Further details about our evaluations

This page gives more details about how each page is evaluated. 

U1) Is the site accessible without a login?

Users should not have to create yet another password in order to access information.  Sites which require paid subscription are given zero for this category;  sites which allow free registration are given 1.

U2) Does the site conform to web Accessibility standards?

This is a very difficult question to assess on a project of this scale.  It is not possible to evaluate it for a given site without testing it out amongst a large number of end-users.

We use an automated validator to identify specific problem areas in the code used on a website.  With this method, we can't rule in 100% accessibility, but we can rule it out.

The validator looks for common errors in web coding which can pose problems with accessibility:  use of Frames, incorrect CSS (Cascading Style Sheets) elements, outdated HTML code, absent ALT tags on images and absent Summary elements in Tables.

For a more comprehensive assessment of Accessibility, you should consult the Web Accessibility Initiative (WAI) website.

U3) Is the site design clear and transparent?

Users must be able to identify the main elements of the page, and predict the site structure from the navigational controls

U4) Is the site design consistent from one page to another?

Consistency of design (font, colours, layout, images) is highly correlated with usable design.  Users become disoriented if they click a link and suddenly see a new way of organising the information on-screen.

U5) Can users find what they need on the site?

We test this out by searching candidate sites for sample prostate cancer queries:

  1. Should I get a PSA test?
  2. What treatment is best for early prostate cancer?
  3. What are the side effects of radiotherapy? 

We would expect to find answers quickly (within 3 clicks) with either browsing or searching.  If we don't, the site will be scored lower and relevant notes made in the assessment.

U6) Is the format of information clear and appropriate for the audience?

It's essential that information is presented in the right way for the audience.  Under this criterion, we will consider:

  • The audience of the information:  is the page too complex for them?
  • Are different media used well?  (e.g. images, animations, videos etc should be relevant and concise, and easy for users to access)
  • Is the complexity of language appropriate?
  • Is the information put into context for the user?
  • Can users see where the information came from?  (e.g. for professionals, are references cited?)
  • Is the page the right length for the topic?  Users should be able to get all that they need about a topic from one page, but each page should not be so long that they can't find what they need on it.

R1) Is it clear who has developed the web site and what their objectives are?

The website should have a clear statement of who runs it, for what purpose and any conflicts of interest which may arise.  This should be located in the About.. section of the site.

R2) Does the site report a robust quality control procedure?

A quality product needs a quality process.  Prostate cancer information should be the end product of a careful system of needs analysis, research, assessment of sources, authoring, checking, review and testing. 

For example, the Department of Health has recently begun testing an accreditation scheme for health information.

R3) Is the page content checked by an expert?

Our user group told us that they would expect prostate cancer information to be checked by a clinically qualified person prior to publication. 

R4) Is the page updated regularly?

Information can rapidly go out of date.  We developed a currency guide for different types of information about prostate cancer, which explains how we assign scores to this criterion. 

R5) Does the page cite relevant sources where appropriate?

Users should be able to tell where a specific piece of information comes from.  In many cases, this will be evident from the production process (R2).  In some cases, such as background information about how cancer grows, it doesn't matter as much as with "foreground" clinical questions.

We assess this with the following approach:  "Can we tell where this information comes from, either from the page or from the site policies? If we can't, does it matter?"

Some patients may not want to be overwhelmed with references to meta-analyses and clinical trials.  However, for highly technical and/or uncertain information, it is essential.

This scheme was developed after an extensive literature review on web site and health information quality.