The CHAs were effective at increasing follow-up among low-income, racially and ethnically diverse women who have had an abnormal Pap smear. The incremental cost per follow-up was $959 (2005 dollars). The total cost for the outreach program for 178 women, including patient time and follow-up care, was $50,132 more than usual care. Although this program would cost less in other parts of the country where labor is less expensive than in the Bay Area, some health care providers may not be able to afford the program. An alternative would be for a health care provider to use usual care for women with ASCUS and AGUS and to use CHAs for LGSIL, HGSIL, or both. The cost per follow-up was lower (more favorable) for women with HGSIL than for all other women. Targeting the program on more severe abnormalities may make sense because women with more severe abnormalities face the greatest risk of cancer and have the most to gain from obtaining follow-up.
In this study, 38% of the women receiving usual care had follow-up within 6 months after the Pap smear. Because we conducted the study in only one health care institution, we can only speculate on why usual care was less effective than what is often reported in the literature (10
). In fact, the study site is unique, which may explain the success of the intervention as well as the poor response to usual care. ACMC is a county facility treating a high proportion of uninsured women who face barriers to care and have low health literacy. Many patients at ACMC have limited financial resources and face cultural or language barriers or both. Some have substance-use problems and practice unsafe sexual behaviors and are at higher risk for human papillomavirus and human immunodeficiency virus. The outreach workers were trained to help the women overcome many of the barriers they face. But until more research is conducted in other settings, it is not known whether this program could be transferable to other county hospitals, public health systems, or managed care programs.
We found that CHAs were successful at increasing rates of follow-up in a health care system that uses conventional cytology. Recent clinical trials suggest that the conventional management of cervical cancer screening creates barriers because multiple visits are needed to screen, conduct follow-up, and treat. Single-visit screen-and-treat approaches were more effective than usual care (30
). In populations in which many barriers inhibit adequate follow-up, screen-and-treat approaches were also cost effective compared with usual care (7
). However, these protocols do not appear to change the long-term screening behaviors of most women. Brewster et al found no differences between intervention and control groups in obtaining a Pap smear a year later except for women with HGSIL/AGUS (31
). Even for women with HGSIL/AGUS in the intervention group, only 63% had a Pap smear a year later; this percentage leaves room for improvement (31
). Consequently, health care systems that adopt screen-and-treat protocols may find value in combining screen-and-treat protocols with educational interventions, such as the use of CHAs.
The primary limitation of this study is that the incremental cost per follow-up is an intermediate outcome and assumes that greater rates of follow-up lead to improved survival and quality of life. Most cost-effectiveness analyses, however, use quality-adjusted life-years (QALYs) for measuring effectiveness (24
). Modeling the incremental cost per QALY was beyond the scope of this study.
A second limitation of this study is the larger than expected difference between the time recorded on the CHA logs and the time reported in the CHA payroll records. These two time estimates were not expected to match perfectly because the CHA logs only record time spent on direct contact with participants. The CHA logs do not record time spent in meetings or in training. We expected the administrative records to show that 10% to 15% of the CHAs' time was spent on meetings and training, but we found that 63% of their time was spent on activities that did not involve direct contact with participants. This information suggests that our cost findings are probably high. It also highlights the fact that CHA workload is not uniformly distributed over time (). Health care systems that adopt this type of intervention should pay particular attention to the volume of abnormal Pap smears. Health care systems with electronic databases of laboratory results can identify the number of abnormal Pap smears over time. They can then make an informed decision about the number of full-time equivalent staff needed to meet the workload and how they manage variations in workload. Health care systems could also look into other ways of paying CHAs. However, alternative payment methods could affect incentive structures for CHAs, which could, in turn, affect follow-up rates.
In summary, the use of CHAs to promote follow-up after an abnormal Pap smear is both more costly and more effective than usual care. The incremental cost per follow-up was $959 (2005 dollars), and this amount would likely be smaller in other areas of the country where labor costs are less than they are in the Bay Area. These findings are particularly relevant for public health care systems where low-income and racially and ethnically diverse women seek care.