The use of Papanicolaou (Pap) smears to screen for cancer and precancerous conditions of the cervix is widely accepted as effective. Despite the effectiveness of this test, some women develop cervical cancer. Most women who get cervical cancer in this country either do not get screened at all or are not screened frequently enough. Furthermore, some women who develop cervical cancer have undergone screening but fail to follow up after having an abnormal test. Improving access to care for the unscreened population and ensuring timely follow-up for women with an abnormal test result should have a substantial impact on this disease. Melnikow and associates studied adherence to follow-up recommendations for colposcopy or repeat Pap smear in women with previously abnormal Pap results.

Study subjects came from three family planning clinics. All women with an abnormal Pap smear who were referred for initial colposcopy and a random sample of women who were referred for a repeat Pap smear were eligible to participate. The intervention included a telephone call, letter or certified letter; the specific intervention used was done without regard to the type of abnormality seen on the Pap smear. Adherence was documented within eight months of an abnormal result. Attempts to contact the patients for follow-up, adherence to recommendations and patient characteristics were abstracted from medical records.

The rate of overall adherence to follow-up recommendations was 56 percent. Adherence to a recommendation for colposcopy was not significantly different from that for a repeat Pap smear. The use of as many as three patient reminders substantially improved adherence to follow-up. The authors conclude that, although adherence to follow-up was low in this population, it was improved by the use of three reminders.

editor's note: Family physicians should establish a system to track patients with an abnormal Pap smear and may need to implement a system to repeatedly remind women with an abnormal Pap smear to present for follow-up.—j.n.