REPORTS

Screening for Breast Cancer: Time, Travel, and Out-of-Pocket Expenses

Roger H. Secker-Walker, Pamela M. Vacek, Gloria J. Hooper, Dennis A. Plante, Allan S. Detsky

Affiliations of authors: R. H. Secker-Walker, G. J. Hooper (Office of Health Promotion Research), P. M. Vacek (Biometry Facility), D. A. Plante (Department of Medicine), University of Vermont, Burlington; A. S. Detsky, Department of Medicine, Mount Sinai Hospital, Toronto, ON, Canada.

Correspondence to: Roger H. Secker-Walker, M.D., 1 South Prospect St., Burlington, VT 05401-3444 (e-mail: rseckerw{at}zoo.uvm.edu).


    ABSTRACT
 Top
 Abstract
 Introduction
 Subjects and Methods
 Results
 Discussion
 Appendix
 References
 
BACKGROUND: We estimated the personal costs to women found to have a breast problem (either breast cancer or benign breast disease) in terms of time spent, miles traveled, and cash payments made for detection, diagnosis, initial treatment, and follow-up. METHODS: We analyzed data from personal interviews with 465 women from four communities in Florida. These women were randomly selected from those with a recent breast biopsy (within 6-8 months) that indicated either breast cancer (208 women) or benign breast disease (257 women). One community was the site of a multifaceted intervention to promote breast screening, and the other three communities were comparison sites for evaluation of that intervention. All P values are two-sided. RESULTS: In comparison with time spent and travel distance for women with benign breast disease (13 hours away from home and 56 miles traveled), time spent and travel distance were statistically significantly higher (P<.001) for treatment and follow-up of women with breast cancer (89 hours and 369 miles). Personal financial costs for treatment of women with breast cancer were also statistically significantly higher (breast cancer = $604; benign breast disease = $76; P <.001) but were statistically significantly lower for detection and diagnosis (breast cancer = $170; benign breast disease = $310; P<.001). Among women with breast cancer, time spent for treatment was statistically significantly lower (P = .013) when their breast cancer was detected by screening (68.9 hours) than when it was detected because of symptoms (84.2 hours). Personal cash payments for detection, diagnosis, and treatment were statistically significantly lower among women whose breast problems were detected by screening than among women whose breast problems were detected because of symptoms (screening detected = $453; symptom detected = $749; P = .045). CONCLUSION: There are substantial personal costs for women who are found to have a breast problem, whether the costs are associated with problems identified through screening or because of symptoms.



    INTRODUCTION
 Top
 Abstract
 Introduction
 Subjects and Methods
 Results
 Discussion
 Appendix
 References
 
There have been numerous estimates of the incremental cost-effectiveness of screening for breast cancer (1-9). Most of these estimates have claimed a societal perspective, namely, measuring all costs and consequences regardless of who bears them. The components of cost that are included in an incremental cost-effectiveness ratio are as follows: direct costs of diagnosis and treatment, indirect costs (a measurement of loss of productivity from illness and time taken for diagnosis and treatment), and intangible costs (estimates of the cost of pain and suffering). Direct medical costs are divided into two components: 1) costs associated with the provision of medical services, which include physician services, hospital services, diagnostic test services, and medications, and 2) patient-incurred costs, such as parking and travel costs. Very few analyses are performed from a patient's perspective, i.e., incorporating only the costs borne by patients, such as out-of-pocket expenses for medical services and travel and parking costs. Even estimates that claim to have a societal perspective often ignore some patient-incurred costs such as travel costs. We are aware of only three earlier reports of the personal costs incurred by women screened for breast cancer; one of these reports was from Australia (10), and the other two were from the U.K. (11,12). These studies concerned the personal costs of participating in mammography screening programs.

As part of a project to determine the cost-effectiveness of a community-based program to promote breast screening, we estimated the costs incurred by women in terms of time spent, distance traveled, and cash payments for the detection, diagnosis, treatment, and early follow-up of both malignant and benign breast problems. We also compared the costs incurred by women for breast problems found by screening with those found because of symptoms.


    SUBJECTS AND METHODS
 Top
 Abstract
 Introduction
 Subjects and Methods
 Results
 Discussion
 Appendix
 References
 
Setting

This observational study took place in four suburban coastal communities in Florida. One community, with a population of 352 051 in 1992 (13), was the site of a multifaceted, community-based intervention designed to encourage women to participate in breast screening through regular breast self-examination, annual clinical breast examinations, and regular mammography (14). The other three communities, adjacent to each other and demographically similar to the first, had a combined population of 358 890 in 1992 (13) and served as the comparison site for evaluation of that intervention.

This study was approved by the University of Vermont's Institutional Review Board, in accordance with an assurance filed with and approved by the U.S. Department of Health and Human Services. Written informed consent was obtained from each subject in the study.

Participant Recruitment

For this study, women who had undergone a breast biopsy 6-8 months earlier were randomly selected from lists of breast biopsy reports supplied by each of the pathology laboratories in the four communities. Because about 25% of the breast biopsy reports indicated breast cancer, our sampling fraction for women with benign breast disease was one third of that for women with breast cancer to obtain similar sample sizes in both groups.

Women were invited to participate in the study by their surgeon. An invitation letter, prepared by our staff on the surgeon's letterhead and signed by the surgeon, was sent to each woman selected. Included with this letter were a response form on which the woman could check whether or not she wished to participate in the study and a stamped addressed envelope to be returned with the form to the woman's surgeon, who then forwarded them to us.

Participating surgeons (n = 27) sent invitations to 941 women. Two hundred fifty women (27%) did not respond to their surgeon's invitation, and 153 (16%) refused to participate. Five hundred thirty-eight women (57%) agreed to be interviewed. Fifteen of these women could not be reached, and five lived outside the study communities, leaving 518 women with completed interviews. Of these, 53 were excluded from the analysis because of a history of breast cancer or of breast biopsies, leaving 465 interviews for analysis: 208 (45%) from women with breast cancer and 257 (55%) from women with benign breast disease. The 53 women excluded did not differ statistically significantly from the study participants in age, educational level, or household income.

The 465 study participants represented 49% of the 941 women invited to take part in this study. Refusal rates did not differ with respect to surgeons or intervention and comparison sites. The only information available to us on the women who refused was their breast biopsy reports: age and breast pathology. There were no statistically significant differences in age distribution or in the proportions of women with breast cancer or benign breast disease among the subjects who refused to participate and those who participated. Among women with breast cancer, there was no statistically significant difference in the proportions of those with in situ, local, regional, or distant disease. Among women with benign breast disease, there was no statistically significant difference with respect to histologic categories of risk for breast cancer derived from their breast biopsy reports (15).

The age and stage distributions for participants who had breast cancer were not statistically significantly different from those for all women with breast cancer from these four communities, as compiled by the Florida Cancer Registry from 1992 through 1994. The stage distribution also was not statistically significantly different from that for women with breast cancer in the Survival, Epidemiology, and End Results (SEER)1 dataset from 1992 through 1994; however, our participants were statistically significantly older.

Interviews

Women who agreed to participate were called by telephone and given a brief outline of the content of the interview, including our interest in costs, and then appointments were made to see them. The women were interviewed by trained interviewers, who first obtained their informed consent. The interviews, conducted from 1992 through 1995, took place 6-33 months after the breast biopsies, with an average elapsed time ± standard deviation (SD) of 11.0 ± 4.0 months and a median elapsed time of 10 months.

The interviews lasted an average of 34 ± 14 minutes. Women were asked how their breast problem had been discovered, what tests were done to establish the diagnosis, what treatment they had received, and how often they were being seen for follow-up. For each of these steps, women were asked how much time they were away from work or their home, their mode of transport, and how far they had to travel. They were also asked about their health insurance coverage and any cash payments made for the various services. The wording of the items in the questionnaire concerning these issues, using mammography as an example, is shown in the "Appendix" section. Cash payments were for procedures that took place from 1991 through 1995 and have not been adjusted for inflation. The average rate of response to questions about time ± SD was 96.0% ± 1.6% and to those about cash payments was 73.7% ± 5.2%. Data concerning the time spent by the subject doing a breast self-examination, having a clinical breast examination, or by family members or others assisting these women were not collected.

Gasoline Cost Per Mile Driven

The average price of gasoline in the United States in 1994, including taxes, was obtained from the Energy Information Administration; it was $1.174 per gallon (16). Average gas mileage ± SD, based on the make and year of 855 vehicles parked in patient and visitor parking lots of all six hospitals in the largest community, was estimated at 25.4 ± 4.8 miles per gallon. This yielded an average cost of 4.6 cents per mile driven.

Definition of Detection by Screening

We classified a breast problem as identified through screening if a woman was without breast symptoms, and her problem was found through regular breast self-examination (17-20), by a clinical examination, or by screening mammography. We classified a mammogram as diagnostic if the woman had breast symptoms that then led to a mammogram.

Demographics

Demographic information including age, education, and household income was also obtained. Table 1Go shows the demographic characteristics for women with breast cancer and those with benign breast disease. Overall, 53% were aged 65 years or older, 55% had more than a high school education, 72% were married, 34% worked outside the home, 52% had household incomes of $30 000 or more, and 17% had no insurance coverage for mammography. Five percent were nonwhite. Compared with women with breast cancer, women with benign breast disease tended to be younger and to have more years of education and larger household incomes; a larger proportion worked outside the home, but a smaller proportion had full insurance coverage. When the demographic characteristics of women whose breast problem was detected by screening were compared with those of women whose problem was found because of symptoms, the only significant difference was in household income—screened women tended to have higher incomes (P = .006).


View this table:
[in this window]
[in a new window]
 
Table 1. Demographics and other characteristics*

 
Analytic Approach

We have chosen to present a set of costs and consequences to represent patient-incurred costs rather than to translate all of the costs into dollar amounts (21,22). Specifically, our set includes out-of-pocket expenses for medical diagnoses and treatment and patient-incurred costs, time spent, and travel distance. We chose to do so for two reasons. First, the majority of participants were retired, making any estimates of monetary value of their time somewhat arbitrary. Second, for those who were working, we did not collect wage or salary data, again making estimates somewhat arbitrary. Thus, although others have put a monetary value on time (10,12), we prefer to use the participants' time estimates as the way to express lost opportunity costs and to use dollar values only for the cash payment estimates for medical services and for gasoline.

Data on time spent, distance traveled, and cash payments are presented in the following three categories: detection and diagnosis, treatment, and follow-up. Data for detection and diagnosis were combined because all women had a breast biopsy. Most women with benign breast disease who reported surgical treatment believed that their breast problem was removed during the surgical biopsy procedure, which they regarded as treatment. For these women, we have attributed the time spent, distance traveled, and cash payments for the surgical biopsy to treatment rather than to diagnosis.

Statistical Analyses

{chi}2 tests were used to compare differences in demographic characteristics between women with breast cancer and women with benign breast disease and between women whose breast problem was detected by screening and women whose breast problem was found because of symptoms. The relationship between women's estimates of the miles they traveled and measured distances was examined by regression analysis. The time spent away from home or work, travel distance, and cash payments were computed for each of the detection, diagnostic, and treatment modalities and for follow-up visits. The time spent away from home or work, travel distance, and cash payments were also summed across modalities to obtain totals for detection and diagnosis, treatment, and follow-up. Because the data were highly skewed, Mann-Whitney tests (23) were used for comparisons between women with breast cancer and women with benign breast disease, as well as between women whose breast problems were identified by screening and women whose breast problems were identified because of symptoms. We have used a two-sided P<=.05 as the level for testing statistical significance.


    RESULTS
 Top
 Abstract
 Introduction
 Subjects and Methods
 Results
 Discussion
 Appendix
 References
 
Methods of Detection, Diagnosis, and Treatment

The methods of detection, diagnosis, and treatment for women with breast cancer and for those with benign breast disease are shown in Table 2.Go Average time and average cash payments for these procedures are also shown in Table 2Go.


View this table:
[in this window]
[in a new window]
 
Table 2. Methods of detection, diagnosis, and treatment for women with breast cancer and for women with benign breast disease

 
For most women, the problem was identified through screening: 73% for breast cancer and 81% for benign breast disease. For a screening mammogram, the average reported time away from home or work was 1.5 ± 1.0 hour (mean ± SD), with similar average times for diagnostic mammograms and for ultrasound. The average reported times for the other procedures have large SDs, especially for the surgical procedures, indicating substantial variability.

Time and Travel Distance

Virtually all travel was by automobile. Only 11 (2%) of 465 women ever used public transportation for any of their visits for medical services. For a small random sample of women (n = 19), we compared their mileage estimate with the most direct route measured on street maps of the study communities. The correlation coefficient between the women's estimates, which ranged from 1 mile to 50 miles round trip, and the measured distances was r = .9 (P<.001). Respondents tended to overestimate the measured distance by 2.0 miles, but there was no tendency for longer distances to be more greatly overestimated than shorter ones.

The time spent and the round-trip distance traveled for detection, diagnosis, treatment, and follow-up are shown in Table 3.Go As would be expected, women found to have breast cancer spent, on average, substantially more time away from home or work (89 hours) and traveled more miles (369 miles) than women with benign breast disease, who spent about 13 hours away from home or work and traveled 56 miles. Time spent and distance traveled for detection and diagnosis did not differ between women with breast cancer and those with benign breast disease. The largest difference for time spent and distance traveled was seen for treatment, although time spent and distance traveled for follow-up was also statistically significantly greater. Overall, for women with breast cancer, the average time spent and distance traveled for detection, diagnosis, treatment, and follow-up was about sevenfold greater than those for women with benign breast disease.


View this table:
[in this window]
[in a new window]
 
Table 3. Time and round-trip distance traveled for detection, diagnosis, treatment, and follow-up*

 
No differences in travel distance were seen by method of detection, i.e., screening compared with discovery because of breast symptoms (Table 3)Go. However, among women with breast cancer, time spent for treatment was statistically significantly lower (P = .013) in those whose cancer was found through screening, 68.9 ± 98.2 hours (mean ± SD) compared with 84.2 ± 62.4 hours for those whose cancer was found because of symptoms. This result was related to the fact that there were statistically significantly more (P<.001) screened women with in situ disease and local disease (87% screened versus 60% symptomatic) and fewer women with regional disease and distant disease (13% screened versus 40% symptomatic).

At 4.6 cents per mile, the average cash payment for gasoline for detection, diagnosis, treatment, and follow-up for women with breast cancer was about $17; for women with benign breast disease, it was about $2.50.

Time Off Work

Most women (67%) were not working outside the home. Among women who took time off work, the proportion not paid for this time varied for each procedure and ranged from 34% of the 88 women who took time off for surgical biopsies to 60% of the 30 women who took time off to have needle biopsies.

Cash Payments for Medical Services

The large SDs for all of the average reported cash payments (Table 2)Go reflect differences in insurance coverage, as well as charges, with many women having no cash payments. This resulted in a median of $0 for all cash payments except for tamoxifen, for which the median cash payment was $5. The average cash payment for a screening mammogram ± SD was $19 ± $35; however, for women with no insurance coverage for this procedure, this average was $82 ± $35.

Table 4Go shows the average cash payments for medical services for women with breast cancer or benign breast disease separated by method of detection, screening, or because of symptoms. No statistically significant differences were seen between women whose breast problems were found by screening and women whose breast problems were detected because of symptoms. To examine overall cash payments related to screening, we combined the data for women with breast cancer and benign breast disease. The cash payments for detection, diagnosis, and treatment (mean ± SD) were statistically significantly lower (P = .045) among women whose breast problem was found by screening, $453 ± $1139 (median = $16), compared with $749 ± $1417 (median = $100) for symptomatic women.


View this table:
[in this window]
[in a new window]
 
Table 4. Average cash payments for medical services for women with breast cancer or benign breast disease by method of detection*

 
When cash payments for methods of detection were combined, payments (mean ± SD) for detection and diagnosis were statistically significantly higher (P<.001) for women with benign breast disease, $310 ± $721 (median = $34), than for women with breast cancer, $170 ± $717 (median = $0). The reason for this difference was that statistically significantly fewer women with benign breast disease had complete insurance coverage for mammograms, ultrasound, and biopsy procedures. Consequently, their average cash payments were higher. For treatment, as would be expected, payments were statistically significantly higher (P<.001) for women with breast cancer, $604 ± $1518 (median = $6), versus $76 ± $203 (median = $0) for women with benign breast disease.


    DISCUSSION
 Top
 Abstract
 Introduction
 Subjects and Methods
 Results
 Discussion
 Appendix
 References
 
As would be expected, time spent and distance traveled for detection, diagnosis, treatment, and follow-up of breast cancer were substantially more than for benign breast disease, primarily as a result of the time spent and distance traveled for treatment and follow-up. Cash payments for treatment were also greater for women with breast cancer but lower for detection and diagnosis, largely because of their more complete insurance coverage. For women whose breast cancer was found through screening, time for treatment was statistically significantly less than that for symptomatic women. Cash payments for detection, diagnosis, and treatment were also lower for women whose breast problem was found through screening.

To put the reported cash payments of women with breast cancer in perspective, we have compared them with recent estimates of first year costs of diagnosis, treatment, and follow-up of breast cancer from Medicare and from two health maintenance organizations in the United States (24-26). Adjusted to 1994 dollars by use of the medical care component of the consumer price index, these estimates range from $10 990 to $19 290. Our estimate of the average total cash payment for medical services and gas mileage for detection, diagnosis, treatment, and follow-up for breast cancer, about $800, amounts to 4%-7% of these first year costs.

For the women in our study, time spent away from home or work for mammography averaged 1.5 hours. Cash payments averaged about $22 for the mammographic examination and gas mileage and amounted to 23% of the average charge for a screening mammogram. These compare to almost 2 hours (117 minutes) and $15 (United States) in 1994 dollars ($16 [Australian] in 1989) for the monetary value of attendance and travel time for an Australian breast screening program (10), which was 14% of the cost per screened woman ($112 [United States] in 1994 dollars or $119 [Australian] in 1989). The Australian costs for screening mammography also included investigations and biopsy for problems detected by screening (10,27). Our participants' cash payments were also greater than the estimate of $15 (United States) in 1994 dollars (£6.20 in 1986) for the monetary value of attendance and travel time from one of the centers in the U.K. trial of the early detection of breast cancer (12). If a monetary value were placed on time and included with the cash payments, the cost for mammography incurred by women in our study would clearly be greater in the United States than either of the estimates from Australia or the U.K.

It is important to realize that our results are based on recall of events that took place, on average, about 11 months earlier and are, thus, subject to error. The reliability of self-reported breast screening information has been reported for low-income women from these same coastal communities (28). For mammography, reliabilities were {kappa} = .82 for "ever had a mammogram" and {kappa} = .68 for "mammogram in the last year" and were comparable to reliability estimates for other health behaviors (29-32).

We have no way of validating the time estimates; however, for procedures such as mammography, ultrasound, needle aspiration, and breast biopsy, the estimates have face validity. For the surgical treatment of breast cancer, the average time spent away from home, 52 hours (2.2 days), compares with a mean time of 3.4 days in the hospital for initial care for breast cancer in the period from 1987 through 1991 at Kaiser Permanente in Northern California (27) and, given the increasing use of ambulatory surgery, lends some credibility to these time estimates.

Distance traveled was based on women's estimates of the round-trip mileage, which, on average, tended to be about 2 miles more than the measured distance.

Although the average response rate to questions about cash payments was 74%, the missing data reduce the reliability of these estimates. The majority of payments were based on recall, although about one third of the women had some record of their payments available for the interviewer. During the time that we were conducting these interviews, average charges ± SD for diagnostic and screening mammography, obtained from each facility in the four counties, were $96 ± $23, although at certain times there were various promotional activities when lower prices were available for screening mammograms. Thus, the uninsured women's recall of their cash payments for mammography, which averaged $82, also have face validity.

It is difficult to estimate how generalizable our observations are. In two respects, age and breast cancer stage, the randomly selected participants were representative of women with breast cancer in the four communities. These participants were also similar with regard to stage distribution but were older than women with breast cancer in the SEER dataset. We note that, in the 1990 U.S. Census, 25% of women in these four communities were aged 65 years or more compared with 12% of women in the United States as a whole. Other factors affect generalization. Patterns of health care show regional variations (33,34). For the Medicare population, these women have been reported as more marked for excision of benign breast lesions than for mastectomy (35). There has also been a substantial increase in managed care throughout the United States since we undertook this study (36). The distance people must travel for health care and the availability of systems of public transportation vary substantially for large cities, predominantly suburban communities, and rural areas. Thus, caution should be used in applying our findings to other populations and settings.

We did not include time for others who might have assisted study participants or the participant's time spent for breast self-examination or clinical breast examination, making our time estimates conservative. For a screening program, in addition to the costs we have estimated, costs for women with normal findings and also for those with abnormal findings that lead to further tests but do not require a biopsy should also be included in a cost-effectiveness analysis.

We conclude that there are substantial personal costs for women who are found to have a breast problem, whether through screening or because of symptoms. Time spent and distance traveled for detection and diagnosis for women with breast cancer and for women with benign breast disease were similar, but cash payments were lower for women with breast cancer. Most of the personal costs for women with breast cancer were incurred during treatment and to a lesser extent during follow-up. Cash payments for detection, diagnosis, and treatment were less for all women whose breast problems were identified through screening. Further consideration should be given to including estimates of women's personal costs when undertaking cost-effectiveness analyses of breast screening.


    NOTES
 
1 Editor's note: SEER is a set of geographically defined, population-based central tumor registries in the United States, operated by local nonprofit organizations under contract to the National Cancer Institute (NCI). Each registry annually submits its cases to the NCI on a computer tape. These computer tapes are then edited by the NCI and made available for analysis. Back

Supported by Public Health Service grants CA46456 and CA22435 from the National Cancer Institute, National Institutes of Health, Department of Health and Human Services.

We thank the many women who freely gave of their time to participate in this study. We are grateful to Violet Hacker, Mary Ann Huber, Danielle Landis, Christine Lapinsky, and Catherine Olivero for their skilled help in conducting the interviews. We also thank Dr. Donald L. Weaver for his helpful advice concerning breast biopsy pathology. We extend our thanks to the following pathologists, surgeons, and organizations and their office staffs for their assistance with inviting women to take part in this study: Mark S. Beatty, M.D., David M. Bernstein, M.D., Eric C. Borock, M.D., Gordon D. Burtch, M.D., Michael J. Costello, M.D., Dale E. Fell, M.D., Paul A. Graham, M.D, Jacob H. Goldberger, M.D., Hiedi D. Gorsuch, M.D., Warren E. Hagen, M.D., Stuart S. Harrison, M.D., William A. Kokal, M.D., Thomas E. Kowalsky, M.D., James W. Large, M.D., Alan Levin, M.D., Jeffrey W. Lewis, M.D., Steven Lofton, M.D., Arun K. Penukonda, M.D., Luis A. Ruilova, M.D., Abraham Sadighi, M.D., Edward J. Scheel, M.D., Roger D. Scott, M.D., Charles P. Shook, M.D., Robert L. Sonn, D.O., Kalkunte R. Suresh, M.D., Michael J. Sweeney, M.D., John H. Terry, M.D., James J. Vopal, M.D., Roger B. Ward, M.D., Leonard Walker, M.D., Virginia A. White, M.D., Gulf Coast Pathology, and Seldenstein, Levine and Associates.


    APPENDIX
 Top
 Abstract
 Introduction
 Subjects and Methods
 Results
 Discussion
 Appendix
 References
 



View larger version (37K):
[in this window]
[in a new window]
 
Appendix

 

    REFERENCES
 Top
 Abstract
 Introduction
 Subjects and Methods
 Results
 Discussion
 Appendix
 References
 

1 Eddy DM. Some results and insights. In: Screening for cancer: theory, analysis and design. Englewood Cliffs (NJ): Prentice-Hall International Inc.; 1980. p. 195-225.

2 Eddy DM. A computer-based model for designing cancer control strategies. J Natl Cancer Inst Monogr 1986:2:75-82.

3 Urban N, Thompson B, Pasket E. Toward effective application of efficacious cancer control interventions. Advances in cancer control: the war on cancer—15 years of progress. New York (NY): Alan R. Liss, Inc.; 1987. p. 39-51.

4 Carter AP, Thompson RS, Bourdeau RV, Andenes J, Mustin H, Straley H. A clinically effective breast screening program can be cost-effective, too. Prev Med 1987;16:19-34.[Medline]

5 Eddy DM, Hasselblad V, McGivney W, Hendee W. The value of mammography screening in women under age 50 years. JAMA 1988;259:1512-9.[Abstract]

6 van der Maas PJ, de Koning HJ, van Ineveld BM, van Oortmarssen GT, Habbema JD, Lubbe KT, et al. The cost-effectiveness of breast cancer screening. Int J Cancer 1989;43:1055-60.[Medline]

7 van Oortmarssen GJ, Habbema JD, van der Maas PJ, de Koning HJ, Collette HJ, Verbeek AL, et al. A model for breast cancer screening. Cancer 1990;66:1601-12.[Medline]

8 Mandelblatt JS, Wheat ME, Monane M, Moshief RD, Hollenberg JP, Tang J. Breast cancer screening for elderly women with and without comorbid conditions. A decision analysis model. Ann Intern Med 1992;116:722-30.[Medline]

9 Mushlin AI, Fintor L. Is screening for breast cancer cost-effective? Cancer 1992;69(7 Suppl):1957-62.[Medline]

10 Hurley SF, Livingston PM. Personal costs incurred by women attending a mammographic screening programme. Med J Aust 1991;154:132-4.[Medline]

11 Breast cancer screening. Report to the health ministers of England, Wales, Scotland and Northern Ireland. By a working group chaired by Professor Sir Patrick Forrest, London (U.K.): Department of Health and Social Security, 1986.

12 Ashby J, Buxton M, Gravelle H. What costs do women meet for early detection and diagnosis of breast problems? Health Economics Research Group, Research Report No. 6. Uxbridge (U.K.): Brunel University; 1989.

13 U.S. Census Bureau at http://www.census.gov/statab/USACounties.

14 Worden JK, Mickey RM, Flynn BS, Costanza MC, Vacek PM, Skelly JM, et al. Development of a community breast screening promotion program using baseline data. Prev Med 1994;23:267-75.[Medline]

15 Page DL, Dupont WD. Benign breast disease: indicators of increased breast cancer risk. Cancer Detect Prev 1992;16:93-7.[Medline]

16 Energy Information Administration. Energy Prices, Petroleum Products at www.eia.doe.gov/oil_gas/pmm/conttxt.html.

17 Foster RS Jr, Costanza MC. Breast self-examination practices and breast cancer survival. Cancer 1984;53:999-1005.[Medline]

18 Baines CJ, Wall C, Risch HA, Kuin JK, Fan IJ. Changes in breast self-examination behavior in a cohort of 8214 women in the Canadian National Breast Screening Study. Cancer 1986;57:1209-16.[Medline]

19 O'Malley MS, Fletcher SW. US Preventive Services Task Force. Screening for breast cancer with breast self examination: a critical review. JAMA 1987;257:2196-203.[Abstract]

20 Worden JK, Solomon LJ, Flynn BS, Costanza MC, Foster RS Jr, Dorwaldt AL, et al. A community-wide program in breast self-examination training and maintenance. Prev Med 1990;19:254-69.[Medline]

21 Torrence GW, Siegel JE, Luce BR. Framing and designing the cost-effectiveness analysis. In: Gold MR, Siegel JE, Russell LB, Weinstein MC, editors. Cost-effectiveness in health and medicine. New York (NY): Oxford University Press; 1996. p. 54-81.

22 Torrance GW, Blaker D, Detsky A, Kennedy W, Schubert F, Menon D, et al. Canadian guidelines for economic evaluation of pharmaceuticals. PharmcoEconomics 1996;27:700-10.

23 Conover WJ. Practical nonparametric statistics. New York (NY): John Wiley & Sons; 1971. p. 224-36.

24 Taplin SH, Barlow W, Urban N, Mandelson MT, Timlin DJ, Ichikawa L, et al. Stage, age, comorbidity, and direct costs of colon, prostate, and breast cancer care. J Natl Cancer Inst 1995,87:417-26.[Abstract]

25 Riley GF, Potosky AL, Lubitz JD, Kessler LG. Medicare payments from diagnosis to death for elderly cancer patients by stage at diagnosis. Med Care 1995;33:828-41.[Medline]

26 Fireman BH, Quesenberry CP, Somkin CP, Jacobson AS, Baer D, West D, et al. Cost of care for cancer in a health maintenance organization. Health Care Financing Rev 1997;18:51-76.[Medline]

27 Gerard K, Salkeld G, Hall J. Counting the costs of mammography screening: first year results from the Sydney Study [published erratum appears in Med J Aust 1990;152:674]. Med J Aust 1990;152:466-71.[Medline]

28 Vacek PM, Mickey RM, Worden JK. Reliability of self-reported breast screening information in a survey of lower income women. Prev Med 1997;26:287-91.[Medline]

29 Shea S, Stein AD, Lantigua R, Basch CE. Reliability of a behavioral risk factor survey in a triethnic population. Am J Epidemiol 1991;133:489-500.[Abstract]

30 Stein AD, Lederman RI, Shea S. The Behavioral Risk Factor Surveillance System questionnaire: its reliability in a statewide sample. Am J Public Health 1993;83: 1768-72.[Abstract]

31 Brownson RC, Jackson-Thompson J, Wilkerson JC, Kiani F. Reliability of information on chronic disease risk factors collected in the Missouri Behavioral Risk Factor Surveillance System. Epidemiology 1994;5:545-9.[Medline]

32 Stein AD, Courval JM, Lederman RI, Shea S. Reproducibility of responses to telephone interviews: demographic predictors of discordance in risk factor status. Am J Epidemiol 1995;141:1097-105.[Abstract]

33 Wennberg J, Gittelsohn A. Variations in medical care among small areas. Sci Am 1982;246:120-34.[Medline]

34 Foster RS Jr, Farwell ME, Costanza MC. Breast-conserving surgery for breast cancer: patterns of care in a geographic region and estimation of potential applicability. Ann Surg Oncol 1995;2:275-80.[Abstract]

35 Chassin MR, Brook RH, Park RE, Keesey J, Fink A, Kosecoff J, et al. Variations in the use of medical and surgical services by the Medicare population. N Engl J Med 1986;314:285-90.[Abstract]

36 Bodenheimer T, Sullivan K. How large employers are shaping the health care marketplace. N Engl J Med 1998;338:1003-7; 1084-7.[Free Full Text]

Manuscript received March 9, 1998; revised January 8, 1999; accepted February 18, 1999.


This article has been cited by other articles in HighWire Press-hosted journals:


             
Copyright © 1999 Oxford University Press (unless otherwise stated)
Oxford University Press Privacy Policy and Legal Statement