We have generated a catalog of utilities that can be used to incorporate health-related quality-of-life effects into cost-effectiveness analyses of alternative surveillance approaches for women who have been treated for high-grade CIN. Other investigators (14
) as well as our group (16
) have reported utilities for health states associated with the management of atypical or low-grade cervical lesions. To our knowledge, this is the first presentation of a comprehensive set of utilities for the process and outcomes of alternative post-CIN treatment surveillance strategies obtained from a sociodemographically diverse sample of women with a range of familiarity with the components of alternative surveillance strategies.
Our findings indicate substantial variation in preferences among women concerning post-CIN 2/3 treatment surveillance strategies. The wide range of mean values we observed for outcomes ending with normal Pap or colposcopy results underscores the importance of the process that women go through to obtain normal results, the most common outcome of screening. It has been suggested that a utility difference of .03 or greater can be construed as clinically meaningful;(18
) the difference we observed in the mean utility associated with having a normal Pap alone versus a normal Pap with positive HPV results was .08. This finding is consistent with results of observational studies suggesting that testing for and treating HPV-associated disease can have adverse psychosocial effects, especially among women treated for high-grade CIN.(19
) Although a recent study in which HPV results were randomly revealed or concealed to women suggested that knowledge of HPV positivity among women with normal cytology had little impact on overall quality of life as measured by a general health questionnaire,(20
) this studies differed in that our utility scores were based on scenarios with and without HPV testing whereas in the other study all women underwent the process and only knowledge of the outcome differed. Interestingly, these investigators did find that women who were told that they were HPV-positive reported decreased sexual satisfaction compared to those who were HPV-positive but were not given this information.
We also noted differences in the mean utility for the scenario of a normal Pap test without HPV testing versus a normal Pap test with a negative HPV test (.037), and we found that while two thirds of the respondents equated receiving Pap normal results to receiving Pap normal/HPV negative results, the remaining third was divided over whether having an HPV test and receiving negative results was a net gain or a net loss. There are many possible explanations for this result. It may be that use of the word “negative” in the context of HPV test results did not provide the same feeling of benefit that use of the word “normal” in the context of Pap test results did. Also, the description of the HPV test made it clear that HPV is a sexually transmitted virus, and it may be that women preferred not to be reminded of the possibility of a sexually transmitted infection in the context of a normal Pap test. Alternatively, it could be that women are relatively satisfied with the confidence of the Pap result and the addition of HPV offers no additional utility to them, or women may prefer having only one test to having two. Future studies should focus on the reasons underlying preferences for one screening or surveillance approach over another.
How do our findings compare with those of other utility studies? Several investigators measured preferences related to management of low-grade cytologic abnormalities and screening and treatment of other gynecologic cancers. In a previous study, we used the standard gamble method of utility assessment (21
) to assess preferences among 170 English- or Spanish-speaking women recruited from family planning clinics throughout Northern California’s Central Valley. We found that the mean standard gamble utility for repeat Pap with spontaneous resolution (0.96) was higher than for immediate colposcopy with normal results (0.93), but somewhat lower utilities were assigned to cryotherapy and cone biopsy following repeat Pap (0.93 and 0.91 respectively) than following immediate colposcopy (0.95 and 0.92 respectively).(16
) In a study of repeat Pap versus HPV triage testing among 73 women residing in Sydney, Australia, who had not recently had a Pap test, no clinically meaningful difference was observed in the mean standard gamble utilities assigned to repeat Pap (0.9972) versus immediate HPV testing (0.9967), but the mean value for abnormal Pap followed by treatment was higher (0.9656) than for HPV positive results followed by treatment (0.9354).(14
) In an analysis of the effect of the duration of time spent in various health states associated with abnormal Pap tests, time tradeoff utilities were obtained from 150 female volunteers at Duke University Medical Center in North Carolina.(15
) These women had higher mean utilities for atypical squamous cells of undetermined significance (ASCUS, 0.94) than for low-grade or high-grade squamous intraepithelial lesion (LSIL/HSIL) and CIN 1(0.91 for both); the mean utility for CIN 2/3 was 0.87. Finally, in a study of utilities for ovarian cancer diagnosis and treatment among 13 ovarian cancer patients and 37 female members of the general public, also conducted at Duke University, the mean time tradeoff utility scores for screening tests ranged from 0.83 to 0.90, while the mean scores for cancer states ranged from 0.81 (for newly diagnosed early ovarian cancer to 0.16 for end stage cancer.(22
Several limitations of our study warrant discussion. FIrst, not all women in the sample had experienced all the surveillance and treatment options they were rating, although many had undergone one or more of the procedures we described and all had received abnormal Pap test results in the past two years. To assess whether undergoing any specific procedure was related to utilities assigned to the various scenarios, we compared the utility values of women who reported having undergone a LEEP procedure to those who did not, and found only one significant difference: women who had undergone a LEEP procedure had a higher mean utility for invasive cancer than women who had not had a LEEP (0.89 vs. 0.68, p-value for Mann-Whitney U test=0.0499). Importantly, there were no significant associations between having had a LEEP and the utility scores reported for normal Pap with and without negative HPV.
Second, given the need to limit the outcomes and the information provided to the participants to a manageable amount, we decided to only include two cancer states (very early (microinvasive) and early or late stage invasive cancer), and we focused on the procedures and treatments associated with these states without providing information on their prognosis. Because the perceived prognosis may play a significant role in the utility of a cancer treatment or surveillance scenario, omitting this information may have artificially improve the utility associated with the invasive cancer scenario and the scenario concerning surveillance of invasive cancer.
Additionally, although we were able to recruit a sociodemographically diverse sample with a wide range of experience with the components of the various screening strategies, the participants were recruited from a single geographic area, limiting the generalizability of our findings. Moreover, the educational attainment of our participants was relatively high, with nearly half of the women (48%) reporting college degrees. 14% of the participants had difficulty with the preference assessments, most of whom had completed the interview in Spanish, did not have a college degree, and/or had household incomes of less than $50,000. Although time tradeoff utilities have been obtained from sociodemographically diverse populations in other contexts,(16
) this finding underscores the difficulty that some groups, particularly those with lower educational attainment, may have in providing this type of preference data and specific challenges confronting efforts to engage women with less education in informed decision making. Finally, the small sample size limited our ability to analyze the factors that underlie some of the variations in preferences we observed.
Despite these limitations, we believe that our study has important implications for clinical guidelines and practice given that differing strategies based on Pap and HPV tests yielded such different utility values. Future comparative effectiveness analyses should consider using these values to assess what the quality-of-life impact of implementing these management strategies may be, in addition to their impact on cervical cancer incidence and overall life expectancy.