Evaluation of a strategy to increase cervical screening through the use of personalised letters

Mullins R

CBRC Research Paper Series No. 15

Aim: The study had two aims:

  • To determine whether there was any benefit in sending personalised invitation letters to unscreened and under-screened women aged 50–69, in terms of them attending for a Pap test.
  • To determine whether the use of PapScreen Victoria (PSV) or Cancer Council Victoria (CCV), or the use of gain-framed or loss-framed messages, was more effective in having women attend for a Pap test.

Method: The Victorian Electoral Commission gave permission to access the names, addresses and dates of birth of Victorian women aged 50 to 69 years old. These details were matched against details held by the Victorian Cervical Cytology Registry (VCCR). Women in this age group, who had no record of having had a Pap test in the past three years and no record of having had a hysterectomy, were sent one of four versions of a letter inviting them to attend for a Pap test. There were 186,109 letters sent between August 2003 and February 2004.

Results: Three thousand randomly selected women were tagged for evaluation purposes – 750 from each intervention group. The overall attendance for a Pap test was 3.2% after three months and 5.1% after six months, and there was no significant difference for either the PSV letterhead compared with CCV letterhead or for the gain-framed letter compared with the loss-framed letter at either three or six months. Slightly more women who received the PSV gain-framed letter had taken action after three months than any other group, but this difference was not statistically significant.
A major issue identified in the evaluation was the proportion of women who received a letter who were not eligible to receive it because they had had a hysterectomy. The VCCR does not have full records of women who have had a hysterectomy, so they could not be excluded from being sent a letter.

The response rate to the letter could be as high as 12.2% if women without a cervix are excluded from the denominator, but this would also indicate that a large number of letters were sent to women who could not respond.

Conclusion: Within six months of receiving the letter up to 12.2% of unscreened and under-screened women had a Pap test, but as there was no control group this cannot be attributed entirely to the letters.

Neither the different letterhead nor the way in which the letter was framed made a significant difference to the outcome. However, there was a trend for the gain-framed letter on PSV letterhead to have more impact after three months. As there is little risk of any negative impact from using a gain-framed message it should be considered in future. Only one sentence of the letters varied, so the framing may not have been strong enough to have made a difference.

A large number of letters would have been sent to women who were unable to act on them as they did not have a cervix, as VCCR does not have complete records of women who have had hysterectomies. The number of letters which will be sent unnecessarily needs to be balanced against the benefits of reaching some otherwise hard-to-reach women when determining whether to run a direct mail campaign.