation (using postcodes) on access to the internet. Descriptive statistics will be used to evaluate questionnaire responses on access to computers and the internet for both groups.

Web-site accessibility, feasibility, acceptability, security and usefulness

From the telephone interview frequencies will be calculated for item responses to specific quantitative questions. Regression analysis will be undertaken to explore the impact of age, sex, disease site and deprivation on participant opinion. A simple content analysis will be undertaken on the general comments from participants on recommendations (31).

Compliance with using the website

Frequency analysis will be undertaken to assess the level of compliance with both modes of administering the questionnaires including the reminders sent (type and number), level of consent achieved (full or part completion of both modes of questionnaire administration). The amount of time and number of times participants accessed the website will be counted.
Score equivalence in patient self-report data collected via the new system compared to traditional paper data methods

Data quality

The quality of the data collected using the two methods of administration will be assessed by counting the number of errors in the final computer and paper databases and by discussing problems experienced by the research team.

Reliability

Mean scores for all dimensions on the QL questionnaires the MOS 36-item Short Form Health Survey and SDI presented on paper and via the website, as well as the individual paired differences in scores between the two versions (website questionnaire scores minus paper questionnaire scores) will be calculated. To examine the effect of mode of administration on the QL scores for the two versions, Pocock’s method for two period cross-over design trials will be used (32). Statistically significant differences between the mean differences in QL scores will be interpreted as evidence of systematic bias at the group level (i.e. patients reporting systematically more or fewer problems on one of the versions). We will consider 90% confidence intervals and both the intention to treat and per protocol populations.
To assess agreement between the website and paper questionnaire scores at the individual patient level we will calculate proportions of exact and global agreement and weighted kappa coefficients.

Ethical considerations

All participants will be offered information and referral on to appropriate agencies if issues are raised during the telephone interview that are of concern.
Data collected for the study will be securely held within the research unit. Audio-recordings will be identified by code number. The research team are honorary members of the medical oncology team, have honorary NHS contracts and are bound by the same ethical code as clinical teams.

 

Project Stages

Stage

Activity

1. Website Development

Build the website according to clearly defined standards with input from web design experts and user partners

2. User evaluation of the website in a focus group study

Build the website according to clearly defined standards with input from web design experts and user partners

3. Website amendments

Dependent upon user evaluation changes will be made to the website. The new version will be shown to the Psychosocial and Clinical Practice Research Group User Partnership Group for comment and further amendments made if required.

4. A randomized cross-over study to formally evaluate the website

This comprises two groups. Follow up patients will be asked if they have access to the internet. For those with no internet access we will ask them to complete a questionnaire on use of computers. Those participants with internet access will take part in a two arm randomized cross over study (N=53 in each arm). They will complete the questionnaire on computer usage followed by randomization. Participants will complete the same questionnaires on paper and on the internet via the website, the order of presentation being randomized. They will be asked to browse the website. Following completion of both paper and website questionnaires participants will take part in an interview over the telephone asking them about what they thought of the website.